<%BANNER%>

Development of Closed-Loop Interface Circuits for Capacitive Transducers with Application to a MEMS Capacitive Microphone

Permanent Link: http://ufdc.ufl.edu/UFE0020083/00001

Material Information

Title: Development of Closed-Loop Interface Circuits for Capacitive Transducers with Application to a MEMS Capacitive Microphone
Physical Description: 1 online resource (173 p.)
Language: english
Creator: Kadirvel, Karthik
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2007

Subjects

Subjects / Keywords: Electrical and Computer Engineering -- Dissertations, Academic -- UF
Genre: Electrical and Computer Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: There has been a trend towards miniaturization and batch fabrication of sensors inspired by a similar trend in the electronics industry using novel fabrication techniques used in micro electromechanical system (MEMS) fabrication. Capacitive microphones, whose common applications include aeroacoustic measurement and cell phones, is one such sensor whose dimensions are being aggressively scaled down. In measurement microphones, miniaturization will facilitate improved measurement precision, and in cell phones, miniaturization will reduce printer circuit board space and complexity. Both applications will benefit from the potential cost reduction brought by the batch fabrication of sensors. As sensor geometry is scaled down, improved circuit techniques are required to measure the sensor output. This is because at small geometries sensor capacitance is comparable to unwanted parasitic capacitance which reduces the transducer sensitivity. Also, at reduced sensor geometry, the voltage required to bias the microphone could cause the microphone plates to pull in. The goal of this work is to design and characterize interface circuits that are suitable for miniature capacitive transducers. To achieve this goal, the performance of existing open and closed-loop interface circuits are investigated. Scaling of the performance metrics of the microphone and interface circuit as sensor geometry decreases is also investigated. A proof of concept closed-loop analog controller for a MEMS capacitive microphone is designed. A test apparatus is developed to characterize the system over the audio range by operating the microphone in a helium medium which increases the bandwidth of the test apparatus. Characterization of the microphone in open and closed loop mode of operation is presented. Results show that stable closed loop operation of the microphone is feasible with increased sensitivity and the potential to address pull-in issues.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Karthik Kadirvel.
Thesis: Thesis (Ph.D.)--University of Florida, 2007.
Local: Adviser: Nishida, Toshikazu.
Local: Co-adviser: Fox, Robert M.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2007
System ID: UFE0020083:00001

Permanent Link: http://ufdc.ufl.edu/UFE0020083/00001

Material Information

Title: Development of Closed-Loop Interface Circuits for Capacitive Transducers with Application to a MEMS Capacitive Microphone
Physical Description: 1 online resource (173 p.)
Language: english
Creator: Kadirvel, Karthik
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2007

Subjects

Subjects / Keywords: Electrical and Computer Engineering -- Dissertations, Academic -- UF
Genre: Electrical and Computer Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: There has been a trend towards miniaturization and batch fabrication of sensors inspired by a similar trend in the electronics industry using novel fabrication techniques used in micro electromechanical system (MEMS) fabrication. Capacitive microphones, whose common applications include aeroacoustic measurement and cell phones, is one such sensor whose dimensions are being aggressively scaled down. In measurement microphones, miniaturization will facilitate improved measurement precision, and in cell phones, miniaturization will reduce printer circuit board space and complexity. Both applications will benefit from the potential cost reduction brought by the batch fabrication of sensors. As sensor geometry is scaled down, improved circuit techniques are required to measure the sensor output. This is because at small geometries sensor capacitance is comparable to unwanted parasitic capacitance which reduces the transducer sensitivity. Also, at reduced sensor geometry, the voltage required to bias the microphone could cause the microphone plates to pull in. The goal of this work is to design and characterize interface circuits that are suitable for miniature capacitive transducers. To achieve this goal, the performance of existing open and closed-loop interface circuits are investigated. Scaling of the performance metrics of the microphone and interface circuit as sensor geometry decreases is also investigated. A proof of concept closed-loop analog controller for a MEMS capacitive microphone is designed. A test apparatus is developed to characterize the system over the audio range by operating the microphone in a helium medium which increases the bandwidth of the test apparatus. Characterization of the microphone in open and closed loop mode of operation is presented. Results show that stable closed loop operation of the microphone is feasible with increased sensitivity and the potential to address pull-in issues.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Karthik Kadirvel.
Thesis: Thesis (Ph.D.)--University of Florida, 2007.
Local: Adviser: Nishida, Toshikazu.
Local: Co-adviser: Fox, Robert M.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2007
System ID: UFE0020083:00001


This item has the following downloads:


Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E20101207_AAAABV INGEST_TIME 2010-12-07T12:27:21Z PACKAGE UFE0020083_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES
FILE SIZE 21118 DFID F20101207_AABEQF ORIGIN DEPOSITOR PATH kadirvel_k_Page_147.QC.jpg GLOBAL false PRESERVATION BIT MESSAGE_DIGEST ALGORITHM MD5
b1854afa951d1f4080bccb81819c92d4
SHA-1
20ea1cc21307a1c70d5a6bc63382d2928bde3a98
743 F20101207_AABEPR kadirvel_k_Page_173.txt
1ba9f15bc94068702a77f4998bb473a9
c66e089989395ae8d4687966287cf60ec288d659
26135236 F20101207_AABDND kadirvel_k_Page_085.tif
737f32c05cdba66255a5dc69516dd76e
b8e3e91bd80a26fe19c9c9c7810f256e7c107086
5912 F20101207_AABDMP kadirvel_k_Page_043thm.jpg
b051aed13ae8e1f6bf5b8c1fd655648a
195b58d0d699d35f6a080ab16bf0ceacf4f14afb
20345 F20101207_AABEQG kadirvel_k_Page_131.QC.jpg
684ada6fdf8be71a1655c9d7d0aa0731
ac2e89fe0c8490bb464bbc2686bf924aaa84cfef
2588 F20101207_AABEPS kadirvel_k_Page_001thm.jpg
9dfccf0daaec5fff8003d3b30a986ba7
8602d3a54eb63603d9a844e569165d8de657105b
201248 F20101207_AABDNE UFE0020083_00001.mets FULL
7e550355d05a70233bf81cf9630bff43
4296dad0d2456c262e9a46dac0dfc60f44fbf528
35540 F20101207_AABDMQ kadirvel_k_Page_172.pro
df826eeac008130e967f8069e6999488
271a94bb9ab6e35c651ede8f8798cff517fe9ed7
6073 F20101207_AABEQH kadirvel_k_Page_023thm.jpg
bdd6f6c98724db9113a3ddd2cf335b93
84f65db35ce2400f20e489f095bbc894b1ddf109
2279426 F20101207_AABEPT kadirvel_k.pdf
62d23b525250da186bfe863683eda7ed
68778a127991b7074a686df746f55c4a56babe2f
2571 F20101207_AABDMR kadirvel_k_Page_169.txt
129d6055b1290386a03189542a1b61ee
dacd70303105c341de705fa2fa4e6e13b54dca81
24810 F20101207_AABEQI kadirvel_k_Page_067.QC.jpg
834ee98b06b48599ee21db2cc0c81078
7cbc2f0c8d67e00aa60666fd5e5c212195b9b2c5
8195 F20101207_AABEPU kadirvel_k_Page_171thm.jpg
8545e97e0bcf0bff07c2a90e55178dc8
05ae10f95219672aa5d9f07f7549a8daae367838
5707 F20101207_AABDMS kadirvel_k_Page_032thm.jpg
1ae4d0947bcb2a688a231be2c55aad0f
b67827bc38542c8442ef945eeec7e21d5aa3367f
15988 F20101207_AABEQJ kadirvel_k_Page_028.QC.jpg
0557efff7a5f404616f610b0c0fd48f6
86ab9e1d10d363170f7e39d09750b1b60e6afb30
5805 F20101207_AABEPV kadirvel_k_Page_104thm.jpg
4865e0267948e000416d236f65cd6c25
9b9597d1b727f38de48332ab8b2e1065a7cda466
27348 F20101207_AABDNH kadirvel_k_Page_001.jpg
ce1369ce5cc6e91baf1c28ee6d2dc279
e75864c5dda7c7026d87f168ed4d0253efab602c
63060 F20101207_AABDMT kadirvel_k_Page_121.jpg
6d5ee6c36692056092d9af4897f4c94b
fe294e1ab2421ad095ab0489eee576bd1dce85bc
8075 F20101207_AABEQK kadirvel_k_Page_049thm.jpg
f79b126aa2f566345f96114d2f4f8eee
4847f34740894987e346820f125f1c324ac7e40d
26112 F20101207_AABEPW kadirvel_k_Page_039.QC.jpg
16a04305efa6e0dec986f472162a167f
8eda5bd94f11a08ce639f695617e1734be2be72f
10870 F20101207_AABDNI kadirvel_k_Page_003.jpg
b1e7db2ddb82af5961527c9873a97c8d
58fae3f4a67f88fef812604a223abf675e5c7560
1089068 F20101207_AABDMU kadirvel_k_Page_034.tif
9275af77373c78256e805ec7cdca60ab
bda6a9eef0c5a960883b761be0c3499ba8211ba9
34186 F20101207_AABEQL kadirvel_k_Page_042.QC.jpg
7d35b866e2d6d3a309933ba290128bf4
b65a1357390b8043c1f3eaec2993c00a98e001ea
22321 F20101207_AABEPX kadirvel_k_Page_035.QC.jpg
9e600eea6f1a71b845da41316cc10535
3645acf7a6af2b3f7982dade26d16108489aa24f
69868 F20101207_AABDNJ kadirvel_k_Page_004.jpg
8150d9e936590fa2467ff31d1dbe3f42
78490a8a0a539406d84035ead4ef94a32e09686a
38758 F20101207_AABDMV kadirvel_k_Page_124.pro
07f213ec502b6b88f71f957f8e83b0e0
189b08bf7ed8f289af1b25d611d637a4fe759aa5
1384 F20101207_AABERA kadirvel_k_Page_003thm.jpg
142e26f2e603b4e7990a189e7cbf2c3d
6d07ac11cb5a9d669e66828f7ffee2cd16e3e7e3
5046 F20101207_AABEQM kadirvel_k_Page_127thm.jpg
a53e0d8fc964886c4dcf8a30212ea45a
c33022ca3676207e5c68438e237943cbc07663c5
37934 F20101207_AABDNK kadirvel_k_Page_007.jpg
91ae33d57df80b59a286f2f1dbb03de2
03f68dc5d4e5d4a0923d7d27b8f3bc69749597c4
19359 F20101207_AABERB kadirvel_k_Page_135.QC.jpg
a3ce388712d3309d19de494c504b9489
0f0b81b34450a741e43f3393a68122b4131917c8
6762 F20101207_AABEQN kadirvel_k_Page_141thm.jpg
02e37fa8f413041cfecf79ee1f4c1eec
ee03a626959fac886b059083aa473cd8aa285ec4
6280 F20101207_AABEPY kadirvel_k_Page_083thm.jpg
c2d5e154146e7abd0be9dd974b2893d5
eb7655aa31ea8453876c33143d03e90b12b5c500
111915 F20101207_AABDNL kadirvel_k_Page_008.jpg
cca986c44818c2dfcbb7d9d9021a9ca0
ae82db1676511fe511c68cfaf97a60832de01c1e
19327 F20101207_AABDMW kadirvel_k_Page_066.pro
b339ee811417a0750577685111b36f81
7c73e161feb5d959bc2427a60a6408c8c6664b7b
2891 F20101207_AABERC kadirvel_k_Page_007thm.jpg
48c0ba14795b4eae975ddbf15740cef6
e69870050050ca4b07ebc669d42f76b01621f721
1806 F20101207_AABEQO kadirvel_k_Page_015thm.jpg
9cc0e8e63615d5fd52b370c08bc7f479
c39b3cae5a204f7fa790f74363be0250771d977d
32298 F20101207_AABEPZ kadirvel_k_Page_013.QC.jpg
5396711fe4cea401056fc66a41d6cab0
91aae4281116be927fdf8f181543ec453c2461d9
21736 F20101207_AABDNM kadirvel_k_Page_009.jpg
fedae3622821cdcb0ada3b9d07d4c53f
2634c932b415be887f02ca90da95a24a63863cd5
F20101207_AABDMX kadirvel_k_Page_079.tif
2c5502bd4c566d83f9463e0e3ccddb5f
d202194f3db09e54448d423c6305cc7c71f6dbda
57024 F20101207_AABDOA kadirvel_k_Page_025.jpg
ce4dbedb8e46dd0e6c1ddc95fc01a77e
9478c4e5e60f5d3bfbf5904edfa2395ec72205f5
16186 F20101207_AABERD kadirvel_k_Page_152.QC.jpg
96c23163efa3e1195eeb95054064ac43
00f70eea912cd291bf23e5812a395d01e109a1d0
16257 F20101207_AABEQP kadirvel_k_Page_156.QC.jpg
77dda963bbfdba0333339f6304257063
d7cb13c3c9097fb3694cb644d72a0b1f8b611cc1
101106 F20101207_AABDNN kadirvel_k_Page_010.jpg
1f071c631b4461447c5f3018360b7f13
28278ad1211669c4f67b76db950d7c53250675a4
2413 F20101207_AABDMY kadirvel_k_Page_151.txt
61807e18650e46a65e55fdf0cf93f6da
751cc7e73a85391a61b82438ca33c2bae6e1fa7d
87006 F20101207_AABDOB kadirvel_k_Page_026.jpg
a87a1c8c48746f64daca71737ac88e09
754ed838164ecee5442e7693b1f49b7f36a17033
8726 F20101207_AABERE kadirvel_k_Page_164thm.jpg
2e3dbe1dca16b7d3e2ff9f3034857676
4536bf2f98076f7dabd795fa66508a19744eb7b3
26355 F20101207_AABEQQ kadirvel_k_Page_141.QC.jpg
9368e0b7313fbc3e530abb0eaa5c3b84
b88725457270a51b0047186f55bf5927adc5e764
114267 F20101207_AABDNO kadirvel_k_Page_011.jpg
e4c437ec7f68119de22793f48817d7e5
a82989d606dcccd133714138d26246515b420e06
28764 F20101207_AABDMZ kadirvel_k_Page_028.pro
d2afe40318ab8dc579c73892315dd766
6ad7021e2eadd10838d103d9e6272f2e7628b498
85122 F20101207_AABDOC kadirvel_k_Page_027.jpg
ffbbe2c82f7a0c9100cc620de11c6793
013d8435ab8b5884d59f233ce36b150d0d510bb3
8386 F20101207_AABERF kadirvel_k_Page_016thm.jpg
4ef5cabfd7a8b85b47d4e523578d6125
7c6776a0601584d6b2c686012929bbf9a5f061d7
1375 F20101207_AABEQR kadirvel_k_Page_002thm.jpg
4f9cc7ad02fcd6465a1da489a7730af5
4de47bae60a1b8bc476ecab117275356025b3fb5
97610 F20101207_AABDNP kadirvel_k_Page_012.jpg
8b9eca86ac79ded6d1cf520ab2a6a14c
4ce1f5ffe2fbbe2ea7d07fdd2c24facbed93fd2b
49399 F20101207_AABDOD kadirvel_k_Page_028.jpg
55e70aed0d8909db9c8286eefb60d0e4
7662ad79c5ceed2045a158fa157303de415287b7
4676 F20101207_AABERG kadirvel_k_Page_034thm.jpg
05c8a2991087394dcc7d113515f7ac23
079b5615a1e7d926399a1ecfb1cc52c3872c082e
19310 F20101207_AABEQS kadirvel_k_Page_046.QC.jpg
763212ee064b6a448938de49262538ac
775d5b78cb0521870c66552c1c7a616d7c1838c2
108883 F20101207_AABDNQ kadirvel_k_Page_013.jpg
1bf1a4c2ecf6038cb56a9f3ab274c6bc
afd73c740d96b8209d976cd3a33ea4da21125bd5
93871 F20101207_AABDOE kadirvel_k_Page_030.jpg
90fd52e64b407e95e371593a858144f1
09e030f0b20dde7425d33fc77ad72f7493d9c9a9
7129 F20101207_AABERH kadirvel_k_Page_068thm.jpg
37ab5662014342171d2be8f14dc7bbf2
7dc39afcd873045fae154eb002fb74bf5d96be32
24981 F20101207_AABEQT kadirvel_k_Page_084.QC.jpg
7d34851ba0408726dd9a8bd40baa4c92
5ec973cb8db66c07a28fdd466a8b6ec1c163f3ab
15661 F20101207_AABDNR kadirvel_k_Page_015.jpg
f4a2afe0465a991545146779cb4b892a
ef3ae903b1ecc33eb9d1152d20c08fd6abc35891
59589 F20101207_AABDOF kadirvel_k_Page_032.jpg
0325c54b6757f789e62f45970f1bbe79
b9fdbbc1e5699ec6410ee0f4d31b14236b620013
6471 F20101207_AABERI kadirvel_k_Page_139thm.jpg
b1a7443220d0c2a9ceca49999a01da02
0a97c287e9352eef6a9593359bc6bc6cae983767
23215 F20101207_AABEQU kadirvel_k_Page_040.QC.jpg
2f8389399e198ec28088f528baabfd1d
30ce14f8f0cd230b0fb714ae9142530fe670d221
103643 F20101207_AABDNS kadirvel_k_Page_016.jpg
bb94f802558fc9e9f74d296cde3befd0
f6d08a81ae8e35f23f3930c6ce6afa9b620ebc2f
81449 F20101207_AABDOG kadirvel_k_Page_033.jpg
df2c649cb49314a2ca3110bfd9747b30
b1f1ddcc5f567311543e5cbd84b3598f40356371
26254 F20101207_AABERJ kadirvel_k_Page_033.QC.jpg
ecce35dfcc6fd186ce2dbed505d68e75
7a5dd7921cc216ad87fe3a5dd20e3c9654cf5c50
5820 F20101207_AABEQV kadirvel_k_Page_048thm.jpg
95ab898eb943db124b605f1ef1a907cc
0caa6c4bdc078d9d41d1225941f0f38384441ea2
109565 F20101207_AABDNT kadirvel_k_Page_017.jpg
0c85542383c3c48f3786eb04b886c615
a27e8201acf2de73ca466989bd9ba63f21bd8f02
46242 F20101207_AABDOH kadirvel_k_Page_034.jpg
52b41b631cb4d13c7481424fffbae093
753071fc8a9c08cf33f651f20af53e3291172c76
6453 F20101207_AABERK kadirvel_k_Page_040thm.jpg
27f63106c51f48d2a962ea90f45a756a
b38a8dbd265c8467b683ed8dbc8d1a64b15e54bf
6448 F20101207_AABEQW kadirvel_k_Page_054thm.jpg
4a7afc687eef18a3a88fdb7181d96589
93949f5868e43c8f6b79d2915c054e48d8d631fe
101216 F20101207_AABDNU kadirvel_k_Page_018.jpg
9d9f768ab22586d79d9ff3a4737bc410
e60a5cf4c58a6434eba0d2186d1d5906f1a9d7e8
72091 F20101207_AABDOI kadirvel_k_Page_035.jpg
b25635e22c54c39416e016362e4ab65b
99965ea42373266e3fef545a63a927a447aaff0b
8023 F20101207_AABERL kadirvel_k_Page_047thm.jpg
3501ec4c6e019e3674a39773601806a7
0c2e1434b2a6ba7487125f414ad9d7b932d5d47e
6879 F20101207_AABEQX kadirvel_k_Page_120thm.jpg
74353879ed91a291d7485ece0c35151a
30de494397559767cea530b6748498ce02b83d26
64699 F20101207_AABDNV kadirvel_k_Page_020.jpg
3c643c1f3eaf54e309b7b99b6864d58a
2a774b21bd5e8dd5664de344dde394f9317e10ad
63820 F20101207_AABDOJ kadirvel_k_Page_036.jpg
fe02e74e9140eca8db32b0957f4b9d6b
ddd78024b33614380ef7c85b579920b5b7e2a31b
32236 F20101207_AABESA kadirvel_k_Page_064.QC.jpg
5384595c709602a3070584154e170d38
dfc5988190a3557d85435e914776ec93bfc32176
10865 F20101207_AABERM kadirvel_k_Page_173.QC.jpg
9d5659c7324547175ba471a43ee61001
a7080e29eefb4e9d2239739d62af3d06b8b807f2
6274 F20101207_AABEQY kadirvel_k_Page_005thm.jpg
88c5b42879e13932b936a58a819c7450
85a6145e37e767943f6590eeab2778c26bb90f18
69973 F20101207_AABDOK kadirvel_k_Page_037.jpg
c0ebf78f85aa5d4369d3319bbe53a6dc
74c13f9b66a89a00f2f543ac5adc96933b4a0cc2
70235 F20101207_AABDNW kadirvel_k_Page_021.jpg
bb4bf1be5ef945446a6154027400b977
05cf9312e47dbb908d4cfb5b290660df2fd37f5d
7214 F20101207_AABESB kadirvel_k_Page_150thm.jpg
a1ea7523f4e2a975110abfd0ea84cbdc
41d2e7c170acecd1dd6ba26f81bbf096da031018
5777 F20101207_AABERN kadirvel_k_Page_163thm.jpg
7a9a9c28fb56ee3cb8323c4ea39f0200
b67e74444fc6030b5d78804537b74640967c7715
81637 F20101207_AABDOL kadirvel_k_Page_038.jpg
128dc0f2b7ccadcd639dd76e7f3616c7
3e6121c105a645051c9eeea6aefa2eec8e5652a0
32344 F20101207_AABESC kadirvel_k_Page_069.QC.jpg
9126ab25c5d0e6e3249e02d929359767
0233b71b7c2db33bcfd2a28ae2038d12a65c84a0
26582 F20101207_AABERO kadirvel_k_Page_144.QC.jpg
066be8252058bf61e0285b2da9798fbd
ebcfbe892aea1d65619f544b39712cdded2da9f1
31610 F20101207_AABEQZ kadirvel_k_Page_061.QC.jpg
3ef68781516f0d510af67690fcc40ce8
1f32cdb7735050f0ecca3130c430f28d2f1882bd
66020 F20101207_AABDPA kadirvel_k_Page_055.jpg
99e2fd5ecd9beccb2dd536601c1c41fa
a1a64f0e4f2dac729d6c46122bdee20f33c2d807
84220 F20101207_AABDOM kadirvel_k_Page_039.jpg
2a8efe943b2565603ca348b8303555ce
4bda43bd35bf0831b5936eed0c36c6b6af701d17
60260 F20101207_AABDNX kadirvel_k_Page_022.jpg
983a9ca99d22829cb65551a60f0dd16a
66e8a033125f86d3c8b660d236f07fef0d935f30
16205 F20101207_AABESD kadirvel_k_Page_071.QC.jpg
097154f9ea6b7786dcf436d03ea27852
17e67e26c88de03182c3a21f4f454c534ffabb80
22686 F20101207_AABERP kadirvel_k_Page_083.QC.jpg
cba0907c3956732b257ea15ecae06cd9
27c4aa8b891d73ca960f78f5778dc0fcb7941e6a
91071 F20101207_AABDPB kadirvel_k_Page_056.jpg
2519e873a9b688fbe5de9f72da61f06a
609c98ded1c5ec043ece270ca476ff8286d46a67
76705 F20101207_AABDON kadirvel_k_Page_040.jpg
44e411ceb85687817762b31e58db4fe7
7e34fe0291d8e627f0e23042dd517896e5f36fb1
72251 F20101207_AABDNY kadirvel_k_Page_023.jpg
4044b83623106bb949db1aa6d8055f2a
21ecdc854f8782b66b032d9d25eff1eab904fe5b
8557 F20101207_AABESE kadirvel_k_Page_062thm.jpg
f7e2a1ae17bb755cbf84326a07c16074
352ef19b85dab4779533e1c3a10e2a3d52e541e5
11689 F20101207_AABERQ kadirvel_k_Page_130.QC.jpg
9e9c990eab2b08694e6c725f942e17f5
c2ef111d3031f93ccfe137b5fa9ab00f5d08f74f
55847 F20101207_AABDPC kadirvel_k_Page_059.jpg
ded8a10ca9facb8d88c50066d2f32096
ecd566e4dcc60508e01754e024baf8bba0d78a67
104499 F20101207_AABDOO kadirvel_k_Page_041.jpg
1ce5f39d6023933eda33fa6a4f768ec1
589ae02378095932c00f55713ea97eb7e85c57ef
63636 F20101207_AABDNZ kadirvel_k_Page_024.jpg
24698157be06632aa8c37ee30b174f15
00b29f1d20184a3b497522b45ad87a226f300688
30030 F20101207_AABESF kadirvel_k_Page_166.QC.jpg
770c37d6638f57b832119486a4ba8e9a
1d6a20523499f477c398724cb3bc16570e24f54d
19753 F20101207_AABERR kadirvel_k_Page_024.QC.jpg
76da3d85f145c3a6e269a058cca1c658
e42ecc33269894ab82d9123e2f61503dd4c3e446
52494 F20101207_AABDPD kadirvel_k_Page_060.jpg
f32eb964b9566286adcfd121d438f71a
0c6a9610480d0d0b7443ce8ed4ce3a6e1bc9d178
110432 F20101207_AABDOP kadirvel_k_Page_042.jpg
754e9e217acd14fce274deb1e6cdfc33
3bb221df55e2000d2ef281e727e6f14fa1175e6e
6644 F20101207_AABESG kadirvel_k_Page_073thm.jpg
9bdffb20f2670fa9af16f917fdfe2a2a
c2d8f8c43631e875093fb1ed138320c6afec8a50
31290 F20101207_AABERS kadirvel_k_Page_170.QC.jpg
1c5942c42a383fb2a96dad3f9fc62a59
8daaaacc18fae064c3178713c1fcc5ddea0be171
102901 F20101207_AABDPE kadirvel_k_Page_061.jpg
acafcc8c1da6a04b31d41884acbb4917
437da858c1fd2fe5e8ba4b6ac0408d2666d7d221
70676 F20101207_AABDOQ kadirvel_k_Page_044.jpg
c3b9e07b9667637761d576b9232d2d73
d2bfdd09287e77d9d8696d1e883f3feb8b1122f6
6353 F20101207_AABESH kadirvel_k_Page_044thm.jpg
82775a86f564bb1a2880d16aea938006
bcdecf98235637b9e07eda240a504f4e87fef4b4
8428 F20101207_AABERT kadirvel_k_Page_086thm.jpg
ef94311229d472a8229a9e0d20ab357a
a765964c346f1b82b91337b86fd8dc6968c26342
106183 F20101207_AABDPF kadirvel_k_Page_062.jpg
9c35cf5e00eace0b5713094641677746
98e9615ebe5c88526e26e2b8374bb25c2e21d954
25688 F20101207_AABDOR kadirvel_k_Page_045.jpg
0a0cdc54a7b849c67e0042e5f9be961f
3ed571b5ad664b647a858d289438522c1cb31686
6705 F20101207_AABESI kadirvel_k_Page_134thm.jpg
358ff45503c92aededdf4e885c6dd708
40e2a2461d13f5a682f0df0611babd1e7eb23601
21561 F20101207_AABERU kadirvel_k_Page_142.QC.jpg
72a3d4fc362326828628d686f900a4d9
62738a8a4520ae123f73c5cb7419ced4c7be8464
79232 F20101207_AABDPG kadirvel_k_Page_063.jpg
34feca3ddf2e66d5a93b9d021e03e51f
cc24cf9767fabf12c90d4e588c23e358b3b88393
62961 F20101207_AABDOS kadirvel_k_Page_046.jpg
c9e2262610dd32f759cc26186e90f46b
5e1d8c50ec6eb9bb6c0979fda6e0c795e9d3c5bc
6242 F20101207_AABESJ kadirvel_k_Page_095thm.jpg
4f6fad3aad40725cbb5ccdeb96d15dc4
1c0f7778ba045534fd9da7db42d89c9a9739a572
30594 F20101207_AABERV kadirvel_k_Page_169.QC.jpg
857ad2adfcc4d80c776199e056705b8e
ba0eb1e66ff61207ab6ee252533c0505d6c463bd
104988 F20101207_AABDPH kadirvel_k_Page_064.jpg
737d5d91bba96bf85b0c0cd80155cfe3
1e64b609e701847c603d76bf9d7afeeb26f5ffc7
62326 F20101207_AABDOT kadirvel_k_Page_048.jpg
53a055ac16b56963b3bfe05f82468e89
48591ea797fdd33da70da06df10fee5c8dfe2233
17429 F20101207_AABESK kadirvel_k_Page_082.QC.jpg
5dc84fdfb760b180ee0ccb6e83fe5c40
5dfea2d8dfd59adb03811ee24595c0c9d633cbd9
5873 F20101207_AABERW kadirvel_k_Page_004thm.jpg
e3ad9a01807bb58fd3094d1e4f6ac874
06f13819387067c1d88dcab56e436dd883fac3e2
79738 F20101207_AABDPI kadirvel_k_Page_065.jpg
e7ccb5fbb93ad7d1b0db8172b185d8a8
4b1f3308c67ffd5c2e4786619c1940fc1f8781cc
100105 F20101207_AABDOU kadirvel_k_Page_049.jpg
742ce7fb500ef542e9f7eb879bdddacd
014a745abb8f7905e41b24959904e4e460d13857
5428 F20101207_AABESL kadirvel_k_Page_082thm.jpg
27bcfcb77bae3e1756c162c12ea52351
c46343e7e7b3710c7027b2ac191196dc0a542c7d
6227 F20101207_AABERX kadirvel_k_Page_122thm.jpg
55fa9b8a4c6822e6057fdfa7aabf478e
47a83254ad05e9993ce5e1d03da50ad5858b3eed
46005 F20101207_AABDPJ kadirvel_k_Page_066.jpg
ccd0e3c4563b7f1e9e41f8408199be3d
609d0eb15776a7c5e51f5616b359503cd4216715
67684 F20101207_AABDOV kadirvel_k_Page_050.jpg
5fbe649af68379a143bff6d2687098dd
1b3929e8f03497eefbe1d3dddf97b820d13be219
21983 F20101207_AABETA kadirvel_k_Page_105.QC.jpg
10fe045517b68e8bea87559d101e495e
94791c908c8a57f61894b9782d9e552b8548ec92
7120 F20101207_AABESM kadirvel_k_Page_014thm.jpg
63ceae4be676d94753f57ad545e405a1
bc77159af39fa5b145e6d8f7c724bc1b2563b8be
34193 F20101207_AABERY kadirvel_k_Page_017.QC.jpg
7109f215906919fd7df35812a7c6de9c
8aae5beb00c83329e4f6f47b644633e95b046cd0
79673 F20101207_AABDPK kadirvel_k_Page_067.jpg
c3e0954a6e0cc1327a82a87c027df372
d8d127c55fc15b15f424ee18bb7645d524a615ab
70187 F20101207_AABDOW kadirvel_k_Page_051.jpg
a2a64e3f880487b208e442c637471b54
64f8c177e0af6d338ca5f8ea7609cc93cf0da452
16243 F20101207_AABETB kadirvel_k_Page_060.QC.jpg
d8d99760a4bfe2e6481dc8437f1ed11a
19aa09349b43de9144d0b9af983f33be5ab0b580
23204 F20101207_AABESN kadirvel_k_Page_073.QC.jpg
eefef9e06304aa93238b5306b25e41de
a214f77ecd5c41e839f23ae864c6dfb49a937260
8225 F20101207_AABERZ kadirvel_k_Page_041thm.jpg
f4016d2ef23bd1157341cf1d002d93f1
9a40d66bcba4342a0cc66328a3b380e809944222
88362 F20101207_AABDPL kadirvel_k_Page_068.jpg
609066542f4d1fd754f6951c64868ec6
a89c33d7af8e82d7fb0833ae282c2bbc6f7c0567
93847 F20101207_AABDOX kadirvel_k_Page_052.jpg
21437dc8c47d74b72cc9037d0442eb65
032fb812698be58788d9d17618ff6deb16411a98
7137 F20101207_AABETC kadirvel_k_Page_112thm.jpg
c7ef551f63303d49132de719dcdd8c4c
71ddc88b65524a92a735c42a4138807939f812c8
6205 F20101207_AABESO kadirvel_k_Page_051thm.jpg
c881adabf2b0c81787da56d3ad7780c6
a086c3e05681a9b6d1f8ff940685f008b35ec1c7
104809 F20101207_AABDPM kadirvel_k_Page_069.jpg
b12511fa5641b09fe677ee46e518db6a
337f14727ad613ef025b0f9d2d057cc26905633d
99699 F20101207_AABDQA kadirvel_k_Page_085.jpg
5c5015a134d6a31856085d3d676ac4bf
8cce997261a02099f99623632d49e0ed8d678cc9
7839 F20101207_AABETD kadirvel_k_Page_008thm.jpg
89eda94c20aeaf12b75d7d9935b8e5e9
46d045780b194c04528b5ff48b3cbe718f695e5b
26739 F20101207_AABESP kadirvel_k_Page_098.QC.jpg
5a6ac6a0303e63ffeaf788f723b2f10e
644593378d997f89418a07f373a567b58220aa15
52491 F20101207_AABDPN kadirvel_k_Page_070.jpg
3625153ec59c1315214b6c008f6c3e83
ae7b1dbb44bb50274d8f873580bae61b7862085d
45499 F20101207_AABDOY kadirvel_k_Page_053.jpg
43c86dcc9014b4e10bdc2646077a808c
2db9ce46411b56373fe16fbd5abd25b60501b4b7
106436 F20101207_AABDQB kadirvel_k_Page_086.jpg
51b839ef2419935dd712d8668db5231a
8d1f2857895af8625306dd0e91bb2000a63d3176
19733 F20101207_AABETE kadirvel_k_Page_048.QC.jpg
002cd6c0665ab882da95b0dd4270f932
616591aa289fea0a02ed4b548ee1cb96777bcdcb
2189 F20101207_AABESQ kadirvel_k_Page_009thm.jpg
a5ede0eeb28222fa745425d8e7d3749b
330c95c571e9ad36cfbacb4aeb5675b6e506c8df
52755 F20101207_AABDPO kadirvel_k_Page_071.jpg
625d6404cfc5c941ab7b6f447a556a5d
f3b045be6e64cb74649ac574361970f8e8d16dfd
67047 F20101207_AABDOZ kadirvel_k_Page_054.jpg
7c6e5403f1ba19498fd8f7450845a5da
8d6be0a05bf72fe4050ee8d39b2c68d589d25887
58837 F20101207_AABDQC kadirvel_k_Page_087.jpg
c29a6dc564fae14f833a4257eb107093
e279da69aeced5986ba07ddc4dc6c10c8cd650b9
5572 F20101207_AABETF kadirvel_k_Page_021thm.jpg
3ecbf90f2effda459f45bec12566ef1a
c05b691ac336e687c484548b0dfb43451a3173c4
5830 F20101207_AABESR kadirvel_k_Page_121thm.jpg
47e06f2bc15440e5717eb79a4aad35da
1732b648d93e8d4dc9b3bf8631adec9abe4a40a7
95513 F20101207_AABDPP kadirvel_k_Page_072.jpg
2995961fffc6d1e3855c6ebe58f3db41
918d88b2ae0009e0cea949f0224686fd1305f97a
75376 F20101207_AABDQD kadirvel_k_Page_088.jpg
081d875efbab60b38c4d2a7fb0ec04c9
47d55a1abbb0afb00566883948621c55bf9da107
7975 F20101207_AABETG kadirvel_k_Page_133thm.jpg
e7d38259c5c50722bff2ace4ccbb2738
e0a0b4fe3fa0bc7dde7f58308796d6db96a1ed32
32498 F20101207_AABESS kadirvel_k_Page_165.QC.jpg
7fbe1da17e2ef09bf214d2197ea49154
dafb55c4a82b7cf03d2d07078dc531a581c111bb
75110 F20101207_AABDPQ kadirvel_k_Page_073.jpg
cba8307c597aae78e2f5377c9d7215dc
5bcbb17cef2494dd8c4530c501e90e9f6f4e58c7
89052 F20101207_AABDQE kadirvel_k_Page_089.jpg
7ee3f77d302e4f0e3e81f844df067e48
1efc51d56fff7c4b8e64b077a689dd77cb2f5f5c
7204 F20101207_AABETH kadirvel_k_Page_012thm.jpg
f5813a9a2a1bfc816052ede59f453ecc
6fabe2381861effbd6c889935102cf1f049e0cb6
21020 F20101207_AABEST kadirvel_k_Page_036.QC.jpg
691a15db5e34f4a46564da1346f43b64
7c4b7e3d5f08847769eda205e1db7a46ba1a9ee7
68143 F20101207_AABDPR kadirvel_k_Page_074.jpg
bd9cafb29c516e480ea63e0cda47033c
fa1e84f55278b29012f19d979a6b7633603c4194
108504 F20101207_AABDQF kadirvel_k_Page_091.jpg
d2fdddef6d243e37eb1c4f6c60c94e50
40e5e46ec98b5e90b9e1606d0e10ddbed3863744
22119 F20101207_AABETI kadirvel_k_Page_031.QC.jpg
fed047d0fe27062bb02384285d2bf09a
e4b4fc0bf59ac43cb4ba6a6539b588b011b9b381
34086 F20101207_AABESU kadirvel_k_Page_164.QC.jpg
e4482aa503751132329070ec2bbfa391
f9b5689716e3ea8bd039fa8112ad06648e9ef2c0
78034 F20101207_AABDPS kadirvel_k_Page_075.jpg
79fed762c4fbd368e9619c7d078df9e3
eeb82b533148db95e555103665a1b7eb8b1a3722
72063 F20101207_AABDQG kadirvel_k_Page_092.jpg
ef170690a9c8c82f82d485d0d9ff4773
def8edc5d27ee51e45e534c62a8d5cb3859e2994
34100 F20101207_AABETJ kadirvel_k_Page_090.QC.jpg
fa074b587a731d8d10de6f599189f6fd
446f6e5f63c2e2be20eff9a2aebca84e6964e2cd
2925 F20101207_AABESV kadirvel_k_Page_149thm.jpg
dbd4a15b2f9937aace1be0764f7e0b8f
4c1162b49e946ba805ae09ecb875b41e438682be
54311 F20101207_AABDPT kadirvel_k_Page_076.jpg
d5cc0bf8a1af9fd3a8168b86f03d71ed
a0d2a529668d8d500ff704d4007716c57b343a30
96283 F20101207_AABDQH kadirvel_k_Page_093.jpg
9ba63e3c264ad8cc79385e408681c7ef
39ddfb766c451fbe6a1a1783370fa8e07731196d
6069 F20101207_AABETK kadirvel_k_Page_097thm.jpg
1a76f4dc97e4f60c738bd94628de57bf
46f22087113fcc046bf5a61dc3057f03ce7c9213
7109 F20101207_AABESW kadirvel_k_Page_161.QC.jpg
09e0655bc3ac30b65c288755c649882c
9e47cf5339947bb8bc48e73cdc47320425d0569c
57695 F20101207_AABDPU kadirvel_k_Page_077.jpg
98abb730aae99c3a780f70a60b943280
b285ca5d26e31e1db19b30600f1e634fba03a2ab
100663 F20101207_AABDQI kadirvel_k_Page_094.jpg
dcde980bc00f8cd540259845f70a97c2
e786eaa913fa7c3138db88863647e6a8a591766b
5349 F20101207_AABETL kadirvel_k_Page_157thm.jpg
08154116120c1ded867707e0e8efc49c
1c305313b3b39473a2adb469a9972b1c30334064
17981 F20101207_AABESX kadirvel_k_Page_155.QC.jpg
7a26c81f2a7ff6da65434a4344e9bc23
4cf2fc5bed2c853244e19079409e90563c22ae78
67244 F20101207_AABDPV kadirvel_k_Page_078.jpg
7590471ea463ebcc4e3730976877d15b
b4e98a1b2264ad270012b85ccfe7e29af3da3de0
72391 F20101207_AABDQJ kadirvel_k_Page_095.jpg
ce2b4b64ee3340e27e0554a8fe7279db
9a3d244c4bdace9260eb7c846dae7e6a74d201ba
5339 F20101207_AABEUA kadirvel_k_Page_135thm.jpg
d931f94ff5b769d39652a210b615d8d6
08718a9d5c8c56a7ab7e66f6adcc7cf6bb2a845e
21822 F20101207_AABETM kadirvel_k_Page_115.QC.jpg
eaa3f094153e3cc54c807296abc52dc3
69fe9e61a32b5233c99659479a8fe252a6de5593
6143 F20101207_AABESY kadirvel_k_Page_009.QC.jpg
b33e3b4748d3a339e7e7292ff6aed3cc
8485b671653dd0313dd36a06a80b9eb934c5b067
96672 F20101207_AABDPW kadirvel_k_Page_079.jpg
42067af1b3ef59786a06781079e5c5bf
6d38e5b7ddfe84a5cefdc35a33329e545aea156b
61284 F20101207_AABDQK kadirvel_k_Page_096.jpg
dd7c318d91c6f954bfe3eac02b103fb8
430826906db29d1eeb3d21ee763b58ae77b6ce6e
6301 F20101207_AABEUB kadirvel_k_Page_142thm.jpg
f8508acad9f746ec3df455178d32504b
62a5a31e667fe6660f2ab5b21c2b61e790c8b9c5
6298 F20101207_AABETN kadirvel_k_Page_031thm.jpg
4acbe01b904a64bebe5b40e4fb44ffa8
7319b0f05dc4d3a470e70db9daa490ec001f41d2
14484 F20101207_AABESZ kadirvel_k_Page_034.QC.jpg
65cd65c3a9ea35b6337aefd42692a07f
eafc15bd0ee91c6715812c211ac00fafd9ebece7
87637 F20101207_AABDPX kadirvel_k_Page_080.jpg
6aa13f0784683915a949c47a723f3d67
9002f365ccf57e8154a4a5547acccf375cd99234
69067 F20101207_AABDQL kadirvel_k_Page_097.jpg
e6423841fc301779b1e8b7045e33806e
0ca48e799ac264694b6580b5dea3c2dee466d743
8515 F20101207_AABEUC kadirvel_k_Page_169thm.jpg
d27aebf493cfe00e67a00b2606465878
b85acc56f34599d4f1baed31e9b8a25a5e79f271
7772 F20101207_AABETO kadirvel_k_Page_072thm.jpg
719f415ada5901c4cee1ae2249bb158d
8cab0fa1065f85133abc753e253c8241fa9dc6d2
70552 F20101207_AABDPY kadirvel_k_Page_081.jpg
ff1ee25a7915bdcc67d85be443a9d75c
ee95df9c3776112bcb1093c1fd8350a1ddfe8ccb
98973 F20101207_AABDRA kadirvel_k_Page_116.jpg
5919053847412cd84755f39aae8c42b3
3f71a0413b6e26a35662036f1a58edf4de0a3610
82961 F20101207_AABDQM kadirvel_k_Page_098.jpg
f055e82f5e21773f542fa5e87682600b
a624a1d2bd5a33c5d338723e0736e549016fa855
24449 F20101207_AABEUD kadirvel_k_Page_075.QC.jpg
80f769f1e2436172ee46f16d87826eb5
c6af253e3218dfe4a6563b0d25650f5804a86cf2
5756 F20101207_AABETP kadirvel_k_Page_147thm.jpg
2feab501a9fb5bb66b70f5cd2c8dc13a
238f68fdf09cd9b4f047174e1ca66f382e360a1a
66886 F20101207_AABDRB kadirvel_k_Page_117.jpg
7f64ce939ff458de3ecf8676626c1a2a
5bcf71cc09a58a26c4ced5b3764a26c5d702dacf
67378 F20101207_AABDQN kadirvel_k_Page_101.jpg
0d6e710d362dd3d0ca705f3a4d120624
163b572e19afb119eac7c79767c2993880feacfd
8216 F20101207_AABEUE kadirvel_k_Page_069thm.jpg
59ef50c323c12c5534b3c249d264489c
57a5b6a08697af4fb30b3b2a392d4fdb7e139130
26209 F20101207_AABETQ kadirvel_k_Page_089.QC.jpg
9a9ff32a44d9d3fcba3af819aad9c7e6
644a5a96a2e60819b1189cb58b6c44b0500ef3f2
73323 F20101207_AABDPZ kadirvel_k_Page_083.jpg
dbbbc48310c71528d21b867941d09795
196529a281f7f2f078e43e722521c50d2820b997
59218 F20101207_AABDRC kadirvel_k_Page_119.jpg
b6baf84eb61b141e353841bbab89e65d
3367ec35070ddf2ea195cff3b543661582b3dc79
72640 F20101207_AABDQO kadirvel_k_Page_102.jpg
e6cdaeb90cde804fc932357dd6410234
f8265d5c25280f8fe0c8f4735e86c3edef85b5cf
29607 F20101207_AABFAA kadirvel_k_Page_052.QC.jpg
38dd80650c674391c68272ea178b6438
a61527242581551ee904f8936d628944421b2fb3
26513 F20101207_AABEUF kadirvel_k_Page_068.QC.jpg
a938bfa37fffb147b451db5cfea7c656
dda24801684dd309e49670338373620d18e0288f
7630 F20101207_AABETR kadirvel_k_Page_056thm.jpg
cc85e1d98aa8bfa63ce72eaa10b36e40
2e194bacc1bc3d596b8dbb4476c51e6758670613
65255 F20101207_AABDRD kadirvel_k_Page_123.jpg
36304c53f6bd3dc2213c7874157429be
2ee672d320040b1c6115052b073a14a49f78172b
67281 F20101207_AABDQP kadirvel_k_Page_104.jpg
093ba0a28b0217b6ba9587e679af22dc
e3b459b8f2787eaf8c9b64faaf942a09290607a9
28810 F20101207_AABFAB kadirvel_k_Page_056.QC.jpg
f03236779045921b4006e14fec921d1b
84302acf7efb5a66f4b5e464d75dc3d01e5569c0
4933 F20101207_AABEUG kadirvel_k_Page_022thm.jpg
6a5bbeafe1e278634ee010b044225f0e
77ccf4f9f15ab2a534bb9f20dad19d9f5faf541e
6981 F20101207_AABETS kadirvel_k_Page_039thm.jpg
7ab52520ef50e4562427ee91bf620857
2c1c06616de371b052c178351aa88d840d4882f7
64661 F20101207_AABDRE kadirvel_k_Page_124.jpg
1fcf28e7d092a212cb560ea622977091
ce094638869bfd389084f4dbdeaab80b19755304
67415 F20101207_AABDQQ kadirvel_k_Page_105.jpg
cede01e73b3d95ec8c118783290b4cb6
5b23e485903265bd51d6eed0e28648342d988b6c
21250 F20101207_AABFAC kadirvel_k_Page_058.QC.jpg
12130aa8f2549173a082abe96b643e0c
37c25f0c1bc2b1f1c8c835d421e1cae65cf21b34
6276 F20101207_AABEUH kadirvel_k_Page_114thm.jpg
ea75b507b0c25b24e0f01634b0fb7f1c
6a8ee7bf995b8672345e52e5d34cf6b06f4352b2
5200 F20101207_AABETT kadirvel_k_Page_159thm.jpg
251b4e5bde49eee88c004117f371af8f
4ab5cd88759c376156169c2fc933a52712448192
70293 F20101207_AABDRF kadirvel_k_Page_125.jpg
44f53ddf3d77b0478927947923e8b9d2
0eb079f6fc74dba052fb15ebdda8191c20002f82
59118 F20101207_AABDQR kadirvel_k_Page_106.jpg
1d267ac10fe81246ae36a8513c3fa4bd
274a7030cbef59dc2a77c597b453edc4df00b836
F20101207_AABFAD kadirvel_k_Page_060thm.jpg
5484d644849d0e914cec92f1a10b3a69
eb50ae6e2d4e045e45fd14e95c56de8b6fd2d3dd
7076 F20101207_AABEUI kadirvel_k_Page_144thm.jpg
604313e8b348b0e0d7a4c08cc2634416
6c615b5053ea541ddf2cd16cd649e0b03ded0417
21139 F20101207_AABETU kadirvel_k_Page_043.QC.jpg
d7be36dabdba95da9e46bf77067a0ee0
1b6e6be138488bda1293b8f2997c3baa671cf29b
83077 F20101207_AABDRG kadirvel_k_Page_126.jpg
4e7e7e1fc5cf4167f116776f3d91925c
e1e2b9bba26eb474e28e46d96822ef93c187f06c
48362 F20101207_AABDQS kadirvel_k_Page_107.jpg
cebc1e6d5943c78bf301b6c3c3175e11
c028165f731c182a873e42784fdfcd8c569adafe
8284 F20101207_AABFAE kadirvel_k_Page_064thm.jpg
8946bb813b8555544820aae28ba25918
ef5b16129828d9bb687696594d1dc166b757b2c1
6088 F20101207_AABEUJ kadirvel_k_Page_029thm.jpg
bed144e5a0ce5f919f0aefc8493bce82
1aeaa07b8fa6d0afe3b0e7f799d840fce4b85453
20793 F20101207_AABETV kadirvel_k_Page_055.QC.jpg
8473390abf3e56ebf27032a669edffab
fe7fcdfdcaf875f7f5f1dffe7a3be1113bc91eb3
61156 F20101207_AABDRH kadirvel_k_Page_127.jpg
a925d431188f2fd44cab97e59137b306
ece9e1ff3a6e7d2b35bbd2829b49e62770c0002f
72915 F20101207_AABDQT kadirvel_k_Page_108.jpg
cd43704146eb3601f5a43c45653397e3
e23d63b83623e1c5cf4c20cf2064e9695dfb66a3
24444 F20101207_AABFAF kadirvel_k_Page_065.QC.jpg
c571a0d5b6bc49d5930e5f18367e5a98
b0b1903d772d8e6301fd0144d6ca5e4a114a38c7
26594 F20101207_AABEUK kadirvel_k_Page_112.QC.jpg
4bc696bfd0bbc44dec2787f0cff8a82a
c14f8889e65dbf1574391c40778911a3ea48b241
24249 F20101207_AABETW kadirvel_k_Page_139.QC.jpg
445d0f882f993e6e525d0fda5dc3896a
b872c07af39d5c34965f68d82bb0f748422261a4
80571 F20101207_AABDRI kadirvel_k_Page_128.jpg
fa63e7da75026db879d89a51efb167d6
bb0ce14d3716aa491de1ef002c4da362f9e99e9f
53863 F20101207_AABDQU kadirvel_k_Page_110.jpg
6f82d6b2b9264273f9d6b7576a624512
892072c84949d69ef92c040c838d165811ccd935
5300 F20101207_AABFAG kadirvel_k_Page_076thm.jpg
b5c928db5df704768f7c9ce611e8f5b7
1d87b76c811aef4dfac00a1f7f075f699fd97db1
16725 F20101207_AABEUL kadirvel_k_Page_076.QC.jpg
11c0e0865be83cc1f6d5caa3ffe2d804
94c58c156341cc66d9fa6fe0ee206c3a6a8e0372
8178 F20101207_AABETX kadirvel_k_Page_168thm.jpg
e2448d2df63f884cf14e65fd943eb2df
c01618b7cecface41dd2099612f059b8dbbaba6f
61605 F20101207_AABDRJ kadirvel_k_Page_129.jpg
3c66dddae5c03502038d6aa6b5eb4a43
4a5a1506cd0f451c043660cafa1c25169acc9204
77205 F20101207_AABDQV kadirvel_k_Page_111.jpg
2b4cda2296f7bdd75f4e5465039f4208
df17e917d8e7d70c1ebfc53e0113ee6a8364c9a7
5491 F20101207_AABFAH kadirvel_k_Page_077thm.jpg
c03359145f05d692e2c60f0ccf64c1b0
d34a2f6d039ff4c3d9497c4b8cd96742ac68a398
16964 F20101207_AABEVA kadirvel_k_Page_118.QC.jpg
42cb72211e625f139b7b7d2cec44d10a
45f0d782f909ca2944251332c725385938030939
20129 F20101207_AABEUM kadirvel_k_Page_121.QC.jpg
222d80f4d45c91b09dd03f1417706097
e2a32db4364536d03af160c3a6b01f2286d8f7b6
21550 F20101207_AABETY kadirvel_k_Page_092.QC.jpg
3825bb7bdd7af4021c5da6ec478f70f0
c7e50137e322a3cd135a454ed6eb584f93508f0d
66978 F20101207_AABDRK kadirvel_k_Page_131.jpg
4706cb2e0babe3fc86279cbf933dfc0f
2eff0805d8b861974b5ca903501a915b32ddefad
85370 F20101207_AABDQW kadirvel_k_Page_112.jpg
d134ba53028ef2dc9d9fa4a72cb90da4
dd064e9b2eb356e9b090a67a4c6712d57d80ed33
7314 F20101207_AABEVB kadirvel_k_Page_143thm.jpg
8fdbf9f04ce6a03df5371973fbfad932
f80547137c69c35597df127ba206864f013a89e0
5237 F20101207_AABEUN kadirvel_k_Page_155thm.jpg
b2b56400aaa5348aa81a4dc00e7ca5d8
aaec9d4ef735ebe2e51f53eb627f88fd91f8dcfe
21580 F20101207_AABETZ kadirvel_k_Page_050.QC.jpg
e004498dee2b21ea144107015366dc2a
a3c20a7498aac9dd25a80e031efcd145426fa619
66539 F20101207_AABDRL kadirvel_k_Page_132.jpg
ccaa5232d1348504499526b066e483f5
317b4907ebfb2f72fece5b5ec43c3f4722bc2c3f
68525 F20101207_AABDQX kadirvel_k_Page_113.jpg
131d78ba24ec587abb68e756b80daade
ce50c015c08f3f69aa0da6596e8dfceaed8bbfce
6133 F20101207_AABFAI kadirvel_k_Page_078thm.jpg
202ffe26ac6032b1e40c5749d61e8853
3368a84bedccba36506edc10372f8f5a55c6b0d4
6081 F20101207_AABEVC kadirvel_k_Page_092thm.jpg
820fc28121df75a80a5bb1e4a9442a1e
ea56eb0b95ed08522539ccd394e3529e9b472cfd
25488 F20101207_AABEUO kadirvel_k_Page_134.QC.jpg
c449fd22d534c29bb8332e3c88379062
3a4f909909f4e029ebf45be94bd346e11b06bfce
102041 F20101207_AABDSA kadirvel_k_Page_148.jpg
08c4fea422fdd6e016c881e84e2e8c5e
9031f06eb8173263fcb04e51f752907d97b9d836
99318 F20101207_AABDRM kadirvel_k_Page_133.jpg
43188d80a049e9d6b1a9864aa7fd040b
8ca89a008dab50811ef0ab75a0c96fad2988d5e8
78119 F20101207_AABDQY kadirvel_k_Page_114.jpg
d4a919d94677e63a9c011a4e9550bd14
b4df582ad7e54f909d15fe1cba79808b4fb1567a
7546 F20101207_AABFAJ kadirvel_k_Page_079thm.jpg
c54a56e71b9624258f7f30b114501273
ba16e9a86fb74e978c77f4c76e142ced37a29159
24611 F20101207_AABEVD kadirvel_k_Page_099.QC.jpg
64549864926f1a0801ed8ea62d56af9c
8853d3e86e5ad13e550476ef9fd0046da029952a
20436 F20101207_AABEUP kadirvel_k_Page_020.QC.jpg
bae4436cad65335203cad39eb6cf2646
579976d17a903e27e943887c7d656b06f9a82bad
27932 F20101207_AABDSB kadirvel_k_Page_149.jpg
5e4132fb3c9d8454c208b7e206144486
b1e52636d5b5cc46b7d0e358cce0ae2409c6e80f
84773 F20101207_AABDRN kadirvel_k_Page_134.jpg
26c5fb86d549601fa76ad14441c30f5d
157cdc09789c4b2b72b9b435f38a0f177bc1ed59
69919 F20101207_AABDQZ kadirvel_k_Page_115.jpg
7d7796dd8d7fe68de114d354a021f9d7
fd1146588fa89583764b0ebb1ec87a025cdaa2a3
31705 F20101207_AABFAK kadirvel_k_Page_085.QC.jpg
ce67dd6b68d7aa77dc9bc52439896202
1534a426bb85708ebcdcc5d8ce22d6b55a002ab0
16495 F20101207_AABEVE kadirvel_k_Page_070.QC.jpg
a220c343acb0def3315c482f9448fe66
7cf8581744cbe56ed98e56522c5171c91f38ae53
31518 F20101207_AABEUQ kadirvel_k_Page_171.QC.jpg
3e1fa06b7d62e8c1b5f1b96b35ead6cd
466a8e9da6a5425daa207ea0116ac6ce9a49b0eb
86545 F20101207_AABDSC kadirvel_k_Page_150.jpg
eb78fcba5093c68df93351ee15fb8aee
b799e24995a1a4447fba810d1e24e516bc06d8b8
62939 F20101207_AABDRO kadirvel_k_Page_135.jpg
e3041098b13682483169e3ac337c87c7
bcdb183872fa5af797dc4ef1ddef6ac05bee69f5
7266 F20101207_AABFBA kadirvel_k_Page_151thm.jpg
2f1e64e511ee181fecd0a1f0327f0dc7
a3c00419d935343de1cc55452583a1ed6a8bfb9b
7996 F20101207_AABFAL kadirvel_k_Page_085thm.jpg
59ae802eb5e7aac6ce87ba1ea831b7fa
791c3260442e144e83d9b0079119d931f740e455
20762 F20101207_AABEVF kadirvel_k_Page_137.QC.jpg
3ed6142f1dd6b1c8afb67831c7579fc3
49757de8c9be36a085d6e7396b5492873c0c502c
5499 F20101207_AABEUR kadirvel_k_Page_146thm.jpg
8c922c8468dcfcb2d8ca57b4c289176c
5d630b8699f8c037314ad0c56a0b9f6c8ad3b680
89953 F20101207_AABDSD kadirvel_k_Page_151.jpg
e587cf4412621aa762a6bbd07acb217b
cbea012b9bff340439389635bd886f2f9fd3f5eb
87122 F20101207_AABDRP kadirvel_k_Page_136.jpg
b37450560058b72feb7f02b1726dcbc0
aa30b56ea92447a65cf3f933d59c6c8ccd8a856f
4783 F20101207_AABFBB kadirvel_k_Page_153thm.jpg
c430d249817e505fa7504241fedb91e6
01e42aa6c45eb4759a533fe3d88d18198628ed7c
29853 F20101207_AABFAM kadirvel_k_Page_093.QC.jpg
e86489186132664d61181744b7187a85
e9ab476f6348ebd019ccfd0e04fc545fe73a5403
18553 F20101207_AABEVG kadirvel_k_Page_127.QC.jpg
27cb19bd838f5df616421387323073ee
d011fe1dd10fce7cc08bc6abe8d05ea0b20c701c
32452 F20101207_AABEUS kadirvel_k_Page_041.QC.jpg
362f5582bc1be3702f7db599ddd12a99
5749741e71305fce5612e08290548ca95b361998
51169 F20101207_AABDSE kadirvel_k_Page_152.jpg
7fdf8c85725b8e537ad029c98b1d2b57
f4c391c5f45789e4ecdcc731caf2dee8508ff471
68717 F20101207_AABDRQ kadirvel_k_Page_137.jpg
281bd5c994ff9c5d459d8b07bedb6c88
3abf82c9716b5d2cebecae908d461d43356551c3
8413 F20101207_AABFBC kadirvel_k_Page_167thm.jpg
6add9b4ffab01e5c0a7a41adbd7d9d12
69367526892cd679f22d2e90bb135601a7c18bd5
6207 F20101207_AABFAN kadirvel_k_Page_102thm.jpg
4b8f0284bd063d76ae22e7b66d2445c7
6074a78112c0c7cbcdf5c5e07e2380c52d0df22b
7192 F20101207_AABEVH kadirvel_k_Page_109thm.jpg
9cfc8981deaa5d3e6e8678ae7186c9bb
8ab30744886dc4a8cb2e0287dca2a703e2720c5f
7201 F20101207_AABEUT kadirvel_k_Page_098thm.jpg
4b9b7381194022bfee84dc63c106d9a0
671f0417cb0cd392d7a6e27ed6879b5decce1bca
47528 F20101207_AABDSF kadirvel_k_Page_153.jpg
a8f16f13fa348f3779ff43a2954599af
c951bbbe89d858a5f0e0d5b969c9371d9ee2f875
71304 F20101207_AABDRR kadirvel_k_Page_138.jpg
71dec7b2d2c136df616253e91e075322
2ca54f7e0c54acc6c4b9af688fe16d9fb360d4cf
4594 F20101207_AABFBD kadirvel_k_Page_172thm.jpg
5f110c0a8e86280cb27f43b63416a9c0
14d9a15e6b53919b4a2dc9d0dc26920aa7dad5cd
6472 F20101207_AABFAO kadirvel_k_Page_103thm.jpg
d60968177bad9847916364981c444e84
3aa54fe60bfaac4cb140574d4f8811ecdb24e2be
6653 F20101207_AABEVI kadirvel_k_Page_100thm.jpg
2dd9cceb6553ca5e96fa24d0b4323460
a2b45cb82680456af6197df2b17bc63a7c7c59ca
29198 F20101207_AABEUU kadirvel_k_Page_072.QC.jpg
cfcb507603d3809304c9ce4ded12bf2f
3158d4944d2ecc7e0799713a166346e361cbd31f
57348 F20101207_AABDSG kadirvel_k_Page_154.jpg
7e100bdb7206e8b537f9ad4ea3d2ff7f
d9d364c283394f86afd9347d3a913c9faa220e38
76836 F20101207_AABDRS kadirvel_k_Page_139.jpg
85def42701c19b35f7fa6db3077cb28d
e37576d09da31ba76b3212c69cb77d352712dc5f
21158 F20101207_AABFAP kadirvel_k_Page_104.QC.jpg
0eab99bf63d18ee6c5a01fa8809f8612
2bf543f6147ca9b0d62825946f08300dfc6df31b
4724 F20101207_AABEVJ kadirvel_k_Page_070thm.jpg
04e6a62e347b133e5a9a27d990b3f52b
d615ae74eb993dd1bf4686d2234a1e285501c93f
3240 F20101207_AABEUV kadirvel_k_Page_002.QC.jpg
675afc3f61e67bd53dc5d5f2418ec504
e50c7bc2265132a3c18dbd06cb23c1ad1d65b83c
56574 F20101207_AABDSH kadirvel_k_Page_155.jpg
be6fe6bcd3953bfdec000c9ddbae84b8
5c96fb64067365a9842a5e76856f1bed0973e367
80997 F20101207_AABDRT kadirvel_k_Page_140.jpg
a4d1974228e2e4e18948388583e468af
de55136d0db74d183826918e056088df67ba5b72
14571 F20101207_AABFAQ kadirvel_k_Page_107.QC.jpg
5d375c2d0e9b9a34b5a26781c382e135
3f95777870c7a52400333eae80f71631c6213db1
5686 F20101207_AABEVK kadirvel_k_Page_162thm.jpg
db01eac3dfee53197c61f85045e65fcc
0a684ebe3a38ab9805f3ab9b36dd58de6e8f9015
21508 F20101207_AABEUW kadirvel_k_Page_095.QC.jpg
4dab3e5c79f3fbcebc0a435f7ac4f0ba
ecf653b7617fbb2d873ad30759e0edf7a90e99fb
54191 F20101207_AABDSI kadirvel_k_Page_157.jpg
26cb04ff469a5638f63708aa53af7f52
c7cc1a5d63fb04f31648604d5dda7a17616a907e
85814 F20101207_AABDRU kadirvel_k_Page_141.jpg
7f8a4a470daffb494d2b5b5cc345e539
86b649645e1abf46b4b3bc45b0449ae6df7a8426
6107 F20101207_AABFAR kadirvel_k_Page_113thm.jpg
b89facafd55957028835689a399ec865
018b157937af5993b5808660851ffb6db1f9a5cf
4059 F20101207_AABEVL kadirvel_k_Page_160thm.jpg
0ca53b13c80532b67fdfe812799514ec
e9a9e818022286ec7d7d0649ca2f295773679a78
13104 F20101207_AABEUX kadirvel_k_Page_160.QC.jpg
69a367db55c55731a1bfd9ef439c9421
9f45a299dcb9ee0332f4246f8a36a6b75a41e446
71959 F20101207_AABDSJ kadirvel_k_Page_158.jpg
d2ba19af4e3cbf808d4d642baad38dd6
679de40da8ac73133e727f74dae1bd8c162e02f2
69818 F20101207_AABDRV kadirvel_k_Page_142.jpg
3ccdc527f66cea1b43ffac95e81cabd8
002c1c1996df9bdbcdb33c7a507aa51ca9b0a7e5
21542 F20101207_AABFAS kadirvel_k_Page_117.QC.jpg
f286ac6cc14e1d08d52c5cdb37ef5c7b
e5b40dbb507de955197bf77dcecc398edceac7b1
31164 F20101207_AABEWA kadirvel_k_Page_047.QC.jpg
4750b4acead41354a7566a888333b7d7
5e845136b93c01ee040a8db58149878e1aa992b0
21603 F20101207_AABEVM kadirvel_k_Page_097.QC.jpg
7be3ba9ea5182bcc9ceeebd9c74797fb
88ffb53649da103727fb04851b950a4eb1791176
8073 F20101207_AABEUY kadirvel_k_Page_011thm.jpg
77cb69039939adff85bf3da78dab7823
9a46e7c83c28132b887ca279a64d6780f815ea10
50549 F20101207_AABDSK kadirvel_k_Page_159.jpg
8eeddc79ef2115fac8b54af81392e17b
5ec93b1b4bf5d61a92fb15c2f66cb397f84bd59f
91969 F20101207_AABDRW kadirvel_k_Page_143.jpg
b65c0437cdf9a5e5783f410d0e1fee97
7d454231ef382fa8a61bf7c4b52d445f87b81385
18546 F20101207_AABFAT kadirvel_k_Page_119.QC.jpg
2974ef8c2a770e7abd4a5513f4c9a01b
18590867c799cc3b4288a0762be68bf65636ed60
6058 F20101207_AABEWB kadirvel_k_Page_074thm.jpg
c2b60de74618c1b1a158905ef80a63d7
28238884337325d1641d820f98a8c9f3e6bb87d6
8849 F20101207_AABEVN kadirvel_k_Page_042thm.jpg
9248a53759557d7de66e0d7e264190c6
1cea4d87c4695971a8b5bf844d822a93183cdcf8
32358 F20101207_AABEUZ kadirvel_k_Page_008.QC.jpg
1b91969d493471ced42b35153549e917
5aaab9630c2b22588ddcf02bd66849d3fca024af
41032 F20101207_AABDSL kadirvel_k_Page_160.jpg
0c9f29625a92f4d3533d38f1fd1c8bdb
c4b1dc494d4ddc27e428e4bc486bab4c0119c7df
90008 F20101207_AABDRX kadirvel_k_Page_144.jpg
277ce72cefab64f716145694cfeb290b
9fdef11228ebe66c7d3ed7be2274410b48726fab
25836 F20101207_AABFAU kadirvel_k_Page_120.QC.jpg
5460407e2aa6cdbc0147a79870134681
13bfd584b72f66829cc260e4c25ec2683ea4f0f5
5668 F20101207_AABEWC kadirvel_k_Page_020thm.jpg
99f07a6f884634189683af04a5506aa4
9ddce34d803578fb7d8fb77a0f95859564b79cdb
25605 F20101207_AABEVO kadirvel_k_Page_019.QC.jpg
4b478202503a1f9c6d2becc656f09e91
b1556dc3fc77de22025fd175c062be193729667f
21788 F20101207_AABDSM kadirvel_k_Page_161.jpg
117f5b8862b363e47294198ffbcee172
4132ececfdc1eb12443a3f46e2bbdaf3a1aa1d1a
84413 F20101207_AABDRY kadirvel_k_Page_145.jpg
320592c0edd0ebe07b27412475ca9d94
2b57830a5ebcb4be76a686dbed0a72a259d2ccb4
708423 F20101207_AABDTA kadirvel_k_Page_007.jp2
cdc2030658e1f122ff2f135d53a08e40
65e95f64ee2d431cfdcd033f1e4e4fbafcf8196e
25616 F20101207_AABFAV kadirvel_k_Page_126.QC.jpg
2afd9e1c059308edfa546da669c9c408
598bbe1e1f5bcf1145e2c75c1947eea25baf11cc
8074 F20101207_AABEWD kadirvel_k_Page_061thm.jpg
a51ff75b1ad2304de999100872a6e155
0455a7a539ea7d2a4423bf916c3e0ff35a7d24f8
6757 F20101207_AABEVP kadirvel_k_Page_063thm.jpg
8c96fb83fea2b82b9d0a73ea0c66ead5
82bb4738e030f347cae849f9d0467c668e4d36f3
61552 F20101207_AABDSN kadirvel_k_Page_162.jpg
1157f488142bc72817e4c5027c87c29e
408fb51ef820674942e26f9f0ee2e96e39fe9b22
68021 F20101207_AABDRZ kadirvel_k_Page_146.jpg
be2c173e687e97d88d5d0f377e9b17c6
456ed0d986a805ea89dd341c7f414fe7cc546155
1087890 F20101207_AABDTB kadirvel_k_Page_008.jp2
435505b58767f20aa2025f12fb432d82
3321ce4a07d45fc5f2c67410c76c8dd5da52c61c
5797 F20101207_AABFAW kadirvel_k_Page_129thm.jpg
7d60cb1fb51d75a447ec75080092e9bc
94506f4558038eefc4c6eb3e8e5875f9093f3e6f
4012 F20101207_AABEWE kadirvel_k_Page_053thm.jpg
5cdfccde78f89f6c3e35465cdf71df01
29b1a7b27680ec63d140c0af43197ebb5398bb52
21071 F20101207_AABEVQ kadirvel_k_Page_122.QC.jpg
147604894426e9ef6ead54e7c87443c4
5923cf8bab30f98d99d7a3a1fa1702c3d60f8e60
118834 F20101207_AABDSO kadirvel_k_Page_164.jpg
45797feb3fa3a86859390316475ce9e7
55ff3c028fd9a1875ab49187959e468279e33018
346089 F20101207_AABDTC kadirvel_k_Page_009.jp2
52452183df026d868172641728605591
d7a10deb8e40f21b8191278a5657c529a9e3a3e9
25730 F20101207_AABFAX kadirvel_k_Page_140.QC.jpg
a96e8ed737bcefe5d720005265c0b3e9
ab0eab26de8b6907cffba3afdcd7c628ce82d547
6028 F20101207_AABEWF kadirvel_k_Page_058thm.jpg
ee21fc227d7e0e65949f8b8441f8718a
cf09d0690e08f72facf49022dd1bf1c76351b295
6899 F20101207_AABEVR kadirvel_k_Page_140thm.jpg
a97c3d66651ddbed22f9d10b5afac718
12fb037483fd75dde7a0f58a5cb8643adc02de7e
106435 F20101207_AABDSP kadirvel_k_Page_165.jpg
55fd7cd1999362c35ee59f134f325505
1959f9a65422d1aa9e30eee17e2be78b54af56d8
1087883 F20101207_AABDTD kadirvel_k_Page_010.jp2
1b750d55af1c7d7cffb3c391a59ab907
234b5ff9bb0ed15394133435f9ac363ae4f83db8
31929 F20101207_AABFAY kadirvel_k_Page_148.QC.jpg
c433dbdfe6df773ae97539d369b83232
0517b9b330e151a824316870fc48f2366bd0d2b8
26137 F20101207_AABEWG kadirvel_k_Page_109.QC.jpg
671650c9e62f462635f2c9d0649dcfe8
955ec5bc5832fe9faaebb61b0c0ec8ce59842b7f
7926 F20101207_AABEVS kadirvel_k_Page_001.QC.jpg
ed61dc210143ae376dc497efff84db22
06f6a2fe9d89a3b876d83dab3e47c178a7bba76b
104645 F20101207_AABDSQ kadirvel_k_Page_166.jpg
2d30b86b943ce79bf10b60d2151954bd
43a852c15b6b702e50df8ad0dd720be1d40a82c0
1087897 F20101207_AABDTE kadirvel_k_Page_011.jp2
b8689fd12012375370c25572e002d0cd
865175adf9a78a8e196235f22c5020c230657e16
27982 F20101207_AABFAZ kadirvel_k_Page_150.QC.jpg
a73a3db1fe3010e4dacbb996f0c0ab4c
f3f52b93595101a482b2fce2b77017e89f706eb3
30653 F20101207_AABEWH kadirvel_k_Page_094.QC.jpg
6409a7d74291813c0f863ec2151eb098
977579232a76bd82751ad7fc256ce9998aa00ef4
4751 F20101207_AABEVT kadirvel_k_Page_066thm.jpg
89ba2325155d3f524c085bfa1cebfb25
12120a01900ebe2aecdd31fb742f404532fcda2f
126371 F20101207_AABDSR kadirvel_k_Page_167.jpg
d50b8d42154c37723224ef46e134b87b
7b14799b57b43324544588e04234e88adb67a453
1087846 F20101207_AABDTF kadirvel_k_Page_012.jp2
bce7406663a73747d86b017fa7dfeb5d
c3be346ea9c4174761523c54c693c5d8b2769db9
6321 F20101207_AABEWI kadirvel_k_Page_057thm.jpg
d02c7eab9dd8ffcd58267a67cffdb185
812760b0826bb99a972b483df2201511b42776dd
14193 F20101207_AABEVU kadirvel_k_Page_053.QC.jpg
ad961d28f052969b40397e6fe430cd76
6daf3be955fdfa3962a3cd8c3be1b52d8e18653c
102289 F20101207_AABDSS kadirvel_k_Page_168.jpg
8d3ac845ed6e42456163374fc259a5ef
793822b50b630308cf42e700804a01e8e28f9b5b
1087896 F20101207_AABDTG kadirvel_k_Page_013.jp2
21372b6beec94e1c914b809cdf14c921
ed32a4d5daaded6184d49fd2bdd313bf0949de12
18804 F20101207_AABEWJ kadirvel_k_Page_146.QC.jpg
c71b4f0cfc7cd801cceb025eea21c6cc
61ea797e9edf69327ef16dddba182bd04d629f2c
6777 F20101207_AABEVV kadirvel_k_Page_084thm.jpg
9aa745487f3a5814b865fc29df08d1a6
14a014f73dcbedc0613c4e51d2771ce274d26208
122823 F20101207_AABDTH kadirvel_k_Page_014.jp2
5f87b4510678b4a1087753e719c71d06
7f8a323c7afeee9a450a71c13196e6d53f05683c
102409 F20101207_AABDST kadirvel_k_Page_169.jpg
8c23ff74f0cbdd1809ff195f9b447ecc
a720f112b65cd8b1c550df6fc98ed036c90066a4
22836 F20101207_AABEWK kadirvel_k_Page_158.QC.jpg
233458f7f4a4b119d8bbef7391609670
746c124ce29727cbda3bc75fad92e754dccf7eb6
6377 F20101207_AABEVW kadirvel_k_Page_158thm.jpg
6dcb008efd90ecf1707f272eab632ed0
cdaa5cce7434d36590561f4bee0c88c23cfaff4d
12566 F20101207_AABDTI kadirvel_k_Page_015.jp2
6fc54c7b32122978023c1c406da03640
1b483aebd5408d07a3a5d44ec0c9295066a2349e
106261 F20101207_AABDSU kadirvel_k_Page_170.jpg
d63bdfaf4f89f6f0f5de6ddd10ab4295
348c78c44ba13fd767ea9141df027f7f69bb4a20
6034 F20101207_AABEWL kadirvel_k_Page_036thm.jpg
371821c35c1ef4b80b59ca4675a66d9c
2a349c12df88276efed22a446dd985b3fe6b1a45
31221 F20101207_AABEVX kadirvel_k_Page_133.QC.jpg
7005f94765471def4eb57712082d1b7d
80add0c7e292ba2ddf78d79de29c341c04febe2e
1087899 F20101207_AABDTJ kadirvel_k_Page_016.jp2
031490d23eff936fc1373ac286a7a164
2a79c86af55a37a8e79a55d1d870670ca47599b3
109452 F20101207_AABDSV kadirvel_k_Page_171.jpg
ec9ac47433075c9fa6e2cfc4298b4e9a
b904f01349bae6682e816ab836de3ef296011591
22532 F20101207_AABEXA kadirvel_k_Page_111.QC.jpg
efef29936de7296e3e3848722aac9c3f
05c5038b87e475567fb0196a1cdf5ed7bd32c122
3051 F20101207_AABEWM kadirvel_k_Page_173thm.jpg
4d83bcc2e2f73e5ce088c5ffa053b8d8
4f8737f68e90f9362da5feb975be032e60f8673e
7461 F20101207_AABEVY kadirvel_k_Page_030thm.jpg
72a4f618ffe2a8ea0325c2b9df8365fd
4495182edf7de8b334ffd7b52bd240803175df41
1087892 F20101207_AABDTK kadirvel_k_Page_018.jp2
669ab4a47d6dd3217ee4e9df48cb08ab
0d4cf16d0a902d621a09385fa84ca1c6511c2b6a
58175 F20101207_AABDSW kadirvel_k_Page_172.jpg
15a74146d159ce21531c836dbc7bb4ae
b2c61a0d2fb60ddb0c9b42514aeb17a2c562fe32
6251 F20101207_AABEXB kadirvel_k_Page_117thm.jpg
4458a4c32e40ac338c9d624d079e3328
e7461712aa97538b05ef551d0a6d1d4588b6929a
18245 F20101207_AABEWN kadirvel_k_Page_106.QC.jpg
d5073ab6aca94cafb09067f304e20fc9
4bd1464e9bcaa0325e2794c2b377bbe3e7244fd4
21605 F20101207_AABEVZ kadirvel_k_Page_123.QC.jpg
70aab1bcc7db530ee4ecb94361d3b2bc
fe9031c59336a4decc877a789efd30964a229f77
1087886 F20101207_AABDTL kadirvel_k_Page_019.jp2
c67b64b09741a4b21d1602df009c2748
6ea875c9c99e43bfb485420f38ac73656d325504
27359 F20101207_AABDSX kadirvel_k_Page_001.jp2
fc33df8fc6d761f7170fd8d6df1110d6
587591876b07f1407847ac75686f2822de5c1516
32977 F20101207_AABEXC kadirvel_k_Page_086.QC.jpg
7bbc4703b9ea24a2a07778a3abf34a81
cbf6da600ac9145cf94fa707d6048d490cd85e18
5440 F20101207_AABEWO kadirvel_k_Page_096thm.jpg
0f74bca9d2ff62a79c646c992df5ba11
30088d2b07a25730665742f356e485ecc38fc4eb
823397 F20101207_AABDUA kadirvel_k_Page_036.jp2
449f8c28a2421ddb22dfc69b414d4ce6
ea220db65617513e2f8bb694a3c3598ad4702239
929336 F20101207_AABDTM kadirvel_k_Page_021.jp2
c0e546d37941b5f6debfeeb3b81c389c
882f50a8bd79cad4dd94e49f7feec3980494792f
94592 F20101207_AABDSY kadirvel_k_Page_004.jp2
830c5513d310f6ee212b57a479ddb374
5b31a0ba7962005fa7ed10b90c1e565a40437a04
2468 F20101207_AABEXD kadirvel_k_Page_161thm.jpg
6e27e179b3fddf7693ea92f1a0cfd039
ea18220ea728ca131edbb5696cec900f7af22881
14595 F20101207_AABEWP kadirvel_k_Page_066.QC.jpg
2f32e904d81e5c6b3eb4c04fc8d3f878
97f2fc4dfc07ba0f5654dddd97b0b1f7b1ea8403
944924 F20101207_AABDUB kadirvel_k_Page_037.jp2
9b98d09b21e259ca48a35b3cc5775178
949c565aeef918bb9a7af70672554b32e7636272
82080 F20101207_AABDTN kadirvel_k_Page_022.jp2
1b713b3f8b7706ff9b89a16a92697afe
099ebd7dea6e4ae8bc3920271511e4c9c623448c
1087900 F20101207_AABDSZ kadirvel_k_Page_006.jp2
fecfd5991d57669a8c5fbac8ac698f6e
fe1c92ce293f594216fce8c75d2d8d4303df624a
5338 F20101207_AABEXE kadirvel_k_Page_118thm.jpg
62008ec147d15a867fdeb9fbd2bea0fe
6d0164acdc9b74d35e823092c96d85faed72cd89
F20101207_AABEWQ kadirvel_k_Page_078.QC.jpg
4ddca6e84986c95e8bf994c1fdef201f
61d47d442ef0c542c99894c9f67283816e4b0f36
1072140 F20101207_AABDUC kadirvel_k_Page_038.jp2
6afe726509678530e4eb1e6f0c6f1e4d
9c29f865f46e938b6935efb6c57235f8a7e184ae
909303 F20101207_AABDTO kadirvel_k_Page_023.jp2
d1ff6a3a03d3570797bce981319dc250
01b5161eb3137e01d2976f3c62b14988adada013
27964 F20101207_AABEXF kadirvel_k_Page_005.QC.jpg
d4a82c3a3cb4eddce3726667f3cc59b1
8238976d42066d8bd637289ece708a1953940bbf
5726 F20101207_AABEWR kadirvel_k_Page_055thm.jpg
2c7edd7f6059fc3faeda9cae03d0e1be
cf7f19c9e9927020b5f95491dbb21f703391a008
1087881 F20101207_AABDUD kadirvel_k_Page_039.jp2
92afa6e1b5baeda09eca4e4551a7d7d0
98b7506efebe9ca0dceb1c8e5a6aa48beb858964
790394 F20101207_AABDTP kadirvel_k_Page_024.jp2
94d3beb9c2827b0e5dae6af4aa18bedc
21309f3adb9a76419f934a81985683877256ef9b
26526 F20101207_AABEXG kadirvel_k_Page_027.QC.jpg
bee2dd3f2813a064863a36ec4e5f22cb
a423e307688aa2d8680b617844aacd9a94978e7d
6696 F20101207_AABEWS kadirvel_k_Page_065thm.jpg
02ed668cd3f309526e0436be3786962f
5ae65189ed38364cb5a23296fac032455e940723
987395 F20101207_AABDUE kadirvel_k_Page_040.jp2
912238043e427370eb377f1a4af80e05
6b2209b5c807d57404882e23fb9489b2c0c22aca
725867 F20101207_AABDTQ kadirvel_k_Page_025.jp2
b5eb3192d0922d2d3624c074f805ea1c
dc9c0a2f9631724452712dc4866889d87cd2459e
31514 F20101207_AABEXH kadirvel_k_Page_168.QC.jpg
c961d6d8f17b234f3ad955e99c0323e3
590214205e84b440ee7c34ea1bbf80ddc2dd4629
22200 F20101207_AABEWT kadirvel_k_Page_044.QC.jpg
7f0f62dc8c5a12f2bd2a9a0409fb5d5b
3bf1a0e5154505029e8f7f58a5bed21cae4c4e22
1087901 F20101207_AABDUF kadirvel_k_Page_041.jp2
9868f0fbef5721a47a49e23c95aaf396
fbc7434b36b98da0a130a92b3fa7e0d74d054742
1087893 F20101207_AABDTR kadirvel_k_Page_026.jp2
a8163c39f907b0dee82393cbad774160
fd3aaac915e5c7a6b39c88df897adc3aac1a6013
F20101207_AABEAA kadirvel_k_Page_035.tif
8f1b752e7b088b95c6bbb14a11f9c89f
23d0aa2a411be9e73ced3afd1b51b4110704918b
16649 F20101207_AABEXI kadirvel_k_Page_059.QC.jpg
b532b0c67d4beaf690ca8dd4558c5590
e89a52552931914c062eb8e64068bb1adaf83816
6630 F20101207_AABEWU kadirvel_k_Page_128thm.jpg
b14a2aafb04eeece8237247634c5a327
c36d55f8cf3522c137792d45ae003929087d889d
1087791 F20101207_AABDUG kadirvel_k_Page_042.jp2
3f0c1cfc0fdf3352f6b5a7e0262e983a
132b8aa9bd8553a7cb91a3f7e1d9d203d1a5eccf
F20101207_AABDTS kadirvel_k_Page_027.jp2
65148b952778035bdbabae29d2022529
188ab0351477eb32fcde34819f426e3d34456178
F20101207_AABEAB kadirvel_k_Page_037.tif
d8e2f00f08024f50c7a2936529e87f80
7a279708cd51a2f5757d777fee65f0f950ef3692
7728 F20101207_AABEXJ kadirvel_k_Page_052thm.jpg
701160f1fa1b530367c1297fd59b5796
4c456f335bb8034b26f9f573eec34650e75bbb76
F20101207_AABEWV kadirvel_k_Page_136.QC.jpg
e50859e2f25da531158b2aa4a84037c0
cd18586af833d42222d5dc1ba92df1c4fc711fac
891141 F20101207_AABDUH kadirvel_k_Page_043.jp2
bbb799918310fe60e4528b00bed2c2fd
c141bb3c7387457df27ecb175ad8d06e4117ba97
646957 F20101207_AABDTT kadirvel_k_Page_028.jp2
dafaadcf56de85c8285f3d269c14681e
c79685f4a2a0141475ceafd420065c01baa934d8
F20101207_AABEAC kadirvel_k_Page_038.tif
816013e2712d9a90da846fae12829773
c627cdef3df4c7e4499a514fadc962f608db3b34
5978 F20101207_AABEXK kadirvel_k_Page_137thm.jpg
85226f91e8a34884e66839056f7dfefb
6fea98d848b1e6e3be979f0ff9a685fe39819a31
6690 F20101207_AABEWW kadirvel_k_Page_075thm.jpg
a28b6666eda01aa2e41f75bdaf68dbbd
8c59b6aff6eee35558a15de0ec048d05138d1cdb
947777 F20101207_AABDUI kadirvel_k_Page_044.jp2
d5d6159617f63543ea50b8cc05c869ef
8d3dc74be2d32627f64713bc8e76f7545a0e7848
887943 F20101207_AABDTU kadirvel_k_Page_029.jp2
0e63db1a31a8c2f9e798746ab9320a90
a216a855e6de937e77b055456d5eff4b4ec95cf5
F20101207_AABEAD kadirvel_k_Page_040.tif
19006a45be4e7538a866b3606bed2950
7a40cd7caa9f807fedccf62cf885de763894f4c9
19369 F20101207_AABEYA kadirvel_k_Page_022.QC.jpg
fa821ef1fe79c9de5429df0d3a95e5ed
79cfaab54d8c0496c8f7a7c8f678e6ce78dad32b
6736 F20101207_AABEXL kadirvel_k_Page_006thm.jpg
c30dc317502d275564c75be1fc0aa688
35a41e617a35b78ece749550fca7b69ceb6e4310
32526 F20101207_AABEWX kadirvel_k_Page_062.QC.jpg
9fd7779eabc5b45c613f93321c8e64d6
182bf51792c47188e36103b8484c139174731791
31431 F20101207_AABDUJ kadirvel_k_Page_045.jp2
c32c2fde3c06f0f60a2d3ede308cb096
4f63c239a961dcb78b9e2215a0cfdb7cea967679
912165 F20101207_AABDTV kadirvel_k_Page_031.jp2
a8ada9cc839a3a11cf6a9128d267b7cf
2ac2aa213140369290be11190fdecf3a59ab3594
F20101207_AABEAE kadirvel_k_Page_041.tif
bb0e9ab3c7792f55a943f482b3c1a79c
0d58575e0552781849595492795df09491fdbb54
22623 F20101207_AABEXM kadirvel_k_Page_037.QC.jpg
4167a3fc4eca80e1475559a513ae97b9
0b240f4d18495c1dde24bf10af74d2fdb2389150
8354 F20101207_AABEWY kadirvel_k_Page_165thm.jpg
3b3e049910b7999e2d1efc648bc010ff
03cb75200bbf0ba8e10038e68849b44c1c191308
786237 F20101207_AABDUK kadirvel_k_Page_046.jp2
094d19f8603c59b185255697b803bea2
753a3963d1b16026be294c98658a75d462cf5db4
777544 F20101207_AABDTW kadirvel_k_Page_032.jp2
e8db8cca904367cd8bda56f5acb6cd6d
3f0a73966b7a203f7e7ae06bbb7a23ada2af289b
F20101207_AABEAF kadirvel_k_Page_043.tif
1115e70f411d036067e5920d7fea47b6
d32abd28e303c3529c1013a4627f33d54e0ea98c
19908 F20101207_AABEYB kadirvel_k_Page_138.QC.jpg
e45e71c8890fabba93703138132f0ef7
7996fa1336850245971ca5fba2d29f2a94869aeb
33964 F20101207_AABEXN kadirvel_k_Page_011.QC.jpg
160d943fd82f87eda1cdf2c8c95758ba
1e8da229d0db00310165ad434bca7d8f3336a43f
22726 F20101207_AABEWZ kadirvel_k_Page_004.QC.jpg
45c876ee617c0e36d8de7fb7a5e9b693
dea904c667571dbe0b55399f51c35770b225f359
1087895 F20101207_AABDUL kadirvel_k_Page_047.jp2
0393495375f63870f9ad6bff7ba5856d
dd1e81699ea142b017cce89162c3ae1556cec2e2
1047955 F20101207_AABDTX kadirvel_k_Page_033.jp2
e676e185263cb51470671a95f0bda628
c27209fa6521b6d6da961a27343fcbc3a26894b6
F20101207_AABEAG kadirvel_k_Page_044.tif
2115b667c1c61b138c68b76cc08a1ebe
51ee34fff0c5e67a4014b8f41d0aa67eef5592bd
F20101207_AABEYC kadirvel_k_Page_094thm.jpg
7736704cc94c949dd2997bbbf07c6911
9af6705a0a194b76515aff74b6743a2073253daa
7354 F20101207_AABEXO kadirvel_k_Page_080thm.jpg
c7a8a3f02d0af15496296a684d294f12
37890591a3a7ad37dfe4a5aee35f7bc84f1bc624
1087889 F20101207_AABDUM kadirvel_k_Page_049.jp2
88916004c94710e264b49f8adcb3570a
cc378fac8ebe4abb26d187e4fce4457ac85f1ea3
64171 F20101207_AABDTY kadirvel_k_Page_034.jp2
f0291806782c31fd076302a0256bc79b
32d6062864619c9375585c22b800b2ccc7d003df
F20101207_AABEAH kadirvel_k_Page_045.tif
61667fc6af8a07bdd1c81a418b94a5b5
079d25cfdeabecbfab63a5c534d519a361b30c61
1082456 F20101207_AABDVA kadirvel_k_Page_065.jp2
8605678cff187736af9e0165a8122ecd
84b6c185c9a543abeec1b38d16b62d996dad9554
21200 F20101207_AABEYD kadirvel_k_Page_125.QC.jpg
5f6e62ec694f3219f8704e9d017c24f8
bdea94f829e809e4d6dd9dfdf8a71452b55f23f2
6103 F20101207_AABEXP kadirvel_k_Page_115thm.jpg
fa704b41e969497b452987fdbac76a8a
3284ab4333e16b6098f1301183e01febd3af67df
936440 F20101207_AABDUN kadirvel_k_Page_051.jp2
e16f5c55885857ca0e14809fd942f731
27a5136a99ce10f01323e80974a032adfde667f1
961437 F20101207_AABDTZ kadirvel_k_Page_035.jp2
24e3d32d96e3d1bafd5f216170b1d54a
aa1d411df6d2f8d3cd5d1e860d2f4bd97500d320
F20101207_AABEAI kadirvel_k_Page_046.tif
e2c78fda7ca9ce8b27aa0768064b962f
e85ee77da5579f52bde86fbfa68758b760c0d590
548983 F20101207_AABDVB kadirvel_k_Page_066.jp2
e7892f60a441c9c75b51b5c68fffbe6d
c11c0b9b8618d51404cf90c664fb22d8537e912e
17385 F20101207_AABEYE kadirvel_k_Page_172.QC.jpg
e00f495ccb64c8a237f0a5b86bafac19
cb08ac05755f25266d45beca1762dca756e5e3fe
23289 F20101207_AABEXQ kadirvel_k_Page_088.QC.jpg
2d169e196947dd3c7c185354be91d0ff
6a00b1b435d0c54c9ded98aa58a96958c453b9dd
F20101207_AABDUO kadirvel_k_Page_052.jp2
68f6d6b7ab0a455366769db0c17f0122
593f26cb6145f26c977fe7eaca79a9df1b953c1f
1047874 F20101207_AABDVC kadirvel_k_Page_067.jp2
fe741da93aa31a4bd4df9139ac4eca5e
19a7f42dc6d42bad85f7143f37ae4a929739414a
7097 F20101207_AABEYF kadirvel_k_Page_067thm.jpg
7253b40614548884e7e0450770db354c
8f6b39edc319e1c9c1ea3bd9bb404cda1cda3c2b
23208 F20101207_AABEXR kadirvel_k_Page_102.QC.jpg
b950b69fa8bbc33fd71c6b6349ea42c5
1c6afd15ace3620b7eb3b21bac539bbfdb8af479
552635 F20101207_AABDUP kadirvel_k_Page_053.jp2
f4e7e45fb4fbd5c4640f56f610982271
5c6c61b850bc99eef83fd019b61d50e969b4c234
F20101207_AABEAJ kadirvel_k_Page_047.tif
83d66d6001c140749bf4bfb7d6ae624b
eae77e85aca83095a548fbf4e7e9fe92b4281a39
F20101207_AABDVD kadirvel_k_Page_068.jp2
1cee7c89d4940b3d8282b076b2d76d02
f333b8d64194b05b19824dd8f45016107ec55969
32915 F20101207_AABEYG kadirvel_k_Page_167.QC.jpg
bdea2d1650f52ae68327fa41d9755800
10dd8c37e706299d2315bf05667fd61caf4ffc50
27759 F20101207_AABEXS kadirvel_k_Page_014.QC.jpg
a548ce38bd6fb048a9f71b3bef75ed84
341708c211c6c66f57cf3b41280b46f3ec24d0a8
930375 F20101207_AABDUQ kadirvel_k_Page_054.jp2
7a490c49293b460da28e8c2a1ab3e3eb
7d2bc0f18a10bd9b61918efa1a261053ffbc7882
F20101207_AABEAK kadirvel_k_Page_050.tif
9aaa2a2a1f74b51d010fee2ed103f606
b196cd0ec2efd7b0af19166066172a3d3b8b0d0c
F20101207_AABDVE kadirvel_k_Page_069.jp2
dfec626a87be07f78e7848f474d057e6
e91ca9c571d0b0f42d0c8127e25d4c951197f0e3
7026 F20101207_AABEYH kadirvel_k_Page_099thm.jpg
8a8b54058e8d22e9051f841fa4fbb55c
2a8ae17a5bdb11fcfcc8301ba96883dc3f88cf7f
29888 F20101207_AABEXT kadirvel_k_Page_079.QC.jpg
db2af2583e35f68ab818015ee9c7c9c3
aed1e55c3e89de2f5a62444bfe8a9c651681eb49
884642 F20101207_AABDUR kadirvel_k_Page_055.jp2
3a371389dc75ae7a4a93044cf0c4fa44
b024c2668e8d85747ef468e99b2294d8dc5fb33d
F20101207_AABEBA kadirvel_k_Page_070.tif
9efd1fa686f95ae5f56f067d01a71b96
5a8be73167e20f8349c6181219399f1154572fcc
F20101207_AABEAL kadirvel_k_Page_051.tif
3177ab2e848b808a9840aec3b8946b65
0e484edeec89507b0147e9b73af01d049700bb0d
65815 F20101207_AABDVF kadirvel_k_Page_070.jp2
2fbdf9b238d749ef9ea912849c88888e
46def82b4ac0d940c21e577ad4047cab306accff
7418 F20101207_AABEYI kadirvel_k_Page_010thm.jpg
2a6cc385decabc3d0b8caac1a57ad5df
9b4065a6af9df449ffa935e9141063f985fc8e63
5863 F20101207_AABEXU kadirvel_k_Page_101thm.jpg
168d0a05ab778f29f8303f38cdf3f26d
60cbd3e74fd523ca4397a296ee721223d546a042
1087891 F20101207_AABDUS kadirvel_k_Page_056.jp2
82971d1158b2c4614866acf748dabf53
b33f02162d35ca7f769ed72bef7078f32907a584
F20101207_AABEBB kadirvel_k_Page_072.tif
2de53291718fe5b4d92baf5bcaf41052
2c6d18159c8fc476876460387e09fab4f6d8d129
F20101207_AABEAM kadirvel_k_Page_052.tif
451543879cf448009ed3b90bfee33c74
3391c13ae47d022e60ea5adff107d3773d2548b5
66393 F20101207_AABDVG kadirvel_k_Page_071.jp2
44b3efcb7ad0ac6fa99547cda50a0f4d
323c578a0cbb71ac9e5d01878825452f316c005a
22856 F20101207_AABEYJ kadirvel_k_Page_100.QC.jpg
2be0e38f861ead36d81d3649ea2550fa
0f7035da5dc409b64ec218508ea5cffe0fb4e8d5
6206 F20101207_AABEXV kadirvel_k_Page_081thm.jpg
a1d076de368da35bdce831013d0810d1
768e192ba43f63cef95eea6bd3fa3592e5e7c41d
1075648 F20101207_AABDUT kadirvel_k_Page_057.jp2
60b19f234bbfd6a076f2502d92aaa40a
f4090e33293d36b872223bc72536602eed0fd714
F20101207_AABEBC kadirvel_k_Page_073.tif
57118df96ccb6e316f2176471ce0344a
e8a25b0c0f172dd084ba119457c2d2b4c126635c
8711878 F20101207_AABEAN kadirvel_k_Page_053.tif
08fc245c1dee15aa2630733fd6c76623
98a0d2efad01b15bcb48badff6a6f87faecf5ac4
F20101207_AABDVH kadirvel_k_Page_072.jp2
31a12244037cb1dcf5e759032a58c1fb
7368ee8aa890e4be86fea008ecdd217c597cccf0
24312 F20101207_AABEYK kadirvel_k_Page_103.QC.jpg
e11679410512179cd6bbd784251d852a
9a9eaf7285f5d27815b60f36490c17c7f6e8902f
8897 F20101207_AABEXW kadirvel_k_Page_149.QC.jpg
d83695b2de441ced34d441e2c52e7a98
5987fcf78c833002e83bcb7d40acd2b6e377541d
972477 F20101207_AABDUU kadirvel_k_Page_058.jp2
183d55d92ce639be14cd12350df7bd1b
540ead00c9fd5ee9e0c804be63acb6778e255e7f
F20101207_AABEBD kadirvel_k_Page_074.tif
a1bdb7b2bb52a1b5c4261832e9b38a27
43aa6062362dcb50e96be5c8a4f4a64316fb9379
F20101207_AABEAO kadirvel_k_Page_055.tif
cd13141d4393f18cc37b06fa0963fea2
1e1e4c864ade83fb7e2f53fe5557ee342afa66b7
933929 F20101207_AABDVI kadirvel_k_Page_073.jp2
6c8df0a0567fa4fd9b6ff9dffca30d2f
5fbee5488aeec9c4e397b71e211bda7d5fd8069a
29400 F20101207_AABEZA kadirvel_k_Page_151.QC.jpg
3b39621ddbe2bf83ca93849f0b35f51e
f8b25330f54027ff09e6e8f45f0182f7b4e747e1
F20101207_AABEYL kadirvel_k_Page_123thm.jpg
7a94edfbbc638f2b8132f6c51ed1e4b1
5bb50fba527bf9d757524f200577f352e5652185
5787 F20101207_AABEXX kadirvel_k_Page_131thm.jpg
fd5315fb7f3c1ff646f92757ac5f54bd
feff69394c6a35545c1418fc74f11b4c182c58fc
72872 F20101207_AABDUV kadirvel_k_Page_059.jp2
1d4bad4702df3f18b653a0ac06f414da
8178bc09723fb33471b37a6c7cbec2e4ca6c6573
F20101207_AABEBE kadirvel_k_Page_075.tif
87ede38d404d967b066aef314ed7770c
3e1b65d39072e34c91ac2676f44c7cfa067bc8ab
F20101207_AABEAP kadirvel_k_Page_057.tif
7729871f8e6f21948fadf091122a00e1
52f5c7c2b82ce5af96fcefff7ff8a005ff2ec0d9
884397 F20101207_AABDVJ kadirvel_k_Page_074.jp2
6ee63075c00c4c0ba8788521beaa2f12
d8143ffff399234e1e83723b50a82164c5d3d6d8
23832 F20101207_AABEZB kadirvel_k_Page_063.QC.jpg
d7fe1c0e02da180506146219a7e6cebb
f60f8db303f261fab8ca32017a3ccc6c6ad22fb6
3368 F20101207_AABEYM kadirvel_k_Page_003.QC.jpg
81022bcd85a7f70bb9c2d221df585ec3
910edad56a76462652be447a51ace08a67d757ad
18277 F20101207_AABEXY kadirvel_k_Page_154.QC.jpg
d75f57a9f1d67a0c6f160acbcf0fd983
3631681431757aefea94772cc077e729540a0947
66673 F20101207_AABDUW kadirvel_k_Page_060.jp2
df4eb6037b21eef2f4fb87dfc43e6f42
669a0425dbfac705e0c342072680a5645f7b5755
F20101207_AABEBF kadirvel_k_Page_077.tif
72c467966484dba633e9ba012ce48d41
b4737069f316c6ba8261a08c85d721c0a5cceefe
F20101207_AABEAQ kadirvel_k_Page_058.tif
a7c2e8e07587e8765c597f67f4a23b6d
cec03a44ef8bf7cdebe093934f94008bfb7508b4
1036131 F20101207_AABDVK kadirvel_k_Page_075.jp2
7d99e99648fa3d26b0870ffa901b87cc
0554a874e2e3bbff31813e0a8bed280a3558c483
21101 F20101207_AABEYN kadirvel_k_Page_132.QC.jpg
be8acfb873ebb8799d151547cda80244
01d2e508e4155b8e4d9b9effbb74ecf31f0add59
22072 F20101207_AABEXZ kadirvel_k_Page_023.QC.jpg
050cd449f94724ce9037f1cd1e529ec2
495b865af32dbdf50aca6fd53a01d172b13a9bee
1087851 F20101207_AABDUX kadirvel_k_Page_061.jp2
7e87fb20faad78febbf0de19b340c1da
446c7e768ee6030d9f17a4d5bdae470de695f184
F20101207_AABEBG kadirvel_k_Page_080.tif
35957e6842446fbab9dd5264bab7cbd0
b85952677d6989f95af4a048ec26bd1f65a87177
F20101207_AABEAR kadirvel_k_Page_059.tif
c5c4e0fcff78ff7cc71e7391a57f88c9
bdf987d1b8eef038f2298b35c04931c5018a1451
682206 F20101207_AABDVL kadirvel_k_Page_076.jp2
a36ababc401f26aec906bf97c661c08c
4496ba8e6d844e5ac0b7efac2ba00037267a43ce
6618 F20101207_AABEZC kadirvel_k_Page_088thm.jpg
4d5ae5e05e9e962a2e786108198b4bc3
fdaad328b09bb2ab0fb46305597e49a825fd5e78
25111 F20101207_AABEYO kadirvel_k_Page_128.QC.jpg
75733d3807faa7c3793cb5cfec4804a0
eebbef8b5d63a16b6bbab69382a0574271eb8589
F20101207_AABDUY kadirvel_k_Page_062.jp2
3c4bd3571bde4649c12a6e766a1744cd
390f1f89c0afe6be08c75b5a095460670440f752
F20101207_AABEBH kadirvel_k_Page_081.tif
5a67418c71d5ba7962ec54227fcf7b10
f0a8a11dbf2eb16c5612f61fd2c749f3418eaf13
1087786 F20101207_AABDWA kadirvel_k_Page_093.jp2
f26994f81ebb0d360d889f37ff847644
39d9d4b82d7c528dc0b6959584fe82856de28814
F20101207_AABEAS kadirvel_k_Page_060.tif
8faaacd5a86d39b2b25906e2a184833d
0a1c02f74d2af969c7ba6fe85a71ee509a0d8f8a
722838 F20101207_AABDVM kadirvel_k_Page_077.jp2
985bc494df7b7a477210312136a47b13
694a735d2abbf0f6ef586d6b97e216871ac32353
F20101207_AABEZD kadirvel_k_Page_018.QC.jpg
f7dfa2ea50f6594550cbbed2c02cd0e5
a36472a8d67429cd2402bcab824176ec362eed13
5070 F20101207_AABEYP kadirvel_k_Page_156thm.jpg
fc8d6e130a57cd5f0384f065fd703a46
d019bb6e440f71d98e4cfa0af75e68aef020509c
F20101207_AABDUZ kadirvel_k_Page_063.jp2
40c283d14a937db657f86544e299e54a
b38ec5ea25986da27e139de3f91f737c3b6f446b
F20101207_AABEBI kadirvel_k_Page_083.tif
479a609a0ab996a5121027988e7cc043
0caf6122c2c8f543f05c2f3838fc5f8bfdb96995
F20101207_AABDWB kadirvel_k_Page_094.jp2
fb5dbe119476ca1e1c503d83be74552f
745007c372e197d641b36a389175f00a2dc594f8
F20101207_AABEAT kadirvel_k_Page_061.tif
1140416bf6d4de0c05c8375edb3afc2e
d5dfe34119cf3960a810dfa4d4bc819963325ff9
842317 F20101207_AABDVN kadirvel_k_Page_078.jp2
9cf2588af247add020c2eac236719a48
7e0525405d8c10085fe011ad7a9d5b143d32d94e
4394 F20101207_AABEZE kadirvel_k_Page_107thm.jpg
6d8b15122ce2e5d80cf1cc7dcb515ef7
cc1286f98038da22378c8c7c01177b401515bdff
21318 F20101207_AABEYQ kadirvel_k_Page_101.QC.jpg
0bac3d029478234251ee2ae97b59823e
eaf3f818e9327950c80f600f7b7f612ca8c8804e
F20101207_AABEBJ kadirvel_k_Page_086.tif
65ff104f759e6f7c3326c0346a015fec
e2f90ab98d98e709270cb24819b7146d10241e63
1030635 F20101207_AABDWC kadirvel_k_Page_095.jp2
87c7908f8d5a03b1221afdaf1141cd03
adb980c8ce219bee2141640c7ddfaa16002f7453
F20101207_AABEAU kadirvel_k_Page_063.tif
4191697f1310ca19e4495f0c30d381f1
6ca6e82c2a079f65721b8b3a139a3d3b8a26f92d
1087879 F20101207_AABDVO kadirvel_k_Page_079.jp2
d07dce72a08711c9c28d3c4afb95ec66
7f16963915e6fe732f24171338d555c1cf6f9b75
5475 F20101207_AABEZF kadirvel_k_Page_106thm.jpg
33404c447a4073d2dd96afed05e6bf08
c965698da1f7045ff53c24ebbb5ae289c9410349
29461 F20101207_AABEYR kadirvel_k_Page_030.QC.jpg
27e5719c2658b99f4b19d90004b308d8
d7829eb94f9476189ca7e70a1144315dc7491746
772158 F20101207_AABDWD kadirvel_k_Page_096.jp2
e9f47232bef74561b1203a7214f26f6e
d429e6a5aab98908321ed146dc2f4526d3196d1c
F20101207_AABEAV kadirvel_k_Page_065.tif
21ed924d5431d30c3f2cdcff1a3fa9cb
44b5b105689ee2379709b7343e9c4f25a151ee9a
929837 F20101207_AABDVP kadirvel_k_Page_081.jp2
9139f713be965bf821247ab92bb51d82
e0fe852af0d15bd2a62b699dac34d59af128b871
4996 F20101207_AABEZG kadirvel_k_Page_059thm.jpg
22ff0defa82a4ea2a7fb918b3c2caf93
ab3cf6b85ecf273c7e43e083d983811ecdbb0e94
4844 F20101207_AABEYS kadirvel_k_Page_152thm.jpg
6995045373f9e979897415a230cd5f31
41d414e2bbeca004abcdd672e905f54aba7e6411
F20101207_AABEBK kadirvel_k_Page_087.tif
af141522d2b6e2495b427f0ea553ff7e
280382a344865889304690bf392e9ce54fe121a0
864837 F20101207_AABDWE kadirvel_k_Page_097.jp2
0d8b5fe05a7638b6d63050e8c43e737e
6c785aab0c708bcbe33414e1511e7ed462a46b8b
F20101207_AABEAW kadirvel_k_Page_066.tif
92f3ba4d441608f38ec2f3abbf79692e
6e8e434d784ce984ebf2986ba01bd58b0023d626
966067 F20101207_AABDVQ kadirvel_k_Page_083.jp2
125bacef7170dc15713f297d0b17794f
19164e98cf0282c9363d0029b216c46b3e088a30
31513 F20101207_AABEZH kadirvel_k_Page_049.QC.jpg
9d96466c17584ad8764278e91dfb79aa
601270210b0badeab2da226d47e3fdc9d78a4e28
19201 F20101207_AABEYT kadirvel_k_Page_162.QC.jpg
38c555192bee0ca1f7c2d1d70a614498
5fc4a47a4a323e2f0512fa9ee5f1243d5c50e533
F20101207_AABEBL kadirvel_k_Page_088.tif
e3321096b2e91a77166d11d152edcd03
d8abe39d8e45b564f33baccd10c664678d3ee144
1087867 F20101207_AABDWF kadirvel_k_Page_098.jp2
335cd50c50ecdae45394c7602c3843a2
446faa190c2ad43027ced9570b2376f81a024c06
F20101207_AABEAX kadirvel_k_Page_067.tif
646480855904b753abb2ec3d49b3a399
c49fb256536f006c842867ff104107f84828b3f7
1049070 F20101207_AABDVR kadirvel_k_Page_084.jp2
e4259b6deb5ea86342c3e804569d9142
0a6d604e926813a6b13b5fc0314709cc9f5fbfad
F20101207_AABECA kadirvel_k_Page_103.tif
45841a9837b729cabf14cba5ef9f0a51
7266905dbccfc23ffc9b5ca875289ed7fd397776
24663 F20101207_AABEZI kadirvel_k_Page_145.QC.jpg
afd76bfc44cea02a4e549ffceb72f477
4480e5484f1c854bf149f827e9d4c8791b18b493
19536 F20101207_AABEYU kadirvel_k_Page_129.QC.jpg
a79c104b21e21be4e5898d03da995dd6
8dc9b9834f7bc5526da7b06fdc4dcf821969d270
F20101207_AABEBM kadirvel_k_Page_089.tif
538561035e9de336cb67032c62000baf
e5af07c5232bfbeb70f7cd5d1e44475592e843fc
1050966 F20101207_AABDWG kadirvel_k_Page_099.jp2
8ead9931b3cd48219186d51429b8ffc5
baf3f412699dce371bc85977e22599baf3bbcbc1
F20101207_AABEAY kadirvel_k_Page_068.tif
867c0a2ada91a51169786818d86a8048
c489cda52db9d3808e250155c4c3f13bc751e94b
1087882 F20101207_AABDVS kadirvel_k_Page_085.jp2
d24e241e3d0ea0ba8e434cefc1235b0c
d764fad978fdc611c25d890cd3661bd46e4a1da6
F20101207_AABECB kadirvel_k_Page_104.tif
9469d6b42f4e12bfa9cf6abb93b595db
e5916b675638f02909208e4674d311dbf959c74e
27799 F20101207_AABEZJ kadirvel_k_Page_080.QC.jpg
2cecb0ccdafbf29ba62f64cded344908
757679af28475bd5105cc08f7c6079cf38a896d5
26779 F20101207_AABEYV kadirvel_k_Page_143.QC.jpg
38fa4a5f83e3daa6ec5691ff68e7b331
00af8e8c7c64e5a1cc5683f5759ba23da90f56b3
F20101207_AABEBN kadirvel_k_Page_090.tif
70e2107052a22104ae261a3e3be8c7b2
b63bc0fff892d7e8d0d2573c5bd7a3a6677ba91f
973936 F20101207_AABDWH kadirvel_k_Page_100.jp2
7de2fe960ae24d428d1abcedaac730a5
d757360c91d1639c58ae9752ba2d0aaf2d3e9021
F20101207_AABEAZ kadirvel_k_Page_069.tif
c9501c7e426ef4d45d50f1be8eaed13d
ecb3de10fe2aae472f0530df507011f014638a68
F20101207_AABDVT kadirvel_k_Page_086.jp2
f550effa0fc914e2e4429d3c926bd1a8
2de22cbaad65bcd003a2d409a53f5feb54a4bccd
F20101207_AABECC kadirvel_k_Page_105.tif
9460980070377778d516a2fd27bb54ff
91d2bf6311bb904c9b800f5dfbd5ad80d594f23c
22160 F20101207_AABEZK kadirvel_k_Page_081.QC.jpg
4423bbd73d254d6326d1617ed175490a
ad7b7bee71321b5b29c4fdbe67f437861296298a
7890 F20101207_AABEYW kadirvel_k_Page_018thm.jpg
d21814ec7ce32fed8b7871dda44f965d
207498a6d1b0e05e73d9e7f0f762ed6786d263eb
F20101207_AABEBO kadirvel_k_Page_091.tif
d234b00b96f183c29deb027e59410598
775000030fd81ec338ccd7a73d066a4ee8c98dff
891692 F20101207_AABDWI kadirvel_k_Page_101.jp2
f3ded737bfac3b7a6aadadbc27b12e1e
adb5f436a410ead37f7214128f0cb90bf6522d9a
756751 F20101207_AABDVU kadirvel_k_Page_087.jp2
b4afaeba9ae734949992d3eb050b0757
c353f69c71f6f384c6a0985bce631c533e851ed1
F20101207_AABECD kadirvel_k_Page_106.tif
66fedc63717773ddaea3ba192311cf12
7258f4d0c5fae42676d0d550e733fc650008e684
21757 F20101207_AABEZL kadirvel_k_Page_021.QC.jpg
43665c02915cee5d67221193027c2472
44bfc7be1db9a8895b7b8fdd224b968cb577b012
22816 F20101207_AABEYX kadirvel_k_Page_051.QC.jpg
d281733e01c5b8ec75b19001d9108b24
aa675b6c47415c46012f884e34da2cb833a090e8
F20101207_AABEBP kadirvel_k_Page_092.tif
4f9828ea3fb77a430af21860a58021d4
38fb40b98341423a6fac99560a3df1c730daffc1
940504 F20101207_AABDWJ kadirvel_k_Page_102.jp2
8262c9aaef522466b4781e2ca0ff375a
71c02a7d256269786cc13bf50273747cdcd42656
989201 F20101207_AABDVV kadirvel_k_Page_088.jp2
8ac31a5e3a27645044b114c67567ddb2
751b50ac44b2d9f66cbb5bf6d54b4a0193ca0d5e
F20101207_AABECE kadirvel_k_Page_107.tif
c26772742daa19e14c48cd61e9478da7
f67730ec021aebce818edfb27c4a135e19299481
5266 F20101207_AABEZM kadirvel_k_Page_087thm.jpg
5220755ccfe076ac978a96a7de110953
e6169349cb753111d531bb00bb65c29555ae0a4b
7820 F20101207_AABEYY kadirvel_k_Page_013thm.jpg
6974d4dcd82d223d6ad456928050fe3d
b831ea8797d798daf88a50447419fccdf6b656ac
F20101207_AABEBQ kadirvel_k_Page_093.tif
641d56f4862b29bcc1857e1398ba2a87
611cdf80db56dcdd07c12ea3521eab4bde7566f4
1007624 F20101207_AABDWK kadirvel_k_Page_103.jp2
619212217079064b0e4fe4b33f17ed59
a46c76ad31a3a96bfa7d34b9a1aabf551590fa62
1087878 F20101207_AABDVW kadirvel_k_Page_089.jp2
d2119c3edc352216be3dd2fc2d76c1e5
2d157bf802709f194ed17595bb5ce54d73bcd858
F20101207_AABECF kadirvel_k_Page_109.tif
e1f2d410c87edf62fb626c23f75114a4
01725a53a8fed94ad76a17df9cd465acd91587d0
260764 F20101207_AABEZN UFE0020083_00001.xml
b0ea81f1f2b2d94fbd50299a67c3efc6
a1cb9b6e674d9eeb6e851dc67b03b6895e04727c
6599 F20101207_AABEYZ kadirvel_k_Page_145thm.jpg
b79f769c73022ef27065e7ff63a80766
2c27f19adbdf6f5eebd4b99896b6de876e2f0cf1
830758 F20101207_AABDXA kadirvel_k_Page_121.jp2
3ae8c2581ca6bcaeeac5f4885213b7db
44a50d33b7d861c73342f052288ca6a1d16e050c
F20101207_AABEBR kadirvel_k_Page_094.tif
7c1acf70869a85b583eda7349a3b41ab
44094ce2f8babcd55856e7e08f0cac8b9e644e99
902924 F20101207_AABDWL kadirvel_k_Page_105.jp2
192d0d68fcdd7ed63545a78c4adb7726
12625df2d351d5a9d64761809afa5ed9f1865248
1087884 F20101207_AABDVX kadirvel_k_Page_090.jp2
ea8535e181a664c1d458c34f284d3070
1ebcf69a6a9f9422d0aa1cccd0e44e9d91722f09
F20101207_AABECG kadirvel_k_Page_110.tif
4c84488bce237616b27b3157e8b867af
8a2ff320bb03c8d25587697279503bbd98f51d23
30843 F20101207_AABEZO kadirvel_k_Page_006.QC.jpg
97efedd5a12af46e413c5d70b1a1f9a0
a200e27e55459173e9589d0499237ac3021df4ef
F20101207_AABEBS kadirvel_k_Page_095.tif
7f950a054afdf94a555e9d29a1910dea
3892ce86598e38613592f834a2a50ce3652036d3
735262 F20101207_AABDWM kadirvel_k_Page_106.jp2
e751ce6cbbd944a70813351c2901ab4c
eeadd1fbbbd199841b9c5ae23ef8b0bd92085ba5
F20101207_AABDVY kadirvel_k_Page_091.jp2
620dff483081d2340aa4b5663f75cf63
4a347ab1c0acce479c5f790bf26cd0a7884d411c
F20101207_AABECH kadirvel_k_Page_111.tif
0e25923ce951af975c03b27831b184f2
6518731ff2ce18061a67bcf96bf3662930aa9dd8
30522 F20101207_AABEZP kadirvel_k_Page_010.QC.jpg
4b5a36923b2053a571b5c1a8a7f1f526
dd56ca31a074a2d422301da9a41c7caf5584a103
878686 F20101207_AABDXB kadirvel_k_Page_122.jp2
271ec1ae449ce6c055bba98b8f1a3167
694844e1e21f519c263b97ba2ff6cd8e6f6f7272
F20101207_AABEBT kadirvel_k_Page_096.tif
14a5d862e186b1d555d329b39d1bb24d
bdbd85a6ac4967b7defbd78b0b2d490d30890c10
556205 F20101207_AABDWN kadirvel_k_Page_107.jp2
7722ccf4d3e7aa009df22cd9557d09c7
89311655ebf64ab3069471c395b06058803dd540
943991 F20101207_AABDVZ kadirvel_k_Page_092.jp2
26819f1cc0d02ea0c504003ede428d48
543d6458bd230a650bccaa798def9d367cc460eb
F20101207_AABECI kadirvel_k_Page_112.tif
b20e1c6c35e9562b7dab9ed145f4552c
ba6b94d0670c68e684f5f248fd3e703ce8238aec
29528 F20101207_AABEZQ kadirvel_k_Page_012.QC.jpg
7e1ab259430660b167b4bfcad69afa7d
ca6d494fcaa7842394410e73e0570442b7d56067
846802 F20101207_AABDXC kadirvel_k_Page_123.jp2
57f766d89a402557aa2056c91858d851
5ff5b5d0bbc690bd97fb5f15700c0d69e7c48857
F20101207_AABEBU kadirvel_k_Page_097.tif
9c8f926a39f1f7d820c0c2c46daca455
801712524eba6965a3654e8847ce4d1caa1230ae
1087809 F20101207_AABDWO kadirvel_k_Page_108.jp2
1794f3b1f83f2a3cf5d67aacfd148a24
64ea0d6636d565aaab42385e7637dda1f1776435
F20101207_AABECJ kadirvel_k_Page_113.tif
670aba95e45a618616b6659153351bc1
acb85de162de4424b688d78f06f4bebc781daa73
8427 F20101207_AABEZR kadirvel_k_Page_017thm.jpg
f8498ca8d331c75751a53d6831a94a53
bf0423acee4fe41bd524f0c8ea0e3b31be862e9d
881253 F20101207_AABDXD kadirvel_k_Page_124.jp2
02e0fe604851b647b2afff9d34e6984a
78b98bcacb5e0ba95f5572e7eeb780478e474452
F20101207_AABEBV kadirvel_k_Page_098.tif
56d42baee4d1dbcdb38f174e4078cc7d
0b042f6fb438e8152753f2c067e7bc0b7e8a744f
1087799 F20101207_AABDWP kadirvel_k_Page_109.jp2
5e3c4c8142d57805f34b84e2eb078a92
f6549956002c537119f411bb0bd6dd8eaa5263d7
F20101207_AABECK kadirvel_k_Page_114.tif
242dc4d1937ecaa1d63e45cb30f4fe0e
72a96318ad8786bf076281b75dd0cff0a3e3ebd2
6917 F20101207_AABEZS kadirvel_k_Page_019thm.jpg
08782049e9f3a09446d486a343941ae3
ecddfee21de2f6e04b680046eadd288c5eddf114
1087841 F20101207_AABDXE kadirvel_k_Page_126.jp2
3d047e56114bad94de0ed6f54893e580
9556e669860089cc83ab10fe80d59642c84b43af
F20101207_AABEBW kadirvel_k_Page_099.tif
bf721bfc44022f822a8ebe3f6104ea4d
0e7a64f91bde880a294a77623cfe2be149f8c4e1
F20101207_AABDWQ kadirvel_k_Page_111.jp2
d5cb2bf5316fe1ac5ace635a28a09c85
393c26d3d990ff37ef7d3621cd842e54389ba035
5738 F20101207_AABEZT kadirvel_k_Page_024thm.jpg
4df212b4d20b6018e7526ae64327cd23
6895faed641423b1d7f5ec94c10cc039fad82c6c
809088 F20101207_AABDXF kadirvel_k_Page_127.jp2
e38133e795b76ad2ad2a1c9c4391aa86
5c76665874720c80a24c2b9b90117d0cb374cb88
F20101207_AABEBX kadirvel_k_Page_100.tif
44692e06ab0fc82fb04fa9513ece7cbf
ff20a503afd0bac8ce29f4da5d5f711b15fb814c
F20101207_AABDWR kadirvel_k_Page_112.jp2
fd8762c0087396b075ebf3b09d47212e
861d440794495da1c18cdb6993ff98d0ea71abf4
F20101207_AABEDA kadirvel_k_Page_134.tif
1d502f61949a72653556ca6d431c5aa9
70e2e016d0254f0ed7936764f3dbf34b8590f5a9
F20101207_AABECL kadirvel_k_Page_115.tif
7b30aaaef5b9b78c6726d7aa5d7d4df8
63ba8b4e886fea7cb4746040c34c59cdd8cf7c08
17556 F20101207_AABEZU kadirvel_k_Page_025.QC.jpg
c824d1fe4322c4f6941d4e426114b475
6ecf5785744263b358277045b93a1c5fa9bc1ffa
1046815 F20101207_AABDXG kadirvel_k_Page_128.jp2
078c7e5b383934c05fb9c69a1d4d89ac
1cf787641cba0af87d76a6a21dfe9a7f8850a3f3
F20101207_AABEBY kadirvel_k_Page_101.tif
5439b1e49f02d85e27257258be8d56c0
7e37c2437b11553a87e25371b6809613321621fe
859030 F20101207_AABDWS kadirvel_k_Page_113.jp2
8c19b939d5a2094ddf52b1cb85b251a0
dc8418c7020484f4a8adff86bcbe4a160b381f44
F20101207_AABEDB kadirvel_k_Page_136.tif
cb786a2c9d5dc696b18909906925088f
d6c974a1002aac45bc41f9aaab2fc72edc1ef1ed
F20101207_AABECM kadirvel_k_Page_117.tif
dca7fa654c6dea3dc792ec418635145e
2e7d4b704e73226d2fced8efd0f63e0b6d704376
6910 F20101207_AABEZV kadirvel_k_Page_027thm.jpg
8f9fc68c662e28bba7cd3482dd229c36
8021ce3da8a834d96c4393fb82746b094eb78cb4
813637 F20101207_AABDXH kadirvel_k_Page_129.jp2
65817322d76e67a052b0bed45676b97f
b5a822fab88233488f738d3ac5c75ee172ff2e50
F20101207_AABEBZ kadirvel_k_Page_102.tif
877d4fe98e722518e14f0db95fec510b
5812ab4fc38597aa18f675e85cd8c5965e021074
1009292 F20101207_AABDWT kadirvel_k_Page_114.jp2
82a2b6feb0aaee062af2d2e4710d318f
57975a650cae91b361148720f57a0e4a98e92da3
F20101207_AABEDC kadirvel_k_Page_137.tif
f77153dd8b2311b8990cbc771144a5f1
3c643deac2baaf893e1f281a96f9db3d9d0a1478
F20101207_AABECN kadirvel_k_Page_118.tif
96f6997eb53d7691118bcf6d44549d1f
4cab73187c1d663b72fbcba27497b984e3374c9f
18917 F20101207_AABEZW kadirvel_k_Page_032.QC.jpg
98bb3f8d7e780b62ed9444ec3cd00be6
df96ffda8ffbaf9c4878b42dd6796c3701a798a2
46223 F20101207_AABDXI kadirvel_k_Page_130.jp2
04b3f5db9e1f18ee9e11aab4c2e1f723
04a28ac7420d6ebf3148431742fe73711987b989
903370 F20101207_AABDWU kadirvel_k_Page_115.jp2
35bc774ba7ba1ea7a001392a5358eda5
c932621fba60e507cb891250e5524b79910adc44
F20101207_AABEDD kadirvel_k_Page_138.tif
09692158938d938d0125637531b77988
e2ee26af801bab8f9942ac65f64fc4d3fe2b87bb
F20101207_AABECO kadirvel_k_Page_119.tif
b337fd38e98ec0c69fed1ccbe17f1b66
81c62c0f95ae56e709cbd197a74403d44f7064d2
6415 F20101207_AABEZX kadirvel_k_Page_035thm.jpg
4247882a812d4da5053d0984a3707cde
1bd106f12861544731c23ccf2c8fc6a13fa3eb29
834460 F20101207_AABDXJ kadirvel_k_Page_131.jp2
3447ef9e8c665c47974d03523b4ebe2f
3cf0590501a88361fa9685718d41381095570dfa
F20101207_AABDWV kadirvel_k_Page_116.jp2
2a9805eb34dc3bfcad32c55ec85e192f
50e89f20f20bccc66853e5ac4a7d33e76d92bc20
F20101207_AABEDE kadirvel_k_Page_140.tif
9e1b69b657f2f02e36cacb68f398db22
29a1ce8677b91d8fa1ec18a3a6de66be4560bc94
F20101207_AABECP kadirvel_k_Page_120.tif
23dbd342b8fd4f23ca5249d73f323d43
babb8b4f18e399ba8a2c86095c48d610b49dd107
7064 F20101207_AABEZY kadirvel_k_Page_038thm.jpg
872eca2c3f8b759b678f8bafc4c8cac6
c275c4dcb33a5dea3507436287fc504b06488a10
962087 F20101207_AABDXK kadirvel_k_Page_132.jp2
94a020294a31e7e1a6e985ff0523ca26
df75436288529a3ef84a4cf7df11ba1bae5c05de
845681 F20101207_AABDWW kadirvel_k_Page_117.jp2
10a13c394ba45a74179bd62b3fed4ec9
cc9e840e8843853013adeb7e3db1d39b9b7e1fbb
F20101207_AABEDF kadirvel_k_Page_141.tif
9f15ec77491a41b6f1ca15802f4925c3
296d3b9782eea90567c1ab0828ba6dd1876780d4
F20101207_AABECQ kadirvel_k_Page_122.tif
3a1791e289070b22d3854dd235462547
a10ca2deb4f7a618814286bb131dad8183fe95d4
2792 F20101207_AABEZZ kadirvel_k_Page_045thm.jpg
f0f4548fca4f8a79a92690b91d3d7c5b
5cf66c2b84ab5e415d251f4ba5ee46875d17615c
F20101207_AABDXL kadirvel_k_Page_133.jp2
90b279d238cab0f58ddcc13bc96bd4da
7d019742313024975e03f73f425aec0e313c21b5
633862 F20101207_AABDWX kadirvel_k_Page_118.jp2
ae6f2dddedcce3657b3e53ecbff38cbf
6b71e7d3f9a4634839f45bf9b53ed0f762b5d271
F20101207_AABEDG kadirvel_k_Page_142.tif
7cc15bf7297cb25ae4b09992e8742940
3aefbe33a78318f75fc72b68569758aefaa9056c
124967 F20101207_AABDYA kadirvel_k_Page_151.jp2
be470e6f9afc2f15f9162951a0b8074e
d6162988f81bcb9c036bdef36285e41eafeffaa4
F20101207_AABECR kadirvel_k_Page_123.tif
ae89502c2992ceda0edeb5889f216c82
a2950423faaeb45805832c81e03a92467fd55da5
1087874 F20101207_AABDXM kadirvel_k_Page_134.jp2
e9df3c221514dc216155ef805d86e611
922eee9ad6f3e0d6ba93e7a7485b5ddde4996938
760858 F20101207_AABDWY kadirvel_k_Page_119.jp2
b7ca31b07a89f1464464e1ad8c4c07b5
96b72ea32a7ef716c687f7e28f6ad1e7c4f819a3
F20101207_AABEDH kadirvel_k_Page_145.tif
5e579545e4ff950ab1f00a860cef653f
da6f8f0b608bf9a2547cca3726267fdba211e9f5
624050 F20101207_AABDYB kadirvel_k_Page_152.jp2
7165e6e99d751c6be7bbe2f60d782a6c
ec1c9adedfc376a96e34710497a0617437ccc09f
F20101207_AABECS kadirvel_k_Page_124.tif
2b9d227a3e364441c6f8ed40cd11dff1
64f276df8d3d65c2ded31ed977d4048bdbd1dff6
910760 F20101207_AABDXN kadirvel_k_Page_135.jp2
de19e6cd685de7e45c3c7d5931d712cd
91bec2d5abc7dcdef5d8ebd62203a14deb2bac6c
F20101207_AABDWZ kadirvel_k_Page_120.jp2
a887e748c68336e09f42f00341de0e24
84d7765aea74b49cfd0caf294a5169532495efcd
F20101207_AABEDI kadirvel_k_Page_148.tif
2a4b2719a0c667d9c80e23b6553b81e2
77e4ce1ba6c11cbf3535094b9b9290309a10b275
F20101207_AABECT kadirvel_k_Page_125.tif
615b053946b916a63facc2daf7ef8632
f859d3a5149aa9f3231a3b1e1bb24b073e495a50
F20101207_AABDXO kadirvel_k_Page_136.jp2
6f5ea0fa4cf33793038d9d21eebd08cd
d5857f10c54551b02e8be51e4ebc8388ebd11f74
F20101207_AABEDJ kadirvel_k_Page_149.tif
ff020474ba2e17d552ee62d35e7a55e1
38537d0270d1897a3ee811f7f07b47890660557a
616273 F20101207_AABDYC kadirvel_k_Page_153.jp2
c8ec22958f7d9b126f7f2fc7a4012fd9
da64b518965c12c29e2fb8e402c453281bc505de
F20101207_AABECU kadirvel_k_Page_126.tif
6340e2826c904fab979096acb1a121c1
124c653111657d0ad79d84d2d2b50d3ec6fcd180
1014492 F20101207_AABDXP kadirvel_k_Page_137.jp2
6a5f2814f26610dcf45d39bc0568ae12
7e0306e0a7d353875ced3edf0839009963ad8ac0
F20101207_AABEDK kadirvel_k_Page_150.tif
4e1374ad66b5201d971205b60433587a
781a430d535bc4a8c18ce711fc5e1bfcd2d109e8
717928 F20101207_AABDYD kadirvel_k_Page_155.jp2
1f64f945f9081ef5d35a4caf270b59b6
0c06f106669a58eaa6dbba6b75b0039412a30992
F20101207_AABECV kadirvel_k_Page_127.tif
9efb138375e9001205a4b114766306f4
c9f613a4e996c29239bb5435f020271b7e82a582
F20101207_AABEDL kadirvel_k_Page_151.tif
f7fc3dda47c24bd243c11db9a2d610e4
005a49d66b63973c8edbb5763a627c1c77c51435
658084 F20101207_AABDYE kadirvel_k_Page_156.jp2
a2d59c4d2b53d9dac0e45d42fafafb9d
1ef35fb37201f7746ae5839c02de0734a933f033
F20101207_AABECW kadirvel_k_Page_129.tif
5b31060b3a896dadd8594b4ae7d0920a
36c3a73b5070d3170e742ed1f68cac58782f4ed5
1087898 F20101207_AABDXQ kadirvel_k_Page_138.jp2
b741a91e9500ee1ad813422ef5aa9fa4
c8f539885e7451639ae24203b71a9be33db87831
F20101207_AABEEA kadirvel_k_Page_166.tif
3f5a10ba714a7533ec7bb82289c5345a
5ffe61ad49e1a0be3e2ba60d747c534a56bc7415
671118 F20101207_AABDYF kadirvel_k_Page_157.jp2
f3be2f4b6b2c51a77c75a71b7308671d
fa92626c38ddde02518cfcf4b0cd962e64ce9354
F20101207_AABECX kadirvel_k_Page_130.tif
660ac51fff96bffbd1e9ef05e0502f71
b524f13ada529d3023935fda3d2455dfe44a891f
F20101207_AABDXR kadirvel_k_Page_140.jp2
6b5217519858efd124d26055ab21de9a
39d95dc29d19a23c5ec3b4c5de325b7e7fd2c7f5
F20101207_AABEEB kadirvel_k_Page_168.tif
9e7f405a0109a4325ee94d922704b856
bd8543a94113fa68147427435175dce75a4211d5
F20101207_AABEDM kadirvel_k_Page_152.tif
bafd79d810939623d18632967556502b
c186817d72c0687cbe787a208686a71976dc41fa
952771 F20101207_AABDYG kadirvel_k_Page_158.jp2
250f0293c5e9550efc4f46605c739fbf
983688c88484ac72d83d1524d1bdeb29f298e13b
F20101207_AABECY kadirvel_k_Page_131.tif
2633de22f2d2b70983bbf5794673576f
81faf2b6e04b6aa942d22d081304587b8b48062d
F20101207_AABDXS kadirvel_k_Page_141.jp2
5df2c54f32a2388c2218a916fdc7cd76
e6fded45b172f868dac2a292dd1f0f1bcf6dbda2
F20101207_AABEEC kadirvel_k_Page_169.tif
d38382455cab61c201d7ada968e0bef4
ce486801d9345ceabedb5c59ad35609d48cfafc4
F20101207_AABEDN kadirvel_k_Page_153.tif
8a5e1efa8855296fc9d4095ff7600c07
dc48c3e03a0d6ef4db5e2dde1d3474cf93d9b84a
623161 F20101207_AABDYH kadirvel_k_Page_159.jp2
bb63af6a21b5bb4a3c44f67cfef71ea1
9809d3e7448f248f7cd375f3428b3653c92d342b
F20101207_AABECZ kadirvel_k_Page_132.tif
3b9ff69c13b10e9956ab9c68cfee230b
5e047c643af5ee7d9db27194756767d78d06425a
F20101207_AABDXT kadirvel_k_Page_143.jp2
81a349239553e3d06674d8e69db7481e
748bbdba423f27938e9f61aaf34a21e7b24ef32b
F20101207_AABEED kadirvel_k_Page_170.tif
d42cdef14731e78f258a1692fbb01c14
152b287fcb9d688e60e2c5ef4120dd66ea8c236a
F20101207_AABEDO kadirvel_k_Page_154.tif
287dbb49ca8b5db2605241b79905dc82
f80c64c54d1cc896f2a45d3f61c176364cea4830
61381 F20101207_AABDYI kadirvel_k_Page_160.jp2
4d8c6b48c8fe3a9046c21f219e100a28
beab562d439389f020c6af44e72054ba919ea4ef
F20101207_AABDXU kadirvel_k_Page_144.jp2
cf5fd59e91b2f7a194db659f15804a40
b0afa2e8cd9c7328b098aeef7d82cc997e52870c
F20101207_AABEEE kadirvel_k_Page_171.tif
3ab295735e4d318cc03a483795a561e3
a1b9c421a9dc0e235dd502fa4e25f31bf5159e6f
F20101207_AABEDP kadirvel_k_Page_155.tif
5500152886089c1db789773098acc924
37c6a73b3f9ba1ee235f75eae6a7f8af77141893
23862 F20101207_AABDYJ kadirvel_k_Page_161.jp2
9c6a0cc0765a2ad6daf8298172569f34
d5621fd606a53da8d1463391d235d51e5f63c15f
1035130 F20101207_AABDXV kadirvel_k_Page_146.jp2
d822c9261ef0b4ef746046967f078907
aff56a03bb3a529fa2ace7da6d159e9321f61a5f
F20101207_AABEEF kadirvel_k_Page_172.tif
90bad15f0dffccc534e794b7be03a019
619b0d8ff0df51763c9c32e40fb182456d03a7b6
F20101207_AABEDQ kadirvel_k_Page_156.tif
fa41e93a338e14527cb19a59f0cf6a18
b4c526d276f251dd38269e1c4500f96f7c7c3ba3
787653 F20101207_AABDYK kadirvel_k_Page_162.jp2
2bb4e9c08918e189fdbbd72e7c5d3e62
a0c68b628d48a312bba80658be87c8a5ec93ae4c
F20101207_AABDXW kadirvel_k_Page_147.jp2
8e1484dd5b6b9f6d6e36dfc5cb0b7ed6
8932259e7551627942cebefcb82bc7651d5c2a73
F20101207_AABEEG kadirvel_k_Page_173.tif
dcb4d255645875b9a358344beafb0597
299961614c5273ece8c0b41040b2c520315aa780
F20101207_AABDZA kadirvel_k_Page_007.tif
738c8510761592d07935c028e4536680
fbc454f2f7993fabd35831c162a9b08cce758974
F20101207_AABEDR kadirvel_k_Page_157.tif
bf72907cfb903d4eda77a4ca8eea5018
51ea45a03050e40af76672fd5df8ee89a62741a6
758091 F20101207_AABDYL kadirvel_k_Page_163.jp2
488994f80d8a1821d5413dda832e687f
0793afb9aa99a1e33101b00e06ade97786140d83
F20101207_AABDXX kadirvel_k_Page_148.jp2
ff77e86917455a8cdfac4e6428c7249a
db401ae23e4f8ac9c27a6eb6d8bf3002bcaeed3b
886 F20101207_AABEEH kadirvel_k_Page_002.pro
643edf80fbb24060cbf602cd1226b8f5
d58254500f23135c87640af908cbfbd93d53f1ac
F20101207_AABDZB kadirvel_k_Page_008.tif
18a41474313271f77c9a55bd4acec080
747092f8732f9c536b77dada2630b22b9e35d10b
F20101207_AABEDS kadirvel_k_Page_158.tif
11f666250e6f2053d27150c861beb64a
65b467b4752b908d4942bc449ab5764869c1f0e2
1087813 F20101207_AABDYM kadirvel_k_Page_164.jp2
5c0dd4e148e5357e48a7cc3b35cd18a8
c657f839011de4e2c02a91bfce68177e69116a86
33107 F20101207_AABDXY kadirvel_k_Page_149.jp2
096f108792ea891199786d1446fa9edc
b645d9b0717361fdbf0cf8531ae4ef4bf3d7f11e
1068 F20101207_AABEEI kadirvel_k_Page_003.pro
8f988c9d38104948d9eef6f1ceabc55f
8d0f793d8c8a5baef1dbebe53733ba8a1e050a92
F20101207_AABDZC kadirvel_k_Page_009.tif
e6c26cc528d28063c7fc23588fca34c8
668696625aaafb0f31297880d3951109738380e7
F20101207_AABEDT kadirvel_k_Page_159.tif
470393cf397ab5f37fddacf5a721b30c
d6ef9eb61dc2600f0004df8d4cd62aafbfa1fad8
162749 F20101207_AABDYN kadirvel_k_Page_165.jp2
a2a2c54c74733a5ca85d220252f730da
e651297d9e6d2e05121a6fda985f6520f9601089
122306 F20101207_AABDXZ kadirvel_k_Page_150.jp2
099cc5cb4babc95aa1653701bcf26536
6474c097dbb5244126c96de2b92a3cc4d3ba6769
44492 F20101207_AABEEJ kadirvel_k_Page_004.pro
192331e0acebe04d0b7cc8e917a98406
f232cef23cbd05024cf8f09a06c72133a102e468
F20101207_AABEDU kadirvel_k_Page_160.tif
91ddc1a24ca0715aae9d508e6f537784
2d84a62eeb9593c812c664ba2b175a4ddeb7e2d4
150382 F20101207_AABDYO kadirvel_k_Page_166.jp2
628c1a37ae840bf0a9efdef132852ff2
54636a7e9dd2c53eabfd806c968d190fe9616e95
76060 F20101207_AABEEK kadirvel_k_Page_005.pro
099d14f29cbc9ee247f61dd4688a4d21
9a0b0274a03365cca2794488c3a59fd30128b560
F20101207_AABDZD kadirvel_k_Page_010.tif
5dc1e0fce2646e8117e9645e765c57f8
f907299f63b263da91100cb7ed48d82b2646ecb7
F20101207_AABEDV kadirvel_k_Page_161.tif
1f6734a47f1ca7cd26063cd427473193
2c261591b446bc89f8d48ccbaeb991117985344f
F20101207_AABDYP kadirvel_k_Page_167.jp2
02a5a27d77e40dc8937665c4ab2c379b
a787afe52855e9ea41e76de1abb1b614fef1ff43
73393 F20101207_AABEEL kadirvel_k_Page_006.pro
dc9edeefc4b0d8b09a2b535332b73bbf
f67f1e5c1a9c6a2febf7cbad1a2c183f99364978
F20101207_AABDZE kadirvel_k_Page_011.tif
f803684a233b5be825ab1c8086543aaa
0e07fac173876bc75e16742c0070d4a0af55b1ae
F20101207_AABEDW kadirvel_k_Page_162.tif
30cdffd196dac6722a06842326d072cd
1f4a9ed57ebbfdcefd6afc5d3e779b9aa662a6e1
162204 F20101207_AABDYQ kadirvel_k_Page_168.jp2
60d4cf26bd24e6ebd723dec8cd8391bd
76202d74e81c82c143ad5740d955ece2bf514914
19898 F20101207_AABEEM kadirvel_k_Page_007.pro
ee630a48fd09465ab715dbed9abec4e2
dfc7f86486343b8d7d2e0a3bd4b655c27b2ced01
F20101207_AABDZF kadirvel_k_Page_013.tif
3b58d6374c4fccd57872ee49bc40215c
6abb8cc67caa5038166bb624d7e56bd05acf2977
F20101207_AABEDX kadirvel_k_Page_163.tif
76987efa05d13bb86cb0f523e0b29310
91a623eb2d5cff96bdf6a0e04f33fb3c677a1231
1087854 F20101207_AABDYR kadirvel_k_Page_169.jp2
8c1dbc10bc96f23711f25996e00a5f48
8ca605ce61ef032434bf4264a003fb4d08b320d6
36304 F20101207_AABEFA kadirvel_k_Page_025.pro
decb56adda654b59f36e7ea221b00122
98a593727d26711ebab054caef576ce36f308090
F20101207_AABDZG kadirvel_k_Page_014.tif
b7561415866e592256769b8ea0d0018c
6a0794bdeb47e3bedc7241eabeba4a4051007709
F20101207_AABEDY kadirvel_k_Page_164.tif
4a03cddd2078daf26b42ac4d3ff3fe83
15335a29cd4b1122da7fa40bf760390c4d0bdcdb
150479 F20101207_AABDYS kadirvel_k_Page_170.jp2
380869a22e573c9869bc64513a6af938
205ed9a659b740a0858e9a88ef775cbb915fe4e6
50927 F20101207_AABEFB kadirvel_k_Page_026.pro
cad129fe2bd56d2422346b9a5c6e5122
3c50620addef5a8a55036b0ab196b27428c239fc
F20101207_AABDZH kadirvel_k_Page_015.tif
0182eb0b32c50f59dc4b2f83a8f33c5e
45715afc34737e2a014d20f071c1751901fe94a2
F20101207_AABEDZ kadirvel_k_Page_165.tif
3f7063f98c9b17efcbef3740ff3c22c3
3fd7807cd1acfe8fd4ae91b198908d02b00d3f29
148432 F20101207_AABDYT kadirvel_k_Page_171.jp2
3a8071f67ca85fba49a5e70176c320c2
afa328391f4720923a9e1cf37c49c116ef1c2999
53542 F20101207_AABEFC kadirvel_k_Page_027.pro
d7bc72635957ffdba2bd0649288afb41
cd19169055efbca30f39853f7c9e33b4f1f719a8
72729 F20101207_AABEEN kadirvel_k_Page_008.pro
a38b6f064a582d6d88558433b85f1815
9fdd6ecb8bb58edae2c0a83da631bbfc00d7e221
F20101207_AABDZI kadirvel_k_Page_016.tif
822dbb1474dab5e4a919e5aa5001e1c4
633ff4456d3bc3c2b3086b234b4e52c54e6c8bc9
77818 F20101207_AABDYU kadirvel_k_Page_172.jp2
f23133a9cda911331be284bb3ae02ffa
d008de084374b0803131c5d8370c99ff4ae7f229
60627 F20101207_AABEFD kadirvel_k_Page_030.pro
cd90d878a55da2ff667bdea37da4e488
f4bbb6dfd8aac2d462bffd3b148304e2ca979fab
73588 F20101207_AABEEO kadirvel_k_Page_011.pro
c7a8fa630e6fffeb1053b16032176fca
654ef036f781ab6ac9ea2be01155a17e40caa589
F20101207_AABDZJ kadirvel_k_Page_017.tif
05928f436565788fd4b8014bbf692a10
3a76f2e7512158fd1b2d384dfcbbefab9de0b22f
41920 F20101207_AABDYV kadirvel_k_Page_173.jp2
6d24c6359b42f2fc195d2bdf313b653a
cc38b111ca90f2cfee1990e0ae7feaf2855c51a1
43470 F20101207_AABEFE kadirvel_k_Page_031.pro
3c65af875bd2c30a3b83cc04e3eae20e
f7839b9132885627fd8be672f9378e2f22dd83ed
75810 F20101207_AABEEP kadirvel_k_Page_012.pro
1c34a9888c9e11f0d32bdcf421f3e65c
306ca725580f353b1f481f0e1f427dc917b95bff
F20101207_AABDZK kadirvel_k_Page_018.tif
fc2eaf407a2c769ca6ca174166ccf841
f450cea8915fbea28faa06077a40828eaacf915c
F20101207_AABDYW kadirvel_k_Page_002.tif
c0c32e46b4abb8831682613466fa2d70
5077c9a7c9ff5ed908297ca4474150e1c1e4de30
38559 F20101207_AABEFF kadirvel_k_Page_032.pro
e8c8d032706fb007b630d65c34507d01
81d70be84443b5112cc58bc29c3fbcf83caf4ba4
74336 F20101207_AABEEQ kadirvel_k_Page_013.pro
9765a0ed8adf1a47eb811b0e15656ed7
551e862a1d53579222502ee21de0fbff0addd864
F20101207_AABDZL kadirvel_k_Page_019.tif
9c7c4fb0511262a03a9f618e57348de5
fe8cec80231ef54b8520a096b5b3e176577a4207
F20101207_AABDYX kadirvel_k_Page_003.tif
ce03e90d17d0e3738c26198a9b4cc38e
9e94f1c690c87c69f069e5ffda01661ef44c26af
32111 F20101207_AABEFG kadirvel_k_Page_034.pro
e3845db3ef43ab9b54e931ff4973a517
ed7c2f6eca54c77ddb8d2fa0957253e802affeda
58359 F20101207_AABEER kadirvel_k_Page_014.pro
29e44b65156833458fb95e17a2db6f8c
3fb021308c6ad705c7d7f6e2266e8533f8889574
F20101207_AABDZM kadirvel_k_Page_020.tif
cae40f1a895ff8b5af20ae4a1c2686d3
01143d27c25551f825da919ddc6147e617e9bd83
F20101207_AABDYY kadirvel_k_Page_004.tif
309933407c578ead6bfdcd87d1ddfe67
7d4181b0428b8e4e6ff4d2f665be4d72a9e1fa7f
44776 F20101207_AABEFH kadirvel_k_Page_035.pro
39bddd2ddd9928a9678ace938342bd9c
d1c6f7e057c2f1fc1cfa8fcb5323a940e77bc1f8
4437 F20101207_AABEES kadirvel_k_Page_015.pro
4acfa598007581f0f097fa4c3ba989b6
1e245fb5b324d40af882f92332c26aa0b9277509
F20101207_AABDZN kadirvel_k_Page_021.tif
f8c0ef3bd15df70d5dc6a7cca7cbec46
da506272dd80389bdab7fcd200efbf5670093d04
F20101207_AABDYZ kadirvel_k_Page_005.tif
57313474a64bcec460f3a865b4b175e6
f23a449030d0b26a2704c9c38b518e7d7412f697
46155 F20101207_AABEFI kadirvel_k_Page_037.pro
ab92ec91fb09f444fd4c7e9f38470264
32c1fcba95d3e569b1c1c201f1cca6b2efabaa22
64944 F20101207_AABEET kadirvel_k_Page_016.pro
df155ed7de7c619ffe3340fed03b8276
68b6f1c959b8ea813bd6e29c6e00cb2733cb2066
F20101207_AABDZO kadirvel_k_Page_022.tif
c3ed057093dc48ebc53da50013e63250
fea38f0ab3b2c68abc4e587ec91824c7e3468a1b
49875 F20101207_AABEFJ kadirvel_k_Page_038.pro
45d83174587bb16a535bf07e366c0af1
27a9cc132dcc00d015d7d4d42f45de1efe947959
64971 F20101207_AABEEU kadirvel_k_Page_018.pro
6c0dd2eed87a7b3c92a9335b88759c3c
d1b60260b3408a1833fcfd0e4433939544ec2b41
F20101207_AABDZP kadirvel_k_Page_023.tif
169b5cc7e9cf166f45adc01982021f0f
a16febaad5c0563a4ae8bb46b850cc736aaad1b0
52452 F20101207_AABEFK kadirvel_k_Page_039.pro
20be85a1488ad5ab63cf139221a1f0de
347da25bd7799edfce09acfe99ef4956c1f5443c
49716 F20101207_AABEEV kadirvel_k_Page_019.pro
1cc95bd7e3df79fa5e56c0f2db8cc25f
3be50d7e05c7928eb99271a4ed3a5931f66fb36f
F20101207_AABDZQ kadirvel_k_Page_024.tif
5b9bd9e36cf7c972ef85017b2ad6d2ec
afe0a7ba5ce1c954a8800e756fc6fc0fab389baf
45092 F20101207_AABEFL kadirvel_k_Page_040.pro
c2b2059abeff08ed78b98daeeb41bec9
499b6160c7f6972595aee9f7b63b0f102e16295b
37754 F20101207_AABEEW kadirvel_k_Page_020.pro
4651f468a32280af90efc10df6e2bfd3
d4b44cdf226318add023a125500d4e8656a74b67
F20101207_AABDZR kadirvel_k_Page_025.tif
2e47cb51eb90bc0b517096d8e86a82ac
9d5791549268661e0662409c5166b8e384f35f04
41082 F20101207_AABEGA kadirvel_k_Page_058.pro
6a680507730c54710783f4b99765a473
f3cbc6f54eae44440d0aa76bee93a0b535bde4a4
66448 F20101207_AABEFM kadirvel_k_Page_041.pro
3b1baef2917be62c4095acee3044cfac
1f7d6c4f73b79a856b3183d1bf5849c12538bb8a
38827 F20101207_AABEEX kadirvel_k_Page_022.pro
049a8083fe82922daff403aef1ebe0d7
70b91993860a8c09982fdb78490cb01ccf8dee86
F20101207_AABDZS kadirvel_k_Page_026.tif
fe7f19079d4048b7068ee32e30b89a68
10b38b585a19b0af3fd4e6d5ef11354666ed9d74
39720 F20101207_AABEGB kadirvel_k_Page_059.pro
c3de0a6b5e29df6c1a703496d66661f5
35e6810dfc4b6c4e19830b7d25ce1348229508b6
70216 F20101207_AABEFN kadirvel_k_Page_042.pro
f5aa632052366f3a49df16bf955e0a62
d651ec374f57475b17181249489cd9cbfa6d9fc8
41865 F20101207_AABEEY kadirvel_k_Page_023.pro
723eee43e8d2c8e58004b1ca485962a0
a646b53f488cbd48447731d7309c6a363f949267
F20101207_AABDZT kadirvel_k_Page_027.tif
83a9c47d2c073c76aaf0671a7b3b6927
5af5c21d9a3dfcbff0e35e398ee064ff5b71b21d
30696 F20101207_AABEGC kadirvel_k_Page_060.pro
d5476645e81135659dbfbee51130cddf
a42460ba6ede10bf9a6bd40e9b261b9b60395ba2
37822 F20101207_AABEEZ kadirvel_k_Page_024.pro
bf642e42c90ec49b37d681bc97f715a4
e92a467ca912a3d187855718d6abe46c4b0a70b8
F20101207_AABDZU kadirvel_k_Page_028.tif
a56a00ad467c4f0d01d4e775c4b800a9
f6d8e7eb2001c41ffb2ff813ffdc0541b881f2d6
63967 F20101207_AABEGD kadirvel_k_Page_061.pro
f4f87450068c889986c13772da08a982
9289416408dea5f4c0b125b0f8d91584af95cb98
41640 F20101207_AABEFO kadirvel_k_Page_043.pro
f956352aa378ef106c66fb127d93601b
74991ddbb63961bb7d7091e6bdc88c5e84951a8c
F20101207_AABDZV kadirvel_k_Page_029.tif
b996211a13b7004837abc73257e037ca
7bdd36ea6a29fa56004c39ec2442da283df7adb3
65210 F20101207_AABEGE kadirvel_k_Page_062.pro
f6da368db25c7bd2990fde0da63df014
ed4eba1114b20b3530e23290ce6a426aa9c6e754
44045 F20101207_AABEFP kadirvel_k_Page_044.pro
886f399b38a65d660779aff77958b740
d228f51f4cb346238d45c446a7ed84809131d1b6
F20101207_AABDZW kadirvel_k_Page_030.tif
32d14a68cae028c86875816e4b3c7859
0554274ec8737b5200d9b410b5cca69b2b04f98f
54736 F20101207_AABEGF kadirvel_k_Page_063.pro
4f51a7ffd3492f716c7e5a3243c50a84
63d0951ff821ae5ddafde2b6cc300ddc5d860776
13204 F20101207_AABEFQ kadirvel_k_Page_045.pro
3d6ff375e4f63085d79f3e38c6115c7e
8b0be0e8cd22f10ff90fda62934d9317162bb503
F20101207_AABDZX kadirvel_k_Page_031.tif
195cc4a1369b2f8724ea44f38f6c50e6
25f42bfffc55224b547898b5233f37cb09c889fa
52311 F20101207_AABEGG kadirvel_k_Page_065.pro
bd86130b2d8b82b1482eeffe72ccd00c
36876fb4927f31ddfdf20f9a7d510e193c48b2a7
65115 F20101207_AABEFR kadirvel_k_Page_047.pro
7143b5cea873dd2089b626222aa09b63
01a8101ce1b9f1c52363ec5cd11c356e16fd7b56
F20101207_AABDZY kadirvel_k_Page_032.tif
8266958450e0becc0a4908380d2b6dad
aeb3d4f498834b72a84177356f932df557e6ceaa
50157 F20101207_AABEGH kadirvel_k_Page_067.pro
9bc54bdf0e53c913de32ff722ec00f00
07dc78b84639b519379392a3cb7ec85fe462e0d1
37072 F20101207_AABEFS kadirvel_k_Page_048.pro
2b747d5ced4f6dc734291924035f8f30
3f4d086e367c21f8242485deeb2b912f9da98f74
F20101207_AABDZZ kadirvel_k_Page_033.tif
a3d7ab8300da3361de0f6970d382f9a0
498bd9ce13f3b0ab49435142cf84cfade0a2c599
54827 F20101207_AABEGI kadirvel_k_Page_068.pro
df88a5b1ebaabe36cca18d26704241c7
a0f1edfc4fc271d0351ffca23a1ab4187950f436
64824 F20101207_AABEFT kadirvel_k_Page_049.pro
69c821be5fd320431ad0b37c7fc5adeb
c50110df3de5348fd3936454fb2b5ba3a27468b5
35547 F20101207_AABEGJ kadirvel_k_Page_070.pro
3f550f1260c46de13aec5acf897a40e1
15d1caf42b109384b245ea2d13f318af7df4112b
45144 F20101207_AABEFU kadirvel_k_Page_050.pro
9ee9b7f97876c1576f8bca1e58401cdc
c70a7ffef69f7506c5b56165c0d8bf3f5f26545c
32553 F20101207_AABEGK kadirvel_k_Page_071.pro
12473ce786ca1293ac1bc7acbccadcd2
56a2bf450886117406cb39e2b1b2762283e89d7b
64593 F20101207_AABEFV kadirvel_k_Page_052.pro
39ebbcb90586a5fac2f1bf468234c308
9e2d4e5b5dfefe832594d54b076dd6c3657c6229
40269 F20101207_AABEGL kadirvel_k_Page_073.pro
2da6def692dcae97853b484e51e027c9
3a7b7340d9d79b594a3da2c804ee161a7f3cce8c
23162 F20101207_AABEFW kadirvel_k_Page_053.pro
ea372dc87906c3cd67109f8d3b6be22d
42c5c02716c10fc3efd986164163841b6e9af142
40366 F20101207_AABEGM kadirvel_k_Page_074.pro
e4343b66f41173d75988d01a198d1ea6
e8408a35946c846847b81b1247408bd3e7857eb1
43123 F20101207_AABEFX kadirvel_k_Page_054.pro
9a32df82afdc50153e3c1b0523d3b82b
e11e6306e1bd3a6f43cf9fe310ccabaa62b9dd0e
62083 F20101207_AABEHA kadirvel_k_Page_089.pro
54015a46f403eba94df911ea99cc1ebe
f27089afbc53f427db9b70346fd5d208eb65f0cd
48797 F20101207_AABEGN kadirvel_k_Page_075.pro
8555edb9662ec06f0eb4a014ee8a5885
f5a51423952eacff8b891cd3ad05b8ae133e5406
41098 F20101207_AABEFY kadirvel_k_Page_055.pro
04625fa24234a9e935deca6d2d011c4d
792f8bbf92b7bdff4c59867eb515b330b8c94b58
68182 F20101207_AABEHB kadirvel_k_Page_091.pro
2fd8a2e594b0f9a26e9a3e415600bd09
bc417ea1193a20b3333ff93c8064fddd0244586a
36894 F20101207_AABEGO kadirvel_k_Page_076.pro
bcf7ea0849262135a27b341f910f0053
cdccd117c94f465258eb755b1e1de5dc5540ab70
58860 F20101207_AABEFZ kadirvel_k_Page_056.pro
63b299914f092404c8d71703ebb15e7a
64ebd974d3292f49fae642977e45d1162c6a7e5e
44559 F20101207_AABEHC kadirvel_k_Page_092.pro
2bf47b56045cfd6fb9de69270f47d86b
7dab1038d8d664c26903e579230eb869b478fc9d
62086 F20101207_AABEHD kadirvel_k_Page_094.pro
de9d8331a7a43b9c1bc65695fabf32a4
079abff7f0e62714cc032993fd6c50a8ac4c01b6
27911 F20101207_AABEGP kadirvel_k_Page_077.pro
3c3b5f11d3d57f96d7bec4f72ec23fb3
7a83efffef37204b7d1b6a3984966c31be467f44
58388 F20101207_AABEHE kadirvel_k_Page_095.pro
c68370e25a07a3819a60cdaed5285472
a602f66f4f38a78bdbc498db80900dd150a1562a
40383 F20101207_AABEGQ kadirvel_k_Page_078.pro
ac2cdbdf317d721405c9d165aae9e939
cb4286755edaf4db402fd18c1e9693043b461e49
33980 F20101207_AABEHF kadirvel_k_Page_096.pro
3eacb525c4f3da1ffb68ad5b3bb43540
05a1190c98a14b88281ec3c4e9405e05b51cd71d
62345 F20101207_AABEGR kadirvel_k_Page_079.pro
45f4b0178e39afffe018e9991479cc5d
4e65bf729e63e50fce3f17a525944f1aed9c7b6c
51922 F20101207_AABEHG kadirvel_k_Page_098.pro
d0f65cbccfe4f3e03960643aea123380
e4aa8c824d282353fa25efc09ca6ae938ae70f07
55137 F20101207_AABEGS kadirvel_k_Page_080.pro
b420cd0adab433c27caeec266e87d20f
73add5a9112889b6f47e725b66946cbf44e66a73
54013 F20101207_AABEHH kadirvel_k_Page_099.pro
a4bf80c7530b35122d680eea6f01f514
f491d797b81b4f59625accb8f6d593262cc1b4f8
41773 F20101207_AABEGT kadirvel_k_Page_081.pro
81870988dc1510608b446ea3f80cbce5
4bee78feba1f9a8c3a84d4161d611315f8120f51
47468 F20101207_AABEHI kadirvel_k_Page_100.pro
5fe11caebec7c5ad6b9bb536bbb2e2c2
59937ff61fe8146ca787a58028e7a5767706bcf1
26891 F20101207_AABEGU kadirvel_k_Page_082.pro
ddf0bdc5a3c65d77af8dc4e722393b5c
5335d342082b0fa0e000914acecff50d14988f36
36038 F20101207_AABEHJ kadirvel_k_Page_101.pro
bb5a9d9de4edadf9df41db9e5950bca2
23bdf4aef0fda5bd3d964b721f733d53d685e72c
47611 F20101207_AABEGV kadirvel_k_Page_083.pro
fc0cebadd9398020272340becb733218
a3ae0bfad3c8ad0885134fd7b3cd8eaacc25c198
31834 F20101207_AABEHK kadirvel_k_Page_102.pro
348763e27e137caf3b704e134db70e63
125196012224ec84ff6ae625ceed065b124ce519
62051 F20101207_AABEGW kadirvel_k_Page_085.pro
d94bafe49c7275f750d71911cb56fe78
4032939cd1f7a3334b16a0fb0a58fbe5f65effde
51257 F20101207_AABEHL kadirvel_k_Page_103.pro
0058044d2e46ad62d433e10caf82fc83
4006a8b0dd570e8583f2ca8fcc7bc5c97a0d15f7
65974 F20101207_AABEGX kadirvel_k_Page_086.pro
f790bf24c053d9d9727f3b6bbc5c85f3
68508ce6689062aaf7a1351da8e2e132cbbde551
29005 F20101207_AABEIA kadirvel_k_Page_118.pro
413351d677877b52b3dd06b8649ff25e
783d40b8df941dcad37472b6ce1be634f53590d1
40127 F20101207_AABEHM kadirvel_k_Page_104.pro
2fc27b69ef4732e198e83627c84dab30
c982bc7f63ce775bdb08699f7446c8a1a8aeed4a
34225 F20101207_AABEGY kadirvel_k_Page_087.pro
650477780ab45c4d46c58ca58ac0b0fd
85435b0c677b0fda5a17c7f24a2637dfcc469f36
37118 F20101207_AABEIB kadirvel_k_Page_119.pro
569f7d58342ec169f30e0938899682c2
65d82611fc05b3a10b3209b751dc2bb3be2eb5ef
40533 F20101207_AABEHN kadirvel_k_Page_105.pro
6265628697b98109706f8951eb75341c
6773c0b878e4c08dce487e3cdce1444bf5ca4a82
46217 F20101207_AABEGZ kadirvel_k_Page_088.pro
7223029fd9af87ac5122a3776dba38b0
40da86c06cb3da65a667a69c3bf7114797c0a02a
52843 F20101207_AABEIC kadirvel_k_Page_120.pro
8952fc06128bf47b99316738b30c80be
7fb4dd9828534b6afefae67bdb6c505908ee6da3
32952 F20101207_AABEHO kadirvel_k_Page_106.pro
697b5d75822d634f7efaf413ebafcc3a
7ea02427734386116c73cb916dacc4543341525f
37116 F20101207_AABEID kadirvel_k_Page_121.pro
f4383e4417176bccb995e9162ad7e887
7175d22ea874f4bdecb99bcd765da6abe9580ef4
22449 F20101207_AABEHP kadirvel_k_Page_107.pro
3d608a32d85e3c12e2f58f30bf4dd3ac
37e0fbf5d30f918fab797c24002af7e5c32f6fad
44873 F20101207_AABEIE kadirvel_k_Page_122.pro
44bc23975952a821b9110a5c809f6fcb
41625e7b125a9a8e32c96460efb0265d0c9a94d5
33487 F20101207_AABEIF kadirvel_k_Page_123.pro
de84c9068f7b285c4eb153107e7dbb1b
ec61a7862baa76fc21c88d37d260920847d51fb9
30491 F20101207_AABEHQ kadirvel_k_Page_108.pro
fb373bfd97cf2cdaf691b99aab68ab49
7712dccb09720d79f3fc15b8243dd84a328424bb
40430 F20101207_AABEIG kadirvel_k_Page_125.pro
3c56966107960c591c60c739fd7b8e9a
5455049748b7d7eb81ba82ce335d464a52754ab0
52147 F20101207_AABEHR kadirvel_k_Page_109.pro
c91a909fd4df0b1e6eaa2937f5c83aac
a0f7d95cd26004311b4718b2c4dcb7276234d7a1
52743 F20101207_AABEIH kadirvel_k_Page_126.pro
8b63523edc80075b2567c82c92c5af74
4bd2adae07acfec78d11552dc914ed533f759779
35560 F20101207_AABEHS kadirvel_k_Page_110.pro
cef968292c0a33e1fef0ae8f815e2792
28acfe9665d2d54284bd646a3d8ae080b3ad1f29
41188 F20101207_AABEII kadirvel_k_Page_128.pro
b1c8a7264f0c9c777dc9456f152f7aa2
7b4d9f4e813eeff6cef352b99b88d28ffdc23e44
37845 F20101207_AABEHT kadirvel_k_Page_111.pro
ecbdb711ac8e91b5ac038a9fc69fabe9
9ef0e4de4d259ab16c9499a55c6b1313a3d807d5
33530 F20101207_AABEIJ kadirvel_k_Page_129.pro
fe1aa9b8856375d4d42e1df3a506f20d
6e4b4a0dcb8f9b4a43a175d5e29f040b708a7fe4
54122 F20101207_AABEHU kadirvel_k_Page_112.pro
b452b8911664006f8e3801951c6f5124
a5b1d74308990b39f07ea8492d88c5580172b95c
F20101207_AABEIK kadirvel_k_Page_131.pro
4caa11eda3ee7fccd562b2fa61b6f735
c67bb09db7fe3b51fbcb0feeb8051942cf1125d7
39206 F20101207_AABEHV kadirvel_k_Page_113.pro
73984ba7c585b321244a00ae2d074d3a
3763786ab5ef17a8d48ca5768dec869797767299
31957 F20101207_AABEIL kadirvel_k_Page_132.pro
a5df2cda1ee1581cbf51ae6de03e0790
6727cf38c47978979141f6b92662c7c024b4b553
62272 F20101207_AABEHW kadirvel_k_Page_114.pro
c008fd431674d5da70640d0520272b43
8d703fb11c2298546484b1019e26c63c14439d32
61388 F20101207_AABEJA kadirvel_k_Page_151.pro
073af78a71991d4a77c620806d692400
83cd6e1dc92fcf81baf76da45d3be00b7b4b1f64
61083 F20101207_AABEIM kadirvel_k_Page_133.pro
38df634bd8125bf9960fb7f78f7c199e
ee7ba01cb90f23c8443da22675218826585ba36b
39870 F20101207_AABEHX kadirvel_k_Page_115.pro
28629126f88b34b228fa6274781ec52a
228d5d2cdfdf0ce4edeca25e8e638c967a704832
26074 F20101207_AABEJB kadirvel_k_Page_152.pro
dbede3c09d412ae3e7bf513e1ab5e1db
652a9b9ef57e66bfd8a2d217eb4e6306e19bec04
46246 F20101207_AABEIN kadirvel_k_Page_134.pro
7e52517c5783b45cf9c7069ac3a60ddd
6b7e0137abaa2ea272534c12405c10303d652d08
62849 F20101207_AABEHY kadirvel_k_Page_116.pro
a5d658e4e97ab599c751cdf1e7ede074
3d9e5724ab9774beb5da8ed31684c7f2c5f87b78
28426 F20101207_AABEJC kadirvel_k_Page_153.pro
bc9ed1d9f591c377a8454d720f0d22fa
f9508fdfd2cf0ef95b6f8c1873b79e82c8bdbaa8
48949 F20101207_AABEIO kadirvel_k_Page_136.pro
2b47c04e9dd7b375be9b4ce6823248d7
07109805fe5d2c936ec706e0e29959c4b59435ce
37231 F20101207_AABEHZ kadirvel_k_Page_117.pro
72df82621a620c63d9d4859844115521
d30b8f577f807eddddeb7cc398abd540b8b70ed3
33919 F20101207_AABEJD kadirvel_k_Page_155.pro
01b14d148ccfc39f3cdccab612170685
9c317f16e7421f59ad90c8abd3c59203790fd135
32496 F20101207_AABEIP kadirvel_k_Page_137.pro
961619c093a5e08b6d830ed3211266e0
0ead419b23f6057d319af1c84b497417de7e3367
F20101207_AABEJE kadirvel_k_Page_156.pro
62dd9a3f325660bd64e575581f36909a
750fc923045e27a0ccfd98e5a14e27b8f9f7cb37
48461 F20101207_AABEIQ kadirvel_k_Page_139.pro
435fef98dccc0cead544db9eef59ecb5
ab39dae7a21899e223c38e9542708316733c242e
45200 F20101207_AABEJF kadirvel_k_Page_158.pro
e42ce99f0bb8a6e5b91c0b2caef4b2f3
06c5c02f67209684308d79f521b9260c4dc6afee
29621 F20101207_AABEJG kadirvel_k_Page_159.pro
078284fe8ded5febefc551a8ae6f72ba
bc7a7440a199a4addf16a1760a8df2bd0c42c5fa
49926 F20101207_AABEIR kadirvel_k_Page_140.pro
8ed64391429c821ce848bfd9fabf9055
6061f5fa959f13345ec3c7eaf00685c1b5afc9ed
30558 F20101207_AABEJH kadirvel_k_Page_160.pro
89bb7f1c678341a6d144be01c96cfcbd
5ab7f04c6d6b1742ddfd6ad2466830b8ac038860
47891 F20101207_AABEIS kadirvel_k_Page_141.pro
741f177b6075c276c58657a7e3c94f11
0e9f950555bcd8f52f5cda4d1f8152120245c499
9703 F20101207_AABEJI kadirvel_k_Page_161.pro
23cdff27694ebd2dfb79ef77847a78e3
e2377d3f18ee5841dd2e10c090f8b60611119127
40350 F20101207_AABEIT kadirvel_k_Page_142.pro
46324503f84892b0fe40d76b9cc1cf6a
79d1fda624523f51a4e6778edc158679d1fae267
34599 F20101207_AABEJJ kadirvel_k_Page_162.pro
8a4ea68a7d84d0b71d01a32475cba0ea
dee66d6b9d98e6d956035201d0756dd8b3dfca69
51440 F20101207_AABEIU kadirvel_k_Page_143.pro
3b91199f58113db4455449793e9817a1
84ff262856c4fe7d2ab03018f7cdfb3986976c22
37884 F20101207_AABEJK kadirvel_k_Page_163.pro
2b963081b14096ba0a52f51d466c74d3
bc5211f1b12d1eb84fe8314aabd59c41bb61ae75
48203 F20101207_AABEIV kadirvel_k_Page_144.pro
5e6c6a53751fa52db9f19cae515eb2fe
405e113775741f3cbe618805e08142c3fd25a238
35189 F20101207_AABEIW kadirvel_k_Page_147.pro
14eb345d528391f3aae49edb5ed38f58
e3cf568bb7aaabf13cf27b29471c4b87f6a04f23
72713 F20101207_AABEJL kadirvel_k_Page_164.pro
f9d9b15cabbd08f80f5139390b3e0c4e
31b9e1af7b0a58c9a29a6b51e4f9716e45bcabee
65284 F20101207_AABEIX kadirvel_k_Page_148.pro
a436e44d9597deaa0e1469722e06b147
a41b10e3d6c1e5f02490743974747e31848ebb03
813 F20101207_AABEKA kadirvel_k_Page_007.txt
7a15606a9ab1572a5b5032b06d179a0c
851137d354bd23400790eb67db601a4f12e1eb07
7954 F20101207_AABDGK kadirvel_k_Page_148thm.jpg
cb20abb14af937ae766356ea343566cb
bffa043d893e8d0c86afd1a2b3af6ebf749c7af5
79425 F20101207_AABEJM kadirvel_k_Page_165.pro
526a3b0c29d597c17f37283a22dd6e75
d55bca5fb5d7c285900ce826edb29d8db4a8dd5d
14392 F20101207_AABEIY kadirvel_k_Page_149.pro
a5e419744c47760abb813790c0362041
991867c9c7672b29d57b9c73b9e2c7fce0cc5d7a
2917 F20101207_AABEKB kadirvel_k_Page_008.txt
37550f6ab5f4f9a19161d00fed10a449
2a4d55b4b28a7e9f7f5b353a2c47d9049cea4e6d
42740 F20101207_AABDGL kadirvel_k_Page_051.pro
258d2e6d482a213e965cd63d15ce55c6
bd35a3b99db96821230ef714ae7626d6c606b6fd
70972 F20101207_AABEJN kadirvel_k_Page_166.pro
ef027947876cb294421cc3a8a0f5b1af
0d2c31362d3ed694129a6e5d5cdc8980175c1de7
58305 F20101207_AABEIZ kadirvel_k_Page_150.pro
c18f28ba03d4e161198e37549424a1e1
f6afa04e8e28b88d7ed31de2ad3c8b440faacf60
916722 F20101207_AABDHA kadirvel_k_Page_104.jp2
582ac054acfb1357889a560e9362dd76
69cf31d6d013b0e3822a200ce3b2ec45a5c85b97
302 F20101207_AABEKC kadirvel_k_Page_009.txt
40c19924ba5a443d68206212ee3bc2dc
9c0590a9a57f945ac59db8c1fc2f50b6153053d7
30907 F20101207_AABDGM kadirvel_k_Page_116.QC.jpg
b7af94635983e368d46c529cb561e9bd
530f8553f5ffc75981eab611ebb1b100a4c6f481
71641 F20101207_AABEJO kadirvel_k_Page_167.pro
1ca2c788383bf7897e43402bb2ce0c04
995c31a916f4da4ee26c5b07a66b6190ec9e6d3b
F20101207_AABDHB kadirvel_k_Page_064.jp2
1224b2aa8265ef255f7b65efe3e48536
a4a8d746d98442a978bbe5fb85b3b95395f59838
2855 F20101207_AABEKD kadirvel_k_Page_010.txt
049b7b12ea19af009d41dbab09cd8e94
67a9de21d59e50368c3027fb43750ebc2897f726
2328 F20101207_AABDGN kadirvel_k_Page_080.txt
bf9c88f88f2a16509bcf1a2db7709cf5
338bf9fce3598ef67ae04d0a36134cb53038ea3b
78210 F20101207_AABEJP kadirvel_k_Page_168.pro
aac1b6b13289712dbac8e789cb6008b0
5bafdc5aaf63a9bf84b1647af0ef4c8d66733375
656268 F20101207_AABDHC kadirvel_k_Page_082.jp2
4c02efa13400c4630fec45cdf8f09936
cad8ec7e8b7cf61f73b79f9f090a7aab2c4a3d44
2898 F20101207_AABEKE kadirvel_k_Page_011.txt
309613af3f046b654ee2d58b7c4b91a5
7f1b14d622038a649efc91aff0a06182f0a884fb
6182 F20101207_AABDGO kadirvel_k_Page_050thm.jpg
8c12a5ef7c2497534ee468a1d1177810
b6d034432e8d34d325e0abb8deb76a1876f274fb
63854 F20101207_AABEJQ kadirvel_k_Page_169.pro
7949dc6ffeb7fb2a31038554924d8788
51e185658f34ec19d3bc61afc1e5b2de70f12b81
18593 F20101207_AABDHD kadirvel_k_Page_087.QC.jpg
58ce65eabb4fe1d0058c88c6a98b94ee
8ddc41f97a7eb236383b304d5f7ea7c48a770252
2962 F20101207_AABEKF kadirvel_k_Page_012.txt
d506389519de0d6778fe8454d0201d6a
bad055014db3a04b4997a4c5d9c8e431bb5fed7e
23606 F20101207_AABDGP kadirvel_k_Page_057.QC.jpg
a5f02e04c6605d875b9a8c2f108e6d2a
bf5df5361c2efcea76445a340ddf25a4d9c34b60
71422 F20101207_AABEJR kadirvel_k_Page_170.pro
f3ea427c9c97833196486cfb0c2e8887
237e28c4f44074062c50897f5cffaf5945674e5b
2317 F20101207_AABDHE kadirvel_k_Page_078.txt
73723dbdbeeb28e0989d942128722859
4b881668106c328287053a763ad345553617c9ab
2470 F20101207_AABEKG kadirvel_k_Page_014.txt
1dc7a2521cc4e20ddcbe9b58712f8902
f2818de5bb4af6aff7252a77011fc30bc8f35e13
64483 F20101207_AABDHF kadirvel_k_Page_069.pro
891bb0a5ccaf7aa2d223c10546a7949a
379949dc6318f9a8be0d3cfe61e252e7032485f2
179 F20101207_AABEKH kadirvel_k_Page_015.txt
5f41915eca413842e796d35a7c125db3
9a2f514b38a1459a07b6fe41c0ec186270327d09
64812 F20101207_AABDGQ kadirvel_k_Page_122.jpg
469118a7696c300277c52ef919472aca
087af86e11a914682b8ca4d495df48033159df80
70368 F20101207_AABEJS kadirvel_k_Page_171.pro
665ff0cb57d95ecffbe7024cda55d199
7b487290e417246cbe13b0a066fda49087e20da4
F20101207_AABDHG kadirvel_k_Page_144.tif
c96d8d5a4de4c787edf9d38eb33cc06e
93409af7987cfd7b26e58e2ef3be2b8c8d021fd6
2615 F20101207_AABEKI kadirvel_k_Page_016.txt
95677ad4fe1d8501f647e2b3da16de8d
0527ce714f489778cea81061512f1e6f0ab2ba56
F20101207_AABDGR kadirvel_k_Page_005.jp2
20005fbf6b9db5b9902fe6eb1d8d220a
f53fdc23e2397f3161c0c3845e2ee05e1bb2ddc7
17776 F20101207_AABEJT kadirvel_k_Page_173.pro
cf7ad9a16cd8f956ead059626e95d8f7
2ef1382975f8bff5aedd845b0abcd0a0ae46810a
F20101207_AABDHH kadirvel_k_Page_039.tif
3a7385adea4ac661db0bf06d0e65bd2c
f237e2e1def913d314ca1bb747c2691f53120016
2682 F20101207_AABEKJ kadirvel_k_Page_017.txt
5a80f30c9631a60f98134a2428dbc1ee
d46c2633f8e1b2c54a54368442f6a4b10d098267
54045 F20101207_AABDGS kadirvel_k_Page_082.jpg
2108299d14edfd99104a1dd1ef74381b
714d72e2554bc2d0bdd8f8a33aa6a28f8f8c369b
509 F20101207_AABEJU kadirvel_k_Page_001.txt
8010969514959645abdb1e95eabd0795
fb86d09e41219629ca451aacc7770eab16157814
2271 F20101207_AABDHI kadirvel_k_Page_031.txt
3702749147d8a268f46d51de2ee7fe88
235c97d41745c8a7ddcf8fc7db1dcd23e7ca2883
2638 F20101207_AABEKK kadirvel_k_Page_018.txt
32f6af16b2ab77810602e2717232483b
cafd6d7de2ecaaf9b4fe8df01e4bc8640f4e4cd0
51837 F20101207_AABDGT kadirvel_k_Page_156.jpg
40174621225e8c3877a1e0201f282872
79b8e72373572a5ddfb7e6356cf200cf2b6b547e
90 F20101207_AABEJV kadirvel_k_Page_002.txt
68bf6190016f8ceb380ed2b64be71d63
9f42783a975b8e95b51ef086767914ab28ce54b3
F20101207_AABDHJ kadirvel_k_Page_062.tif
382de13ddda6c9e8523e71962d412666
e66a2abb3627487ecb381bd0bd4a4eb7b2eb2627
2203 F20101207_AABEKL kadirvel_k_Page_019.txt
06894d4558b59d63e6315c2e5c673d07
a7a8502f824e05394db21a0bd83cd7ced54dc221
50081 F20101207_AABDGU kadirvel_k_Page_084.pro
9d2034201455ba77c265302e3603bc62
95517a9544d95551e339253f6e0d9f7b802d1e46
103 F20101207_AABEJW kadirvel_k_Page_003.txt
f725df7a6b9e73672c83d04aa0bc5e36
b9752b9880a3ee4561dabb8bcca8e9ec24ac8508
1906 F20101207_AABELA kadirvel_k_Page_036.txt
80b00a4c2579d9ea07f0119e661a9e48
3b2aead6841d9c8355158d381aba74dfabff3547
83944 F20101207_AABDHK kadirvel_k_Page_109.jpg
20e0e48112a2893f2a55179261ece552
f7bccaafbe6b4d6a56cbab5bf3519c341dd7a414
2927 F20101207_AABEKM kadirvel_k_Page_020.txt
bbe9894bb6fef962cd0251b1cc72514f
e3a36c85b504be4d1a5707ac233c4c59971fb0e7
5444 F20101207_AABDGV kadirvel_k_Page_025thm.jpg
c0bac583a8576198698daa7449bc642a
38b413c21abfaf5836c7549d51952636ce3f1d3d
1824 F20101207_AABEJX kadirvel_k_Page_004.txt
b0570c07fc09608e52e46518daabb394
8545af8b2bd3e41e323f02bd61f345016df29777
2680 F20101207_AABELB kadirvel_k_Page_037.txt
a58258037d60b333fd9a05fbcce97489
8d9e50791f04751ca44a35ec36ab249256bffa87
F20101207_AABDHL kadirvel_k_Page_146.tif
63c4f6ac4de9270b8bf545b8789255d6
40cf17887ed583ad8ea62dbfe5977db0f8c05a1f
2419 F20101207_AABEKN kadirvel_k_Page_021.txt
088387ddecf5aff9124c1f4d26f45ecc
b762c5b3463805a2ca33f3034247961b89779822
24470 F20101207_AABDGW kadirvel_k_Page_114.QC.jpg
93a6110d137810da4b9f3c7499d25b2a
649a153d5c65357751c266f9df62432856e3371c
3040 F20101207_AABEJY kadirvel_k_Page_005.txt
062f36d3cec3a7ce086d4c522fbbed34
f8de856f38de6e82acd8b45d503760adbde03c74
2430 F20101207_AABELC kadirvel_k_Page_038.txt
0db8f51f536f9fdc96e2ac1aa4837716
a6a3f6054c05f10463d182cf448c6a27d7127f08
68628 F20101207_AABDIA kadirvel_k_Page_043.jpg
990a47d79c0aabeef0f80fdd8755863e
4681d71fa6cd320404cf183099c262ca2a98152a
8182 F20101207_AABDHM kadirvel_k_Page_170thm.jpg
9fe674183c2ba95aa7c9e6aecb5dc6ad
5f5c4a35a393d1167709cdc6159024dfa7ec24d1
1553 F20101207_AABEKO kadirvel_k_Page_022.txt
70661825537f6ef29e160032f5fd9e95
2da0daa61e3d217eb8daeb0763e8cdcbbad3856c
36836 F20101207_AABDGX kadirvel_k_Page_130.jpg
771cccb9da76c5288840fc13696ab8c2
291ca4a9a032d28820e6b79757ad58b0201c7a9c
3043 F20101207_AABEJZ kadirvel_k_Page_006.txt
dd23b75c8c17d58520a5b9717edb3767
624dd0a85cb08b116622c5f242211d686a0d29a3
2536 F20101207_AABELD kadirvel_k_Page_039.txt
0b027d70cbb63e0a0266f8f24b80320d
02b5833af63348c5f9e634353ad663e1cd0855a2
34348 F20101207_AABDIB kadirvel_k_Page_091.QC.jpg
e5e343e8958c7666616a148c138527e6
eaa7c149bbbe3436694bccbce0d2fdff66aee61c
7586 F20101207_AABDHN kadirvel_k_Page_009.pro
6e5499a21c02c0b2b252afeddfd6a0c3
8963221311faa15db4a4e08bf801185dc4e91002
2448 F20101207_AABEKP kadirvel_k_Page_023.txt
70a79059b66d7c769bf16224648effc6
231ee32f0d56782cb1fba965c72585f5a05353bb
8691 F20101207_AABDGY kadirvel_k_Page_090thm.jpg
63792cda198c3578d356c0b34d0944ab
18604f0e3eb4433c24e81b5fb7f7d5368c6e50b4
2510 F20101207_AABELE kadirvel_k_Page_040.txt
203c018ad2cea7bb548716869c9eafc9
20c146a958de7ef762b3043566b5fdbaa816ec90
2065 F20101207_AABEKQ kadirvel_k_Page_024.txt
6e0d94521b67aacd2ea81749771baa30
626877634055f002cdb7c32a57a18557e7092ebd
6700 F20101207_AABDIC kadirvel_k_Page_126thm.jpg
c1d6f811a517a357b3ab252c3809a396
4d0acf8bb3666c96db73166b7c2aa845e25c42b4
5438 F20101207_AABDHO kadirvel_k_Page_119thm.jpg
5b92f1da34b1ce8320ac7dbc2e2508f3
16d46a7af4ad1a42c0572f5d127aa4673683951b
21936 F20101207_AABDGZ kadirvel_k_Page_054.QC.jpg
9678a585952c4a8df96ea03d8dd529fe
f7a26294fe75456c86a08f585dd5fe696ea35a62
2635 F20101207_AABELF kadirvel_k_Page_041.txt
42beac6df9e51983d7844b13f2672b12
84da4271bf1dacef6a25d90342041428df4ddaa5
2098 F20101207_AABEKR kadirvel_k_Page_025.txt
f555d086c27ee5cae2d6c5adab831b0f
4bcb8f472263f7dc4fc142b503135a34da4b9abf
6124 F20101207_AABDID kadirvel_k_Page_132thm.jpg
f8a48a8c95d3794c993c0e70365c15ed
2133c2bd40db786b263a5123e525cd4758b10e28
F20101207_AABDHP kadirvel_k_Page_084.tif
cd9cff35a1cd53d6716d1ad0954b528b
6317ebcae1a117a4ac73f9026f6882801e1d38e1
2787 F20101207_AABELG kadirvel_k_Page_042.txt
247a8225dee2d00fad6ccbfe025a2157
2d691483a5d145435531cf9d13c71c10fbf41c49
2242 F20101207_AABEKS kadirvel_k_Page_026.txt
948cd7530aa168ae2f89df8b3295bfec
7a5dfd471dac100ca988c85f3efcaff63649cc2f
F20101207_AABDIE kadirvel_k_Page_133.tif
ade27a179d852f8d506863ad354db58c
d91158478a089bd5c71261640f9e9916e0aee978
25245 F20101207_AABDHQ kadirvel_k_Page_038.QC.jpg
884d7a9ff067edcecb4fc20e2442735b
8e26e685797ba39cb033710e00e698ab91a43326
2342 F20101207_AABELH kadirvel_k_Page_043.txt
bcd8fc0363296215f52acc9b9d203ce8
e3c1011386058b77583e241af4e1ad2c24e37997
65364 F20101207_AABDIF kadirvel_k_Page_064.pro
9fe73b3a4bfce4c356e4977c6704b95f
409895a289b863f5a81b79a6d907cff78233e8e6
2319 F20101207_AABELI kadirvel_k_Page_044.txt
1157009e287b8b65f14cc2fd42a55990
323856f76bb7c881ee56d8d2ed1874f5403b559a
2425 F20101207_AABEKT kadirvel_k_Page_027.txt
535ace2d299751e003e0f5bb015bb63e
9e6556fbd02c68defd6af5948c11b1e34bf6d475
42692 F20101207_AABDIG kadirvel_k_Page_029.pro
ad4aaffdfcc4d9d8ea37cbcf4710f0a4
4056cfa27660ba94de271a33b1ca6e7c515c1425
1830 F20101207_AABDHR kadirvel_k_Page_160.txt
5796d7f430741f0651488abbcf6a4c73
a7c35f67840e63ec56f11903c78a0e9ed0eb842b
894 F20101207_AABELJ kadirvel_k_Page_045.txt
1cbcaf2c1fc7c04675c59d50065c3b31
9709e354dfb04a776d7fb310b2624a82b81a5f5f
1710 F20101207_AABEKU kadirvel_k_Page_028.txt
0c52af2d9fdc687fbecdfbd8ab5c7e04
280297e0964e68aff54a07cd38883e050ea13040
73793 F20101207_AABDIH kadirvel_k_Page_100.jpg
d288abd55357bff044f949ae90c88562
28aeb7d0bd94a06f74da72c1b36507ca02e51e40
2702 F20101207_AABDHS kadirvel_k_Page_029.txt
96d068cdde2c41e73f7ef23f8e9927a8
3c74851f7d20115790f02c57b7b34db6e604732a
1267 F20101207_AABELK kadirvel_k_Page_046.txt
aa8347613bc4c2a7a4214c6650936589
165335bb1138e60047971949d03538b4a50b044c
2535 F20101207_AABEKV kadirvel_k_Page_030.txt
dc33b6d5b6a69f5fe52bc1184c8a3046
5a6434c3fa2c55572e21e16c378a859651e1229d
7106 F20101207_AABDII kadirvel_k_Page_033thm.jpg
9399f62ab634408bf55be497d07356c3
9a6c3d3fb0053ad2293f983d52396e07d7df1152
F20101207_AABDHT kadirvel_k_Page_128.tif
629babf77b47b64a7ea04bf3611dde2d
d3ba9c18f78a970ed9ed33af7cbbe5454b818759
2604 F20101207_AABELL kadirvel_k_Page_047.txt
7c2863a775ea302f62a18d2aae34c3ad
3abfcd93d1b611dbe9b64e14e1dcd6bdbdb4a2ca
2521 F20101207_AABEKW kadirvel_k_Page_032.txt
0ce71b9d771c135ae18ec9c333ccd45c
4d4a0d918981eee167dab9bbcac4a40fb5caf380
79250 F20101207_AABDIJ kadirvel_k_Page_099.jpg
2cf0f21bff0a3da874e904ad9efc0dd6
b421760e5efd20352f3c3f20db154924c4c1957c
5448 F20101207_AABDHU kadirvel_k_Page_154thm.jpg
dbd23a38b609746a8f20ebab1a3c88b5
0ded012397122a520a246aa933e120b433334ea6
2570 F20101207_AABEMA kadirvel_k_Page_062.txt
8904b183215390c8e42ce33d75e9acd8
9ea66095a0d5413d1346f253f2cda2106e7f1ad4
1917 F20101207_AABELM kadirvel_k_Page_048.txt
bfd80abdc9e4147295e7eed8a4f9137f
1ac681831c626c22310f7e349ea309b6a2b7b3d2
2310 F20101207_AABEKX kadirvel_k_Page_033.txt
e93351d1c92a2bcd333e36e65f7769af
cbaf0025dd3aa48f5b81c9ee7f2970227a52ec84
59246 F20101207_AABDIK kadirvel_k_Page_163.jpg
7cb6b1510c93e79ac3a6b946b65ae672
4a8a3dda2e3df4352903bd930985da10abb050e9
71651 F20101207_AABDHV kadirvel_k_Page_010.pro
54f6fc0b53c18bf623cabe27d5ebf269
e14dbbdd3822e48cbd3f41cd8437086bef47086d
2562 F20101207_AABEMB kadirvel_k_Page_064.txt
f78a7f8afb4a43259d2c21d2592052fd
13c526904c6a18531aa63087073c450b713fe69e
2721 F20101207_AABELN kadirvel_k_Page_049.txt
f91ee6d90e20eb5b4fa6c1e01ce7fcd8
e89b8282e04792004f2968ab3b2ba69203739f5b
2066 F20101207_AABEKY kadirvel_k_Page_034.txt
d3762d4de2c48902bc6722c8052cc242
9c00e7e90d41fcc7266d12eea23d23b3b38d6b8d
43945 F20101207_AABDIL kadirvel_k_Page_057.pro
cf590b4485a31d43a287760359bcebec
20cebd872fd8a534f972095abb1f7c7a1c013d8c
6503 F20101207_AABDHW kadirvel_k_Page_037thm.jpg
d4304078705b91d7cd6b85ca047915a7
27c9b959a3366f0b59ed9f3133c6000ab1e2dc34
2559 F20101207_AABEMC kadirvel_k_Page_065.txt
3191cea6bcfbafa2ac13926595aa32c4
bf9d8e9308c10e7f5606adc82c2e3579272d84b6
2248 F20101207_AABELO kadirvel_k_Page_050.txt
3580548c6d55bdb3f991b03809954fc2
1b8172c37a5a4af870d5854b77ad401db78b5b90
2172 F20101207_AABEKZ kadirvel_k_Page_035.txt
cc860f461f86db2f98ac4f5c6cfbc4b8
92bdb8ff9f6f6f6b2af2033b71ebcaab99d10697
7261 F20101207_AABDIM kadirvel_k_Page_136thm.jpg
e942f038e5d4750868db29ed5b2ebd69
3f27a14c0c13d7f0a293005209df98034c7dfcb5
68229 F20101207_AABDHX kadirvel_k_Page_031.jpg
726a7c4637e380a11af24452541a17c0
73fdd7405fcbce64e3f672728ab125df9c657978
108225 F20101207_AABDJA kadirvel_k_Page_090.jpg
8aacc5aa0611ab62e3ad1fd16749f782
187d42759d79b9bc67e6838d2ebb86bf9ba9bbda
1230 F20101207_AABEMD kadirvel_k_Page_066.txt
e8793aa5f1938f2390eba1a7ec663024
e9a57db994bd1da26c1e7b2c97e7df9462458b48
2282 F20101207_AABELP kadirvel_k_Page_051.txt
afab8b15f1ff990cf39747f90dc0bb8f
3f8de0244a978c289a51f0f06be1ae675affc990
F20101207_AABDIN kadirvel_k_Page_006.tif
3d3d4c383bf8d234f81f8e0eb3c4b4ea
f695e439eb756a90d531690f0a44609ea2456b20
20596 F20101207_AABDHY kadirvel_k_Page_113.QC.jpg
72c14023d4f2d82db4105f8f007c6921
2edbac7a98222b4a1618728b7e63a4aade74a0e0
1087843 F20101207_AABDJB kadirvel_k_Page_145.jp2
6ae8719113f46aa0a82ad723f48d137e
9beb0aed95586ab3b33567e007954beafd4300da
2480 F20101207_AABEME kadirvel_k_Page_067.txt
e2ff96ea349024a414cbdf56244fc082
6149349288971ce371095297ae1a93bd7a92714f
2961 F20101207_AABELQ kadirvel_k_Page_052.txt
bf82548f2aec7c41cd23d3d62e662df1
2f99d56f82caa398fcdd0e9d0e9429d7241b0008
8516 F20101207_AABDIO kadirvel_k_Page_091thm.jpg
1b6271750d7cc048c07adf39351b9558
e81875d084a0225db9b4eb456425f0e8adf28349
1966 F20101207_AABDHZ kadirvel_k_Page_101.txt
009d5e22d7587bfe0d0ba1937a3b495f
c4c9e1f52a1a5bd97ef36bc26ea69fc8827de01f
35490 F20101207_AABDJC kadirvel_k_Page_036.pro
3fec8b059f49639166261dd15a3fa608
26674cd9a30cad713669b7c91ee879cad9ad28dc
2688 F20101207_AABEMF kadirvel_k_Page_068.txt
8cf8752eead86b7e09c2f384d07b7ad3
e24c7ac8e498921d572f0fb1639dd33cb598f54a
1273 F20101207_AABELR kadirvel_k_Page_053.txt
3a28d5f293c0e1af2248d158c287d00c
1a915501466ea0176c416b7f0aefe8683c9b3c4c
41128 F20101207_AABDIP kadirvel_k_Page_097.pro
7ce419e6b193c1604889c6338a4f7649
706a255df7d179354290433939b885d798623ace
22251 F20101207_AABDJD kadirvel_k_Page_108.QC.jpg
683f56776e628007a477a6db687fe292
9d3e57e8bcc3cbf6e8a4dba9ed52c6f68845dafa
F20101207_AABEMG kadirvel_k_Page_069.txt
99091872d1ad494e8dbbc9d6b48bd1cb
4a8f8eb788a78a2c0a27f1a384b243d66388f525
2097 F20101207_AABELS kadirvel_k_Page_054.txt
62b70985a6af9e75761d2604676974b5
84fd6b3cd165990d4df68d8bb9bbaead87ec0248
18670 F20101207_AABDIQ kadirvel_k_Page_163.QC.jpg
2ad99a15d20ead0d6101bd7353df3896
eaaf0d5466ae532c478e0a42f89c9f9a07e6cab6
36149 F20101207_AABDJE kadirvel_k_Page_135.pro
24cb56decc211c69295243accf76d124
6c8546075f894c17b0159d5ed05256523a6fdbc3
1560 F20101207_AABEMH kadirvel_k_Page_071.txt
f2b94ba4b0c5e7a4fc8bbdb2f015da04
f65e6cbe12f1dbb760c323cc72567a88d9c23cef
1950 F20101207_AABELT kadirvel_k_Page_055.txt
a80b34f356cb600c5ae03d95eea34cc4
018cc31a08fb1b557de1a849cc86fedb492af622
3461 F20101207_AABDIR kadirvel_k_Page_130thm.jpg
ef1f318e46e169fe911b9d066c98aa69
d6429e5223be4610102bd3f392d315f3ad369d4b
2963 F20101207_AABDJF kadirvel_k_Page_013.txt
de0060c259f5a071f08aed41dd90b65c
b17da2452549eac9472da310ab10be7abe083f98
2915 F20101207_AABEMI kadirvel_k_Page_072.txt
38f5eb5094b27b20c259bf2411949b08
3211f76f72daad31d8bc089c02ffd856399f5865
1482 F20101207_AABDJG kadirvel_k_Page_138.txt
29c35d3aa876d994a133c2ce4a36a541
3f7ae1005c1597d34859de5ac157e53dafaafdf1
1840 F20101207_AABEMJ kadirvel_k_Page_073.txt
6fda5d6c053f1922ca2b9af44695d48c
4166f4b4ecb0e26350b058e070321384fd997e03
F20101207_AABELU kadirvel_k_Page_056.txt
7eed7b7ce9ee7021d42162d70c82fa9f
a635ec0a16d045b7ccb1a383732caed642665565
F20101207_AABDIS kadirvel_k_Page_076.tif
c648fbd7a595af56bb87878fcb25ced2
976d01613e775822d556b290d1a40e739deeaa41
F20101207_AABDJH kadirvel_k_Page_143.tif
0fec2feac06fb256b6b0528410800ed5
d9cc02fa94bc0369ef46288da6fecd7b77f1be0f
1909 F20101207_AABEMK kadirvel_k_Page_074.txt
24cd62aeb0621f7923bd4e8b98495e93
4492aac894579fedbaea6f985f30f8c8a99423bb
2377 F20101207_AABELV kadirvel_k_Page_057.txt
1916025bc769a71a7153ef1f7deac350
7042b990915e76f05965a453e16ccc7983f4035a
F20101207_AABDIT kadirvel_k_Page_139.jp2
548d20d781809b06feb6ddd8791e957e
1612ecfe585b2367cbc9b23a925f353a0414d587
F20101207_AABDJI kadirvel_k_Page_048.tif
314e31fabdc47bac47eb4003f10046f5
b2890bb109a15ab12c4edf814eec562708efaeac
2281 F20101207_AABEML kadirvel_k_Page_075.txt
33a34f8daa6f7a8a09cc28ccb3947b0e
1d699619290aed8341f6e31abfc8336240c1f33e
F20101207_AABELW kadirvel_k_Page_058.txt
533b5f73de25633d3f1871b67842111b
e6323de6aba6f9f23242ea43b895cc623cf5daf1
5039 F20101207_AABDIU kadirvel_k_Page_028thm.jpg
f9ba4bec3f6de3f77a33c5b0bd8c6587
d4aaac9b7523833dc2c162e5ce8870242778949e
F20101207_AABDJJ kadirvel_k_Page_108.tif
8e4544a22bcab8bcd45852b6040630c3
a822a36320ba56ed27a7542d875bf330c1051b05
2714 F20101207_AABENA kadirvel_k_Page_094.txt
a8b4e371f0f61b9baa2d98594cb9d544
60796097e01b71e489e1d82667322f03a5a091b5
2429 F20101207_AABEMM kadirvel_k_Page_076.txt
ee12b95148809d2b028b0adb677352d0
dd5e1adfd5e89f9bc54b7740504df0c59d69ae6e
2484 F20101207_AABELX kadirvel_k_Page_059.txt
ecccd1f56725041bc2eeea96cc7dc464
877ebe245811047e65bc363759ce5978c2d6f54f
F20101207_AABDIV kadirvel_k_Page_001.tif
31380d7edfe78b372fe19934893fa156
08ab37b0f1becaa494772ef2d3a73835166b738f
938942 F20101207_AABDJK kadirvel_k_Page_050.jp2
70b6adf026d930e4f5dbffa257eb7e6d
d4c688d7ae5f92e58cedaac298a0f39bd960c705
3013 F20101207_AABENB kadirvel_k_Page_095.txt
d9711ec84700779e2f35107cd8534ef3
df68223ebd573e5a12d2758aa8923165468f6bc1
1441 F20101207_AABEMN kadirvel_k_Page_077.txt
f2711fcbc5183b27a9cf1ee7edf56136
1522d9d09f72ba70b0c0345f0c5e6a7ef5bdbad7
1494 F20101207_AABELY kadirvel_k_Page_060.txt
ca2d8cd45139589d32b1f127b31e237a
5f58e3d47fe4af65f251e318508188380a2e17f4
74410 F20101207_AABDIW kadirvel_k_Page_057.jpg
1375329f5b937e2f4be04337c2ec92ed
bea15b9757daceb25bd1d6087799bb447f53eec2
29415 F20101207_AABDJL kadirvel_k_Page_046.pro
a764a9184edd61f676052814b2e6e2ed
657fe9c93e22e8f2f77ecf2595dcbb7727c106d3
1411 F20101207_AABENC kadirvel_k_Page_096.txt
f407151d1fe73625fead3fb6997f38bc
64761fbb2f800c01510f5c6f99e19bb2b0d3188d
2789 F20101207_AABEMO kadirvel_k_Page_079.txt
eb73f13a874633a35a5baf87a04edbad
4a4ad095fab89f719badaf09aebf9fc180030017
2665 F20101207_AABELZ kadirvel_k_Page_061.txt
261dc28cddbbce273ca7686bb5d2dc92
d12a7c582bc212d1118160327ece3f1b563881d9
49656 F20101207_AABDIX kadirvel_k_Page_033.pro
c81f33c7539bb011cf0ff6841a20a8de
93ac781951f6da6747d3b2b53bf7aabdd355b421
5642 F20101207_AABDKA kadirvel_k_Page_003.jp2
470ac63846ca735c58decd45ce81d7ea
3ce5d174f9a88f9ef523b0be2bf9044b2017a914
33609 F20101207_AABDJM kadirvel_k_Page_146.pro
bcfd2761c34fd87ddfbbbde58cc62101
e16c34613b16ef8c98deb628dcbddccaef543233
F20101207_AABEND kadirvel_k_Page_097.txt
aae851a1a346f8838efa94e134831b02
b408106fd1fd15b1beac65f12c28d1420a13e0f2
2100 F20101207_AABEMP kadirvel_k_Page_081.txt
7b5bfbae2a8ccee819810e1928c3c506
5745bdf28c3d567dbdc0a6f9377f61084396ffcc
F20101207_AABDKB kadirvel_k_Page_147.tif
84d2708c260aa4e9203953f22fb6ca78
2863493903f4317053d06700a809a9231679aed9
1355 F20101207_AABDJN kadirvel_k_Page_137.txt
cff527f07b48cb62fb6be705f1ef9482
d68d8b1f8ac68c6b06ae65fdf40589c173e11c65
1087855 F20101207_AABDIY kadirvel_k_Page_080.jp2
a1ba0d860940ed80f193d9df1597a207
8542e09a9946dd591b409917397e7bc192cc6058
2389 F20101207_AABENE kadirvel_k_Page_098.txt
3a968bccc921f11b2e69a8b4a6250961
555c86f409ea2e70b16832f6ca0ccfe0408394e2
2450 F20101207_AABEMQ kadirvel_k_Page_084.txt
865b2d73d54c9336a35623de9ba6c2b7
a2af1dee1a039cc068f2100a7be54b157da8aeea
66623 F20101207_AABDKC kadirvel_k_Page_029.jpg
16851d7a462d9249518898ff874ff8d2
df097494365fc80aeacb23982eef25d1ae77ab9e
70794 F20101207_AABDJO kadirvel_k_Page_058.jpg
aba828e6ac27bdeb484df58c0bf14180
5fa3b23b5e83593bdc61b5fcdd5bf15b7cb1ca1f
740771 F20101207_AABDIZ kadirvel_k_Page_154.jp2
0d0e2e80d5611dd5a290188680383d21
6034f60bce4aa6da07d31b761fd228e705994b56
2308 F20101207_AABENF kadirvel_k_Page_100.txt
93c729b77bc38d524e9336be13d89b9e
565d2a549fa98bc60b87e8730a0eb41667c22e9d
2616 F20101207_AABEMR kadirvel_k_Page_085.txt
40d9f7532cec384d3f1a8e1098bd8c91
4f8e169900c5e76c4d1610a474f5fef6d91e9c6e
5990 F20101207_AABDKD kadirvel_k_Page_125thm.jpg
b98f6cd4266b3c76c82144482ff36d1c
cdd01c731e2e88c446f69f4f25ee686cc1274910
21403 F20101207_AABDJP kadirvel_k_Page_029.QC.jpg
98b2904fba336901b1b525f5d6efcb89
881c4029374f1d117acdc931d7a1a4932f562bc3
1303 F20101207_AABENG kadirvel_k_Page_102.txt
767d8589e433d6666590e9cd86742e48
ad38f2c583848ef84a33a39d7d3cb6535c0407d2
F20101207_AABEMS kadirvel_k_Page_086.txt
d49105f144a1524cbf24059b4bdc51b7
b81a730cb3fea6d5971fa9be87273ebdafc36a4f
F20101207_AABDKE kadirvel_k_Page_064.tif
ec9f87ba2d2d92cc6a720c2980be92c7
6aa94476182e0306101329a99812d4ea705f937a
F20101207_AABDJQ kadirvel_k_Page_054.tif
260a484e976081b444cf74223b7f339f
ef60739b3714b3153ceab656e8198738a40364c9
3018 F20101207_AABENH kadirvel_k_Page_103.txt
de2b2caeed63d4b8780114b55fe31a40
e24fd317b9350aecbad8ed993402ab6bbfe4bda9
2090 F20101207_AABEMT kadirvel_k_Page_087.txt
90a4a9c4ceb7cb49c8bc6899d7248467
bb597f9c3631ef15a80e0753f2ceb6abf1b8f19b
10400 F20101207_AABDKF kadirvel_k_Page_007.QC.jpg
bfddde23013618f0f20adede1594d8ad
41f1c3b5303bbf38737b63f33b74ab9b52f2d29a
F20101207_AABDJR kadirvel_k_Page_036.tif
c64a881f218affef771bdd5e8850c8b7
7e90c48377aed8f87ff9be7b72bf2e6668b32e16
2103 F20101207_AABENI kadirvel_k_Page_105.txt
ebb1471707a63d6eab8ff97266987db8
7afb43880c089f7bf0d4f56c6408bfa48a007dd3
2326 F20101207_AABEMU kadirvel_k_Page_088.txt
b8bf7f4f470b18a609a0c078b174ebf4
dfed90e761a4735d024da26cd843f5e26ca92f42
112797 F20101207_AABDKG kadirvel_k_Page_006.jpg
06c9c2416e1bfea6eb8d0ad232d8585c
aada030da8c377d7552b355983ca15250f7efa9d
68074 F20101207_AABDJS kadirvel_k_Page_090.pro
8c8eeb74a654f2a53a3eed4006db8de3
ba4362a5ad81c0764974fbd5864a029aa94df58a
1793 F20101207_AABENJ kadirvel_k_Page_106.txt
cf7115263dcc4e8444ea1f21740ffe6d
e8230d2b62e64196bb2b0ff3231abe631199c61c
2338 F20101207_AABDKH kadirvel_k_Page_120.txt
6cd3024a783879dd9d4f37b440c333b7
c790cc82d1617279f4c966f33a26edadf728a277
1200 F20101207_AABENK kadirvel_k_Page_107.txt
d2fa80c527d8747c6a6685918087c3b1
2b8f48abda915e17974f4e3543ccc04b4ff82053
F20101207_AABEMV kadirvel_k_Page_089.txt
b7743ae7c06e0860c29065442266abf5
06fd9c0941b44239633f99caab88cf07f903d953
71245 F20101207_AABDKI kadirvel_k_Page_147.jpg
09cc1015149e9f444672bfc990a9b0a8
6987455954a945b489c0fd83363009e80fac9092
60915 F20101207_AABDJT kadirvel_k_Page_072.pro
111830084cd1396502324ad937cbba4b
6b10439f510cc7c5c0a7a3550367ccf4d6ecba07
1290 F20101207_AABENL kadirvel_k_Page_108.txt
36a55843bb87d79bc6b0be1ead6ab1c6
f82fdde0c1803ab42893381932a0130798b89ce8
2726 F20101207_AABEMW kadirvel_k_Page_090.txt
390604255398af9f268052cf2808b860
0f3f81d023714db87432a6f8484738bd1e240de9
4757 F20101207_AABDKJ kadirvel_k_Page_015.QC.jpg
d505375ed12d865207b2dfbc0d66f902
7885faaa23df7ce0c80e17cd79f15feaf1f11a4d
68437 F20101207_AABDJU kadirvel_k_Page_017.pro
27534c91f5f94968acbd262b8773c960
a75e5595be45de64a4b25dc4e5ff982415e05674
1949 F20101207_AABEOA kadirvel_k_Page_124.txt
dfd1fd3e50e272c830ccf2ac99f4fc11
d0e90394c982cd6f03f21218391fd0d7fd60e77f
2497 F20101207_AABENM kadirvel_k_Page_109.txt
5290ebffb9f1e7be517049df5f464274
878e0034a550a2a16a8c08d21ecc6a2e49c9425c
2661 F20101207_AABEMX kadirvel_k_Page_091.txt
4f20f3b634b02444b30240367f7cea77
5e3dee2592094202bd7ae30ad621898cfe2da853
16630 F20101207_AABDKK kadirvel_k_Page_157.QC.jpg
c8c03437ffc39b5bbcafe896f1e2eb97
68e0595e99cf4fbc24102d3fe6242f6e90c1f6c3
44645 F20101207_AABDJV kadirvel_k_Page_021.pro
1937aeffd49ce159b9eb2650beb0c15e
645aa4e04d6d941af78d6706c7742df0a7269432
2154 F20101207_AABEOB kadirvel_k_Page_125.txt
a376e650ce4c9437dbfc02e9f0ebd083
17d07c8d92a3235fa73e93f2be669e8f9bd3432d
1533 F20101207_AABENN kadirvel_k_Page_110.txt
ebb08aff3abba38796e7ac1f47dca855
c07ad4c4daf1a083fad06eeaf3994f4811556750
2095 F20101207_AABEMY kadirvel_k_Page_092.txt
01c097f6a9b8f2d4210f39a009d5b1b5
1ba6c060ccd8fb09de72b06c99d8bf6e2d7c34a6
F20101207_AABDKL kadirvel_k_Page_121.tif
0a7dad769a6302f5a9a725c5f4866b03
7cf38d20b802404a18e79f8081b999b2bb678020
943237 F20101207_AABDJW kadirvel_k_Page_142.jp2
9a440bd0ad92731bc476353aa43a5471
463876e6f8042fd9bb4f1ee665a7f334ce614fd7
2731 F20101207_AABEOC kadirvel_k_Page_126.txt
9c9ff358fe0e6b7978d4112aa457fb77
6b5beb5ba7d1ca7ff938992375b75b88bb74103f
2093 F20101207_AABENO kadirvel_k_Page_111.txt
71aa7c671f535f9e14044995090b827a
9f3f30bd5e2475ece34065a7b890fcdb308740a9
2575 F20101207_AABEMZ kadirvel_k_Page_093.txt
2c13b4dabafc0c337340c6c8f6750d52
5ad749f7cccc7e111baa7c19012ca43257e91dd1
2004 F20101207_AABDKM kadirvel_k_Page_070.txt
7ee79b4200e4916c76b25abbf29ded56
247c3729004b5677e48e208d7125e9353c6a9cb2
5403 F20101207_AABDJX kadirvel_k_Page_046thm.jpg
77011afd36f928a4b3d230a447605c91
5af2ae6fa6d8ff3baa4432d89f78c9637549090a
9370 F20101207_AABDLA kadirvel_k_Page_001.pro
46a6bf7740c709491551dd5274818c6f
b0845bd6eb34ca81ab3d3aeab6dfd7d8f221f556
1795 F20101207_AABEOD kadirvel_k_Page_127.txt
1c4312d5ded444ca91c624e770cbdc34
d78295a148321770f5eade1418084a6c821efd94
2270 F20101207_AABENP kadirvel_k_Page_112.txt
ed0948b8474eb6040cd02a9e906f07f0
ab79c2f707ccebf8dba91d566816a391e4d70077
82241 F20101207_AABDKN kadirvel_k_Page_120.jpg
d4bb7fd1addd982451bfa5e99cf4fc01
e117eebe584acee54ecfab4d7cb41934c1a5440e
79319 F20101207_AABDJY kadirvel_k_Page_084.jpg
88166cb40f8af2cf529d8160672dd355
b4a932da95db6a7fe34d38737e3336c4ae5315bd
35701 F20101207_AABDLB kadirvel_k_Page_154.pro
c3dd808a9cbd104deb9c7afd236b2745
699abc9447ddfac36a19f45808b56973b2f0b35a
F20101207_AABEOE kadirvel_k_Page_128.txt
7640944b7b3f9b0f9d4f4c9a91988123
763e935f63bb8e0b15bd6c959bfe855f7f996955
2051 F20101207_AABENQ kadirvel_k_Page_113.txt
e0cf1bcb196d5dc50aa6db7b644febe2
b69ed9352eb84b52f5e9eb9adf64ef1b5d73d1e6
F20101207_AABDKO kadirvel_k_Page_082.tif
a44bf8b4b50da4156f7a4644e1bd7d46
72de23e6b9296687afa94ba5e0883c88a12def7f
10685 F20101207_AABDJZ kadirvel_k_Page_002.jpg
a99be1b1fa8dfbbfdeefb849f1dd61e6
fc4774f381cf19ab4ed4ea2dea9a98ebe84fb966
7058 F20101207_AABDLC kadirvel_k_Page_026thm.jpg
4fac0b2450c2974feee4150c4835d93f
eeb1bbfd84477a5d867047106f886c9be7552500
1902 F20101207_AABEOF kadirvel_k_Page_129.txt
1d1867cbab4331a50fa83764318c077a
4f5318fee4c9d56ef1d2bb29d2ffc2c0416b7a49
2720 F20101207_AABENR kadirvel_k_Page_114.txt
f003f2805c69c84536ba75c6826dba73
93b3ee889eab639f59a02bdacfd8d037ed56539b
34537 F20101207_AABDKP kadirvel_k_Page_173.jpg
7d9061ce0b8dd885d46738c089954f34
cbe40462f161d9ef6e718c374307cda5feaaad05
F20101207_AABDLD kadirvel_k_Page_116thm.jpg
afc62f7098df40b7b03b880dcd960f5f
1e55174c04779adbde41ce7566e40f2dc173a7d1
866 F20101207_AABEOG kadirvel_k_Page_130.txt
77d45b3a2397a680f2c63f8467499f28
a5247106146d6af504a1a7302223f0dbaf5c1d59
2027 F20101207_AABENS kadirvel_k_Page_115.txt
539bcfd32123ad5456132fbd27c72012
de1b9336f6334579e79d0b9ae5320de7f1329613
31175 F20101207_AABDKQ kadirvel_k_Page_127.pro
27b756dc0c8ab59d6bbbdea81aac5273
edf96200e6a93d73b021d2784564361d13b5a55b
52530 F20101207_AABDLE kadirvel_k_Page_118.jpg
7ce231b36042d858945ac2e0bc3489e0
ac291eb72bb0f303eeb6a3454b5e0ede9bca8c57
2060 F20101207_AABEOH kadirvel_k_Page_131.txt
8551467900ad65a2b672d7b9c4586933
331b05fd90a7c7a87c293c669baa1dea9ab43157
2532 F20101207_AABENT kadirvel_k_Page_116.txt
9a02d43ce6d92f5592c2d9ede4bfc78a
9bd8d402be9077e69ef8e2494eaba1f1bbd8b4c9
45567 F20101207_AABDKR kadirvel_k_Page_145.pro
98ce8d3546894a4e6274338caf4e26a6
10f754755a3efb2c0aa8f8452b97eb3498a61952
5462 F20101207_AABDLF kadirvel_k_Page_138thm.jpg
58eba992a3c35ff8622c3afabeec43b4
c9d2bc69624b2409e0e7aeaa6a282f678b6b5ce2
1310 F20101207_AABEOI kadirvel_k_Page_132.txt
fc69a8534779e534c917c90cc47e9ce0
65624beee281c6a37a2bdfdce437869008f28277
2087 F20101207_AABENU kadirvel_k_Page_117.txt
1916a1ab6ad34a52d30cc7f5567895ea
13b5c12087efde60672de680564d443c6a9a3162
F20101207_AABDKS kadirvel_k_Page_012.tif
cc72323463cd5fac6e56e506abbdec27
02d060871c15bd1af77d61da588e6eb9f2a1c8d8
27101 F20101207_AABDLG kadirvel_k_Page_026.QC.jpg
e34978ee7a73ade77b8064d8e01837b3
415ebb032f5aebdbf300836fed4bfdeadf0cb037
2574 F20101207_AABEOJ kadirvel_k_Page_133.txt
37923d09a1f51cf4eb7203e04c3404b0
26d07f7e445459f02033a7ea8fc615f2565eda24
1611 F20101207_AABENV kadirvel_k_Page_118.txt
654fb73af282558cc93404b7b3233728
a480bd5ff56687739d0ae6188c38efed4663b8e0
F20101207_AABDKT kadirvel_k_Page_078.tif
e6d5377fdabc7a6ac26061bcdd17d527
f3e5f93802fb86fda44e9ded8f96afee30edac83
76854 F20101207_AABDLH kadirvel_k_Page_103.jpg
209c661bb45bf422a029d9c463d732ad
a463d9f9ee69b1da1c18003c718b00f12964044b
2063 F20101207_AABEOK kadirvel_k_Page_134.txt
6ac3bc1a44e65ad5db6ffd6869ad8437
5fa8d3813c28a41332b1c676d425647a1851bf11
F20101207_AABDLI kadirvel_k_Page_116.tif
66ea2e62409ca6b6a5bbc9ebf93b3380
378b4b01b723544562deef09436e865d0809e6df
2367 F20101207_AABEOL kadirvel_k_Page_135.txt
279f4b87fe88a0266566d13bb083adc2
2c64b4ce584a0c44ae10ac5c409f351c61ddc0a4
F20101207_AABENW kadirvel_k_Page_119.txt
502e737fe68f7fb44e9071120b5ca8f1
ab564556db986e654d5cacadec9df837a92aeeca
821201 F20101207_AABDKU kadirvel_k_Page_020.jp2
f605fe270eef1aa05f9c0343a9a18081
f695b55bbdd7833fb73e1912634131ee58b95194
61289 F20101207_AABDLJ kadirvel_k_Page_093.pro
33a0f095d895e02b3bd793929637d7cd
4da1e0ccbcf3eb3a30c11230ce4db0b888d4c5e0
1394 F20101207_AABEPA kadirvel_k_Page_153.txt
f931140b47f262c5b80a720bb0a9dbde
4140243933c7753c60a581c67cb0e28c68b6566c
2404 F20101207_AABEOM kadirvel_k_Page_136.txt
2ac99dda48e8224b4443f64e94604470
3557b0e28897f6ebca3d539e3a4661e2df4b2187
1565 F20101207_AABENX kadirvel_k_Page_121.txt
b2d0ed208cfd68b9e1b3d8a5b8b2a26d
5476ac9b240c257378bcf60a8226f07a255dfac0
801714 F20101207_AABDKV kadirvel_k_Page_048.jp2
b6ccad5552249a3ddf290d4d4e438075
94d94651b1d004a5e8a894c5b834ca4b97054065
4690 F20101207_AABDLK kadirvel_k_Page_110thm.jpg
6fba7bee2524c4bb5f1326fd982f8068
51270a57184c30467870e0807c151a60422870d4
2267 F20101207_AABEPB kadirvel_k_Page_154.txt
88ddd21ce4c49a35f2b4459d1789a7dc
053a1a7b8455a0f9177405b88d76f6af15e18a3b
F20101207_AABEON kadirvel_k_Page_139.txt
6a9de2e5d2ed0916c9c379c985c26733
65c0bf5e23e0260cbf29e5d20c8ed084dc9e3c13
2325 F20101207_AABENY kadirvel_k_Page_122.txt
910ff330f1285fd60e1c0862cd1de3e2
3f05a23a66010c2a65097a179924df507771d550
101851 F20101207_AABDKW kadirvel_k_Page_047.jpg
bebfb4f48779d1a9148f26c86d84b132
071b737aa3c7e462c472b1869472b8389b267d6f
1454 F20101207_AABDLL kadirvel_k_Page_172.txt
22baa4954c31d1604e1c1f7c63b14b6d
9318340567c0725318717ac4cc6cbcd2d5752845
1895 F20101207_AABEPC kadirvel_k_Page_155.txt
e66a0b1437d51cd3f37e09360d6f6b73
d89a64ab18d682aee10ae4fbf0a778f437354876
2293 F20101207_AABEOO kadirvel_k_Page_140.txt
70aaa0e1f9d464f1c667042a862097fe
22dddbd7ec847d1cc60f781d14fb62a1426e3ab0
1376 F20101207_AABENZ kadirvel_k_Page_123.txt
185fd198366f02a0d9ad72a233ee1773
506e0418a86a7efa7a0f7e306788098178c96f85
2585 F20101207_AABDKX kadirvel_k_Page_063.txt
091701c1d1bbff39cbdcc34c9aaf5c88
a27f119cc0775308e7526e4be1bb8e5272410915
F20101207_AABDMA kadirvel_k_Page_139.tif
e8f505137ded6a186181f6999f7725a9
d96681e205a12805bb85a9f03caa701d8b14dbba
7252 F20101207_AABDLM kadirvel_k_Page_089thm.jpg
755a7f8a483a7e351e9c1c20bfab4a87
e3c99a7571af657f2663c7838e91307ab916831a
F20101207_AABEPD kadirvel_k_Page_156.txt
eff1fd59095a454f14376934497ddbe5
f11afc49f4c7aee007690bad4a8a9fbd52d75284
2250 F20101207_AABEOP kadirvel_k_Page_141.txt
29ad996e1dcc07f1739d248862af1a2f
1e37209647f131bb6dbb3dce262f52a22614f366
20723 F20101207_AABDKY kadirvel_k_Page_124.QC.jpg
fcef0c668ca748082887400df981c6e7
8064f2fad295c18fa3e2335bc96984382d120d49
1269 F20101207_AABDMB kadirvel_k_Page_082.txt
83580910cf9f0f43810a8b738a0cbf9b
5d4d93ba2aadcdfa2683246949c71f6cfb34bfeb
6228 F20101207_AABDLN kadirvel_k_Page_105thm.jpg
a6dd53dc909888227b7e1e65ee21bb12
43842dabd6dcc07950f87ed23a22dbd8fd265f59
1544 F20101207_AABEPE kadirvel_k_Page_157.txt
6f85193cfdf4345568008da65c346f1a
8680772bd3263790c8be53371231f281d3ac71f6
2374 F20101207_AABEOQ kadirvel_k_Page_142.txt
073da94fb63649b152915d483f7f4fc7
227ce6116120008393a7b24d701d0b8b62c79305
85215 F20101207_AABDKZ kadirvel_k_Page_019.jpg
055a8acfd801f07906d994ff60c6353f
34bca570afec9c79bc4dd6240c5b182c7cf1a825
22213 F20101207_AABDMC kadirvel_k_Page_074.QC.jpg
cafd710d2112542674568be6da5f46b5
4b66cc847951af40c40ef8b4262bb6c8b53437a8
5486 F20101207_AABDLO kadirvel_k_Page_002.jp2
8b44a681a4d277a7390300aecf5dc2e7
9d36086cf379e7219a9d96f7a537cf9d24396e49
2219 F20101207_AABEPF kadirvel_k_Page_158.txt
5f626657517b515e16d7be267f544bb9
eeccb28a993656de556f319d298024cd65b07c72
2456 F20101207_AABEOR kadirvel_k_Page_143.txt
e0176cae009bf617e2df62dc835494ce
fe0f030326280c255a038110c42ddfb55bf91407
6142 F20101207_AABDMD kadirvel_k_Page_111thm.jpg
814ada710f8f09fae17dc8298071d3c9
1de250cf7c622e819f15cfaac1b7a1c86a0d28d4
15950 F20101207_AABDLP kadirvel_k_Page_159.QC.jpg
c1b82ef1499677a2aa168625055b5266
92892f3561dd9e133f867badf36a24d78ba268e3
1466 F20101207_AABEPG kadirvel_k_Page_159.txt
06ce8915cbae3b0003f0a00eab29ac77
7417dc09d371733940dcc750bbd9805d2749cce5
2324 F20101207_AABEOS kadirvel_k_Page_144.txt
756316482404536d7197f1bf46560d60
509d88c69ca8793441bba8b7a378b087bcb2bfca
16239 F20101207_AABDME kadirvel_k_Page_110.QC.jpg
aeae9b93041cbb87e22ba9eff17edb06
36f3da83d79efc4aa2c1a369494ac0ed862b2a17
85296 F20101207_AABDLQ kadirvel_k_Page_014.jpg
42fe7f727e4ffadb35ccff432f396781
484b64b4cc7ffcab6a53512745034c237af9d956
456 F20101207_AABEPH kadirvel_k_Page_161.txt
05d5721ec9b8497b5cf77d54618f1dfa
99eea63e301f2034a86ba170aa21cb3a7e1e6102
2221 F20101207_AABEOT kadirvel_k_Page_145.txt
84b67dcdaa60b030988205f3b13b2830
86a007041e95f2a7fd77e8d18306dbb7b31ff6d3
99503 F20101207_AABDMF kadirvel_k_Page_005.jpg
2314e4f61a35447933f60b58b15d7ffa
b489933d1cef28df74662671843c2fd6972b0e2e
F20101207_AABDLR kadirvel_k_Page_071.tif
d66b8208945d4c63aedd012bae626841
7e809d74c01d633af949d793b8ae86e17732c177
F20101207_AABEPI kadirvel_k_Page_162.txt
a35a032f1946aa3cdc1c90d34cf23dfc
3614451ea08136db8d3dc5d8674756040cf34294
2190 F20101207_AABEOU kadirvel_k_Page_146.txt
40d2d8b9be22fa424851a8f4c6908341
f693846ba7e28984915f4e3394594ef0ba8c44f0
8614 F20101207_AABDMG kadirvel_k_Page_045.QC.jpg
2d522793fe8040c48d09ca4fba16564e
74aabd58d8e0ab6648c30b7dc7bba5d4a11f95cf
2543 F20101207_AABDLS kadirvel_k_Page_083.txt
c01b413789a9e2bc4ed5a67b8b706b06
2abb8370b02b11fd6287de27b2c1392cc77ea5f3
2128 F20101207_AABEPJ kadirvel_k_Page_163.txt
7fe60e3b228b7d7536f4fa834bed8567
acba7d7afb35d2d03d79accfc28c7eb2021635d3
2002 F20101207_AABEOV kadirvel_k_Page_147.txt
5c8c3a154d1239a6b7d4e9b7d408e81c
175fabd1881024e153053cd48c34fa95d729fbca
34144 F20101207_AABDMH kadirvel_k_Page_138.pro
c094870b9625dd3a186a6bcb29235fe7
ba6c0b70a17f571ff50e1e18dac0176d4fae19a9
17703 F20101207_AABDLT kadirvel_k_Page_077.QC.jpg
55b418c5bb5c91b23e8b968bd7420760
2c04e8d2972893b27e6db354ebdf1c66ef42aa0c
2912 F20101207_AABEPK kadirvel_k_Page_164.txt
38ad275ddd94acc64b9f39d293a69849
16f5f5779db91952f8b2c856417774e4f927e750
2852 F20101207_AABEOW kadirvel_k_Page_148.txt
97fe6971f69ba15c44f0801cd624e75f
2d91ca4a658e557c0350f98488c96b88f9a800ff
F20101207_AABDMI kadirvel_k_Page_167.tif
36b714d8d13b43dc016cc4645f57da9d
313088676fe457c757c345f12aad046ef672aea7
F20101207_AABDLU kadirvel_k_Page_049.tif
4d1a810a8cfb8585e7f5a0bf7aef4c46
55d0cfa8084eb0c56802c3ab456e13c1b301ca1f
3189 F20101207_AABEPL kadirvel_k_Page_165.txt
794523fda387a3efbb17518fb393835a
54d2e4e0c397af8cc28b7cecb260fad365024d41
F20101207_AABDMJ kadirvel_k_Page_042.tif
3c211be6d3bc910c300435f6e7f157d6
2f2ed0b54c9495d193296b90b723a5f24f0931de
4586 F20101207_AABEQA kadirvel_k_Page_071thm.jpg
ef972f1ffa2dd517bb7a131a524ff138
20d234751332c726304e6895dde525ea1ff8b0a7
2846 F20101207_AABEPM kadirvel_k_Page_166.txt
96a5abc9089e43eb3c7d4debd429dfcc
cbf94031c8ce86a066936345b3057416356abc46
617 F20101207_AABEOX kadirvel_k_Page_149.txt
b1cb309857e52ad41a147935660e38e9
bdc438a9882be4fa4f68b99f2b26755b3e865399
20548 F20101207_AABDMK kadirvel_k_Page_130.pro
5813cd707827b54d48efe4675b044634
0e98f5aec0bef2d006dc094ffa22964070ef7583
2003 F20101207_AABDLV kadirvel_k_Page_104.txt
452d0a8b09996ac36090bdc74c700ef8
29e0ba20473ae401314ad5243c53f349a6cfc1a3
7840 F20101207_AABEQB kadirvel_k_Page_093thm.jpg
4231e2f73e869f76b28773e6ff2db039
8416b8af2de2d2bfe5f2e843d028958f057706de
2876 F20101207_AABEPN kadirvel_k_Page_167.txt
a9fc23ae3977b414c7d4890395ca465c
826ed9df476aa74923dedcb3d6ef567ceb969cb2
2445 F20101207_AABEOY kadirvel_k_Page_150.txt
d6b6cebb876ee5031187537e9d2354ff
0a9a9d88da0c71cb6e4ec4eac663ad0eaee93e0e
F20101207_AABDML kadirvel_k_Page_135.tif
679da031ffb37c250b0328b4ba0f3944
049e922de64f9d5eec9f970e66212fe3d650a79e
73457 F20101207_AABDLW kadirvel_k_Page_110.jp2
38844a594686b13c002c8553c4fe9dd6
f4f739193fad6b0504af99612c7d6215ae3300d9
19158 F20101207_AABEQC kadirvel_k_Page_096.QC.jpg
94b845212a0fbf7e25a7a679c3bdfab4
a0e0700a7113f3a0249012b64948ca167a1f47b6
3134 F20101207_AABEPO kadirvel_k_Page_168.txt
9e220cf5df52124cfc8acf98bc26ea49
2e3f10f7ae6a3a507621948950d97c8d291442ee
1330 F20101207_AABEOZ kadirvel_k_Page_152.txt
2293ff7de26f16a892b6902032857e5d
dfc623d2c4260ee74325456ce2c8c1bde4d001a0
8274 F20101207_AABDNA kadirvel_k_Page_166thm.jpg
233cbae632aba9a3806205024dd038fb
db6547ef282df3a3cbb9761a415c8aa64eba53e5
1087848 F20101207_AABDMM kadirvel_k_Page_017.jp2
501ffb526b00ab2bb6fb248738288ee2
a0e1b12fb791e6fc62360e59e0968223786fec42
875834 F20101207_AABDLX kadirvel_k_Page_125.jp2
906a8b23109306becd6ec4586eb5d456
fa86282f988e0cf45ca5bce316d8fb4d2582d80c
32509 F20101207_AABEQD kadirvel_k_Page_016.QC.jpg
2e216324114553e59d72fbd24b15f845
b4b2ebb0ed6f7df8ecef0f35649833eb74c51465
2881 F20101207_AABEPP kadirvel_k_Page_170.txt
79b657b4cd69332338361e3aad2c35d8
4e115f0cd2a2cd29f02b6ef1d3a4af47ab0312e2
27561 F20101207_AABDNB kadirvel_k_Page_157.pro
eb3265fa173069a4bc268be7db1cd3a0
9f6103e3cdbcf663c8e573e0553f964cc9253a0d
1087871 F20101207_AABDMN kadirvel_k_Page_030.jp2
59a5c907f79950664cdebb5a2976395a
06c6bb28da73581d4ca2d3cb50e521ad08d313de
6309 F20101207_AABDLY kadirvel_k_Page_108thm.jpg
04b0dcfe01c39075764a4c1fe06bb12e
a449ed44970cb19dd48a6af42a837ddebd92d0df
6294 F20101207_AABEQE kadirvel_k_Page_124thm.jpg
149dc3e71f69a94f5b01e60482ae37bc
b72b68bee4d1306533de5d58370b8f8c9c11fbed
F20101207_AABEPQ kadirvel_k_Page_171.txt
c0920b174c4d0b4f40a30e1590214f1d
17de50271d33ce29a324910626fcc3f7db89b519
2825 F20101207_AABDNC kadirvel_k_Page_099.txt
841c08521f1c1df9fe49a8e5ddd6eb10
be63f643b32899ac0dba1f20b7600aa405915e45
F20101207_AABDMO kadirvel_k_Page_056.tif
59b77241c1936e00a06682d0d060304a
a8757d5ddaa6c7aa210dc1cd9e3a66a479442f48
14747 F20101207_AABDLZ kadirvel_k_Page_153.QC.jpg
84f791cf9644e230c020ee485da7c8ad
33d1fda4b4ca55683428a13a034b871a2eaaff98







DEVELOPMENT OF CLOSED-LOOP INTERFACE CIRCUITS FOR CAPACITIVE
TRANSDUCERS WITH APPLICATION TO A MEMS CAPACITIVE MICROPHONE






















By

KARTHIK KADIRVEL


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2007


































@ 2007 Karthik Kadirvel





































To my parents, sister and Sona









ACKNOWLEDGMENTS

I would like to thank my advisor and mentor Dr. Toshikazu Nishida for his constant

motivation and encouragement to bring out the best in me. This research would not have

been possible without his logical approach to problem solving and emphasis on fundamental

principles. Dr. Mark Sheplak and Dr. Robert Fox provided indispensable help on MEMS

processing and circuit design respectively at various points during this research The many

discussions with Dr. Louis N. Cattafesta on control theory provided valuable information at

crucial junctures of this research.

I would like to thank my project colleagues David Martin for providing me with a MEMS

microphone to use as a test vehicle for the circuit development, Jian Liu for the help in modeling

using Matlab and Joel Fuster for the various board level circuit design ideas.

I would like to thank all my friends at the Interdisciplinary Microsystems Group for their

friendship and support during the course of my graduate program. Special mention goes to

Robert Taylor, Erin Patrick, and Stephen Horowitz for all their help.

Mr. T.K Sundaram, my high school mathematics and physics instructor, deserves special

mention here. I would like to extend my heartfelt thanks and gratitude for providing me with the

strong foundation on which all my achievements stand.

The help provided by Al Ogden, Ivan Kravachenko and Bill Lewis of the UF Nano

fabrication facilities and Ken Reed of TMR Engineering is gratefully acknowledged.

Finally, and most important I would like to thank my family and my wife for believing in

all my goals and aspirations and, for their love, encouragement, and constant support in all my

endeavors.











TABLE OF CONTENTS


page


ACKNOWLEDGMENTS ........... ...... .....

LIST OF TABLES ...........................

LIST OF FIGURES ...........................

ABSTRACT .............................

CHAPTER

1 INTRODUCTION ........................

1.1 Background ... .. ... ... ... .. ... ... ..
1.2 Research Objectives .....................
1.3 Dissertation Overview ....................


. . 18
. . . 20
. . . 2 1


2 CAPACITANCE TRANSDUCERS ................... ....... 23

2.1 Introduction ...................... ............. 23
2.2 Canonical Capacitive Transducers ......... ................ 23
2.2.1 Single Backplate Capacitive Transducers ........ ........ 24
2.2.2 Canonical Dual Backplate Capacitive Transducer . . ... 33
2.3 Structure of Microphone .................. ........... .. 39
2.4 Lumped Element Model ............... . . 41
2.5 Summary ................... ..... .... ........ 44

3 INTERFACE CIRCUITS ................. . . . ......46


3.1 Introduction.......... . . .
3.2 Performance Metrics of Capacitive Transducers . ...
3.3 Analog Open Loop Sense Techniques . ........
3.3.1 Sensing Using A Voltage Amplifier ...........
3.3.2 Sensing Using A Charge Amplifier . . ....
3.3.3 Sensing Using Synchronous Modulation/Demodulation .
3.4 Digital Open Loop Sense Techniques ..............
3.5 Closed-Loop Sense Techniques .................
3.5.1 Sensing Using Closed-Loop Analog Interface Circuit ..
3.5.2 Sensing Using Digital closed-loop Interface Technique
3.6 Scaling Analysis of AY Interface Techniques .........
3.7 Summary ...... ......................

4 DESIGN OF ANALOG INTERFACE CIRCUIT . .....

4.1 Introduction . . . . . . .
4.2 Sim ulation . . . . . . . .
4.2.1 Simulink M odel .................. ...
4.2.2 Simulation Results . ..............
4.3 Design of Analog Interface Circuit ..... . . .


4.3.1 Design of Amplifier for Top and Bottom Plate
4.3.2 Input Buffer Amplifier Design . ..


Voltage Gen


. . 46
. . 46
. . 51
. . 52
. . 65
. . 72
. . 80
. . 85
. . 86
. . 9 1
. . 93
. . 96

. . 97

. . 97
. . 97
. . 98
. . 102
. . 103
aeration . 103
. . 105










4.3.3 Demodulator Design .. ................
4.3.4 Compensator Design .. ................
4.4 Integration of Interface Circuit Components on a PCB .......
4.5 Noise model ............................
4.6 Limitations of the Current Implementation of the Interface Circuit
4.7 Sum m ary . . . . . . . .

5 DESIGN OF DIGITAL CLOSED LOOP INTERFACE CIRCUIT .

5.1 Introduction . . . . . . . .
5.2 Design of AY Interface Circuit ...................
5.3 Component Design .. .....................
5.3.1 Switched Capacitor Amplifier ................
5.3.2 Compensator Design .. ................
5.3.3 Comparator Design .. ..................
5.3.4 Bias Network Design .. .................
5.4 System Level Timing .. ....................
5.5 System Level Simulation .. ..................
5.6 Circuit Implementation .. ...................
5.7 Sum m ary . . . . . . . .


6 EXPERIMENTAL SETUP AND RESULTS . .

6 1 Introduction


6.2 Acoust
6.2.1
6.3 Experi]
6.3.1
6.3.2
6.3.3
6.3.4
6.3.5
6.3.6
6.3.7
6.3.8
6.3.9
6.4 Discus
6.5 Summa


. . 131


ic Pressure Coupler ........................
Design of Acoustic Pressure Coupler ..............
m ental Results . . . . . . .
Characterization of Pressure Coupler ..............
Open Loop Frequency Response of Microphone with dc Bias ..
Open Loop Frequency Response of Microphone with dc+ac Bias
Linearity of Open Loop Microphone ..............
Noise Floor of Open Loop Microphone .. ............
Compensator Characterization .. ...............
Closed Loop Frequency Response .. .............
Closed Loop Linearity ......................
Closed Loop Noise Floor .. ..................
sion . . . . . . . . .
ary . . . . . . . . .


7 CONCLUSION AND FUTURE WORK .................. ...... 150

7.1 Conclusion ................... . . ... 150
7.2 Future Work ................... . . ... 150
7.2.1 Simulation .................. . . . .. 151
7.2.2 Hardware Implementation .......... . . . 151

A PERFORMANCE METRICS FORMULATION .................. .. .152


A.1 Derivation of Sensitivity for a Voltage Follower . . .
A.2 Noise Analysis for a Capacitive Microphone with a Voltage Follower .
A.3 Derivation of Sensitivity for a Charge Amplifier . .


. . 152
. . 154
. . 156









A.4 Noise Analysis for a Capacitive Microphone with a Charge Amplifier . 157
A.5 Sensitivity for Synchronous Modulation and Demodulation Interface Circuit 158

B FORMULATION USED IN SWITCHED CAPACITOR CIRCUITS ......... .162

B.1 Bandwidth Considerations for Amplifiers in Switched Capacitor Amplifiers 162
B.2 Analysis of Noise in a Switched Capacitor Amplifier . . . .... 162
B.2.1 Switch Resistance Noise .................. .. ...... 163
B.2.2 Wideband Amplifier Noise ......... . . ......163

REFERENCES ...................... . . . 164

BIOGRAPHICAL SKETCH ................. . . . ..173









LIST OF TABLES


Table page

1-1 Comparison of audio and aeroacoustic microphone specifications . . ... 21

2-1 Dimensions of aeroacoustic and audio microphone[64]. ............... ..40

2-2 Material properties of microphone structure. .................. .. 40

2-3 Theoretical formulation for lumped element model parameters. . . .... 43

2-4 Lumped element model parameters of aeroacoustic and audio microphone. ...... ..43

2-5 Second order system parameters of microphones. ................ 44

3-1 Parasitic capacitance for various packaging technologies. ..... . . 54

3-2 Specification of Voltage Amplifiers for Bruel and Kjaer Microphones. . ... 61

3-3 Review of interface circuits used in MEMS microphones published in literature. . 63

3-4 Specification of a representative sample of operational amplifiers recommended by
manufacturers for use as voltage followers for capacitive transducers. . ... 65

3-5 Review of charge amplifier interface circuits used in capacitive microphones. . 72

3-6 Specification of a representative sample of general purpose operational amplifiers
recommended by the manufacturer for use as charge amplifiers using a feedback resistor
and capacitor. .................. ............. .. .. 72

3-7 Summary of previous synchronous modulation/demodulation based open loop capacitive
transducer interface circuit. .................. .. ......... 79

3-8 Summary of previous synchronous modulation/demodulation based closed-loop capacitive
transducer interface circuit. .................. .. ......... 89

3-9 Frequency requirements of AZ interface circuits with increasing resonant frequency. 94

3-10 Feedback voltage requirements of AZ interface circuits with increasing dynamic range. 94

3-11 Summary of previous AZ based closed-loop capacitive transducer interface circuit. 95

4-1 Feedback voltage requirements of aeroacoustic and audio microphone for force balancing
maximum input pressure. ................. . . 105

4-2 Element values for resistors and capacitors used in compensator. . . ... 107

4-3 Magnitude of parameters used in noise analysis ............... .. ..110

5-1 Clock signals used in the various components of the AZ modulator. . ... 125

5-2 Function performed during each time interval. .................. ..127

5-3 Parameters used in closed loop simulation. ................... . 128

5-4 Frequency requirements of AZ interface circuits with increasing resonant frequency. 129









5-5 Feedback voltage requirements of AY interface circuits with increasing dynamic range. 129

6-1 Comparison of simulated and experimental sensitivity. ..... . . ..... 139

6-2 Characterization results of microphone in open and closed loop mode of operation. 148









LIST OF FIGURES


Figure page

1-1 Microphone classification based on sensor structure. ................. ..19

1-2 Common condenser microphone structures. .................. .... 20

1-3 Microphone classification based on sense technique. ................. ..21

2-1 Overview of Chapter 2 ................... . . ..23

2-2 Schematic of a single backplate capacitive transducer. . . . 24

2-3 Schematic of a single backplate capacitive transducer connected to a voltage buffer. .. 26

2-4 Schematic of a single backplate capacitive transducer connected to a charge amplifier. 28

2-5 Schematic of a single backplate capacitive transducer showing electrostatic force and
restoring force .................. ................... .. 31

2-6 Schematic of a differential capacitive transducer. ................. ..33

2-7 Schematic of a dual backplate capacitive transducer connected to a voltage amplifier. .36

2-8 Schematic of a dual backplate capacitive transducer connected to a charge amplifier. .. 36

2-9 Schematic of a dual backplate capacitive transducer showing forces on the movable
plate........... ......... ..................... ...... 37

2-10 Cross section view of dual backplate microphone [64]. ............... ..40

2-11 Low frequency LEM of dual backplate microphone. ................. ..43

2-12 Theoretical normalized frequency response of audio microphone. . . ... 44

2-13 Theoretical normalized frequency response of aeroacoustic microphone. . ... 45

3-1 Overview of Chapter 3 ................... . . ..46

3-2 Characteristic frequency response of a capacitive transducer. . . ..... 48

3-3 Characteristic ideal and actual response of a capacitive transducer. . . ... 48

3-4 Characteristic noise floor of a capacitive transducer using a voltage amplifier interface
circuit .. . . . . . .. . . .... .... 50

3-5 Noise model of resistor. .................. . . ...... 51

3-6 Noise model of opamp showing voltage and current noise source. . . ... 51

3-7 Plot of noise PSD of amplifier. ............... . . 52

3-8 Organization of each interface circuit discussed in Chapter 3. . . ... 53

3-9 Schematic of a dual backplate microphone connected to voltage buffer. . ... 53

3-10 Noise model of dual backplate microphone connected to voltage buffer. . ... 55









3-11 Variation of sensitivity as a function of total parasitic capacitance for various sense
capacitances ... ... .. ... ... ... .. . . . 57

3-12 Power spectral density of individual noise contributors and total output noise in a voltage
amplifier based interface circuit .............. . . .... 58

3-13 Plot of output noise for voltage amplifier with a fixed bias resistor (10MQ) ..... ..59

3-14 Plot of output noise for voltage amplifier with a fixed parasitic capacitance of (10pF). 60

3-15 Schematic of dual backplate microphone connected to a charge amplifier. ...... 66

3-16 Noise model of dual backplate microphone connected to charge amplifier. ...... 66

3-17 Power spectral density of individual noise contributors and total output noise in a charge
amplifier based interface circuit. .................. ........ 68

3-18 Plot of output noise for charge amplifier with a fixed bias resistor (10MQ) ...... ..70

3-19 Plot of output noise for charge amplifier with a fixed parasitic capacitance of (10pF). 71

3-20 Schematic of capacitive microphone with a synchronous modulation and demodulation
technique using a voltage amplifier based circuit. .................. ..73

3-21 Schematic of a capacitive microphone with a synchronous modulation and demodulation
technique using a charge amplifier based circuit. .................. ..73

3-22 Plot of spectrum at various nodes of Figure 3-21 .................. ..74

3-23 Noise model of synchronous modulator and demodulator using voltage amplifier. ... 77

3-24 Noise model of synchronous modulator and demodulator using charge amplifier. .... 77

3-25 Switched capacitor implementation of correlated double sampling. . . ... 81

3-26 Equivalent circuit of switched capacitor amplifier during amplify and reset phase. 82

3-27 Noise model of switched capacitor amplifier during amplify phase. . . ... 82

3-28 Schematic of analog closed-loop force feedback scheme. ..... . ..... 87

3-29 Schematic of first order AZ modulator. .................. ...... 92

3-30 Schematic of second order mechanical AZ modulator. ................ ..92

4-1 Overview of Chapter 4.................. . . . ..97

4-2 Block diagram of analog closed-loop system showing microphone and interface circuit. 98

4-3 Bode plot of compensator. .................. ............. 101

4-4 Bode plot of loop gain with and without compensator. .. . . ..... 102

4-5 Simulink model of analog closed-loop control system. ..... . . ..... 102

4-6 Subsystems used in Simulink model. .................. ....... 103










4-7

4-8

4-9

4-10

4-11

4-12

4-13


4-14

4-15

4-16

5-1

5-2


5-3

5-4

5-5

5-6

5-7

5-8

5-9

5-10


5-11

5-12

5-13

5-14

6-1

6-2

6-3


Schematic of second order mechanical AZ modulator. . .

Transistor level schematic of amplifier. . .......

Schematic of amplifier during amplify phase. . .....

Transistor level schematic of compensator. . .....

Compensator schematic during the two phases of operation. .

Transistor level schematic of comparator . ......

Schematic of bias network . ..............

Flowchart showing the sequence of operation of the overall AZ r
interface circuit . ...................

Schematic of AZ modulator circuit . .........

System level timing diagram over one complete cycle. . ..

Simulink Model of the AZ modulator based interface circuit. .

Power spectral density of output AZ modulated bitstream. .

Overview of Chapter 6 . ................

Photograph of acoustic pressure coupler. . ......

Plot of input pressure signal as a function of frequency . .


.... . .. 115

. . . .. 117

. . . .. 118

. . ... 121

. . . 122

. . . 123

. . ... 124


nodulation


based


. . . 125

. 126

. ... . . 127

. . . 128

... . . 129

. . ... 131

. . ... 132

. ... . .134


Diaphragm displacement with and without feedback. ... . . 104

Diaphragm displacement with and without feedback. . . . ...... 104

Schematic of summing amplifier that generates top and bottom plate voltages. . 105

Schematic of voltage amplifier. ............. . . . 106

Schematic of analog multiplier (AD835) showing relevant pins. . . ... 106

Schematic of two stage compensator. ................. .... ....... 107

Annotated photograph of PCB 1 showing buffer amplifier, breakout headers and optical
port. . . . . ... .. . . .... ....... 108

Annotated photograph of PCB2. .................. ........ 108

Schematic of forward path of the closed loop system ... . . .109

Plot of theoretical noise floor of the open loop system and forward path system . 111

Overview of Chapter 5.................... . . . ... 113

Flow chart for design of AE modulator component parameters from system level
specification . ................... ................ 114









6-4 Comparison of frequency response of pressure coupler with air as medium and with
helium as medium ................... . . ... 135

6-5 Block diagram describing characterization procedure used. . . ..... 135

6-6 Frequency response with dc bias (5V) with air medium. .. . . ..... 136

6-7 Frequency response with dc bias (5V) with helium medium. . . .... 137

6-8 Schematic of capacitive microphone with a synchronous modulation and demodulation
technique using a voltage amplifier based circuit. .................. ..137

6-9 Open loop frequency response measured at the output of low pass filter. . ... 138

6-10 Open loop frequency response normalized to the peak ac voltage measured at the output
of low pass filter. ................... . . .... 138

6-11 Open loop linearity of the open loop system. ................. .139

6-12 Linearity of the open loop system normalized to peak ac bias voltage measured at the
output of the demodulator. .................. ............. 140

6-13 Noise floor of microphone for two different ac bias voltages measured at the output
of the demodulator. ................ . . .... 141

6-14 Theoretical and experimental frequency response of compensator. . . ... 142

6-15 Block diagram of analog closed-loop system showing microphone and interface circuit. 142

6-16 Closed loop frequency response measured at the output of demodulator . ... 143

6-17 Closed loop frequency response measured at the output of demodulator with ac bias
of 6Vp. ................... ........... ........ 144

6-18 Closed loop frequency response measured at the output of demodulator with ac bias
of 9V,. ................... ........... ........ 145

6-19 Closed loop frequency response measured at the output of compensator. . ... 146

6-20 Linear range of system in closed loop configuration as measured at the output of the
compensator. ................... . . .... 146

6-21 Linearity of the closed loop system normalized to peak ac bias voltage measured at
the output of the compensator. .................. .......... 147

6-22 Closed loop noise floor. ................. . . ....... 147

A-1 Noise model of dual backplate microphone connected to voltage buffer. . ... 152

A-2 Noise model of dual backplate microphone connected to voltage buffer. . ... 155

A-3 Noise model of dual backplate microphone connected to charge amplifier. ...... .157

B-1 Noise model of switched capacitor amplifier during amplify phase. . . ... 162









Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

DEVELOPMENT OF CLOSED-LOOP INTERFACE CIRCUITS FOR CAPACITIVE
TRANSDUCERS WITH APPLICATION TO A MEMS CAPACITIVE MICROPHONE

By

Karthik Kadirvel

December 2007

Chair: Toshikazu Nishida
Major: Electrical and Computer Engineering

There has been a trend towards miniaturization and batch fabrication of sensors inspired

by a similar trend in the electronics industry using novel fabrication techniques used in micro

electromechanical system (MEMS) fabrication. Capacitive microphones, whose common

applications include aeroacoustic measurement and cell phones, is one such sensor whose

dimensions are being aggressively scaled down. In measurement microphones, miniaturization

will facilitate improved measurement precision, and in cell phones, miniaturization will reduce

printer circuit board space and complexity. Both applications will benefit from the potential cost

reduction brought by the batch fabrication of sensors.

As sensor geometry is scaled down, improved circuit techniques are required to measure the

sensor output. This is because at small geometries sensor capacitance is comparable to unwanted

parasitic capacitance which reduces the transducer sensitivity. Also, at reduced sensor geometry,

the voltage required to bias the microphone could cause the microphone plates to pull in. The

goal of this work is to design and characterize interface circuits that are suitable for miniature

capacitive transducers. To achieve this goal, the performance of existing open and closed-loop

interface circuits are investigated. Scaling of the performance metrics of the microphone and

interface circuit as sensor geometry decreases is also investigated.

A proof of concept closed-loop analog controller for a MEMS capacitive microphone

is designed. A test apparatus is developed to characterize the system over the audio range

by operating the microphone in a helium medium which increases the bandwidth of the test

apparatus. Characterization of the microphone in open and closed loop mode of operation is









presented. Results show that stable closed loop operation of the microphone is feasible with

increased sensitivity and the potential to address pull-in issues.









CHAPTER 1
INTRODUCTION

There has a been a major trend towards miniaturization and batch fabrication of transducers

inspired by a similar trend in the electronics industry. With the advent of micromachining [1]

techniques used in microelectromechanical system, (MEMS) manufacturing, fabrication of

transducers with very small sizes have been made possible. Micromachining also facilitates the

batch fabrication of devices which possess nearly matched structural and electrical properties.

Batch fabrication of devices also can reduce the cost of manufacturing if the volume is large

enough.

One key transducer leading the way in this trend are capacitive transducers. Capacitive

transducers are devices that respond to physical phenomenon via a change in capacitance

that is measured using various interface circuits. Commercially successful applications that

utilize capacitive transducers include accelerometers used in air bag deployment system for

cars and laptop hard drives by Analog Devices [2], projection technology for large screen TVs

using Texas Instruments Digital Light Processing (DLP) processors [3], pressure sensors by

Freescale Semiconductors [4], and audio microphones for cell phones by Knowles Acoustics

[5]. Other capacitive transducers published in the literature include accelerometers [2],[6] -[11],

micro-mirrors [12], microphones [13] [21], pressure sensors [22] [25], oscillators [26] -

[31], and gyroscopes [32] [35]. In all these applications, the capacitance change of the sensor

is converted into a usable voltage using interface circuits [36] such as the voltage buffer [37] -

[40], charge amplifier [41], ac bridge [6], force balancing circuits [6, 8, 42], and oscillator based

circuits[43] [45].

The miniaturization of capacitive transducers results in new challenges for the design of the

interface circuits. Some of the issues that are exacerbated at small sensor geometries that make

the design of interface circuits more challenging include a parasitic capacitance comparable to

the sensor capacitance, reduced sensitivity, and pull in [46]. All of these factors can degrade the

performance of micromachined capacitive transducers. As we scale down aggressively from

micro scale devices in MEMS systems to nano scale devices, an understanding of the tradeoffs of

interface circuits is important to obtain the full performance benefits of the miniature capacitive

transducer.









The overall topic of my work is the study of interface circuits for a specific class of

capacitive transducers namely capacitive microphones. Traditionally, capacitive microphone

interface circuitry consisted primarily of voltage amplifiers For example, commonly used

instrumentation grade capacitive microphones for aeroacoustic applications, the Briiel and

KjaerType 4134 and Type 4138 series of microphones [38], use a voltage buffer amplifier as the

interface circuit. Capacitive microphones used over the audio bandwidth (20Hz 20kHz) such

as the Briiel and KjaerType 4135 microphone also use a voltage buffer as the interface circuit.

Commercial MEMS microphones that use voltage buffers include the SiSonic microphones [5]

manufactured by Knowles acoustics and DigiSiMic microphones [47] by Sonion Technologies.

Though voltage amplifier interface circuits are commonly used, there exist other interface

techniques that may possess advantages for high performance capacitive transducers in certain

applications. For example, since capacitive accelerometers are similar in operating principle and

structure as compared to capacitive microphones circuits used for accelerometers may be adapted

for use in capacitive microphones. A closed-loop interface interface circuit technique used in

some accelerometers, [2], may provide benefits such as improved pull-in performance, higher

sensitivity, and direct digital output. This provides the motivation for investigating the potential

advantages and disadvantages for closed-loop interface circuits for capacitive microphones.

To achieve these goals, a thorough understanding of the theoretical performance of existing

capacitive interface techniques is required. Although previous work [48], [49], review interface

techniques such as voltage buffer, charge amplifier, synchronous modulation and demodulation,

they do not provide detailed theoretical formulations to compare the various techniques. In

addition, closed loop interface circuits are not discussed in detail.

The goal of my work is to design and evaluate capacitive interface circuits with an emphasis

on closed-loop interface circuits for MEMS capacitive microphones. To achieve this goal, the

performance of existing open and closed-loop interface circuits are investigated. Theoretical

formulations for the performance metrics of the interface circuits are derived and compiled

into a common framework for comparison. Using the information obtained, a proof of concept

closed-loop analog controller for a MEMS capacitive microphone is designed.









1.1 Background

This section begins a brief introduction to the operating principle of a capacitive microphone.

Various commonly used microphone structures are then described. A preliminary overview and

a brief literature review of the microphone interface circuits is presented in this section. This

section concludes with a flow chart showing the various interface circuits which forms the

framework for Chapter 2. A detailed literature review along with theoretical formulations for

each of the interface circuits is presented in Chapter 2.

A microphone is a transducer that converts input acoustical energy into corresponding

electrical energy. Capacitive microphones consist of a flexible diaphragm that is exposed to

the pressure fluctuations of the acoustic field of interest and one or more fixed backplates. The

pressure induced structural deflection of the diaphragm causes a variation in the capacitance

between the vibrating diaphragm and the fixed backplates which is detected and measured.

There are two main ways of measuring the capacitance change. The two approaches are

based on the fundamental equation that relates the charge Q stored on the capacitor plates to the

capacitance C and the voltage V across the plates of the capacitor,


Q = CV. (1-1)


In the fixed charge scheme [14], [50] [52] the, charge is held constant and the voltage change

associated with the capacitance change is measured. The charge is held constant by using

a dc voltage source to bias the capacitor or by using a special dielectric material (electret)

such as polypropylene, mylar, or silicon dioxide that can hold charge for long periods of time

measured in years. The electret material is bonded either to the movable diaphragm or to the

fixed backplate which is charged with a high potential of hundreds of volts using an external

dc source. The main advantage of electret microphones is that they do not require an external

power supply and are thus useful for portable applications [53]. The fixed voltage that can be

generated by the charge on the electret is much higher than what can be achieved via a battery

which leads to a higher sensitivity. The main disadvantages of micromachined electrets are that

the micromachined electret materials lose their charge over time when exposed to humidity [16].

In the fixed voltage scheme, a fixed voltage is maintained using an external voltage source while









the charge transfer associated with the capacitance change is measured. The externally applied

voltage is known as the 'polarization voltage'.

Capacitive microphones can be broadly classified based on the structure and on the detection

mechanism. Figure 1-1 shows a classification tree of the common capacitive microphone

structure. Based on the structure, they can be classified as single back plate and dual backplate

microphones. Various authors have reported single backplate microphones [17, 21], [54] [60]

and dual back plate microphones [19],[61] [64] in the literature. A schematic of the common

microphone structures is shown in Figure 1-2.

Most commercial MEMS [5, 47] and non-MEMS [38] microphones use a single backplate

design. In these microphones, a single variable capacitor is formed between the fixed backplate

and the movable plate that is measured by the interface circuit. In dual backplate and dual

diaphragm microphones, two capacitors that vary differentially are formed. Dual backplate and

dual diaphragm microphones have certain advantages such as increased sensitivity as compared

to single backplate microphones, a higher bias voltage which increases the sensitivity, improved

pull in performance, and increased linearity but are more complex to fabricate and have higher

cost.

Condenser
Microphone



Single Dual
Backplate Backplate




Single Dual
Diaphragm Diaphragm

Figure 1-1. Microphone classification based on sensor structure.


Figure 1-3 shows a classification tree of the various capacitive microphone interface

techniques. Based on the sensing scheme, microphones can be broadly divided into open

loop [41, 59, 65, 66] and closed loop [6] [8],[10, 60], [67] [69] categories. Both of these

categories can be further divided into analog and digital techniques based on the presence or











-- Diaphragm
Air Gap

Backplate

Acoustic Hole
Backchamber
SPressure
I Equalization Vent
(a) Schematic of single backplate single diaphragm
microphone[55].

-- Top Diaphragm
-,, --Air Gap

Bottom Diaphragm
Backplate
Acoustic Hole
Backchamber
Pressure
I Equalization Vent
(b) Schematic of single backplate dual diaphragm
microphone[55].

-- Top Backplate
Air Gap

S- Bottom Backplate
Diaphragm
Acoustic Hole
Backchamber
SPressure
I Equalization Vent
(c) Schematic of dual backplate single diaphragm
microphone[55].

Figure 1-2. Common condenser microphone structures.


absence of a clock signal, respectively. A detailed literature review, principle of operation, and

performance metrics of each of these circuits is discussed in Section 2.2 to Section 2.5.

1.2 Research Objectives

The goal of this work is to design a closed-loop force feedback interface circuit for a MEMS

capacitive microphone. Capacitive microphones with two different target applications namely,

aeroacoustic applications and audio applications, are the theoretical testbeds for characterizing

these circuits. The target specification for the two microphones are listed in Table 1-1. These

specifications provide typical values for the various parameters. The actual values depend

upon the application. For example, telephony applications require audio microphones with

20










Condenser
Microphone



Open Closed
Loop Loop

Digital Analog Digital

Synchronous Sigma
Switched Mod/Demod Delta
nalg Capacitor

Voltage Synchronous Charge
Buffer Mod/Demod Amplifier

Figure 1-3. Microphone classification based on sense technique.


a bandwidth of 300Hz 3.4kHz while hearing aid microphones have a bandwidth of

20 Hz-20 kHz. Also, in the table, it is assumed that the dynamic range is 100dB. To aid in


Table 1-1. Comparison of audio and aeroacoustic microphone specifications
Property Audio Microphone Aeroacoustic
Microphone
Max Pressure 120 dB 160 dB
Bandwidth* 20 Hz-20 kHz 45 Hz-160 kHz
Noise Floor 20 dB 60 dB
*100dB dynamic
range is assumed


the design of the overall system, a detailed model of the microphone and the interface circuit has

been developed. Theoretical formulations for the sensitivity, frequency response, and minimum

detectable signal (MDS) of the microphone along with the interface circuit system have been

derived. The theoretical parameters are compared with simulation and experimental results by

characterizing the microphone and interface circuit in a acoustic pressure coupler.

1.3 Dissertation Overview

This dissertation is divided into seven chapters. Chapter 2 is an in-depth study of the

various capacitive transducer interface circuits. For each of the interface circuit techniques, the

literature is reviewed and the principle of operation and performance metrics are explained. Since

the design of the circuit depends on the characteristics of the capacitive transducer, Chapter 2









discusses the structure and specifications of two capacitive microphones used as the test vehicles

to theoretically evaluate the performance of the interface circuit. To aid in design and prediction

of the performance specifications, an electrical model of the overall system consisting of the

sensor along with the interface circuit is described. Furthermore, a reduced order lumped element

model of the microphone is described for use in multi-domain simulations.

Chapter 3 discusses various open and closed loop interface circuits for capacitive

transducers with an emphasis on capacitive microphones. Theoretical formulations for the

performance metrics of the interface circuits are developed.

Chapter 4 and Chapter 5 detail the design and simulation of the analog and digital

closed-loop interface circuits respectively. The two interface circuits are compared in terms

of performance specifications, design complexity, ease of implementation, and cost. Chapter 5

concludes by presenting the reason for implementing an analog closed-loop interface circuit over

a digital closed-loop interface circuit.

In Chapter 6, the test apparatus used to characterize the performance of the interface circuit

is explained. Results of the characterization of the test apparatus is presented first followed by the

characterization results of the microphone with the analog open and closed loop interface circuit.

Chapter 7 discusses the results obtained along with the future work.











CHAPTER 2
CAPACITANCE TRANSDUCERS

2.1 Introduction

In this chapter, canonical capacitive transducers are discussed followed by a discussion

of the microphone under consideration. Lumped element modeling of the sensor to facilitate

multi-domain simulation of the sensor along with the interface circuitry in the following chapter

is also discussed.

This chapter begins with a discussion of a canonical single backplate and dual backplate

capacitive transducer. Theoretic formulation of the voltage, charge, capacitance, and force

relations for the two canonical transducers are derived. Linearization of the formulation is also

presented. A discussion of the performance metrics of capacitive transducers is then presented.

The structural and material properties of the microphone under consideration are discussed

next. A lumped element model (LEM) is then developed for the microphone. Using the LEM, a

reduced second order system is developed which is used to extract the sensor resonant frequency

and damping coefficient. This chapter concludes with MATLAB simulations of the reduced order

models. A graphical overview of this chapter is shown in Figure 2-1.


Introduction
ISingle backplate
Canonical
capacitive
transducers
Dual backplate
Structural and
material properties
of microphone
Audio microphone

Lumped element
modelling
4 Aeroacoustic
microphone
Summary



Figure 2-1. Overview of Chapter 2.



2.2 Canonical Capacitive Transducers

In this section, the principle of operation of a single backplate and dual backplate capacitive

transducer is presented.









2.2.1 Single Backplate Capacitive Transducers

Structure and Principle of Operation. Figure 2-2 shows a simple single backplate

capacitive transducer. It consists of two plates that are at a nominal gap xo between them. One

plate is fixed and the other plate is free to move in response to an input physical phenomenon.


Movable Plate [ I r,
MovablePlate-- ------------------I x


xo
x0


Fixed Plate I I

Figure 2-2. Schematic of a single backplate capacitive transducer.


The plates are so arranged such that only the distance x between the plates can vary and not the

area of overlap A between the two plates. Under the application of an external pressure signal,

the diaphragm moves from its equilibrium position.

The distance between the plates can be expressed in terms of the nominal gap distance, xo,

and the instantaneous displacement from the nominal position, x', by the equation


x = Xo '. (2-1)


The instantaneous capacitance C, between the plates is given in terms of the permitivity of free

space co, relative permitivity cr, area of overalap A, and the distance between the plates x, by


C c(2-2)


For air gap, the relative permitivity is approximately unity. Substituting Equation 2-1 into the

previous equation, we obtain

C =cOr A i ] (2-3)
Lr 1 o -
Defining the nominal capacitance Co as


Co c, (2-4)
X:









the instantaneous capacitance C can be rewritten as


C = Co ,1 (2-5)
1-

i.e.,

Co o t (2-6)

From Equation 2-5, it can be observed that there exists a non-linear relationship between the

capacitance and the distance between the plates. For small displacements (x' << xo), using a

Taylor's series expansion, Equation 2-5 can be linearized as

C =Co 1 + (2-7)


If the change in capacitance is denoted by AC, then


C = Co + AC. (2-8)


Substituting for C from Equation 2-2 and Co from Equation 2-4 in Equation 2-8, we obtain


AC = Co-. (2-9)
Xo

This shows that for small displacements, the change in capacitance is a linear function of the

change in gap distance. The next equation to consider is the relationship between the voltage V

across the plates of the capacitor and the charge Q stored across the plates of the capacitor which

is given by

V=Q. (2-10)
C

Substituting Equation 2-5 in the previous equation, we obtain


V ]- (2-11)
C'o ro

From Equation 2-11, it can be seen that


V oc Qx'. (2-12)


This shows that the voltage across the plates of a capacitor is a product of two time dependent

variables. Hence, the voltage across the plates of the capacitor is a non-linear function of the









displacement and charge stored on the plates of the capacitor. Equation 2-12 equation can

be linearized by using either a fixed charge or a fixed voltage to measure the displacement.

These two linearization methods determine the choice of interface circuit. For the fixed charge

technique, a voltage amplifier is used. This is because for the fixed charge technique


V oc x'. (2-13)


For the fixed voltage technique, a charge amplifier is used. This is because for the fixed voltage

technique, the output is charge which will be shown to be linear with displacement after

integration by a charge amplifier. A preliminary discussion of the two interface techniques is

presented next. A detailed review of these two techniques along with detailed performance

metrics and a literature review is given in Section 3.3 and Section 3.3.2.

Linearization Using Fixed Charge (Voltage Amplifier Interface Circuit). In this

technique, a fixed charge is stored on the plates of the capacitor. The fixed charge can be

provided by a dc voltage source to bias the capacitor or by using a special material known as

an electret. An electret is a material that can store charge for long periods of time (measured

in years). The charge is applied onto the electret material using a technique called polarization

whereby a large (few 100s of volts) is applied to the electret material. Figure 2-3 shows a single

backplate capacitive transducer linearized using a fixed charge technique by connecting to a dc

voltage source Vdc. Vdc is connected to one end of the capacitive transducer. A bias resistor Rdc

+Vd




+ out

Rdc





Figure 2-3. Schematic of a single backplate capacitive transducer connected to a voltage buffer.


is connected between the other end of the capacitive transducer and ground and is used to set the

dc operating point. A unity gain voltage amplifier is connected to the common point between the

26









resistor and the capacitor as shown in Figure 2-3. The fixed charge on the transducer is given by

the relation

Qdc VdcCo. (2-14)

The RC network consisting of Rdc and C sets up a high pass filter with a cut in frequency of

1 For all frequencies above the cut in frequency, the voltage on the plate is free to change

which is buffered by the voltage amplifier and can be measured. Substituting Q with Qdc in

Equation 2-11, the voltage across the plates of the capacitor is given by


V (2-15)
Co 7Xo"

The small signal output voltage Vot is given by


Vout = Vdc- (2-16)
Xo

From Equation 2-15, it can be seen that


Vout OC x', (2-17)

which is a linear relation between the voltage and the change in gap distance. It should be

noted that in this discussion, the parasitic capacitances are neglected. The effect of parasitic

capacitances are discussed in Chapter 3.

Linearization Using Fixed Voltage (Charge Amplifier Interface Circuit). Figure 2-4

shows a single backplate capacitive transducer connected to a charge amplifier. In this technique,

the fixed voltage is provided by a DC voltage source Vdc which is connected to one end of the

capacitive transducer. The other end is connected to the inverting input of an inverting amplifier.

The inverting terminal of the amplifier is held at virtual ground due to negative feedback. A

feedback capacitor Ct is connected between the output of the amplifier and the inverting input.

This circuit integrates the charge flowing out of the transducer. A bias resistor Rd, is connected

in parallel to the feedback capacitor and is used to set the DC operating point. The RC network

consisting of Rdc and C(,, sets up a high pass filter with a cut in frequency of 21 For all

frequencies above the cut in frequency, the charge flowing from the transducer is integrated by

the feedback capacitor. The relation between the input and output of a charge amplifier above the









Rdc


+V


C7


out
-+




Figure 2-4. Schematic of a single backplate capacitive transducer connected to a charge amplifier.


cut in frequency is given by
C
Vout = Vdc (2-18)

Substituting for C from Equation 2-8 and Equation 2-9, Vout can be expressed as

Co Cox'
Vo.t = Vd CoxVd (2-19)
Cint "0

From Equation 2-19 it can be seen that


Vout oc x', (2-20)


which is a linear relation between the output voltage and the change in gap distance. The effect

of parasitic capacitance on the output voltage is discussed in Chapter 3.

Force Relations. We now consider the force between the plates of the capacitor. By the

law of conservation of energy, the sum of the incremental change in electrical potential energy

dEe and the incremental change in mechanical potential energy dEm is zero, i.e.


dEe + dE, = 0. (2-21)


The incremental change in mechanical potential energy is given in terms of the force between the

two plates F and the incremental change in gap distance dx' by


dEm = -Fdx'. (2-22)









The negative sign denotes that the force increases as gap distance decreases. Rewriting

Equation 2-22, we obtain

F = d (2-23)
dx'

Substituting Equation 2-21 in Equation 2-22, we obtain

dE
Fe (2-24)
dx'

The electrical potential energy stored across the plates of the capacitor E, is dependent on the

biasing technique used. If the capacitor is biased with a fixed voltage Vd, then the energy stored

is given by the relation

Ee= CV2. (2-25)
2

Substituting for C from Equation 2-2 and using Equation 2-25 in Equation 2-24, we obtain an

expression for the magnitude of the electrostatic force acting on the movable plate biased at Vd

1 V2 cA (2-26)
Fvc 2 (xo X')

i.e.,
1
SFelectrostatic I| (X- c 2 (2-27)
(Xo X')
From Equation 2-27, it can be seen that the force between the plates of a capacitor is non-linearly

related to the gap. The direction of the electrostatic force is towards the fixed plate. The movable

plate comes closer to the fixed plate, the force will increase and approach infinity as gap distance

decreases. The force relation can be linearized for small displacements (x' << Xo) using a

Taylor series expansion as

1 eA x'
Felectrostatic V 2 1 + 2- (- ). (2-28)
Vdc 2 x xo

If the plates are biased with a fixed charge, then the energy stored is given by the relation


Ee (2-29)
2C"

Substituting for C from from Equation 2-2 and substituting Equation 2-29 in Equation 2-24 we

obtain

F c (-2) (2-30)
Qdc 2eA
29









This can be rewritten in terms of Vdc as


1 eA
Felectrostatic 2 dc (2-31)
Qdc 2 xo

From this equation, it can be seen that the force is independent of the displacement of the

movable plate for the fixed charge technique.

Quasistatic Pull-in. When the plates of a capacitive transducer are biased with a fixed

bias, it is possible for the plates of the transducer to collapse into each other if the potential

between them is large enough [46]. The reason for this is as follows. When a voltage is applied

to the plates, an electrostatic force attracts the movable plate towards the fixed plate. At small

voltages, this is balanced by the restoring force of the diaphragm. As the voltage is increased,

the electrostatic force between the plates keeps increasing. At some critical voltage, known as

the pull-in voltage Vpi, the two plates will snap in. The study of the pull-in characteristics if

important to determine the maximum voltage Vdc that can be applied on the transducer plates.

The actual voltage used is much lower than the pull in limit to ensure stable operation of the

transducer. There are two main types of pull in.

1. Quasistatic Pull-in: In this type of pull-in the movable plate is initially at rest. No time

varying external force is applied on the movable plate.

2. Dynamic Pull-in : In this type of pull-in, a time varying external force is acting on the

movable plate and/or a time varying bias voltage [70].

In this discussion, only the quasi static case is considered. The pull-in voltage and the critical gap

at which pull-in occurs is derived next for the constant voltage and constant charge techniques.

Quasistatic Pull-in for Single Backplate Transducer with Fixed Voltage Biasing.

Figure 2-5 shows the fixed plate and movable plate along with the direction of forces. It can

be seen that the force due to electrostatic attraction FeIectrostatic and the restoring spring force

Fsprin, oppose each other. The electrostatic force is always attractive and tries to bring both

plates together. The restoring force always tries to bring the movable plate back to its equilibrium

position. Based on the sign convention shown in the figure an increase in the electrostatic force

causes a decrease in gap distance.










Movable Plate I .- -


electrostatic X A


Fixed Plate I I 0

Figure 2-5. Schematic of a single backplate capacitive transducer showing electrostatic force and
restoring force.


The restoring force in the diaphragm Fr is given in terms of the mechanical compliance

Cm,d of the diaphragm by

Fspring (1/Cm,d) x'(I). (2-32)

Using Equation 2-1 Fspring can be expressed as a function of the gap distance, x as

1
Spring = + (Xo x)(). (2-33)

The Felectrostatic due to the electrostatic attraction between the plates is given by Equation 2-25

which is repeated below.
-* 1 eA
Felectrostatic = Vdc2 ). (2-34)
2 X2

The net force acting on the diaphragm F,,t is thus given by the sum of the restoring spring force

given by Equation2-33 and the electrostatic force given by Equation 2-25

1 2CoA 1
IFnt -Vdc2 +- ( x) = 0. (2-35)
2 X2 Cm,d

Initially, when the bias voltage is small, the movable plate will be displaced by a small

amount from its equilibrium position. As the voltage increases, the gap distance decreases. At

any given voltage, the stability is determined by applying a small perturbation to the diaphragm

and analyzing the subsequent diaphragm motion. If a small perturbation causes the diaphragm

to start moving towards the fixed back plate the equilibrium is unstable, else the equilibrium is

stable. To determine the critical gap distance and critical voltage at which the system transforms

from stable to unstable, we partially differentiate the net force Fnet with respect to the gap









distance x. When the plates are biased with a fixed voltage Vdc, OFnet is given by

( 1 coAVdc2\
OFt =- + aA3 Ox. (2-36)
\( Cm,d x3

If the diaphragm is perturbed towards the backplate, i.e., ax < 0, then to ensure stable operation

the quantity in parenthesis must be positive so that F,,et is also negative. Thus,

1 cAVdc2
S + C- < 0. (2-37)
Cm,d X3

Denoting the critical gap distance by xp, and the critical voltage by Vpi, Equation 2-37 becomes

1 ~ Vp12
=+ 0. (2-38)
Cm,d Xp13

By substituting Equation 2-38 into Equation 2-36 and evaluating at xp, and Vpi, the critical gap

distance is
2
xpi = 2Xo, (2-39)
3
and the pull-in voltage is

VP- =A (2-40)
27C,aeA"
Quasistatic Pull-in for Single Backplate Transducer with Fixed Charge Biasing. Now

we derive the critical gap distance and pull in voltage for a single backplate capacitive transducer

biased with a fixed charge biasing technique. For this technique, the electrostatic force if given

by Equation 2-31 which is repeated below


Felectrostatic = 2 ((x)). (2-41)
2 xo

It should be noted that the electrostatic force due to a fixed charge does not depend on the

position of the movable plate. The net force is given by


1 2 A 1
I Fnt Vdc + (xo x) = 0. (2-42)
2 X2 Cm,d
Differentiating the net force,

OFet a-x. (2-43)
Cm.d









If the diaphragm is perturbed towards the backplate, i.e. Ox < 0, there is an opposite restoring

force ensuring that the equilibrium point is always stable. To determine the equilibrium position

Xeq, we set Fnet = 0. The equilibrium position is thus given by

1 eA
Xeq 0 -C,dVdc2 62 (2-44)
2 xo

If too large a voltage is chosen, then Xeq < 0 which implies that the two plates are touching each

other. Furthermore if Vdc is too large then the initial deflection of the movable plate will be too

large and the transducer might operate in the non-linear region. To determine the pull-in voltage

we set Xeq = 0 and solve for the voltage. This is given by


VPI = (2-45)
VP1 C,.,d6A"

Comparing the pull-in voltage between the constant charge and constant voltage biasing

techniques, it can be seen that the pull-in voltage limit is higher for the constant charge

technique.

2.2.2 Canonical Dual Backplate Capacitive Transducer

Another commonly used structure in capacitive sensors is a a dual backplate structure [71]

as shown in Figure 2-6. The need for a differential structure is motivated by a need to linearize

Fixed Plate I I


C2=C20- AC2

I+ AC
f__CC 10 1


Movable Plate


Xo

X x
xo0


Fixed Plate II

Figure 2-6. Schematic of a differential capacitive transducer.


the system using structural modification. Also the differential structure has twice the change

in capacitance of the single backplate structure. The improvement in sensitivity and linearity is

discussed later in this section.

This structure consists of a movable plate sandwiched between two fixed plates. Two

capacitances C1 and C2 are formed, one between the top backplate and the movable plate and

the other between the bottom back plate and the movable plate. Under the application of an input










force on the movable plate, one of the capacitances increases while the other one decreases from

the nominal value. Under zero input force, the nominal capacitance of the two capacitors are

denoted by Clo and C20. If the change in capacitance of capacitor C1 is denoted by AC1 and the

change in capacitance of capacitor C2 is denoted by AC2, then


C1 = C1o + AC1 (2-46)

C2 = 20 AC2.


It is assumed for the initial discussion that the nominal capacitance and change in capacitance are

matched i.e.

Clo C20 C (2-47)

and

ACI AC2 = AC. (2-48)

The effects of capacitance mismatch will be discussed later in this chapter. From the geometry,

C1 and C2 is given by

coA
C1 = (2-49)
Xo X'
coA
C2 cA
Xo + X'

The total capacitance change AC is given by


AC = Ci C2. (2-50)


Substituting for C1 and C2, we obtain

coerA coerA
CI 2 C=cA C (2-51)
Xo xo + X'

This can be simplified to

C1C2 c2 2 2r x' (2-52)
X0 X
For small displacements i.e [x' < Xo], x'2 can be neglected and the above equation can be

simplified as

CI C2 2 xrA'. (2-53)









Using Equation 2-4, Equation 2-53 can be simplified to


C, C2 = 2Co-. (2-54)


From Equation 2-54, it can be observed that the capacitance change is a linear function of the

change in gap distance for small displacements. Comparing Equation 2-54 with Equation2-9 it

can be observed that the change in capacitance for the dual back plate structure is twice that of

the structure with a single backplate.

We now consider the voltage across the plates of the two capacitors. If the voltage across

capacitors C1 and C2 is denoted by V1 and V2, respectively, and the charge stored is Qi and Q2,

respectively, then


V Q1 (2-55)
coA

V2 X+ 2,
coA

Equation 2-56 show that the voltages across the plates of a capacitor are a non-linear function of

the displacement and charge stored in the plates of the capacitor. Similar to the case with a single

backplate, the above equations can be linearized using either a fixed charge or a fixed voltage.

This is discussed next. i.e,


V1 oc Q1x' (2-56)

V2 OC Q2'.


Fixed Charge Technique (Voltage Amplifier Interface Circuit). In the fixed charge

technique, the top and bottom backplates of the microphone with top and bottom capacitances

given by C1 and C2 are biased at a DC potential of +Vdc and -Vdc volts as shown in Figure 2-7.

A unity gain voltage amplifier is connected to the middle plate. Rdc is a bias resistor that sets the

DC operating point. The linearized relation between the input and output of a charge amplifier

above the cut-in frequency is given by


Voxt 2Vx'. (2-57)


The detailed derivation of this equation is provided in Appendix B.









+V


C1




-v
C2 < RDC

-Vdc

Figure 2-7. Schematic of a dual backplate capacitive transducer connected to a voltage amplifier.


Fixed Voltage Technique (Charge Amplifier Interface Circuit). In this technique, the

top and bottom backplates of the microphone with top and bottom capacitances given by C1

and C2 are biased at a DC potential of +Vdc and -Vdc volts. Figure 2-8 shows a dual backplate

capacitive transducer connected to a charge amplifier. A feedback capacitor Ct is connected

+Vdc DC


C,


o ut
C2 Cp
CL

Vdc

Figure 2-8. Schematic of a dual backplate capacitive transducer connected to a charge amplifier.


between the output of the amplifier and the inverting input. This capacitor integrates the charge

flowing out of the transducer. A bias resistor Rdc is connected in parallel to the feedback

capacitor and is used to set the DC operating point. The RC network consisting of Rdc and Cmt

sets up a high pass filter with a cut-in frequency, of 21RC For all frequencies above the cut

in frequency the charge flowing from the transducer is integrated by the feedback capacitor. The

linearized relation between the input and output of a charge amplifier above the cut in frequency

is given by
2Co Vdc
Vout x (2-58)
X0 3 int
36









The detailed derivation of this equation is provided in Appendix B.

Force Relations. We now consider the force on the movable plate for the differential

structure. Figure 2-9 shows the schematic of the dual backplate capacitive transducer and the

force acting on the movable plate. The force on the movable plate depends on the type of biasing

+VC
dc

Fixed Plate I
-X
Movable Plate x
-_ \x--' F"
V------------- ---------- --X

Fixed Plate I
8-v
Vdc
Figure 2-9. Schematic of a dual backplate capacitive transducer showing forces on the movable
plate.


technique used. We first consider the fixed voltage biasing technique. The net electrostatic force,

Felectrostatic, on the movable plate can be expressed in terms of the individual forces Fi and F2 by


Felectrostatic = FI + F2. (2-59)


Using Equation 2-26, F1 and F2 are given by

-1 A
F, -- ( (2-60)
2 dc (xo 2 (-) (2 60)

1 A2 c
F v2 ). (2-61)
2 C(xo + ')(261)

The two forces operating on the movable plate are acting in opposition. This is because the

force between the movable plate and each of the fixed plates is attractive. Fi is acting in the

negative x direction. The total force Felectrostatic is given by the sum of F1 and F2. Substituting

Equation 2-60 and Equation 2-61 in Equation 2-59 and simplifying, Felectrostatic is given by


Felectrostatic 2 AXdc o 2 (--2 ). (2-62)
(c (xo2 X'2)

Comparing Equation 2-62 with Equation 2-26, it can be seen that the net electrostatic force with

a constant voltage bias is lower for a dual backplate capacitive transducer as compared to a single









backplate capacitive transducer. The force will still tend towards infinity as the fixed plates starts

moving towards the movable plate. At rest, there is no net force acting on the movable plate.

We now consider the case when the two plates are biased with a fixed charge. In this

case there are two components to the force. They are the force between the top backplate and

the movable plate and the force between the bottom backplate and the movable plate. Using

Equation 2-30, the two forces on the diaphragm are


=A (-_J) (2-63)
2cA
F2
2cA

The total force Feiectrostatic IQdcis given by the sum of the two forces given by Equation 2-63 i.e


Felectrostatic Qd= F1 + F2 = 0 (2-64)

which is zero for constant charge case for matched dual backplates.

Quasistatic Pull-in. We now derive the pull in voltage and gap distance for the dual

backplate capacitive transducer structure.

Quasistatic Pull-in for Dual Backplate Transducer with Fixed Voltage Biasing. The

derivation of the pull-in voltage and pull-in gap distance for a dual backplate microphone with

fixed voltage biasing is similar to that of the single backplate case. The net electrostatic force is

given by Equation 2-62 and the restoring force is given by Equation 2-32. The net force is given

by
( xox
S= -2Vc2 x. (2-65)
-2v (xo2 2)2 Cm,d

In the case of a dual backplate structure, there are two gap distances. Hence, we analyze the net

force by differentiating it with respect to the distance moved from the nominal gap x' instead of

the gap distance. Differentiating Fet with respect to x', we obtain

(2V c22Axo x +3x'2
8Fet = [ 2 2) ( 3 ) ] a X'. (2-66)
(\O [X ) ) \ 0
For the system to be stable, a small perturbation in the change in gap distance, i.e Ox' > 0, must

cause a decrease in net force i.e. OF,, < 0.









This implies
( 2V eAxo \ 12 2
(2V co20Axo + 3x'P2 < 0. (2-67)
) P 2 / 1 2 m ,d
Defining the voltage at the transition point from stable to unstable region of operation by Vpi and

the displacement from equilibrium position by x'p1, we have

(2Vp 2 x, ( + 3x? 1
2V 0Axo x + 32= 0. (2-68)
2( /2 2 2 X /2 2m,d

Substituting Equation 2-68 in Equation 2-65 and solving for x';p, we have


x'I = 0. (2-69)


This shows that the equilibrium position for a dual backplate structure biased with a fixed voltage

is at the rest position of the diaphragm. The pull in voltage is given by

x03
Vp1 = (2-70)
S 2Cm,~eA

Quasistatic Pull-in for Dual Backplate Transducer with Fixed Charge Biasing. The

net force on the dual backplate transducer with a fixed charge biasing as given by Equation 2-64

is zero. Hence the equilibrium position is x' = 0. Based on the model used, there is no upper

bound on the charge that can be placed on the transducer plates. This is not practically true as

the model did not account for non-idealities such as curvature of the plates, or other dynamic

conditions. Based on the physical realization of the transducer, there will be limits on the

maximum charge that can be applied on the plates.

2.3 Structure of Microphone Used as Test Vehicle for Open and Closed Loop Interface
Circuit Design

The overall goal of this work is to investigate open loop and closed loop interface circuits

for capacitive microphones. A dual backplate capacitive microphone was used as the test vehicle

to characterize various interface circuits. To facilitate the design of the interface circuit, a

model of the microphone that accurately captures the dynamics is required. In this section, the

microphone structure and material properties are described from which a model can be developed

. The actual model development is discussed in the next section.










The microphone consists of a solid diaphragm sandwiched between two porous backplates

over a large cavity. A vent is provided to equalize the pressure on both sides of the diaphragm.

The cross section of the microphone is shown in Figure 2-10. The microphones were fabricated

using the Sandia SUMMIT V process [64]. This five-layer poly process allows the fabrication of

ultra-planar poly layers with low in-plane stress. Two microphone structures were designed for

use in aeroacoustic application and for audio application. The material properties and dimensions

of the two microphone structures are listed in Table 2-1.

Acoustic holes Top backplate
Dia hragm
-T- C,

Cavity Bottom
backplate
Substrate

Figure 2-10. Cross section view of dual backplate microphone [64].


Table 2-1. Dimensions of aeroacoustic and audio microphone[64].
Property Aeroacoustic Audio
Radius of Top Backplate 256/pm 756/pm
Thickness of Top Backplate 2.25/ m 2.25/ m
Radius of Bottom Backplate 213/pm 713/pm
Thickness of Bottom Backplate 2.5/pm 2.5/pm
Radius of Diaphragm 230/pm 730/im
Thickness of Diaphragm 2.25/pm 2.25/pm
Thickness of Top Gap 2/pm 2/pm
Thickness of Bottom Gap 2/pm 2/pm
Radius of Holes 5/pm 5/pm
No. of holes in Top Backplate 557 5030
No. of holes in Bottom Backplate 367 4361
Depth of Cavity 650/pm 650/pm
Radius of Cavity 186/pm 686/pm



Table 2-2. Material properties of microphone structure.
Property Magnitude
Young's modulus of polysilicon (e) 1.6 x 1011Pa
Poisson's ratio of polysilicon (v) 0.22
Density of polysilicon (pp) 2.2 x 103kg/m3









2.4 Lumped Element Model

In order to develop an interface circuit that will meet the overall system level specifications,

we need to develop a model that accurately captures the dynamics of the microphone as a

function of the dimensions and material properties. We use the technique of lumped element

modelling (LEM) to achieve this goal. LEM is valid when the length scale of interest is much

smaller than the characteristic wavelength of the physical phenomenon [46]. In this case, the

length scale of interest is the diaphragm diameter and the characteristic wavelength is the

acoustic wavelength, A. Using LEM, a theoretical formulation for the resonant frequency and

damping coefficient is developed. The general technique for developing the LEM for a transducer

will be discussed next followed by the application of the technique to the microphone under

consideration.

To develop the LEM, we consider the power flow between the various mechanical elements

using the principle of conjugate power variables. In this technique, power is denoted as the

product of two generalized quantities, effort and flow. In the electrical domain, the effort variable

is the applied voltage and the flow variable is the electric current. The product of these two

quantities is the power. In the acoustic domain, the effort variable is pressure, and the flow

variable is volume velocity. Using the effort and flow variables in the system, we next discuss the

energy storage and dissipative elements in the system.

The ratio of the effort variable to the flow variable represents the generalized impedance in

the system. In the electrical domain, the ratio of the applied voltage to the current is equal to the

impedance which if in phase represents the electrical losses in the electrical domain. Similarly,

the ratio of the input pressure to the volume velocity is the acoustic impedance. A real acoustic

impedance is denoted by a resistor and it represents the energy loss mechanisms in the acoustic

domain. The physical cause of the energy loss in the acoustic domain is viscous damping and

radiation resistance.

The energy storage elements are discussed next. In the electrical domain, the capacitor

stores potential energy (energy associated with the effort variable) and the inductor stores kinetic

energy (energy associated with the flow variable). In the acoustic domain, the kinetic energy

and potential are stored by the motion and deflection of the elastic diaphragm respectively. The









distributed elastic deflection of the diaphragm is lumped into a rectilinear piston having the

same deflection as the center of the diaphragm and a spring. The equivalent mass of the piston is

determined by equating the kinetic energy of the distributed diaphragm motion into the kinetic

energy of the lumped piston motion. Similarly, the compliance of the spring is found by equating

the potential energy stored in the distributed diaphragm deflection to the potential energy in the

lumped compliance.

Finally, we discuss the rules that govern the interconnection of the various elements. In the

impedance analogy, elements that share the same flow and displacement variable are connected

in series while elements that share the same effort variable are connected in parallel. Also, for the

impedance analogy, by Kirchoff's laws, the sum of all the efforts around a closed loop is zero,

and the sum of all the flows entering a node is zero.

The LEM of the dual backplate microphone structure is discussed next. All the LEM

parameters discussed are in the acoustic domain. Ideally, the top and bottom backplates are

rigid. However, physically they do deflect by a small amount. Hence, they are modelled as

capacitors Cbpl, Cbp2 which represents the storage of potential energy. The main loss mechanism

in the backplates is the viscous damping loss [72]. These are represented by Rbpl, Rbp2 which

are connected in parallel to the corresponding compliance as they both share the same effort

variable. The diaphragm responds to the input pressure Pi, and deflects storing potential

energy and is modelled as a capacitor Cd. The diaphragm motion stores kinetic energy and

is modelled as an inductor 3 ,. These elements share the same volume displacement (flow

variable) and are thus connected in series. Viscous losses associated with the diaphragm are

neglected. The cavity impedes the motion of the diaphragm by storing potential energy and

is modelled as a compliance Cca,. The cavity between the top and bottom back plates and the

diaphragm is neglected. The resistance of the vent is modelled by as a resistor R,. Theoretical

formulation for each of the model parameters is listed in Table 2-3 [73, 74, 75]. Using the

formulation in Table 2-3, and the structural and material properties listed in Table 2-1 and

Table 2-2 respectively, the model parameters for the two microphones are calculated. The results

are listed in Table 2-4.

The LEM of the microphone with approximations discussed is shown in Figure ??. In the

frequency range of interest, the top and bottom plate compliances and the vent resistances can be

42








Table 2-3. Theoretical formulation for lumped element model parameters.
Parameter Symbol Expression
Diaphragm Mass Mm 9p
Diaphragm Compliance Cma '1-E )
Cavity Compliance Ccav p- 2
Backplate Hole Resistance Ra 12mXB(A)
Vent Resistance R, 2

Table 2-4. Lumped element model parameters of aeroacoustic and audio microphone.
Parameter Symbol Aeroacoustic Audio
Microphone Microphone
Top backplate compliance (m4s2/kg) Cbp1 2.8 x 10-17 1.9 x 10-14
Bottom backplate compliance (m4s2/kg) Cbp2 6.9 x 10-18 9.8 x 10-15
Diaphragm Compliance (m4s2/kg) Cd 1.5 x 10-17 1.6 x 10-16
Cavity Compliance (m4s2/kg) Ccav 5.0 x 10-16 6.8 x 10-15
Top backplate hole resistance (kg/m4s2) Rbpl 1.6 x 109 1.7 x 108
Bottom backplate hole resistance (kg/m4 s2) Rp2 1.3 x 109 1.6 x 108
Vent Resistance (kg/m4 2) Rv 7.5 x 1012 7.5 x 1012
Diaphragm mass (kg/m4) Md 5.7 x 104 5.6 x 103

neglected. Under this condition, the microphone model can be simplified into a simple second
order LCR circuit as shown in Figure 2-11. The LCR system can be described by a second order


Rbpl Md d Rbp2



( )P
cav



Figure 2-11. Low frequency LEM of dual backplate microphone.


displacement to pressure transfer function H(s) [76]

H(s) = 1 2, (2-71)
S2 + 2 x ( x 27Tfr X s + (27Tfr)2

where the resonant frequency fre of the system is given by the expression

1 1
fres (2-72)
Cv+Cd











and the damping coefficient ( by the expression


(= (2-73)
Rbpl+ Rbp2 C (27

Substituting for the various parameters in Equation 2-72 and Equation 2-73, we obtain the

Table 2-5. Second order system parameters of microphones.
Quantity Aeroacoustic (Theoretical) Audio (Theoretical) Aeroacoustic (Experimental)[76]
frs 174kHz 31kHz 230kHz
< 0.018 0.15 0.025


20
0
-20
S-40
S-60
-80
-100
100 102 104 106 108



0
-50

S-100

-150

100 102 104 106 108
Frequency (Hz)

Figure 2-12. Theoretical normalized frequency response of audio microphone.


resonant frequency and damping coefficient for the two microphone designs. The results are

listed in Table 2-5. A plot of the normalized theoretical transfer function of the audio and

aeroacoustic microphone is shown in Figure 2-13 and Figure 2-13 [76].

2.5 Summary

In this chapter, canonical capacitive transducers were introduced. The charge, voltage,

capacitance, and force relations for the single and dual backplate microphone were derived.

The structure and material properties of the microphone that are used as the test vehicle to

characterize the open and closed loop circuit described in Chapter 3 were presented. A lumped

element model of the microphone was then developed. A reduced second order model of the

microphone was derived from the general lumped element model from which the resonant

frequency and damping coefficient were extracted. Simulations results of the frequency response

of the microphone model were also presented.















0)

-50


-100
100 102 104 106 108


0

S-50-
o
S-100-

-150

100 102 104 106 108
Frequency (Hz)


Figure 2-13. Theoretical normalized frequency response of aeroacoustic microphone.



To design the closed loop circuit, a model of the interface circuit is required. Chapter 3


presents an overview of the various capacitive transducer interface circuits along with models


that will be used in Chapter 4 and Chapter 5 for the development of the overall closed loop


system.









CHAPTER 3
INTERFACE CIRCUITS

3.1 Introduction

This chapter discusses the various open-loop and closed-loop capacitive transducer interface

circuits with an emphasis on circuits for capacitive microphones. For each technique, the

principle of operation is first explained. Then theoretical formulations for the input/output

characteristics, sensitivity, dynamic range, bandwidth, and minimum detectable signal are

developed. A review of commercial transducers and research prototype transducers reported


Figure 3-1. Overview of Chapter 3.


in the published literature using these interface circuits is presented. This chapter concludes

with a summary of the various techniques. A graphical overview of this chapter is provided in

Figure 3-1.

3.2 Performance Metrics of Capacitive Transducers

In this section, the main performance metrics [77] of capacitive transducers with an

emphasis on capacitive microphones is introduced. The sensitivity, S, of a capacitive transducer

is defined as the change in the output voltage for an input physical phenomenon-derived change









in capacitance. For capacitive microphones, the input phenomenon is a time varying pressure

signal, and the microphone sensitivity is measured in Volts/Pascal. Similarly, for a capacitive

accelerometer, the input is an acceleration commonly measured in g. Therefore, the units of

sensitivity for an accelerometer is Volts/g. The sensitivity of the transducer can be expressed

as the product of the mechanical sensitivity (Sm) and electrical sensitivity (Se). The mechanical

sensitivity denotes the transfer function between the change in capacitance for an input change

in physical phenomenon. The electrical sensitivity refers to the change in output voltage for an

input capacitance change.

The emphasis of this chapter is on interface circuits which measure capacitance. Hence,

the electrical sensitivity of the interface circuit is expressed in V/F. It should be noted that the

sensitivity is defined as the magnitude of the flat region of the frequency dependent transfer

function. For a linear transducer, the transfer function H(w) is defined as the ratio of the Fourier

transform of the output of transducer Y () to the Fourier transform of the input of the transducer

X(w), i.e.

H(w)= (3-1)

The transfer function is in general a complex quantity. Hence, the complete description includes

both the magnitude IH(w)l and phase ZH(w) I as a function of frequency. The magnitude of the

transfer function, as defined in Equation 3-1, is thus a plot of the sensitivity of the transducer as a

function of frequency. For an ideal transducer, the phase must be constant and zero.

The typical frequency response of a capacitive transducer is shown in Figure 3-2. It can

be seen that over a certain frequency range, the frequency response magnitude is flat within

a given tolerance, say 3dB. For a transducer, this is the range of frequencies over which it

can be used and is defined as the bandwidth of the sensor. The lower and higher roll off of

the transfer function is caused by both the transducer structure and the interface electronics.

For a microphone, the lower end of frequency range (denoted by fi), also known as the cut on

frequency, is determined by either the biasing resistor of the interface circuit or the acoustic

resistance of the vent channel and cavity compliance [53]. Similarly, the higher end of the

frequency range (denoted by fh), also known as the cut off frequency, is determined by the












I Resonance Peak


r-



SFlat frequency response fh Frequency
(bandwidth)



S0
-o
0--90
-180

Frequency (Hz)

Figure 3-2. Characteristic frequency response of a capacitive transducer.


bandwidth of the interface circuit or the mechanical resonance of the sensor structure which in

turn is determined by the diaphragm mass and compliance.

The next performance metric to be considered is the linearity of the sensor. The linearity

of the sensor is a measure of how close the output versus input calibration curve approximates

a straight line at a given frequency. A plot of the ideal and actual linearity of a transducer is

shown in Figure 3-3. The slope of the straight line provides the sensitivity of the transducer at


... Ideal response
Actual response








Input Pressure


Figure 3-3. Characteristic ideal and actual response of a capacitive transducer.



that frequency. It can be seen that at high input amplitudes, the output of the transducer deviates

from the ideal straight line curve. The lower and higher ends of the linear range are determined

by both the sensor interface circuit and the sensor. The lower end is limited by system noise

such as thermal noise, 1/f noise, and mechanical noise. The higher end of the linear range is

determined by structural non-linearities such as spring stiffening [78] or by circuit non-linearities









such as clipping [79]. The dynamic range of a capacitive transducer is defined as the ratio of the

maximum and minimum input signal of the linear range.

The deviation of the output from the ideal linear curve causes distortion in the microphone

output. Distortion can be defined in two different ways based on the input signal. When the

system is excited at a single frequency, distortion is defined as the minimum input amplitude that

causes the output to deviate from linearity by a fixed percentage. For a measurement microphone,

the deviation is usually chosen to be :'.. When the transducer is characterized over a frequency

range, the distortion is defined in terms of the total harmonic distortion (THD). THD is defined

as the ratio of the sum of the power in all the harmonic frequencies (Uc) to the fundamental

frequency (wo).
0" p2(L )
THD o= n (o (3-2)
P2( o)
The output noise floor of a transducer is the noise mechanism generated output for zero

input signal. Referred to the input, the input noise floor represents the minimum detectable signal

(MDS). The output referred noise voltage power spectral density (PSD) of the interface circuit

denoted by Sv,, is expressed in V2/Hz. The electrical noise at the output of the interface circuit

can be referred to the interface circuit input by dividing the output noise by the square of the

electrical sensitivity. The input signal PSD of the interface circuit is expressed in F2/Hz. Thus,

the output noise power in a system is dependent on the bandwidth of operation. The total output

noise power or mean square output noise voltage is obtained by integrating the output noise

voltage PSD over the measurement bandwidth i.e.


-Vo j20 Sv0df. (3-3)


and the rms output noise voltage is n,rms = 0. A characteristic noise PSD curve is shown

in Figure 3-4. The overall system noise includes noise contributions from both the sensor and

the interface circuit. At low frequencies, the noise PSD is dominated by the flicker noise [80]

of the interface circuit which has a 1/f spectral shape. It should be noted that flicker noise is

present only when a dc current flows. At high frequencies, the noise is dominated by the thermal

noise [80] above a comer frequency. The corner frequency is the point at which the asymptote of

the 1/f noise region meets the thermal noise floor. Thermal noise is present in all systems that











-N-

0 1/f Noise region

SCorner frequency
T5
z
Thermall/f Noise region
Frequency (Hz)

Figure 3-4. Characteristic noise floor of a capacitive transducer using a voltage amplifier
interface circuit.


are in thermodynamic equilibrium. The magnitude of the thermal noise PSD is proportional to

the dissipation in the system.

To study the effect of the interface circuit noise on overall system performance, the noise

models developed in this chapter consider only the electronics noise of the interface circuit

elements and the spectral shaping of these noise sources by the interface circuit elements.

The commonly used elements in the interface circuit are resistors, capacitors and operational

amplifiers (opamps).

The noise model of the resistor is considered first. The 1/f noise of a resistor is dependent

on the fabrication method used to manufacture the resistor. Metal film resistors have the lowest

1/f noise and have a flat frequency response across the entire spectrum. Figure 3-22 shows the

voltage and current noise representations of a resistor. The PSD of resistor voltage and current

noise representations is given by

SR = 4kTR (3-4)

and
4kT
SiR = (3-5)
R

respectively.

Ideally, capacitors do not have any noise associated with them. In practical applications,

capacitors have an equivalent series resistor (ESR) which has a noise associated with it. The

ESR of a capacitor is small in comparison with other physical resistors present in the system.

Hence the noise of the physical resistor dominates over the noise due to the ESR. Hence, noise in

capacitors are not considered in the discussions in this chapter.










R

R
4-

SvR SiR
(a) Voltage noise model (b) Current noise
of resistor, model of resistor.

Figure 3-5. Noise model of resistor.

The noise model for the opamp is discussed next. Figure 3-6 shows the noise model of an

opamp consisting of a voltage and current noise voltage source at the non-inverting input of

the amplifier. The fundamental physical phenomena which cause noise in the opamp are the

thermal noise, shot noise, and 1/f noise in the transistors and resistors inside the opamp. The

typical noise voltage PSD (S,) and current noise PSD (Si) is shown in Figure 3-7. It can be

seen that at low frequency the PSD has a 1/f shape and at large frequency, the PSD is constant.

The point at which the asymptote of the 1/f noise meets the thermal noise floor is the comer

frequency and is denoted by f, and fji for the voltage and current noise PSDs respectively. At

any given frequency, the voltage and current noise PSD, S,(f) and Si(f) respectively, can be

approximately expressed in terms of the corresponding voltage and current thermal noise limits,

S,0 and Sio by

S (f) S0o 1j+ f (3-6)

and

Si) So 1l + f1 (3-7)


Sva
+ Vout
out



Figure 3-6. Noise model of opamp showing voltage and current noise source.


3.3 Analog Open Loop Sense Techniques

The function of the sense electronics is to measure the capacitance change associated with

the deflection of the diaphragm and provide a corresponding voltage signal. In the open-loop

scheme, the change in capacitance is measured directly by voltage or charge biasing either the

51










backplates or the diaphragm and measuring the voltage or charge signal generated on the other

plate. No feedback signal based on the measured output signal is applied to the microphone

plates to null the deflection caused by the input phenomenon. The open-loop techniques can be

further classified into analog and digital schemes which are discussed next.

N S'

> /f Noise region 0 1/f Noise region

SCornerfrequency W Corner frequency
0
0 z
z / z
Z
I s ,0 ............... .............h.......
S .......... Thermal Noise regional Nose region

Frequency (Hz) Frequency (Hz)
(a) Plot of voltage noise PSD of (b) Plot of current noise PSD of
amplifier, amplifier.

Figure 3-7. Plot of noise PSD of amplifier.


The performance metrics discussed so far will be derived for the various open and

closed-loop interface techniques. The analog open-loop sense schemes are characterized by the

absence of a clock signal as compared to the digital open loop sense techniques. The commonly

used analog open-loop techniques sense the voltage change using a voltage amplifier under a

constant charge bias, sense the charge using a charge amplifier under a constant voltage bias,

and sense using synchronous modulation/demodulation. Each of these techniques is discussed

in detail. For each of these interface techniques, the principle of operation, formulation for

performance metrics, tradeoffs, and a literature review is presented. Figure 3-8 shows a flow

diagram for each section. The derivation of the equations for each of the interface techniques is

presented in Appendix B.

3.3.1 Sensing Using A Voltage Amplifier

Operating Principle. In the simplest case, the top and bottom backplates of the

microphone with top and bottom capacitances given by C1 and C2 are biased at a dc potential

of +Vd, and -Vd, volts. A unity gain voltage amplifier is connected to the middle plate. Rde is

a bias resistor that sets the dc operating point. Figure 3-9 shows a schematic of a dual backplate

capacitive microphone connected to a voltage buffer including parasitics. Cp represents the

total parasitic capacitance at the input to the amplifier due to the packaging. The parasitic

capacitance of commonly used packaging techniques is presented in Table 3-1. For a single chip

52




























Figure 3-8. Organization of each interface circuit discussed in Chapter 3.

+VCl
dcI"

C,/

Vout
C -- 02 \ DC




Figure 3-9. Schematic of a dual backplate microphone connected to voltage buffer.

solution in which both the microphone and interface circuitry are on the same die, the sensor and
interface circuitry are electrically connected using metal or poly lines and the parasitic or input

capacitance is < lpF. For two chip solutions, the parasitic capacitance is on the order of a few

pF depending on the interconnect used such as bond wires or PCB tracks. Ci represents the total
input capacitance of the amplifier. For an ideal interface circuit with no parasitic capacitance, the

output of the circuit, for a given change in capacitance, AC, is given by,

Vd
V,, = VAC, (3-8)
Co









Table 3-1. Parasitic capacitance for various packaging technologies.
Packaging Parasitic Capacitance
n or metal On die interconnect < 10fF/um (Layout
dependent)
[81] Interconnect between unpackaged 30fF/mm


PCB Trace (FR4 PCB)

BNC cable
(RG164[82]])


dies
Interconnect between packaged dies

Interconnect between different
boards


1 2pF/cm (Layout
dependent)
20.6pF/ft


which is the same as a single backplate microphone. The derivation of all the equations is

provided in Appendix B. For a real interface circuit with parasitic capacitances, the output

voltage is given by

Vout = 2AC.
2Co + C, + C


(3-9)


Equation 3-9 can be rewritten as


Vdc c 2Co 1 Ac
Co 2Co+Cp+QC .

Defining the overall voltage attenuation factor He as


H = 2 CiO
2Co + Cp + C,


(3-10)





(3-11)


Equation 3-10 can be written as


Vot = H dAC.
co


(3-12)


Comparing Equation 3-12 and Equation 3-8, it can be seen that the output voltage is attenuated

by the factor He that is dependent on the parasitic capacitance.

Performance Metrics. The first performance metric that will be derived is the sensitivity

of the circuit. Using the definition of sensitivity discussed in Section 3.2 and using Equation 3-9,

the electrical sensitivity Se of the circuit is given by


2Vec
2CGo + + C


(3-13)


From this equation, it can be seen that the sensitivity of the circuit is reduced with increased

parasitic capacitance.

To determine the total output noise, the noise model shown in Figure 3-10 is used. The PSD


Technique
Polysilicoi
lines
Wire bond

















C2 "1 -I ia LL
TT sVR



Figure 3-10. Noise model of dual backplate microphone connected to voltage buffer.


of the circuit Sv,, is determined by the amplifier noise and the thermal noise of the bias resistor

and is given by
Z T) 2 Z 2
SVz Sa + S i2i 2 S R, (3-14)
Zi + Rde Zi + Rdc
where Zi = j 12Co p+C
jwC(2Co+Cp+Ci)
To obtain the minimum detectable capacitance per Hz, the noise PSD equation is divided

by the square of the electrical sensitivity Se followed by a square root. The minimum detectable

capacitance Cmin is given by

(2Co + Cp + Ci ZiRdc zi
Cmin. (= [ +C+- S+SCa2 2 + S2R [F//Hz] (3-15)
2Vd V Zi + Rdc Zi + Rdc

From the equation for Cmin, it can be seen that the minimum detectable capacitance can be

decreased by increasing the bias voltage. It can also be seen that the parasitic capacitance

increases the minimum detectable capacitance.

The upper end of the circuit dynamic range is limited by the maximum output voltage of the

amplifier and is dependent on the topology of the amplifier used. For a given power supply Vdd,

the maximum output voltage Vomax is determined by a circuit dependent voltage drop Vdro, below

the supply rails Vdd and is given by the expression


Vomax = Vdd Vdrop. (3-16)

For example, for an actively loaded single stage amplifier, the minimum Vdrop is given by

the overdrive voltage across the load transistor. Dividing the above equation by the electrical









sensitivity, the maximum input capacitance Cmax that can be measured is

2Co + Cp+C
Cmax 2C0 + C [Vi Vdrop (3-17)
2Vd

This shows that the upper end of the dynamic range can be increased by decreasing the top and

bottom plate voltage and increasing the sense capacitance.

The next performance metric that we derive is the circuit bandwidth. The lower limit of the

frequency response is set by the total capacitance at the inverting node and the bias resistor. [83].

It is given by the expression
1
fi = (3-18)
27TRdc (Cp + C, + 2Co)

The upper limit of the bandwidth fh of this circuit is determined by the amplifier. For a folded

cascode amplifier, the bandwidth is determined by the unity gain frequency, and is expressed in

terms of the load capacitance CL and transconductance gm by the expression [79]

1 g,~
fh -t (3-19)
27 CL

Tradeoffs. The main advantage of this circuit is its simplicity. It has very small

component count and can be easily implemented on an IC or on a PCB. However, the circuit

has some cons which are discussed. The overall sensor sensitivity is reduced by the parasitic

capacitance which is not easily controllable. A plot of the electrical sensitivity as a function of

parasitic capacitance is shown in Figure 3-11 for different sensor capacitances. It can be seen

that as sensor capacitance is scaled down, the effect of parasitic capacitance is exacerbated. The

noise floor of the interface circuit given by Equation 3-14 is dependent on the filtering of the

amplifier and resistor noise by the sense and parasitic capacitance. Furthermore, the choice of

amplifiers also makes a difference as amplifiers are optimized for either low current noise or low

voltage noise and not both. Figure 3-12 shows a plot of the noise contributions by the various

noise sources and the total output noise.

To understand the effect of the various circuit elements, on the output noise PSD,

Figure 3-13 and Figure 3-14 are plotted. Figure 3-13 shows a set of plots of the output noise

PSD as the sensor capacitance is scaled from 100pF to lpF for various parasitic capacitances

assuming a fixed bias resistor of 10MQ.












lpF
0 5pF
+ 10pF
50pF
l100pF
0
0
0
O
CL 0
c O0
+ 0,


10 + Decreasing Sense Capacitance
z ~++










0 5 10 15 20 25 30 35 40 45 50
Parasitic Capacitance (pF)

Figure 3-11. Variation of sensitivity as a function of total parasitic capacitance for various sense
capacitances


Figure 3-13(a), Figure 3-13(c), and Figure 3-13(e) show these plots for an amplifier with

low voltage noise and Figure 3-13(b), Figure 3-13(d), and Figure 3-13(f) show these plots for a

low current noise amplifier. It can be seen that for a fixed bias resistor, an amplifier optimized for

low current noise has a lower noise floor as compared to an amplifier with low voltage noise. It

can also be seen that this is true for all combinations of sensor and parasitic capacitance.

Figure 3-14 shows a set of plots of the output noise PSD capacitances as the sensor

capacitance is scaled from 1pF to 100pF for various bias resistors assuming a fixed parasitic

capacitor of 10OpFQ. In Figure 3-14(a), Figure 3-14(c), Figure 3-14(f) show these plots for

an amplifier with low voltage noise and Figure 3-14(b),Figure 3-14(d), Figure 3-14(e) show

the plots for low current noise. It can be seen that for a fixed parasitic capacitor, an amplifier

optimized for low current noise has a lower noise floor as compared to an amplifier with low

voltage noise. It can also be seen that this is true for all combinations of sensor capacitance and

bias resistor.

Literature Review. The voltage amplifier is the most common interface circuit used

in measurement microphones and audio microphones both for commercial and research











Voltage Noise
** ...... Current Noise
Resistor Noise
10-10 Total Noise











10-25
10 : :

S10-0



10 -2 0 2 4 6 8 1



10o30 --------------
10-2 100 102 104 106 108 1010
Frequency (Hz)

Figure 3-12. Power spectral density of individual noise contributors and total output noise in a
voltage amplifier based interface circuit. (Low Voltage Noise Amplifier (OP471)
So = 0nV/ ,Hz, Sio 1= .6pA/HTz, f,, 10Hz, fi =- 10Hz, Rd = 44MQ,
Cp = 10pF,C, = 2.6pF, 2Co = 1.7pF)


applications. Table 3-2, Table 3-3, and Table 3-4 list the properties of commercial and research

prototype voltage amplifiers used for capacitive microphones. One of the commonly used

commercial measurement microphone for aeroacoustic applications is the Bruel and Kjaer Type

4138 microphone [38].The nominal capacitance of this microphone is 6.2pF. The interface

circuit used is the Type 2670 voltage amplifier.

One of the commonly used commercial measurement microphone for aeroacoustic

applications is the Bruel and Kjaer Type 4138 microphone [38].The nominal capacitance of

this microphone is 6.2pF. The interface circuit used is the Type 2670 voltage amplifier. The Type

2670 amplifier has a flat frequency response from 15Hz 200kHz and a noise level of 14pV

integrated over a 30Hz to 300kHz. The input capacitance of the preamplifier is 0.5pF. These

microphone are biased with 200V using a Bruel and Kjaer Type 2804 microphone power supply

unit.




















-Cs= 1pF
. Cs = 10pF
---Cs= 100pF


1 0 1
102 100 102 104 106 10 1
Frequency (Hz)

(a) Low Voltage Noise (Cp + C, = pF)


106
Cs= 1pF
....... Cs = 10pF
10 ---Cs= 100pF


10 -10o -




10 *\
10 / -.\






10
10-16 /


10-18 ---------------------
102 100 102 104 106 108 1
Frequency (Hz)

(c) Low Voltage Noise (Cp + C,= 10pF)


10-
Cs= 1pF
....... Cs= 10pF
10-8 \ ---Cs= 100pF


10-10- ^




10 \


10-14
10-16^ s______


-Cs= 1pF
..Cs = 10pF
---Cs= 100pF


10



Z 101



10-16



101
010 10 2 100 1 10 104 106 108 101
Frequency (Hz)

(b) Low Current Noise (C + C, = 1pF)


10
Cs= 1pF
.-Cs = 10pF
S---Cs= 100pF


10 --
0102 10 102 10 10 10 0 10 10s 1
Frequency (Hz)

(d) Low Current Noise (Cp + C = 10pF)


10
-Cs= 1pF
.--Cs = 10pF
S---Cs= 100pF


10 10
10 2 100 102 104 10 10 1010 10 2 100 1 10 104 106 108 1010
Frequency (Hz) Frequency (Hz)

(e) Low Voltage Noise (C + C, = 100pF) (f) Low Current Noise (Cp + C, = 100pF)



Figure 3-13. Plot of output noise for voltage amplifier with a fixed bias resistor (10MQ).

(Low Voltage Noise Amplifier (OP471) So = OnV/ ,Hz, Sio

1.6pA/ z, f, 10Hz, fi = 10Hz, Low Current Noise Amplifier (TL2274)

So = 10nV/ /Hz, So O.0pA/ Hz, f, 10Hz, fi 10Hz)




















-Cs= 1pF
* Cs= 10pF
---Cs= 100pF


10 10U 10, 104
Frequency (Hz)


- Cs= 1pF
*. Cs = 10pF
---Cs= 100pF


10, 10 u 10 10 10U 10, 104
Frequency (Hz)


10, 10, 10


(a) Low Voltage Noise (Rdc = 100kQ)


(b) Low Current Noise (Rdc = 100kQ)


-Cs= 1pF
* Cs= 10pF
---Cs= 100pF


-Cs= 1pF
..Cs = 10pF
---Cs= 100pF


10 10 10 10 10
Frequency (Hz)

(c) Low Voltage Noise (Rd = 10M2)




10 -




o-,
10-15




10 .



102 100 102 104 106
Frequency (Hz)

(e) Low Voltage Noise (Rdc = 1GQ)


10 10 10 10 10 10 10
Frequency (Hz)

(d) Low Current Noise (Rdc = 10MQ)


-Cs= 1pF
* Cs= 10pF
--Cs= 100pF


-Cs= 1pF
...Cs = 10pF
---Cs= 100pF


10 10 10 10 10 10 10
Frequency (Hz)

(f) Low Current Noise (Rdc = 1GQ)


Figure 3-14.


Plot of output noise for voltage amplifier with a fixed parasitic capacitance of

(10pF). (Low Voltage Noise Amplifier (OP471) So = 10V/ Hz, So

1.6pA/ /H, f,, 10Hz, fi = 10Hz, Low Current Noise Amplifier (TL2274)

So = 10nV//Hz, So O.OlpA/ Hz, f, 10Hz, fi 10Hz)


10 10


10 10









Table 3-2 shows a list of commonly used Briiel and Kjaer amplifiers suitable for aeroacoustic

measurement and their characteristics.

Table 3-2. Specification of Voltage Amplifiers for Bruel and Kjaer Microphones.
Model Number Input Capacitance Noise Floor Bandwidth
(A-weighted)
Type 2660 0.3pF 0.8pV 20Hz-200kHz
Type 2669B 0.3pF 1.9pV 20Hz-200kHz
Type 2670 0.3pF 4.0pV 30Hz-300kHz


Commercial MEMS capacitive microphones are now available from Knowles acoustics and

Sonion Technologies. These microphone use an integrated voltage buffer along with an analog to

digital converter (ADC) to provide a digital output. The target application for these microphone

is in the high volume cell phone market where price is the important metric. Also for integration

with the rest of the system, a digital output is an important feature.

Knowles Acoustics sells commercial MEMS based microphones for audio applications

namely, the SiSonic series of microphones [5]. They consist of a MEMS microphone wirebonded

to a voltage amplifier. The interface circuitry operates on a 1.5V source. The high voltage (11V)

required for biasing the microphone is obtained by an on-chip charge pump. The circuit has a

bandwidth in excess of 10kHz. The performance metrics of the amplifier itself are not provided.

Sonion Technologies sells MEMS microphones for cell phone applications namely the

TC100 series microphones [47, 84]. It consists of a PMOS buffer amplifier cascaded with

a NMOS gain stage. The chip is powered by a 1.8V supply. The output of the amplifier is

processed via a ADC and thus the microphone has a digital output.

The National Semiconductors LV1OXX series of amplifiers [85] is specially designed

for electret condenser microphones and consists of a high gain, low input capacitance stage

followed by an optional sigma delta based analog to digital converter thus providing both analog

and digital outputs. The circuit runs on a 1.8 to 5V supply. The input capacitance is 2pF over

the entire power supply range. The gain variation is 8V/V to 25V/V over the entire product

family and power supply options. These amplifiers have a flat frequency response from 100Hz

to 110kHz. The input noise of these amplifiers is approximately 500nV/ Hz. The primary

application of these interface circuits is for microphones used in cell phones.









Table 3-3 presents an overview of voltage amplifier circuits used in research prototype

capacitive microphones. It has been observed that in these publications, the overall microphone

sensitivity in V/Pa is reported and not the electrical sensitivity (V/F). Hence, the gain reported

in Table 3-3 is the sensitivity of the amplifier alone and is expressed in V/V. The general trend

in the input stage of voltage amplifier implementation using custom ICs is the use of unity gain

single transistor CMOS stages This is done to minimize input parasitic capacitance and input

current noise.

In the published literature, one of the earliest semiconductor voltage follower based

capacitive microphone interface circuit was reported in a patent by Killion et al. in 1970 [86].

The circuit consisted of a JFET voltage follower. Back-to-back diodes were used for biasing the

microphone capacitance.

In 1989, Murphy et al, [87] presented a silicon electret microphone that uses a CMOS

preamplifer. The amplifier had a reported gain of 0.9V/V. The overall device had a bandwidth

of 15kHz. The bias voltage used was not reported. In 1991, Bergqvist et al presented a single

backplate microphone with a perforated backplate. The microphone capacitance was 5pF and

was biased with a 5V dc battery. A modified Briiel and KjerType 2619 was used as the interface

circuit. The input parasitic capacitance of the interface circuit was 2pF, and the bandwidth was

from 20Hz to 20kHz. In 1995, Ning et al. [88], presented a voltage buffer based interface circuit

for a micromachined capacitive microphone to be used in cell phones.

In 1995, Ning et al. [88], presented a voltage buffer based interface circuit for a micromachined

capacitive microphone to be used in cell phones.The nominal capacitance of the sensor was

9. lpF. The input parasitic capacitance of the cables and the interface circuit was 12pF with the

amplifier contributing 2pF. The microphone was biased at a dc potential of 6V.

In 1996, Bernstein and Borenstein presented the first MEMS microphone with onchip

interface circuitry [18] in 1996. It consisted of a JFET source follower with an input parasitic

capacitance of 0.5pF.

In 1998, Pedersen and Bergveld reported a MEMS amplifier [59] with integrated CMOS

amplifier and dc-dc converter. The amplifier input consisted of a p-MOSFET based voltage

follower. A p-MOSFET was used because of its lower noise floor as compared to an n-MOSFET.











Table 3-3. Review of interface circuits used in MEMS microphones published in literature.


Author


Gain Bandwidth


Comments


NR NR JFET input custom amplifier,
Back-to-back diodes to bias
microphone plate
0.9 15kHz Custom CMOS amplifier


Killion,
1970[86]*

Murphy et
al,
1989[87]
Sprenkels
et al.,
1991[89]
Kuhnel et
al.,
1992[90]
Bergqvist
et al,
1994[91]
Ning et al,
1994[88]

Bernstein et
al, 1996[?]
Hsu al,
1997[14]

Kabir al,
1999[14]
Baker et al,
2003[92]


Kern et al,
2004[93]*
Furst et al,
2005[84]*
Loeppert et
al,2006 [94]


Microphone
Capacitance
NR


NR


3pF


lpF


5pF


9. lpF


NR

16.2pF


2.4pF

NR




NR

NR

0.5pF


Parasitic
Capacitance a
NR, NR


NR, NR


lpF, NR


0.25pF, NR



2pF, NR


2pF, lOpF


0.5pF, NR

2pF, 10pF


NR, NR

NR, NR




NR, NR

< 10pF, NR

< lpF,
0.25pF


JFET input amplifier


Commercial Briiel and
KjarType 2633 preamplifier

COTS** Briiel and KjaxrType
2619 amplifier

Custom amplifier.
Microphone used in cell
phones.
Custom JFET source
follower.
COTS** HPA-COP4012
amplifier. Input impedance is
10G0.
Custom JFET source
follower.


upto Uses the current output of a
OkHz JFET follower. Used for
microphones in cochlear
implants
NR Large back to back diodes are
used to bias amplifier
NR PMOS unity gain stage
driving a NMOS gain stage
NR Unity gain buffer with an
0 20dB gain opamp in
cascade. The output of the
first stage drives the shield.


a Input Capacitances, Interconnect Capacitance
* Patent


In 1998, Hsu et al reported a single backplate polysilicon condenser microphone [95]. The

device had a zero bias capacitance of 16.2pF. The interface circuit used was a HPACOP4012

preamplifier. It had an input impendance of 2.5GQ. The input capacitance is not reported. The

microphone was polarized at 10V.

In 1999, Kabir et al. reported a MEMS microphone that used a doped p+ poly as the

membrane and gold backplates. The device had a zero bias capacitance of 2.4pF. During


0.9 NR


1 20kHz


1 20Hz -
20kHz


< 1 10

NR
1


)OHz-
9kHz









normal operation, the amplifier was biased with a 9V dc supply. The bias resistor was a 16GQ

poly resistor. The microphone output was buffered by a JFET input voltage follower that was

fabricated on the same chip. The microphone had a flat frequency response from 100Hz to

9kHz.

In 2003, Baker et al reported a JFET input preamplifier that used the current output of a

source follower as opposed to the traditional output voltage [92]. The amplifier was designed

for microphones used in cochlear implants. The amplifier was optimized for high power supply

rejection ratio (PSRR) which is a requirement for low noise design. The circuit operated on a 3V

power supply and had a flat frequency response up to 10kHz.

In 2004, Kern et al obtained a patent for a microphone interface circuit. The circuit biases

the microphone using large resistors from back-to-back diodes. Other performance metrics of the

circuit were not disclosed in the patent.

In 2005, Furst et al [84] of Sonion Technology obtained a patent for a two stage preamplifier

for an electret condenser microphone. It consists of a PMOS unity gain source follower driving

a NMOS gain stage. The circuit has an input capacitance of less than lOpF. Using a two stage

topology allows the independent tuning of the amplifier gain and the input capacitance. The

preamplifier is designed for use in telephony applications.

In 2006, Loeppert et al [94] presented the first commercial MEMS microphone. The

interface circuit consists of a voltage buffer with a 0.25pF input capacitance followed by a gain

stage. The microphone is biased by a 11V dc signal that is generated by a charge pump that is on

the same die as the amplifier.

General purpose operational amplifiers can also be used as interface circuits for capacitive

microphones using any of the packaging techniques listed in Table 3-1. Compared to custom

ICs, the input parasitic capacitance of these amplifiers are comparatively higher and is further

increased by the packaging technology used to connect the amplifier to the microphone. These

amplifiers are usually provided in open loop configuration. Thus, careful circuit design, which

may result in adding more components such as resistors and capacitors is required for the proper

operation of these amplifiers. These extra components may introduce parasitic capacitance

that have to be taken into consideration during the interface circuit design. A set operational









amplifiers recommended by the manufacturer for use as a voltage amplifier for microphones is

listed in Table 3-4 along with their performance specifications.

Table 3-4. Specification of a representative sample of operational amplifiers recommended by
manufacturers for use as voltage followers for capacitive transducers.
Model Number Input Voltage Noise Current Bandwidth
Stage (Sv, fcF ) Noise
Capacitance (Si, fi)
AD8079 1.5pF 3.2nV/ Hz, NR 20MHz
(Analog 2kHz
Devices)
LM6211 5.5pF 5.5nV/ VHz, NR 20MHz
(National 400Hz
Semiconductors)
TL2274 (Texas 8pF 50nV/vHz, 0.5fA/ Hz, 2.25 MHz
Instruments) 200Hz 10Hz
OP285(Analog 2.6pF llnV/ Hz, 1.7pA/ Hz, 6.5 MHz
Devices) 10Hz 10Hz


3.3.2 Sensing Using A Charge Amplifier

Operating Principle. A charge amplifier consists of a opamp circuit with a capacitor in

the feedback path as shown in Figure 3-15. The inverting input is connected to the middle plate

of the microphone and the non-inverting input is grounded. RDC is a bias resistor that sets the dc

operating point of the amplifier input. The middle plate is biased at OV by the principle of virtual

short between the non-inverting terminal and inverting terminal of the amplifier. The charge

flowing through the middle plate is integrated by the feedback capacitor C(,t. The top and bottom

plates are biased at +Vd and -Vd respectively. The input-output relation of the circuit for a

capacitance change is given by the equation


Vout = Vd 2AC. (3-20)




Performance Metrics. From the input-output relationship given by Equation 3-20, the

electrical sensitivity of the circuit is given by the equation

2V4
Se = (3-21)


which is twice that of a single backplate microphone. From the sensitivity equation, it can be

inferred that the sensitivity can be increased by increasing the bias voltage or decreasing the









RDC


+V


C1

Cint
V
+ ~o -ut



-Vdc TT

Figure 3-15. Schematic of dual backplate microphone connected to a charge amplifier.

integrating capacitor. To determine the total output voltage noise PSD Sv,,, the noise model

SR



RCDC




CS > --_
C2 L



Figure 3-16. Noise model of dual backplate microphone connected to charge amplifier.


shown in Figure 3-16 is used. Defining Ctot 2C0 + Cp + Ci Zi and
jw(2Co+Cp+Ci)
Zf R, |I the total output noise PSD is given by




For frequencies above the cut on frequency i.e. > R the output voltage noise PS

equation can be simplified to


Sv~, Sa 1 + Ct 2 + a S) (wnt)2 (3-23)
Cn (sva UCit)









Dividing Sv,, by the electrical sensitivity of the charge amplifier squared and taking the square

root, we obtain the minimum detectable capacitance per IHz is given by the equation


Cmin 2= -V S, 1 + (Cot + (Si + SiR) 2 (3-24)
2Vac C\nJ (VQnt)

From the equation for Cmin, it can be seen that the noise floor is increased by the parasitic

capacitance. Furthermore, the current noise PSD due to the bias resistor and the current noise of

the amplifier is shaped as 1/w2 which increases the noise at low frequencies considerably.

The upper end of the dynamic range is limited by the maximum output voltage of the

amplifier and is dependent on the topology of the amplifier used. For a given power supply Vdd,

the maximum output voltage Vomax is determined by a circuit dependent voltage drop Vdro, below

the supply rails Vdd and is given by the expression


Vomax = Vdd Vdrop. (3-25)

Dividing the above equation by the electrical sensitivity, the maximum input capacitance that can

be measured is
C, mt
Cmax = [Vdd Vdrop] (3-26)
2Vdc
This shows that the upper end of the circuit dynamic range can be increased by decreasing the

top and bottom plate voltage and increasing the integrating capacitance.

The lower limit of the frequency response is set by the feedback capacitance and the

feedback resistor [83]. It is given by the expression

1
f 2 (3-27)
2TTRdcint

The upper end of frequency response of this circuit is determined by the bandwidth of the

amplifier and is dependent on the amplifier topology. For example, for a single stage folded

cascode amplifier with a transconductance gm, the bandwidth is given by the expression

1 gm
fh 2 CL + (3-28)
27 CL + Cint

From the equation for fh, it can be inferred that the bandwidth is also increased when the Cent is

decreased.










Tradeoffs. The main advantage of the charge amplifier based circuit is the immunity of

the sensitivity to parasitic capacitance. The circuit sensitivity can be increased by reducing Cint

without increase in the noise floor. However, another consideration when reducing Cnt is that

the bandwidth of the amplifier must be increased to ensure that there is sufficient loop gain for

the stable operation of the amplifier. Decreasing Cit to increase the sensitivity also reduces the

upper end of the dynamic range.

The main disadvantage of this circuit is the need for a large dc bias resistor Rdc because

the noise PSD due to this resistor is 1/f2 shaped which considerably increases the noise at

low frequency. Choosing Rdc is a trade off between noise and the lower end of the frequency

response fl. Implementing a large Rdc resistor is non-trivial in an integrated circuit process.

Options include thick film resistors, back-to-back diodes, or low resistance poly resistors. Each

of these techniques has its own advantages and disadvantages[8]. Figure 3-17 shows a plot

105
Amplifier Voltage Noise
.. .. ** Amplifier Current Noise
10 \ Resistor Current Noise
L Total Noise

10- "'-

S10-20
0


10-25 "


10-30 1


10- 5 0 5 10 15 20
105 10 10 101 105 102
Frequency (Hz)

Figure 3-17. Power spectral density of individual noise contributors and total output noise in a
charge amplifier based interface circuit. (Low Voltage Noise Amplifier (OP471)
So = lOnV/ ,Hz, Sio = 1.6pA//Hz, f,, 10Hz, fi = 10Hz, Rd = 44MQ,
Cp = 10pF,C = 2.6pF, 2Co = 1.7pF)


of the PSD of various noise contributors in a charge amplifier and the total output referred

noise. It can be inferred that at low frequencies the resistor current noise is the dominant

noise mechanism and at higher frequencies the amplifier voltage noise is the dominant noise

mechanism. Figure 3-18 and Figure 3-19 plot the effect of scaling various circuit elements, on

68









the output noise PSD. Figure 3-18 shows a set of plots of the output noise PSD as the sensor

capacitance is scaled from 100pF to IpF for various parasitic capacitance assuming a fixed bias

resistor of 10MQ. Figure 3-18(a), Figure 3-18(c), and Figure 3-18(e) show these plots for a low

voltage noise amplifier and Figure 3-18(b), Figure 3-18(d), and Figure 3-18(f) show these plots

for a low current noise amplifier. It can be seen that, for a large fixed bias resistor(10AOM ), an

amplifier optimized for low current noise has a lower noise floor as compared to an amplifier

with low voltage noise. It can also be seen that this is true for all combinations of sensor and

parasitic capacitances. Figure 3-19 shows a set of plots of the output noise PSD as the sensor

capacitance is scaled from 100pF to IpF for various bias resistors assuming a fixed parasitic

capacitor of lOpF. Figure 3-19(a), Figure 3-19(c), and Figure 3-19(f) show these plots for an

amplifier with low voltage noise and Figure 3-19(b), Figure 3-19(d), and Figure 3-19(e) show

the plots for a low current noise amplifier. It can be seen that, for a fixed parasitic capacitance,

an amplifier optimized for low current noise has a lower noise floor as compared to an amplifier

with low voltage noise where the difference in large for a large bias resistor.

Literature Review. Charge amplifiers are commonly used for capacitive accelerometers,

gyroscopes, and pressure sensors when the sensor is located at a distance far away from the

circuitry. To overcome the parasitics of the long parasitic leads, a charge amplifier is used.

Traditionally, in microphones, the output of the amplifier is buffered as close as physically

possible to the sensor. Therefore, very few capacitive microphones use a charge amplifier

interface circuit. Another common reason for the usage of voltage amplifiers over charge

amplifiers is the increased noise at low frequencies due to the 1/f2 shaping of the current noise.

A summary of the literature survey is given in Table 3-5.

In 1999, Amendola et al. [96] reported a charge amplifier-based interface circuit for a

capacitive microphone.

The target applications for the microphone was measurements inside a resonant cavity

which had a narrow bandwidth centered at 25kHz. The amplifier had a simulated bandwidth

from 19Hz to 1.4MHz. The simulated sensitivity of the circuit was 2V/pC and a measured

sensitivity of 0.3V/pC.




















-Cs= 1pF
*. Cs = 10pF
---Cs= 100pF







-- ~''~'


10- 10U 10, 10"
Frequency (Hz)


10, 10,


(a) Low Voltage Noise (Cp + C, = pF)


Cs= 1pF
*.*.** Cs = 10pF
05 ---Cs= 100pF
105




> 1010 "
o
5 z
. . . .


10' 10U 10, 104
Frequency (Hz)


10, 10,


(c) Low Voltage Noise (C + C, = 10pF)


-Cs= 1pF
_: Cs= 10pF
S- ---Cs= 100pF


> 10-10-


10 10 10 10
Frequency (Hz)


10 10'


(e) Low Voltage Noise (C + C, = 100pF)


-Cs= 1pF
.-Cs = 10pF
---Cs= 100pF


10 10U 10, 10"
Frequency (Hz)

(b) Low Current Noise (Cp + C,




10 5


10 10U 10, 104
Frequency (Hz)


10, 10,


lpF)


-Cs= 1pF
....Cs = 10pF
---Cs= 100pF


10, 10,


(d) Low Current Noise (Cp + C, 10pF)


-Cs= 1pF
...Cs = 10pF
---Cs= 100pF
1015





S......................
z
">10----------------


10 10 10 10
Frequency (Hz)


10 10


(f) Low Current Noise (C + C, = 100pF)


Figure 3-18. Plot of output noise for charge amplifier with a fixed bias resistor (10MQ).

(Low Voltage Noise Amplifier (OP471) S0o = OnV/ Hz, Sio

1.6pA/ z, f, 10Hz, fi = 10Hz, Low Current Noise Amplifier (TL2274)

So = 10nV/Hz, So O.OlpA/ Hz, f, 10Hz, fei 10Hz)





















-Cs= 1pF
- -- Cs = 10pF
---Cs= 100pF


10' 10U 10, 104
Frequency (Hz)


10, 10,


(a) Low Voltage Noise (Rdc = 100kQ)


10' 10U 10, 10
Frequency (Hz)


(b) Low Current Noise (Rdc = 100k2)


-Cs= 1pF
.-.- Cs = 10pF
---Cs= 100pF


10 10 10 10
Frequency (Hz)


10 10


(c) Low Voltage Noise (Rdc = 10MQ2)


-Cs= 1pF
....... Cs= 10pF
10 ---Cs= 100pF





10



1015


10 10 10 10
Frequency (Hz)


10 10 10 10
Frequency (Hz)


(d) Low Current Noise (Rdc = 10M2)


10 10-


10 10 10 10
Frequency (Hz)


(e) Low Voltage Noise (Rdc = 1GQ2)


(f) Low Current Noise (Rdc = 1GQ)


Figure 3-19. Plot of output noise for charge amplifier with a fixed parasitic capacitance of

(10pF).(Low Voltage Noise Amplifier (OP471) So = 10nV/Hz, So

1.6pA/ /H, f,, 10Hz, fi = 10Hz, Low Current Noise Amplifier (TL2274)

So 10nV/ Hz, So O.OlpA/ Hz, f, 10Hz, fi 10Hz)


-Cs= 1pF
..Cs = 10pF
---Cs= 100pF


10, 10,


-Cs= 1pF
..Cs = 10pF
---Cs= 100pF


10 10


-Cs= 1pF
...Cs = 10pF
---Cs= 100pF


10 10


.... .


----------------


.\ .. .... .... .... .... ....










In 2005, Martin et al. [97] reported a dual backplate MEMS capacitive microphone for use

in aeroacoustic applications that used a commercial-off-the-shelf (COTS) Thermoptics DN620

charge amplifier which has a sensitivity of 1V/pC and a bandwidth of 100kHz. The feedback

elements consisted of a lpF capacitor in parallel with a 1GQ resistor.

Table 3-5. Review of charge amplifier interface circuits used in capacitive microphones.
Author Type of Sensitivity Feedback Bandwidth Comments
Sensor/Sensor capacitor
Capacitance
Amendola Microphone 0.3V/pC with NR 19Hz JFET input custom amplifier.
et al. 1999 0.7pF a 10V bias, 1.5MHz Uses a switched capacitor
[96] technique to obtain stable
operation
Martin et Microphone 1V/pC with a lpF 300Hz- COTS Thermoptics DN620
al. 2005 differential 9V bias 20kHz [98] charge amplifier
[97] capacitor with
lpF,0.7pF
nominal
capacitance


Commercially available general purpose operational amplifiers can be used as a charge

amplifier by employing a feedback resistor and capacitor. A representative sample of general

purpose amplifiers that are recommended by the manufacturer for use as a charge amplifier by

using a feedback resistor and capacitor is listed in Table 3-6.

Table 3-6. Specification of a representative sample of general purpose operational amplifiers
recommended by the manufacturer for use as charge amplifiers using a feedback
resistor and capacitor.
Model Number Input Stage Noise Floor Corer Bandwidth
Frequency
OPA128 (Texas DiFET 27nV/ /Hz 300Hz 1MHz
Instruments) [99]
OPA111 (Texas DiFET 8nV/ /Hz 300Hz 800kHz
Instruments) [100]
AD745 (Analog Devices) FET 3.2nV/ Hz 2kHz 20MHz
[101]


3.3.3 Sensing Using Synchronous Modulation/Demodulation

Operating Principle. The fundamental idea behind this technique is to improve the

noise performance [65, 46] of the circuit by modulating the capacitive signal to a high carrier

frequency above the low 1/f corner frequency of the amplifier. This high frequency signal is

then amplified and demodulated. During the demodulation and subsequent low pass filtering

process, the original signal is recovered and the noise is filtered out.

72








+Va sin(wt)


Vac sin(wt)


C RD I > out
C/l it-



02 Multiplier Lowpass Filter
(Demod)
-Vac sin(wt)

Figure 3-20. Schematic of capacitive microphone with a synchronous modulation and
demodulation technique using a voltage amplifier based circuit.

In this technique, the top and bottom plates are excited by a sinusoidal signal of equal
magnitude (Vac) but 180 degrees out of phase. The excitation frequency (Uc) is chosen to be
much higher ( 100 x) than the resonant frequency of the sensor. This is done so as to prevent the
electrostatic excitation of the sensor by the ac waveform. The interface circuit used can be either
a voltage amplifier or a charge amplifier. A schematic of the system using a voltage amplifier and
charge amplifier interface circuit is shown in Figure 3-20 and Figure 3-21 respectively.
Rdc

+ Vc sin(wct)


01--/- Vx 1A Io



C \~ C- Multiplier Lowpass Filter
-- -- Charge Amp (Demod)

-VCc sin(wVt)

Figure 3-21. Schematic of a capacitive microphone with a synchronous modulation and
demodulation technique using a charge amplifier based circuit.

The amplitude modulated (AM) signal is converted back to a baseband signal using a
synchronous demodulator. The carrier frequency used to demodulate the output of the charge
amplifier is phase locked to the modulating a signal applied to the top and bottom backplate. The

73













o


ff
Frequency (Hz)
(a) Spectrum of various signals involved in
synchronous modulation/demodulation
technique


a



f, f-f, fo+,
Frequency (Hz)
(b) Spectrum of signals at the middle plate
of the microphone.


F/


fm r
Frequency (Hz)
(c) Spectrum of signal at output of demodulator


r\
f
Frequency (Hz)
(d) Spectrum of signal at output of
low pass filter.


Figure 3-22. Plot of spectrum at various nodes of Figure 3-21


synchronous demodulator consists of an analog multiplier followed by a low pass filter. This

circuit overcomes the noise due to amplifier 1/f noise and the 1/f2 shaped current noise of the

bias resistor and current noise of the amplifier by modulating high frequency carrier frequency

with the low frequency acoustic signal. This signal is then amplified, demodulated, and low pass

filtered to obtain the original signal. During the demodulation process the low frequency 1/f and

1/f2 noise are modulated to the carrier frequency which is then filtered out while the acoustic

signal at the carrier frequency is translated to the baseband and passes through. The relation

between the capacitance input and the low pass filtered output of this technique is given by [46]


Vt = H, c AC. (3-29)
CoJ

for a voltage amplifier and by

Vout I= AC. (3-30)

for a charge amplifier where U = 1V is a scale factor. The dimension of Equation 3-29 and

Equation 3-30 needs consideration. In practice, the demodulator [102] is implemented as an an









analog multiplier with a transfer function given by


XY
Z (3-31)


where Z is the output of the multiplier measured in Volts, X and Y are the inputs to the

multiplier measured in Volts and U is a scale factor also measured in Volts. Thus, the output

of the demodulator has the dimension of Volts and Equations 3-29 and 3-30 are dimensionally

consistent.

Performance Metrics. From the input output relationship given by Equation 3-29 and

Equation 3-30, the electrical sensitivity of the circuit is given by the equation


Se= [V/F]. (3-32)
U Co

for a voltage amplifier and by
V2
S = [V/F] (3-33)
UCint
for a charge amplifier.

From the above equation, it can be inferred that the sensitivity can be increased for both

interface circuits by increasing the peak ac bias voltage, Vac. The bias voltage can not be

increased without bound. There are two main factors that determine the maximum value of

the ac signal. They are the electrostatic pull-in limit which was discussed in Chapter 2 and the

actuation of the microphone plates by the electrical ac voltage instead of the input acoustic

signal. Although the frequency of excitation is 100 x higher than the resonant the resonant

frequency, if the input amplitude is large, the diaphragm can vibrate at the high carrier frequency.

For the charge amplifier based interface circuit, the sensitivity can also be increased by

decreasing the value of the integrating capacitor. The minimum value of the integrating capacitor

is determined by the limitations of the fabrication process and the overall bandwidth of the

circuit.

For illustration, the sensitivity of a voltage amplifier based interface circuit using square

waves for modulation and demodulation is derived. Let the top plate voltage vt, and bottom plate









voltage vtp be given by


Vtp = +Vdc Vac(t)

Vb = Vd + Vac(t), (3-34)

where

1 2r
Vac(t) = Vac 0 < t < --

ac(t) 0 < t < (3-35)
2 we 2c

Also, let us assume that the demodulator voltage Vc(t), is given by

1 27r
VC(t) = 0 < t < --
2 w~
1 2r 2xr
Vc(t) 0 2 < t < (3-36)
2 We c C

The output of the demodulator, Vdemod(t), is given by the expression


Vdemod(t) = Vmod(t) X Vc(t). (3-37)

The output of the demodulator is low pass filtered. From Equation 4-4, it can be seen that

demodulation is a non-linear operation. The entire process of demodulation and low pass filtering

can be linearized by ignoring the harmonics at the output of the low pass filter as they will be

attenuated by the filter. Thus, ignoring the harmonics which are attenuated by the low pass filter,

the output of the filter can be shown to be

ACH, (V c + c) 2Vc 2V, Va (3-38)
Vt(t) = dc + )+ (3-38)
Co U 2 2 x 7

for a voltage amplifier and

ac 1 ac ) + 2Vc 2Vac
Vou (t) = c + )+ (3-39)
Cint U 2 2 7 7

for a charge amplifier. The derivation of the above equation is provided in Appendix A.5. Thus,

the sensitivity is given by

H 1 ( ( V\ 2 2r (3-40)
Se = t Vdc + + ,V (3-40)
GO U 2 2 7









for a voltage amplifier and


1 1 (( V c2 2V2 Va\
S'e Vdc + )+ V2V 2Vc (3-41)


For the reminder of this section, we use the simplified sensitivity. To determine the

total output noise, the noise model shown in Figure 3-23 is used for a voltage amplifier and

Figure 3-24 is used for a charge amplifier. For the voltage amplifier interface circuit, the total


(Demod)


Figure 3-23. Noise model of synchronous modulator and demodulator using voltage amplifier.


SiR


Vacsin(wt)


CL Multiplier
1 T (Demod)
Charge Amp


Lowpass Filter


Figure 3-24. Noise model of synchronous modulator and demodulator using charge amplifier.


input noise is given by


2 Zi 2
SVo, = Sa + Sia (Z,| R dc) + SvR + Smultiplier, (3-4
Zi + Rdc

where Z i j2Co +1 and SmuItiplier is the output referred noise PSD of the multiplier. For

the charge amplifier interface circuit, the total output noise is given by


Svot SVa 1 + Cint + Smultiplier. (3-4
( Cint ) I


2)


3)









Dividing the total output noise, Equation 3-42 and Equation 3-43, by the corresponding

electrical sensitivity, we obtain the minimum detectable capacitance and is given by the equation

(2Co + C~ + C) U 2
Vmin SVa + Sia (Zi Rd, ) + Z Rd SR + Smultiplier. (3-44)
Vaoc Z Z + Rdc

for a voltage amplifier and


Cmin = 2 S + + Smultiplier, (3-45)


for a charge amplifier. The upper end of the dynamic range is the same for both interface circuits

and is given by

Vomax = Vdd Vdrop. (3-46)

Dividing the maximum output voltage equation by the corresponding electrical sensitivity, the

maximum input capacitance that can be measured is given by

(2Co + C, + C) U
Cmax = + [Vdd Vdrop] (3-47)
ac

for a voltage amplifier and
Cint U
max [Vdd Vdrop] (3-48)
Vac
for a charge amplifier. This shows that the upper end of the dynamic range can be increased by

decreasing the peak ac bias voltage, Vac for both interface techniques.

The lower limit of the frequency response is the same as the dc voltage amplifier and charge

amplifier techniques considered previously. It is given by the expression

1
f 27Rd, (2Co + Cp+ C)(349)

for a voltage amplifier and
1
f 7 (3-50)
2TTRdCint

for a charge amplifier. The upper limit of the frequency response of this circuit is determined by

the cutoff frequency flpf of the low pass filter and is given by the expression


fh = fipf (3-51)










Tradeoffs. This technique has all the advantages of the voltage and charge amplifier

techniques with the additional benefit that the 1/f and 1/f2 noise components are attenuated. It

does not suffer from the problem of aliasing of the broadband white noise of the amplifier to the

baseband as is the case in switched capacitor based circuits. The circuit is purely analog in nature

and reduces the demands placed on layouts as routing mixed signal circuits is avoided.

There are two main disadvantages with this technique. A large bias resistor is needed for the

operation of the circuit. Implementing this resistor is a non-trivial problem. For example, in the

ADXL-50, a 3MQ resistor was used and was fabricated using a special resistor layer. Another

solution is to use back-to-back diodes. The issues in using back to back diodes are leakage

currents and poor transient performance. The second major disadvantage of this technique

is the use of the low pass filter. The low pass filter will cause phase delay which needs to be

compensated when used in closed-loop operation as will be shown in Section 3.5.1. Furthermore,

the multiplier noise, increases the output noise.

Literature Review. In this section, a review of capacitive transducers, especially

capacitive microphones and accelerometers, that utilize synchronous modulation and demodulation-based

interface circuits is presented. A summary of the literature survey is provided in Table 3-7.

Table 3-7. Summary of previous synchronous modulation/demodulation based open loop
capacitive transducer interface circuit.
Author / Type of Transducer Nominal Carrier Sensitivity Comments
Transducer Bandwidth Capacitance Signal
Open Loop
Kraft et al. 56 7.2pF 5Vpp 700mV/g Charge amplifier
1998 [104](acceleromet 1MHz interface circuit
Li et al. 2000 [103] NR NR NR 75V/g Charge amplifier
interface circuit
Yakabe et al. NR NR NR NR Charge amplifier
2006 [105] (patent) circuit using a two
opamps.


In 2000, Li et al. 2000 [103] reported an open loop accelerometer using synchronous

modulation/demodulation as the interface circuit. They reported a sensitivity of 75V/g when

the input was in 0. 19 range. The characteristics of the carrier signal and the bandwidth of the

device was not reported.









Kraft et al.[104] reported an analog open loop accelerometer that used 2Vp-_ 1MHz carrier

signal. The device had a reported sensitivity of 0.7V/g and a bandwidth of 300Hz. The open

loop bandwidth of the device was 56Hz.

Yakabe et al. [105] obtained a patent for a modified form of a voltage amplifier synchronous

modulation based interface circuit that allows for reduced sensitivity to parasitic capacitance.

This is accomplished by placing the sensor output signal, an extra feedback capacitor, and an

auxiliary opamp in the feedback path of a primary amplifier. No details of the performance

specifications are presented.

3.4 Digital Open Loop Sense Techniques

Digital sense techniques are interface circuits in which the continuous time output of the

sensor is converted to discrete time signals by sampling the sensor output and then processing

the sampled data. These circuits are implemented using switched capacitor circuits [79]. In these

circuits, a resistor is replaced with appropriately connected MOSFET switches and capacitors

and driven by non-overlapping clock waveforms that simulate a resistor. Switched capacitors

thus eliminate the need for large resistors. Switched capacitors using the technique of correlated

double sampling (CDS) [106] can attenuate dc offsets and 1/f noise from the interface circuit.

The fundamental idea in the CDS scheme is to measure the noise dc offset voltage at some

instant of time and subtract it from the measured signal during the next instant. This is performed

in two phases which are time multiplexed. During the sampling phase, the noise and offsets are

measured and stored. Afterwards, during the signal processing phase the offset free amplifier is

available for operation. The operation performed by the system can be expressed in terms of the

output signal il[, the input signal x[n], and the 1/2 period previous input signal x[n 1] by the

equation

[,1] =x[n]- x[n- 1/2]. (3-52)

Using the z-transform, the transfer function H(z) of the above system can be expressed as


H(z) =1 z-. (3-53)









Substituting z = e T and simplifying, we can obtain the magnitude transfer function as


H(ejw) = 2sin (T (3-54)
( 4 )

It can be seen from this transfer function that the CDS operation puts a zero at dc (i.e w = 0)

and has a high pass response for increasing frequencies up to the Nyquist frequency. Thus,

theoretically, CDS eliminates dc offsets and high pass filters the input signal.

Sensing Using Switched Capacitor Amplifier. Figure 3-25 shows one implementation

of a switched capacitor amplifier that implements CDS. The switched capacitor amplifier

Crest
+AV 0 V v


1T / gain
Vout

C2 L

-AT T


Figure 3-25. Switched capacitor implementation of correlated double sampling.


consists of two capacitors Cain and Creset and four switches arranged as shown in Figure 3-25.

The output common mode voltage, Vim, is chosen so as to maximize the output range. For a

circuit with 0 5V supply rail, Vm is chosen to be 2.5V. The switches are driven with two non

overlapping clocks 01 and 02.

The equivalent circuit during each clock phase is shown in Figure 3-26. During the reset

phase (01), the noise voltages are measured and stored on the integrating capacitor. During the

amplify phase (02), the circuit performs the actual measurement. To perform the measurement,

step voltages of equal magnitude AV but opposite in sign are applied to the top and bottom

plates respectively. The input output relation during the amplify phase is given by [79],


Vot =[ 2AC. (3-55)
I^gain\















S-- Vout
0 + CL 0 CL



I T
(a) Reset Phase 91. (b) Amplify Phase 02.

Figure 3-26. Equivalent circuit of switched capacitor amplifier during amplify and reset phase.

Performance Metrics. From the input output relationship given by the above equation,
the electrical sensitivity of the circuit is given by the equation,


2AV
Ce in
Gain


(3-56)


From the above equation, it can be inferred that the sensitivity can be increased by varying the
height of the sampling pulse or decreasing the integrating capacitor.
The noise analysis of the switched capacitor network is done using a simplified model for
the amplifier noise based on the work by Gobet et al [107]. The complete noise model is shown
in Figure 3-27. The detailed derivation of the noise sources is in Appendix B. The noise analysis


C =C0




C2=C 0


Vout


Figure 3-27. Noise model of switched capacitor amplifier during amplify phase.









of this technique is based on the work by Wongkomet [108]. The switched capacitor interface

circuit is an oversampled system. The main sources of noise are the kT/C noise associated with

the switches and the wide band white noise of the amplifier that is aliased to the baseband. The

total capacitance CT at the input of the amplifier that is sampled by the switches is


CT = 2Co + p + Cgai. (3-57)


The total kT/C noise at the output can be expressed in terms of the sampling frequency, fs, the

total capacitance, CT, and the gain capacitance Cgain, and is given by

1 CT 2 kT
SkT/c =-t [ 2 kT (3-58)
fs Cain_ CT

The total sampled wide band noise at the output vrp noise at the output can be expressed

in terms of the sampling frequency, fs, the total capacitance, CT, the gain capacitance Cgavi, and

the unity gain bandwidth f, of the amplifier by


Svopamp Q n (3-59)
C gain f s 2.

Adding Equation 3-58 and Equation 3-59 we obtain the total output noise as

1 C 12 kT + CT 12 fT
SVo fs 7g+ = + S (3-60)
fs Cgain CT Cgain fs 2

Equation 3-60 can be divided by the sensitivity to determine the minimum detectable

capacitance per \Hz and is given by the expression

1 CT 2 kT CT \ 2 f
Cmn+ = [T r SV_ (3-61)
msi ft 2AV CT 2AV fs 2

The upper end of the circuit dynamic range is limited by the maximum output voltage of the

amplifier and is dependent on the topology of the amplifier used. It is given by the expression


Vomax = Vdd -Vdrop (3-62)

Dividing the above equation by the electrical sensitivity the maximum input capacitance that can

be measured is given by
CC = n -
Cmax [Vdd V- Vdrop] (3-63)
AV









This shows that the upper end of the circuit dynamic range can be increased by decreasing the

height of the sense pulse and increasing the integrating capacitor.

The bandwidth of this circuit is determined by the closed-loop bandwidth of the amplifier.

Based on the discussion in Appendix B, for a switched capacitor amplifier, the closed-loop

bandwidth f, can be expressed in terms of the duty cycle of the sense phase m(< 1), number of

settling time constants n,, and the sampling frequency f, [41]
-1
f,. I 1 (3-64)
27x nf,

The amplifier core is made of a folded cascode amplifier whose open-loop bandwidth famp is

given in terms of the total output capacitance CLT,


CLT = CL + Cgan (Cp + 2Co), (3-65)

by the expression
1 9m
famp 2g CLT (3-66)
27T CLT
The closed-loop bandwidth f, is related to the open-loop bandwidth famp by the feedback factor


CT (3-67)
Cgain + C/T
by the expression

fu = Ofamp. (3-68)

Tradeoffs. The main advantage of this technique is that it eliminates the need for the

bias resistor, making this technique easily amenable for standard IC processes. It has the

advantage of the synchronous modulation/demodulationtechnique in that it reduces the 1/f

noise components. Furthermore, it removes the dc offset which is not done by the synchronous

modulation/demodulation technique.

The main disadvantages are as follows. Being a sampled data system, the wideband noise

of the amplifier is aliased into the baseband thus increasing the noise floor in the bandwidth of

interest. This problem is not present in the synchronous modulation/demodulation technique.

To mitigate this effect, the minimum allowable closed-loop bandwidth must be chosen for the

amplifier. Based on the discussion in Appendix B, the minimum closed-loop bandwidth is given











-1
f,. (3-69)
27r n, f,
Literature Review. In 2001, Yamada et al. [66] reported a switched capacitor interface

circuit for capacitive pressure sensors with a sense capacitor of InF.. The interface circuit was

implemented using discrete components. The clock frequency used was 10kHz with a reset

capacitor of 3.9nF and a reset capacitor of 101.2pF.

In 2004, Potter et al. of Knowles Acoustics received a patent [109] for a switched capacitor

based interface circuit for a capacitive microphone. It is designed so as to reduce the input offset

voltage and 1/f noise. Performance specifications were not specified in the patent.

In 2006, Jawed et al [110] presented the simulation results of a switched capacitor based

interface circuit for a single backplate capacitive microphone. The nominal capacitance of the

microphone was 4.5pF and the pull in voltage was equal to 3.5V. The parasitic capacitance was

on the order of a few pF. The microphone had a bandwidth extending from 50Hz to 20kHz.

The circuit is a simpler version than the structure shown in Figure 3-25. It has only one feedback

capacitor and uses a fixed on-chip poly-poly dummy capacitor to form a pseudo-differential

structure.

3.5 Closed-Loop Sense Techniques

The fundamental concept behind the closed-loop schemes in capacitive sensors is to use an

external force to balance the force due to the input phenomenon. In capacitive microphones, a

voltage source generates an electrostatic force which counters the pressure induced deflection and

restores the diaphragm to its equilibrium position. The voltage used to generate the electrostatic

force is thus a measure of the input acoustic signal. This section is based on the work developed

by Hunt [111].

In closed-loop schemes, there are two phases of operation, namely sense and feedback.

During the sense phase, the capacitance change is measured and during the feedback phase, an

electrostatic force is applied using a voltage to either the backplate or the diaphragm to null the

deflection. These two phases are separated in time in the case of digital closed-loop schemes

and in frequency in the case of analog closed-loop schemes. This separation is essential and is

done to prevent the excitation signal used to the measure the capacitance change from interfering









with the feedback signal [69]. To implement the force feedback scheme, there are two common

approaches. In one approach, a separate set of plates for sense and force feedback [68, 67, 11] are

used while for another approach, the same plates can be multiplexed for the sense and feedback

phase [2, 112].

The main advantages of the closed-loop approach is increased linearity and dynamic range

[11, 42, 113]. Bandwidth improvement beyond the resonant frequency is also predicted for the

analog closed-loop interface circuit approach. Since the feedback electrostatic force keeps the

plate at the center position, the diaphragm becomes much stiffer than the simple open-loop case

which results in increased bandwidth. Force feedback also allows one to use higher voltages by

reducing instability due to pull-in voltage and thereby increasing the sensitivity.

The advantages provided by the closed-loop schemes come with their own limitations

and constraints. Since the sensor is a feedback loop, the stability of the loop must be carefully

considered during the design of the electronics. Large voltages, which might not be easily

available in current CMOS technologies, may be required to balance large input sound pressure

levels. For example, to force balance a 160dB acoustic signal, approximately 45V is required

which is not a commonly available voltage in standard IC processes.

The detailed operating principle of the various closed-loop techniques and their performance

metrics are discussed next.

3.5.1 Sensing Using Closed-Loop Analog Interface Circuit

In the closed-loop analog force feedback technique, the sense and feedback signal are

separated in frequency. In this technique, the sense phase is similar to the open-loop synchronous

modulation and demodulation technique discussed earlier. Thus, for the sense phase the top

and bottom plate are biased with a high frequency ac modulation signal Vacsin(" I) of equal

magnitude but of opposite polarity. For the feedback phase, a low pass filtered version of the

demodulated output signal denoted by vf is used. The voltage required to produce a net feedback

force is generated by summing vf to a dc bias voltage Vd of equal magnitude but of opposite

polarity applied to the top and bottom electrode respectively.

The design of the dc bias voltage is discussed later in this section. Thus, three different

voltages are summed and applied to the top and bottom plates. The top and bottom plate voltages









vtp and vbp is thus given by


Vtp = +VacSin(Wct) Vd + Vf (3-70)

bp = -VacSin(Wct) + Vdc + Vf.


Figure 3-28 shows a schematic of the analog closed-loop technique.


Vdc-Vac (t)+V



C RD
C,< Multiplier
-- < -- -1-- T Lowpass Filter
S 1 1 (Demod)
-V d+vac (t)+Vf
Feedback signal at base band

Compensator

Figure 3-28. Schematic of analog closed-loop force feedback scheme.


The net force acting on the movable plate AF can be expressed as the difference between

the force, Ftp, acting between the top plate and the middle plate and the force, Fbp, acting

between the bottom plate and the middle plate and is given by the expression


AF =Ftp Fp. (3-71)


Assuming that the middle plate is kept at ground potential because of negative feedback of the

amplifier and using Equation 2-61, Ftp and Ftp are given by the expressions

1 v2
Ftp 2 -A tP 2, and (3-72)
2 [xo x']
1 v2
S- A V 2
F 2 [x x']2

Substituting Equation 3-70 and Equation 3-72 in Equation 3-71, neglecting high frequency

terms as they will be filtered by the low pass filter, and assuming x' << Xo, we obtain

Iox' ((Iva)2 + [vdc+ ) ) + v) Vdc + 2) -r yV
AF = 2cA v4 (3-73)
xo









During steady state operation, the displacement of the middle plate is highly attenuated i.e.


x' w 0. (3-74)


Substituting Equation 3-74 in Equation 3-73, for small displacements we obtain

2A c + -
AF 2cA d 2 V. (3-75)
xo

Equation 3-75 shows that for small displacements (i.e. x' << Xo), the net electrostatic force on

the middle plate is a linear function of the feedback voltage.

The design of the bias voltage is based on the force corresponding to the maximum pressure

Pmx that the sensor has to balance. By equating the work done by the electrostatic force

between the plates of a capacitor to the mechanical work done for an applied pressure Pmax, we

obtain the voltage required Vd,

Vdc = (3-76)

Performance Metrics. The sensitivity of this technique is the same as that of the

synchronous modulation/demodulation technique using a voltage amplifier. It is given by


S H 2V= r2 (3-77)

The noise floor is given by


SvO = Sa + Sa (Z || Rdc) 2 + Z+ SR. (3-78)
Zi + Rd,

The low frequency cut off is the same as that of a voltage buffer based interface circuit and is

given by
1
f 2 =27 (2Co+C + C) (3-79)

The upper end of the bandwidth is determined by the closed loop response of the system. As the

closed loop system is non-linear, a closed form expression is not available. An estimate of the

system bandwidth can be determined graphically by plotting the magnitude frequency response

of the linearized closed loop system.

Literature Review. Table 3.5.1 provides a review of microphones and accelerometers

that use a synchronous modulation and demodulation based interface circuit. It can be seen that










this technique is commonly used for low bandwidth applications i.e. mainly accelerometers.

The reason for this is that simultaneously meeting carrier frequency requirements and feedback

voltage requirements is a challenging problem with current semiconductor technology. One way

in which the high frequency required for sensing can be overcome is to sense the diaphragm

position using optical techniques [114] and use electrostatics only for the feedback phase.

Commercially the synchronous modulaiton/demodulation technique has been used in the Analog

Devices (ADXL-50) accelerometer.

Table 3-8. Summary of previous synchronous modulation/demodulation based closed-loop
capacitive transducer interface circuit.
Author / Type of Transducer Nominal Carrier Sense / Comments
Transducer Bandwidth Capacitance Signal a Feedback
Open Loop / capacitors
closed-loop
van der Donk et al 30Hz / NR NR Common
1992 [115] (single 400Hz 1MHz plates


backplate
microphone) (Design
1)
van der Donk et al 400Hz /
1992 [115] (single 7kHz
backplate
microphone) (Design
2)
Hadjiloucas et al., NR / 50kHz
1995 [116] (single (Theoretical)
backplate
microphone)
Hall et al., Diaphragm
2005 [114](single resonance at
backplate 44.8kHz
microphone) (Theoretical)
Analog Devices 14kHz /
1993 [2] 1kHz
(accelerometer)
Kampen et al 1994 [7] NR / NR
(accelerometer)
Park et al. 1998 [112] NR / 350Hz
(accelerometer)
Kraft et al. 56 / 300Hz
1998 [104](acceleromet
Luo et al. 8.9kHz /
2002 [117](acceleromet 400Hz

a Peak to Peak Voltage,Frequency


NR Common
1MHz plates


NR NR NR Common Optical sensing
plates using Fabry-Perot
interferometer.

NR NR NR Common Optical sensing
plates using Michaelson
interferometer.


NR 5Vpp
2.5MHz

NR 2Vpp
1MHz
0.4pF 1Vpp
50kHz
7.2pF 5Vpp
1MHz
0.064pF 1Vpp
2MHz


Common
plates

Common
plates
Common
plates
Common
plates
Separate
plates


Commercially
available




Sensitivity is
39mV/g
Sensitivity is
700mV/g
Sensitivity is
2.2mV/g Noise
Floor is 1mg/ Hz


In 1992, van der Donk et al. [115] reported analog closed loop control for a single backplate

microphone. The microphone consisted of interdigitated electrodes with acoustic holes on the

89









backplate. Two different microphone designs were tested. One of the microphones consisted of

a two fingers with 16 acoustic holes and an open loop -3dB frequency of 30Hz. The second

microphone consisted of 10 fingers with 81 acoustic holes and an open loop -3dB frequency

of 400Hz For both microphones, a bias voltage of 7V were used. A 1MHz carrier signal used.

In closed-loop operation, the bandwidth of the two microphones increased to 400Hz and 7kHz

respectively. The feedback loop was compensated using a simple proportional gain controller.

In 1995, Hadjiloucas et al. [116] reported a force feedback microphone that uses optical

sensing and electrostatic force feedback. The movement of the microphone diaphragm is

detected using a Fabry-Perot interferometer. Since the diaphragm position is detected optically,

no ac signal is required for the operation of this sensor.

In 2005, Hall et al. [114] reported a measurement microphone that uses optical sensing with

electrostatic force feedback. The sensing is done using a Michaelson interferometer. The laser

required for sensing is integrated on the sensor itself using a vertical cavity surface emitting laser.

The backplate of the microphone is perforated and integrates the diffraction gratings. The paper

does not report any closed-loop results.

Next, closed loop accelerometers are discussed. In 1993, Analog Devices ADXL-50 [2]

produced a commercial accelerometer that uses closed-loop control with a 2Vpp, 1MHz carrier

signal. It uses the same plates for both sense and feedback.

In 1994, Kampen et al. [7] reported a closed-loop accelerometer that uses a 2Vp_p 1MHz

carrier signal. In this design, the same capacitor was used for both sense and feedback. No

performance metrics of the accelerometer is provided.

In 1998, Kraft et al. [104] reported a analog closed-loop accelerometer that uses a 2Vp_p

1MHz carrier signal. The device has a reported sensitivity of 0.7V/g and a bandwidth of

300Hz. The open loop bandwidth of the device is 56Hz.

In 1998, Park et al. [112] reported a closed loop accelerometer that uses a 11V,_ 50kHz

carrier signal. The sensitivity of the accelerometer is 39mV/g and the closed-loop bandwidth is

350Hz. This sensor uses the same plates for sensing and feedback.

In 2002, Luo et al. [117] reported a lateral accelerometer with a linear range of 13g and

a noise floor of lmg/ Hz. The carrier signal used was a 11,_, 2MHz signal. The sense and

demodulation circuitry is on the same die as the sensor. In this sensor, separate capacitors are

90









used for sense and force feedback. The sensor resonant frequency is 8.9kHz. A compensator

consisting of a single pole low pass filter at 400Hz is introduced in the forward path after the

demodulator which limits the bandwidth.

In 1990 Joyce et al. [118] reported a capacitive force for atomic force microscopy that

uses analog closed-loop force feedback. The nominal capacitance of the sensor is 0.4pF and

the resonant frequency is 30kHz. The sensor was excited with a 2.5MHz carrier signal. They

report a minimum detectable force of 10-8N which corresponds to a displacement of 0.5nm for

a bandwidth of 1kHz.

3.5.2 Sensing Using Digital closed-loop Interface Technique

In the digital technique for closed-loop force feedback, the sense and feedback signals are

time-multiplexed. The AZ modulation based force feedback technique is used. In this section,

the general operating principle of a first order AZ modulator is first explained. This principle is

then extended to a case study of a AZ modulator interface technique for the microphone.

AZ modulation is a technique used in high precision analog-to-digital converters (ADC).

AZ modulators are inherently oversampled systems and trade precision in time for precision

in amplitude. This allows one to use coarse analog circuitry with precise digital circuitry for

analog to digital conversion. The principle of operation of the AZ modulation is illustrated in

the block diagram of a first order AZ modulator shown in Figure 3-29. It consists of a discrete

time integrator whose output is quantized using a one-bit quantizer. The input to the integrator

is the error between the analog input signal and the quantizer output. The system is operated at

a sampling frequency, fs, that is much higher than the Nyquist frequency fN of the input signal.

Over many periods, the average error signal e is driven to zero by the feedback loop assuming

that the loop is properly compensated. The input information can be recovered from the output

bitstream by low pass principle is used as an interface technique, the discrete time integrator is

replaced by the microphone which is modelled as a second order system. Figure 3-30 shows

a schematic of the mechanical AZ modulator. The forward path is represented via three

physical blocks namely, the microphone, a charge amplifier, a discrete time compensator and a

comparator. The feedback path is represented via two functional blocks (digital to analog (D/A)

converter and voltage to force (V/F) block) which represent the electrostatic feedback force










Integrator Quantizer


Analog


First Order Delta
Sigma Loop


Figure 3-29. Schematic of first order AZ modulator.


applied on the plates at an instant of time. The summing node computes the difference between

the force corresponding to the input acoustic pressure signal and the electrostatic feedback force.

A feedback force is applied over time that nulls this error signal.


Input Digital
Pressure Bitstream
Area X(+ x/C --Am] p *] f -

Sensor Compensator
Comparator

V/FI~ _- D/A .9

Mechanical Electrical
Domain Domain

Figure 3-30. Schematic of second order mechanical AZ modulator.


Operating Principle. The operation of the system is divided into the sense phase during

which time the capacitance is measured and the feedback phase when an electrostatic force

is applied to the backplates to null the diaphragm deflection. During the sense phase, the

displacement of the diaphragm is first converted into a capacitance change by appropriately

biasing the top and bottom plates. This capacitance change is measured using the switched

capacitor technique discussed earlier in the open-loop techniques. The output of the charge

amplifier is given as the input to a two tap finite impulse response (FIR) filter. The compensator

is used to stabilize the second order loop. The output of the filter is quantized via a comparator

which produces a high speed single bit output signal at the over sampling frequency. The one bit

output is used during the feedback phase. This signal drives the logic that generates the feedback









signal which is then applied to the top and bottom plate. This one bit signal is decimated and low

pass filtered to obtain a multibit word that represents the input acoustic signal.

Performance Metrics. The noise floor of the AZ modulation based interface is

determined by the quantization noise. Assuming that the noise floor P, is dominated by the

quantization noise, it can be expressed in terms of the oversampling ratio M by [79]

A2 4
P, 6 --M-5 (3-80)
60

The AZ modulator is a non-linear closed-loop system. Linear systems theory cannot be

applied for the analysis of these systems. Some of the existing work in the literature [119, 6]

do not discuss the bandwidth variation with closed-loop AZ techniques. Kraft [69] discusses

an analytic method using the describing function technique [120] to qualitatively predict the

bandwidth improvement. The conclusion based on his work is that the feedback voltage increases

the sensor bandwidth as long as the input signal frequency is below the limit cycle frequency

predicted by the describing function analysis.

Tradeoffs. The main advantage of this technique is that the time averaged feedback force

is derived from the number of fixed height, fixed width pulses applied to the top and bottom

plates and not the amplitude of the feedback voltage. This provides a linear feedback voltage to

force relationship which was not possible with the analog closed-loop technique.

The main disadvantage of this technique is the complexity of the circuit. The other

disadvantages of this technique as we scale device geometries down are discussed in the next

section on scaling.

3.6 Scaling Analysis of AZ Interface Techniques

In this section, the circuit requirements for the AZ modulator as we scale device geometry

is analyzed. In general, for a given material, as the size of the mechanical element is reduced,

the resonant frequency increases. Therefore, to obtain the same performance using the AZ

based interface technique, the oversampling frequency must be increased. With the increased

oversampling frequency, high bandwidth amplifiers are required to meet the settling time

constraints. High bandwidth amplifiers imply larger device geometry which in turn leads to large

parasitics that increase the noise floor.









To get an understanding of the frequency consideration, Table 3-9 shows a comparison of

the sampling frequencies required for sensors with increasing resonant frequencies. The target

dynamic range is 120dB. To achieve the target dynamic range, a second order modulator needs

an oversampling ratio of approximately 725. This is rounded up to 1000X to provide a margin

of safety. To determine the closed-loop bandwidth f,, we use Equation 3-69 with n, = 7 (i.e.

0.1 settling), and m = 0.25 (i.e. 25' of time spent in sense phase). The next issue to consider

Table 3-9. Frequency requirements of AZ interface circuits with increasing resonant frequency.
Property Accelerometer Audio microphone Aeroacoustic
microphone
fres kHz[8] 20kHz 200kHz
f 1MHz[8] 20MHz 200MHz
f, 4.5MHz 90MHz 900Mhz


for high dynamic range capacitive sensors using feedback techniques is the voltage required

to balance the input at maximum input signal level. Table 3-10 shows the voltage required to

balance the maximum pressure for audio grade and aeroacoustic microphones. From Table 3-9

Table 3-10. Feedback voltage requirements of AZ interface circuits with increasing dynamic
range.
Property Audio microphone Aeroacoustic microphone
Maximum input 120dB 160dB
Feedback voltage 6V 45V


and Table 3-10, it can be seen that designing a AZ modulation based controller for aero acoustic

application is a non-trivial problem. This is because current semiconductor manufacturing

process technologies do not have 45V processes which support amplifiers with a 900Mhz gain

bandwidth product. One method to overcome this problem is to use separate voltages for sensing

and feedback using off chip drivers.

Literature Review. Table 3.6 provides a review of capacitive transducers using AE

based closed-loop techniques. This technique has been reported extensively for capacitive

accelerometers. The use of this technique for closed loop force feedback in high bandwidth

operations is very challenging due to the large simultaneous gain bandwidth and voltage

requirements for these applications.

The first digital closed-loop accelerometer was reported by Henrion et al.[42] in 1990. The

system operates at a clock frequency of 512kHz. The reported system sensitivity is 1V/g. The

94










Table 3-11. Summary of previous AZ based closed-loop capacitive transducer interface circuit.
Author / Type of Transducer Nominal Clock Sense / Comments
Transducer Bandwidth CapacitanceFrequency Feedback
Open-Loop / capacitors
Closed-loop
Henrion et al., 266 Hz / NR 512kHz Common Sensitivity is 1V/g
1990 [42] 350 Hz plates
(accelerometer)
Yun et al., 1992 [6] Theoretical NR 1MHz Separate Only open loop
(accelerometer) resonant plates mode of operation
frequency of was tested thought
sensor the chip has the
15kHz electronics for
closed-loop.
Wu et al., 1992 [121] 25kHz NR 5MHz Common
(microphone) plates
theoretical paper
Kraft et al., NR 7.2pF 500kHz Common
1998 [104] plates
(accelerometer)
Lemkin et al., Theoretical 101fF 500kHz Common Sensitivity is 1V/g,
1998 [122] resonant plates Noise floor is
(accelerometer) frequency of 1 10g/Hz (Only
sensor x-axis data is
3.4kHz presented)
Moon et al., Theoretical 0.lpF 1MHz Separate
2000 [123, 10] resonant plates
(accelerometer) frequency of
sensor
15kHz
Edelson 2001 [124] NR / NR NR NR NR NR
(microphone)


resonant frequency of the sensor was 266Hz and the measured resonant frequency with the

electronics use 350Hz.

In 1992, Yun el al. [6] reported the design of a digital closed-loop accelerometer. In this

design separate capacitors was used for sense and feedback. The fabricated device was tested

only in the self test open loop mode using electrical excitation.

In 1994, Wu et al. [121] presented a theoretical paper on the design of a AE based

closed-loop controller. The proposed interface circuit uses interferometric optical sensing to

determine the diaphragm position of a dual backplate microphone and electrostatic feedback. The

circuit is designed to have a overall system bandwidth of 25kHz. The oversampling frequency

used is 5MHz.









In 1998, Kraft et al. [104] reported a digital closed-loop accelerometer that uses alOkHz

clock signal. The open loop bandwidth of the device is 56Hz and the closed-loop bandwidth is

300Hz.

In 1998, Lemkin et al. [122] presented a three axis MEMS accelerometer. It uses a switch

capacitor amplifier for detecting the capacitance change. The circuit was characterized at a clock

frequency of 500kHz. The x-axis accelerometer has a sensitivity of 1V/g with a noise floor of

11Opg/ Hz.

In 2001, Edelson et al [124] of Borealis Technical Limited obtained a patent for a

AZ modulation based force balanced microphone. The microphone described is of the

electrodynamic type but the patent covers electrostatic devices also. In this patent, the same

plates are used for both sensing and for feedback.

3.7 Summary

In this chapter, the commonly used open and closed-loop interface techniques were

reviewed. Theoretical formulation for the performance metrics of the various techniques were

developed. Tradeoffs for each of the techniques was explained. Requirements of AZ modulation

based interface circuits as device geometry is scaled was then addressed. Chapter 4 and Chapter

5 present a detailed design procedure for the design of the analog and digital closed-loop

controller, respectively.











CHAPTER 4
DESIGN OF ANALOG INTERFACE CIRCUIT

4.1 Introduction

In this chapter, the design and simulation of an analog closed-loop interface circuit is

presented. This chapter begins with the development of a model for the closed-loop system. This

model is based on the lumped element model of the microphone developed in Chapter 2 and

the model of the synchronous modulator/demodulator developed in Chapter 3. Open and closed

loop time domain simulation of the model for sinusoidal and step input is presented next. This is

followed by the details of the circuit implementation This includes the design of the interface

circuit components namely the amplifier, demodulator and compensator. This chapter concludes

with a discussion of the limitations of the circuit. A graphical overview of this chapter is shown

in Figure 4-1.


Introduction
Matlab model


Simulation Open loop
Simulation

Amplifier design Closed oop
Closed loop
Design of interface simulation
circuit
Demodulator
design
Implementation of
Compensator interface circuit
design

Limitations of the
interface circuit


Summary



Figure 4-1. Overview of Chapter 4.



4.2 Simulation

The simulation of the complete system involves the development of a model that accurately

captures the non-linearities of the demodulator and voltage to force transfer function in Simulink.

This model is then simulated for both open-loop and closed-loop mode of operation. A block

diagram of a closed-loop microphone is shown in Figure 4-2. The input to the system is a









Pressure np Net Interface Compen

Sensor Displacement
(Force to to voltage
displacement)

Electrostatic Voltage to
force d Force

Figure 4-2. Block diagram of analog closed-loop system showing microphone and interface
circuit.


time varying pressure signal generated by the acoustic source. The input pressure acting on the

constant area of the microphone diaphragm produces an input force. The net force acting on

the diaphragm is the difference between the input force and the electrostatic feedback force.

The diaphragm displacement is electrostatically transduced as a change in capacitance by

appropriately biasing the microphone plates [37]. Synchronous modulation/demodulation [125]

is used as the interface circuit to detect the capacitance change. The output of the circuit after

compensation provides the feedback voltage. This voltage causes an electrostatic force which

opposes the input force and acts to reduce the diaphragm displacement.

4.2.1 Simulink Model

The second order model of the microphone in the acoustic domain developed in Chapter 2

is used for the microphone model. A linear model for the interface circuit and voltage-to-force

block is developed using the small displacement approximation. Using the lumped element

microphone model, interface circuit model, and voltage-to-force transfer function model, the

compensator is designed.

The microphone is modeled as a second order system given by the equation

1/m
1Hsen=2 (4-1)
S2 + 2( (27fe) S + (27 frf2)2

developed in Chapter 2.

The interface circuit measures the change in capacitances C1 and C2 induced by the input

pressure acting on the diaphragm. The top and bottom plates of the microphone are excited

by a dc voltage (Vdc) superimposed on a square wave ac carrier (va) of opposite polarity. The

excitation frequency (u ) is chosen to be much higher than the resonant frequency of the sensor









to prevent the ac signal from exciting the microphone diaphragm. The top and bottom plate

voltage, vtp and Vb, with no feedback is given by


Vtp = +Vdc Vac(t) ; Vb = -Vd + Vac(t. (4-2)


Cp represents the parasitic capacitances due to the interconnection between the microphone and

interface circuit, and C, is the amplifier input capacitance. Rdc is a bias resistor that sets the dc

operating point at 0V. A unity gain amplifier is connected to the middle plate of the microphone.

The output of the unity gain amplifier Vmod is given by

Clo + C2o x'
Vmod = Vtp, (43)
r CiO + C2O + C20+ C / ox 9' (4-3)
CIO + C20 +Cp + X

where xo is the nominal gap distance between the plates of the microphone, x' is the displacement

of the diaphragm from its equilibrium position, and Clo and C20 are the equilibrium microphone

capacitances. In Equation 4-3, the ratio c',o+c20 is defined as the attenuation factor and is
Clo+C20+Cp+Ci
denoted by He. The output of the amplifier is connected to a demodulator which multiplies the

modulated middle plate voltage with a carrier signal (vc(t)) which is in phase with the top plate

voltage. The output of the demodulator, Vdemod, is given by the expression


Vdemod = Vmod X Vc(t), (4-4)


and has signals at the base band, carrier frequency, and multiples of the carrier frequency. The

output of the demodulator is low pass filtered and is given by

x' 2 V 2 Vac
Vout = He 2 (4-5)
o0 7 2 7 2

Using Equation 4-5, the transfer function of the interface circuit, Hint is given by

x' H, -2 Vc2 Vac
Hint = (4-6)
Vout xo r 2 r 2

The voltage-to-force block models the transduction of the feedback voltage vf to an

electrostatic force acting on the diaphragm. The forces between the top plate and the diaphragm,

Ftp, and between the bottom plate and the diaphragm, Fp, are given by

1 v2 1 v
Fp -A- 2 ; F;, A 2p (4-7)
2 [xo x']2 2 [xo x'(]2









where the top and bottom plate voltages vtp and vbp are expressed in terms of the feedback

voltage vf by the expression


Vtp = +Vd Vc(t) + f ; Vbp = -Vd + Vac(t) + f. (4-8)

The feedback voltage vf is generated by the compensator. The design of the compensator is

discussed in the next section. The net electrostatic feedback force on the diaphragm F, is given

by the difference of Ftp and Fbp i.e.,


F Ftp Fbp. (4-9)


For small displacements (x' w 0), F, can be linearized to


Fe 2cA VVf. (4-10)


Equation 4-10 represents the linearized input-output relation between the feedback voltage and

the force on the microphone diaphragm. Using Equation 4-10, the feedback transfer function,

Hfb is given by

Hfb F- 2cA V. (4-11)
vUf x
The function of the compensator is to filter the high frequency modulating signal and

provide appropriate gain and phase so that the resultant closed loop system is stable. The

compensator consists of two stages. The first stage is a high-gain low-pass filter that sets a

dominant pole at low frequency. The second stage consists of a lead-lag section that provides

primarily phase lead to obtain sufficient phase margin. The loop gain of the system without the

compensator, T,,(s), is given by


Tc(s)= Hmic x Hit x Hfb. (4-12)

From the uncompensated system Bode plot in Figure 4-4, it can be seen that the system has

very low loop gain at dc and a high Q resonant peak due to the complex conjugate poles of

the microphone. To improve the loop gain and provide sufficient phase margin, a two stage

compensator given by the transfer function H(s)

F( I o 27w- 1- 103 s + 27 -230- 103
H(s) = 750 2. 10 L s+2.30 103' (4-13)
s + 2 1100 s+ 30 03
100










is used. The bode plot of the compensator is shown in Figure 4-3. It can be seen that the

200

S 100
0-


-100
100 102 104 106 108



100

0
"o


-100-

100 102 104 106 108
Frequency (Hz)

Figure 4-3. Bode plot of compensator.


compensator has a high gain low frequency pole at 1kHz. A high frequency zero at 230kHz is

used to provide phase margin. The loop gain of the system with compensator, T(s), is given by


T(s) 1/M (4-14)
S2 + 2((27Tf,) + (27Tfrs)2

H, (2Vc2V.a)
Xo 7r 2 7r 2

x 2cA (4.15)

S+27- 1- 03 s+27-30- 03j


A Bode plot of the compensated loop gain is also shown in Figure 4-4. It can be seen that the

system has a phase margin of 420 and a dc loop gain of 87.

The complete closed loop Simulink model is shown in Figure 4-5. The microphone

and compensator are modeled using a "s-domain" transfer function block which implements

Equation 4-1 and Equation 4-13 respectively. The synchronous modulator/demodulator interface

circuit model and voltage-to-force subsystem model implement Equation 4-4 and Equation 4-9

respectively using multipliers and gain blocks. Thus, a time domain model of the closed loop

system has been developed. Open and closed loop operation for step and sinusoidal inputs were

simulated and is presented next.

























S-100

-200

-300
100


Frequency Hz

-****** without compensator
~ -with compensator

t ".,


104
Frequency (Hz)


6 ----
106 108


Figure 4-4. Bode plot of loop gain with and without compensator.


Figure 4-5. Simulink model of analog closed-loop control system.


4.2.2 Simulation Results

In this section, the open and closed loop simulation results of the capacitive microphone are

presented. To demonstrate closed loop operation, a 2Pa (100dB re. 20/Pa), 2kHz sinusoidal

pressure input is used as the test signal. From linear control theory [126], during closed loop

operation, the diaphragm displacement at a given frequency will be attenuated by the loop gain

evaluated at that frequency. From the time domain simulation (Figure 4-7), it can be seen that the

output displacement for a 2kHz input is reduced by a factor of 40 which matches the attenuation

in diaphragm displacement predicted by the loop gain (Figure 4-4). To further examine closed

loop operation, a lm drop test was simulated. The pressure corresponding to a lm drop acting

on the diaphragm is simulated by a 1ms, 0.029Pa(63dB) pulse. The diaphragm displacement is













vO p X
vac vtp X 1/Kmul 1
2 multiplier utpeZ
mutliplier
attenuation
vdc
vdc (b) Simulink model of demodulator.
(a) Simulink model of
top plate voltage
subsystem.


r O 5"eO"A/((dO-u(1))^2) 4 -
displacement
1 function nt
Force Product-- ^,)*u(l) 1 2
top_plate_voltage
volsquared1
Bottom Plate'
Voltage
Gain

Product vol_squared2
0 5*eO*A/((dO+u(1))^2) 4 feedback_voltage

function
(c) Simulink model of voltage-to-force subsystem


Figure 4-6. Subsystems used in Simulink model.


attenuated from 6.2pm to 7.1 x 10-2pm during the pulse (Figure 4-8). This matches the dc loop

gain of 87(38dB) obtained from the loop gain Bode plot (Figure 4-4).

4.3 Design of Analog Interface Circuit

The design of the analog closed-loop controller circuit involves the design of the summing

amplifier for top and bottom plate voltage generation, input buffer amplifier, demodulator, low

pass filter, and compensator. The design of each of these circuits is discussed next.

4.3.1 Design of Amplifier for Top and Bottom Plate Voltage Generation

The top and bottom plate voltages are generated using a matched pair of summing amplifier.

The two main requirements of this amplifier are as follows. First, the unity gain bandwidth of

the amplifier must be higher than the modulation frequency. For the aeroacoustic microphone

with a measured resonant frequency of 230kHz, the modulation frequency was chosen to be

one decade above the resonant frequency at 2MHz. Second, the amplifier must accept a dc input

greater than the voltage required for force balancing at the maximum input sound pressure level.

The voltage required for the two microphone structures discussed in Chapter 2 is repeated

in Table 4-1. Based on the table, it can be seen that we need an amplifier operating at a voltage












2-


a -2
0 0.5 1 1.5 2
x 103
0.5
S-- Without feedback
E 0.

CL a -0.5
S 0 0.5 1 1.5 2
Time (s) x10-3
0.5
"E --With feedback
E E 0

Sa--o.5-
S 0 0.5 1 1.5 2
Time (s) x 103


Figure 4-7. Diaphragm displacement with and without feedback.


e 0.05
C)



i o
0


S0.01

1E E 0

a a -0.01
S 0


0. u

E .


Time (s) x 103


ul
Ul---------------------- i
-With Feedback

0
55 0 0. 1 .

01----------------------


S 0 0.5 1 1.5 2 2.5 3
Time (s) x 10-3


Figure 4-8. Diaphragm displacement with and without feedback.


greater than 45V with a unity gain bandwidth of at least 2MHz. Such amplifiers are not


available commercially. The maximum operating voltage of high bandwidth commercially

available amplifiers is 15V. An OP275 dual operational amplifier which has two matched


amplifiers with a unity gain bandwidth of 11MHz at an operating voltage of 15V is used. To


prevent power supply noise from entering the system, the top and bottom plate dc voltages are

provided by batteries regulated using a linear regulator. The maximum pressure that can be force


balanced is approximately 134dB.

The schematic of the summing amplifier circuit that generates the top and bottom plate


voltages is shown in Figure 4-9. In this figure Vdc and -Vdc represent the dc voltages used









Table 4-1. Feedback voltage requirements of aeroacoustic and audio microphone for force
balancing maximum input pressure.
Property Audio microphone Aeroacoustic microphone
Maximum input 120dB(20Pa) 160dB(2000Pa)
Feedback voltage 5V 45V


for force balancing, Vac(t) and -vac(t) are the voltages required for synchronous modulation

and demodulation, and vf is the feedback signal generated by the circuit. The ac voltages are

generated by two Agilent 33014 signal generators that are exactly phase shifted by 180 degrees.

The amplitude of the ac signal is 5Vpp.

R
R
Vd

-Vcsin(wt) R V
S> tp





R
R'R
-Vdc -
Vacsin(wt) R V
R 0
Vf > V




Figure 4-9. Schematic of summing amplifier that generates top and bottom plate voltages.


4.3.2 Input Buffer Amplifier Design

The input amplifier is a simple unity gain follower and is shown in Figure 4-10. The

unity gain amplifier was implemented using a low noise, high bandwidth amplifier, TLE2074

[127] configured for unity gain. A quad opamp was chosen so that the four aeroacoustic

microphones on each sensor die could be interfaced using a single chip. This amplifier

operates on a 15V power supply, has an input capacitance of 10.6pF, an input referred

noise voltage of 1lnV/(Hz) and an input referred noise current of 1.6 fA/(Hz) The

middle plate of the microphone is biased using two surface mount 22MQ resistors in a









1210 (3.2mm x 2.5mm x 0.6mm) package wired in series, i.e,


Rdc = 44MQ. (4-16)


The total parasitic capacitance at the input node is given by the sum of the PCB trace capacitance

and the input capacitance of the amplifier. The total parasitic capacitance is approximately 20pF.

4.3.3 Demodulator Design

The output of the interface circuit is demodulated by the the Analog Devices AD835 four

quadrant analog multiplier IC [102]. This IC is powered with a 5 power supply and has a

200MHz full power bandwidth. If the two input signals to the multiplier are X and Y

+
RDC!



Figure 4-10. Schematic of voltage amplifier.


respectively and the output signal is W, then the transfer function of the multiplier is given

by
XY
W = (4-17)

A schematic of the demodulator IC is show in Figure 4-11.

Input Demodulated
Signal x- w Output
-E x2 W2 11_
AD 734
Carrier
Signal
F7 Y2


Figure 4-11. Schematic of analog multiplier (AD835) showing relevant pins.


4.3.4 Compensator Design

The compensator was implemented using two opamps. The schematic of the compensator

for the first stage and second stage is shown in Figure 4-12. The transfer function of the circuit









shown in Figure 4-12(a), Hcl, is


H, = 1RflCf
Rn k1+RRfCfIs '


(4-18)


and the transfer function of the schematic shown in Figure 4-12(b), Hc2, is

2 f2 1 + RC12 (4-19)
RH12 I + sRf2 Cf2 (4

For the transfer function given in Equation 4-18 and Equation 4-19, the component values used

are listed in Table 4-2. The various circuitry used in the closed loop operation are soldered on

two different PCBs. This is discussed in the next section.

Table 4-2. Element values for resistors and capacitors used in compensator.
Element Value Element Value
R1 Ikt C11 680pF
Rfl 5.lkQ Cjl 1000pF
R12 1lk
Rf2 1590kQ Cf2 IOOpF


Vi
0--


R11


(a) Schematic of first stage of compensator.


(b) Schematic of second stage of
compensator.

Figure 4-12. Schematic of two stage compensator.









4.4 Integration of Interface Circuit Components on a PCB


Figure 4-13. Annotated photograph of PCB 1 showing buffer amplifier, breakout headers and
optical port.


Summing amplifier
for top and bottom
plat voltage Demodulator Compensator











Figure 4-14. Annotated photograph of PCB2.


The complete system consisting of the microphone and interface circuit is characterized

in an acoustic pressure coupler (APC). The details of the APC are provided in Chapter 6. To

facilitate mounting of the microphone in the APC, the circuitry is divided into two separate

printed circuit boards (PCBs). The microphone and buffer amplifier are soldered onto the

microphone PCB (denoted as PCB1) which is mounted onto the APC while the remainder of

the circuitry consisting of the demodulator, compensator, and top and bottom plate voltage

generation circuitry PCB are mounted on another PCB (denoted as PCB2). The two boards

are electrically connected with each other using a ribbon cable. This cable couples the top and

bottom plate voltages generated by the PCB2 to the microphone plates on PCB1. This cable also









connects the buffer amplifier output from PCB1 to the PCB2. Figure 4-13 and Figure 4-14 shows

an annotated picture of the PCB 1 and PCB2. The headers for the ribbon cable are shown in the

figures.

4.5 Noise model

In this section, the noise model of the interface circuit is developed using the noise

parameters of the components used in the system. The noise floor of the forward path system

is determined by the the buffer amplifier, gain amplifier, demodulator, and compensator and

is shown in Figure 4-16. The total output referred noise of the forward path is determined by


RRf,
R d R1 R
+VmI Vdmod R1 R12
CO ~^ Multiplier + +I
Buffer n (Demod) C ,,
Amplifier Amplifier C sato Compensator
Stage 2

Figure 4-15. Schematic of forward path of the closed loop system


summing the noise at the output of each stage of the forward path with the appropriate gain. This

noise model is valid in the bandwidth of operation of the closed loop system which extends from

300Hz to 20kHz. In the bandwidth of interest, the noise spectrum is determined y the thermal

noise of the various components and is flat over the spectrum. Hence in this analysis, the shaping

of the noise by the various components is not considered.

Noise at the output of the buffer amplifier. The noise PSD at the output of the buffer

amplifier, S1, is determined by the voltage noise PSD of the buffer amplifier, Svai, and is given

by

S1 = Sva. (4-20)

Noise at the output of the gain amplifier. The noise PSD at the output of the gain

amplifier, S2, is determined by the output noise PDS of the previous stage, S1, voltage noise PSD

of the input resistor R1, voltage noise PSD of the feedback resistor, Rf, and voltage noise PSD,

Sva2, and current noise PSD, Sia2, of the amplifier and is given by


S2 S, (Rf 4kTR1 2 + 4kTRf+Sa2 Rf2 + Si2 (Rf)2. (4-21)
U \HR1) R












Table 4-3. M
Parameter
Voltage noise of input
amplifier (S,vi)
Voltage noise of gain
amplifier (Sa2)
Current noise of gain
amplifier (Sia2)
Voltage noise of multiplier
(Smultiplier)
Voltage noise of compensator
first stage amplifier (Sva3)
Current noise of compensator
first stage amplifier (Sia3)
Voltage noise of compensator
second stage amplifier (Sva4)
Current noise of compensator
second stage amplifier (Sia4)
Input resistance of gain
amplifier (R1)
Feedback resistance of gain
amplifier (Rf)
Input resistance of
compensator first stage
amplifier (R11)
Feedback resistance of
compensator first stage
amplifier (Rf1)
Input resistance of
compensator second stage
amplifier (R12)
Feedback resistance of
compensator second stage
amplifier (Rf2)
Sfwdpath
SOL


Noise at the output of the multiplier.


agnitude of parameters used in noise analysis
Magnitude Units
18 nV/ Hz


3.8

0.6

50

7

1.5

7

1.5

1

4

1


5.5


1


1590


1
5


nV/ /Hz

pA/lHz

nV/ Hz

nV/ Hz

pA/lHz

nV/ Hz

pA/ Hz

kQ


mV Hz
uV Hz


The noise PSD at the output of the multiplier, S3, is


determined by the output noise PSD of the previous stage, S2, the internal voltage noise PSD of

the multiplier, Smutiplier, and the demodulator voltage Vref,and is given by


S3 = S2 x Vref)2 + Smultiplier.


(4-22)










Noise at the output of the first stage of compensator. The noise PSD at the output of

the compensator first stage, S4, is determined by the output voltage noise PSD from the previous

stage, S3, voltage noise PSD of the input resistor R11, voltage noise PSD of the feedback resistor

Rfl, and voltage noise PSD Sva3 and current noise PSD Sia3 of the amplifier and is given by

SI ) 2 + 2 4 \ 2 SR \ 2 (Rf1)2
S4 =3 + Sva3 l 4kTRj f + 4kTRf + Sva3 + Sia3 fl

(4-23)

Noise at the output of the second stage of compensator. The voltage noise PSD at the

output of the compensator second stage, S5, is determined by the voltage noise PSD from the

output of the previous stage, S4, voltage noise PSD of the input resistor R12 amplifier, voltage

noise PSD of the feedback resistor Rf2, and voltage noise PSD Sv43 and current noise PSD Sia4

of the amplifier and is given by


S5 S4 2 +Sa4 + 2 2 + 4TR2+ Sva44 Rf + Si4 ()f2 2

(4-24)

Theoretical noise floor. The magnitude of the various components in the above equation

listed in Table 4-3. Substituting the values in the Equation 4-24 we obtain

101
SForward path noise
Open loop noise


100



10-1 : : : : :
0)




0-3
(/)

10-2




10-3 2. 3 4 5
102 103 104 10
Frequency

Figure 4-16. Plot of theoretical noise floor of the open loop system and forward path system









SOL = 3 = 5uV/ Hz)( (4-25)


and

Sf dpath S5= lmV/ ( Hz) (4-26)

Thus, the noise at the output of forward path of the demodulator is high due to the components

that are required to generate the large gain in the forward path. It can be seen that the forward

path noise is very high. The reason for the high noise floor is the need for large gain in the

forward path to achieve a large loop gain. One solution to reduce the noise and still maintain the

large gain would be to use amplifiers with lower noise. In this work, all the ICs utilize the same

15V power supply whereas lower noise ICs usually are single ended and operate at 3.3V and

1.8V. Hence, to use these lower noise components, multiple power supply rails are required.

Also, as the ICs are single ended, appropriate level shifting needs to be performed to obtain

stable operation.

4.6 Limitations of the Current Implementation of the Interface Circuit

In this section, the two main limitations of the interface circuit designed in this chapter

are reiterated. The interface circuit was designed to meet the dynamic range and bandwidth

specifications of an aeroacoustic microphone of 60dB 160dB and a bandwidth in excess of

160kHz.

The voltage required to force balance the maximum input pressure of 160dB as shown in

Table 4-1 is 45V. This cannot be met due to power supply limitations of 15 for the commercial

off the shelf operational amplifiers that are used in the design. The bandwidth of the overall

system is determined by the compensator. The compensator design was limited to an overall

bandwidth of only 20kHz because of the gain bandwidth constraints of the amplifiers used.

4.7 Summary

In this chapter the design, simulation and implementation details of the analog interface

circuit was presented. Some of the limitations of the current design was discussed. The

characterization results of the complete system is presented in Chapter 6. Chapter 5 discusses the

design and simulation of the digital closed-loop interface circuit










CHAPTER 5
DESIGN OF DIGITAL CLOSED LOOP INTERFACE CIRCUIT

5.1 Introduction

In this chapter, the design procedure for a AZ modulation based interface circuitry is

presented. This chapter begins with a top down design methodology for determining system

level AZ interface circuit parameters starting from microphone system specification. This is

followed by the design of the interface circuit components, namely the amplifier, compensator

and comparator. Closed loop time domain simulation of the model using the Simulink tool in

MATLAB is presented next. A graphical overview of this chapter is shown in Figure 5-1.


Introduction




Design procedure



Amplifier design tom
Component design

Compensator
design ---
design Simulation Matlab model

Comparator design C l o
Closed loop
Summary simulation


Figure 5-1. Overview of Chapter 5.


5.2 Design of AY Interface Circuit

In this section, the design of the complete AY interface circuit is discussed. The design

procedure for the key parameters of each of the blocks of the AS modulator from overall system

level specification is also explained. The transistor level design of each of the components is

presented in Section 5.3. The sequence of operation of the various blocks and the overall system

level timing is explained in Section 5.4. Section 5.5 describes the system level simulation of the

microphone and AZ system using Matlab.










The system level block diagram of the AZ modulator is repeated in Figure 5-3. The design

of the interface circuitry involves the design of the components inside the box labeled "electrical

domain". The AZ modulator based interface technique is a sampled data system. Hence, all the

blocks need a clock signal for their operation. Therefore, the first step in the design procedure

of the interface circuit is to determine the sampling frequency f,. The sampling frequency is a

function of the resonant frequency of the mechanical system and the dynamic range specification

of the overall system. This is illustrated in Figure 5-2. The blocks in ellipsoidal boxes represent

the system level input specification and rectangular boxes represent the assumptions used for the

design and the parameters derived from the design procedure. The dynamic range is determined

(Dynamic Range) (Bandwidth




Effective feedback Sampling





consideration (Closed loop BW)
voltaFigure 5-2. Flow chart for design of modulator component parameters from system level







specification.


by the noise sources in the system. The main sources of noise are the sampled thermal noise of

the amplifier, kT/C noise of the switches, and quantization noise. For an ideal AZ modulator,

the noise floor is dominated by the quantization noise. The sampled thermal noise and kT/C
/ Nose Settling time


















noise is assumed to be below the quantization noise floor. Assuming that the noise floor is
FCompensator Amplifier g.







dominated by the quantization noise, the dynamic range DR, in dB, of a one-bit second order

Figure 5-2. Flow can be expressed in terms of the oversamponent paratio meters f/2rom system level
specificatiorelation [79]

DRby the noise sources in the system. The main sources of noise are the sampled thermal noise of
the amplifier, kT/C noise of the switches, and quantization noise. For an ideal AS modulator,

the noise floor is dominated by the quantization noise. The sampled thermal noise and kT/C

noise is assumed to be below the quantization noise floor. Assuming that the noise floor is

dominated by the quantization noise, the dynamic range DR, in dB, of a one-bit second order

AS modulator can be expressed in terms of the oversampling ratio M(- fs/2fr..) by the

relation [79]

DR -7 + 50log1o(M). (5-1)










Using the above relation, the sampling frequency of the system is determined.


Mechanical Electrical
Domain Domain

Figure 5-3. Schematic of second order mechanical AE modulator.


Once the sampling frequency is determined, the next step is the partitioning of the time

available during one clock cycle T(= 1/ f) into the sense and feedback phase. To maximize

the dynamic range and bandwidth, the time spent in the sense phase must be minimized and the

time spent in the feedback phase must be maximized. The time spent in feedback is determined

by the maximum pressure that must be force balanced and the maximum voltage Vmax available

in the system. The voltage required to force balance the maximum input pressure Pmax can be

expressed in terms of the nominal capacitance Co, the capacitance area A, and the nominal gap

distance of the capacitor xo by the expression

2P nmxAxo
Vfb,max 2Pm xo (5-2)
Co

This voltage must be available during the fraction of time rfb spent in feedback. If the maximum

voltage available in the system is Vmax, then

Tfb Vfb,mazx (
(5-3)
T Vmax

Using the above relation, the time spent in feedback can be determined. Then, the time spent in

sense -Tsens is given by

Sense = T Tfb. (5-4)

Next, we consider the partitioning of the time spent in the sense phase. The time spent during

sense Tsene is determined by the settling time requirements of the amplifier Tamp, compensator









Tcompen, and comparator Tco-par and is related by the expression


Sense = Tamp + Tcompen + Tcompar r (5-5)


The individual blocks must to satisfy two conflicting criteria, namely high speed (settling time

requirements) and low noise. To achieve this, the circuit is first designed for low noise and

then tested for settling time requirements. If the low noise design does not satisfy the speed

requirements, the circuit design is reiterated by changing the gain capacitor to meet the settling

time requirements, and the noise performance is reevaluated. Theoretical formulation for the

noise and settling time for each of the components are described in Section 5.3.

Once the individual components are designed, the sequence of operation of the individual

parts of the sense and feedback phase is designed. This is explained in Section 5.4. For this

work, the timing signals are generated off chip using a Alter EPM7128 [128] complex

programmable logic device (CPLD). This allows one to have better control of the sense and

feedback phase during the initial testing phase. Based on this information the system level

simulations (Subsection 5.4) are performed.

5.3 Component Design

In this section, the transistor level design of the amplifier, compensator, and comparator

is explained. For each component, the procedure to design the component with low noise is

described. A theoretical formulation for the noise and the settling time is also developed.

5.3.1 Switched Capacitor Amplifier

The design of the switched capacitor amplifier involves the design of the amplifier core so as

to reduce the thermal noise. A folded cascode topology with a PMOS input stage was chosen so

as to reduce the noise of the input amplifiers and improve the output voltage swing. The switches

are arranged such that the output common mode is centered around 2.5V while the input is at

ground. Centering the output voltage at the midpoint of the power supply rails allows one to

maximize the output dynamic range of the amplifier. Similarly, setting the input common mode

voltage of the amplifier at ground maximizes the voltage that can be applied across the top and

bottom capacitors during the sense and feedback phase. An input common mode voltage of zero

(ground) is within the common mode voltage range of the PMOS input folded cascode amplifier.










The schematic of the amplifier core and the switched capacitor amplifier is shown in Figure 5-11.




M1l M7 M8


M5 M6

Creset


2 V-
-o M3 M4
C ,
gain


SVbn
0ov + M9 M10


(a) Switched capacitor amplifier. (b) Amplifier core.

Figure 5-4. Transistor level schematic of amplifier.


The design of the amplifier closely follows the work by Wongkomet [108]. The following

assumptions are made:

The amplifier core is a simple single stage amplifier that can be treated as single pole

system which is common in switched capacitor circuits.

The gain capacitor Cgai is arbitrarily chosen to provide a closed-loop gain of 10 20dB.

This value may be changed to optimize the noise performance.

The total parasitic capacitance at the input node is fixed and can be estimated.

The amplifier closed-loop bandwidth is designed to be at the minimum value required

based on the settling time constraints. This prevents excess noise from aliasing into the

baseband.

The amplifier schematic during the amplify phase used for the rest of the analysis is shown

in Figure 5-5. The thermal noise of the amplifier can be expressed in terms of the amplifier

transconductance gm and the ratio of the total transistor noise to the input transistor noise nf by

the expression[79]
2
S,, 4kT i nf. (5-6)
3./.










reset

OV


Cc
0Cgain



OV
Co 0-- --
C C
SCp,ext p,ampT





Figure 5-5. Schematic of amplifier during amplify phase.


For a well designed amplifier, only the input transistors contribute noise, and nf is equal to two.

In practice it can be greater. From the discussion in the previous chapter, the total output thermal

noise of the amplifier is
CT 1 2 ^r5_
Svopamp S .v (5-7)

In the above equation, CT is given by


S 2Co + Cp+ Cgain (5-8)


The parasitic capacitance, Cp, can be split into the input capacitance of the amplifier, Cp,amp and

other external parasitics Cp,xt. Therefore, CT can be expressed as


CT 2Co + Cp,amp + pxt + Cgain. (5-9)


For any transistor, the transconductance gm, can be expressed in terms of the input capacitance

Cp,amp, and the unity gain frequency, fT, as


gm = 27TfTCp,amp. (5-10)


Substituting, Equation 5-6 and Equation 5-9 in Equation 5-7, we obtain

S2Co + Cp,amp Cp,ext + Cgain 2 8kT 1 f, 7r
Svopamp Cgain 3 27frTCp,p (f 2-11









This equation is optimized for noise by varying the input capacitance of the amplifier.

Differentiating the above equation with respect to Cpamp and equating to zero we obtain


Cp,amp = 2Co + Cp,ext + Cyain. (5-12)

From the previous equation, it can be inferred that to minimize the noise of the amplifier,

the input capacitance must be equal to the total capacitance at the amplifier summing node.

Substituting Equation 5-12 in Equation 5-10, gm can be designed. From the transconductance,

the W/L of the transistors of the amplifier can be calculated.

Once the transconductance is designed, the load capacitor is designed to meet the bandwidth

requirement. As discussed in the previous chapter, the closed-loop bandwidth of the switched

capacitor amplifier is given by

fi t (5-13)
27 nrfs

The closed-loop bandwidth f, is related to the open-loop bandwidth famp by the feedback factor


CT
C (5-14)
Cgain + CT
and the open-loop gain,

famp (5-15)
2r7 CLT'

by the expression

fu. famp (5-16)

The total load capacitance CLT, is given by the expression


CLT = CL + Creset + Cgin|| [2Co+Cp,amp + Cp,ext] (5-17)


Substituting Equation 5-16 in Equation 5-15, we obtain

1 9-
f. = (5-18)
27 CLT
1 gmj3
2w CL Cr + gain [2Co+Cp,amp+Cp,ext]
SCgain+2Co+Cp,amp +p,ext









Rearranging the above equation, we obtain


CL Yai grn (2 + Cgain + Cp,amp + Cp,ext)] reset.
2Co + C-gin + Cp,amp + Lp,,,t 2 f
(5-19)

Substituting for Cp,mp from Equation 5-12 and f, from Equation 5-13 into the previous

equation, we obtain

CL = 1 n eset (5-20)

This can be simplified to

CL -- 7 gainfT C reset. (5-21)

This is the load capacitance required to satisfy the low noise requirements and settling time

constraints. If the actual load capacitance which is formed by the compensator is larger, then a

larger integrating capacitor can be used and the procedure is reiterated. Thus, the amplifier has

been designed to meet the settling time constraints and minimize noise.

5.3.2 Compensator Design

The function of the compensator is to provide feedback to stabilize the second order

feedback loop. One of the proposed solutions in the literature [119] for compensating a AE

modulation based capacitive sensor is to use a two tap discrete time finite impulse response

(FIR) filter. The filter transfer function can be expressed in terms of a design parameter a by the

general z-domain transfer function,


K(z) = (5-22)

The design parameter a is designed so as to reduce the total system noise and provide stable

compensation. For the initial design, a was chosen to be 2 using [119] and verified via simulation

and not via analytical methods. The settling time of the compensator is determined by the RC

time constant of the compensator capacitance and the on-resistance of the switches.

The actual transistor level implementation of the compensator is shown in Figure 5-6. The

circuit consists of only switches and capacitors, and the transfer function is determined by the

principle of charge sharing. Since this circuit does not have any amplifiers, the settling time is

considerably reduced. It consists of three capacitors C1, C2, and C3 connected to switches that









are driven by the clock signals generated from the Sampl, Eval, and Chop signals. Capacitor C1

always holds the current information while capacitors C2 and C3 hold the information from the

previous clock cycle alternatively. The Chop signal which runs at a clock frequency equal to half

the system frequency is used to alternate between C2 and C3 for storing the information from the

previous cycle. The compensator has two phases of operation namely, Sample and Evaluation


Vin C
Sampi Eval


Sampl*Chop



Eval*Chop

Vcm
Sampl*Chop


Eval*Chop

Vcm


Vout


Vcm
C2 Eval*Chop



Sampl*Chop

Vcm
C3 Eval*Chop



Sampl*Chop

Vcm


Sampi

Eval
Chop


Figure 5-6. Transistor level schematic of compensator.


phase. During the Sample phase, the output of the amplifier is sampled on to capacitors C1 and

one of the two capacitors C2 and C3 as determined by the chop signal. The derivation of the

transfer function of the circuit is done by analyzing the two half cycles of this circuit as shown in

Figure 5-7. Figure 5-7(a) shows the case when capacitor C1 and C3 are being charged with the

information from the current cycle while C2 holds the amplifier output from the previous cycle.

The charge on the three capacitors during the sampling phase at some instant of time k is given

by


(5-23)

(5-24)


Q1 = C1 [V (k) V,,] ,

Q2 C2 [v,(k 1) V,] ,












Q3 C3 [Vn(k) Vm]. (5-25)


Vin C, Vout Vi C Vout
/- Sampl Eval
Sample Eval
SVcm Vcm
Sampl*Chop C2 Eval*Chop Sampl*Chop C2 Eval*Chop


vaho Sampl*Cho EvalChop Sampl*Chop

Vm Vm Vm Vm
Sampl*Chop C3 Eval*Chop Sampl*Chop C3 Eval*Chop


Eval*Chop / Sa. p Eval*Chop /Sa
1Sampl*Chop Sampl*Chop

Vcm VVcm Vcm
(a) Sample phase. (b) Evaluation phase.

Figure 5-7. Compensator schematic during the two phases of operation.


During the Evaluation phase of the same cycle, shown in Figure 5-7(b), the plates of

capacitor C2 are reversed and connected to the output. Applying charge conservation, we equate

the final charge on the capacitors connected to the output node to their respective initial charges.

This can be expressed as


(Ci + C2) [Vot(k) V,,] = C, [V1,(k) V,] C2 [VI(k 1) Vm]. (5-26)


Taking the z-transform of the previous equation and rearranging the terms, we obtain,

Ci C(2-1 (71
Vot (k) C- (Vk) C'V6,. (5-27)
CI + C(2 1 + C2

From Equation 5-27, it can be seen that the transfer function of the compensator has a DC

offset. When this compensator is used in closed-loop, this offset will cause the center plate to

be displaced from the center position. This does not affect the operation of the circuit or its

performance metrics. Neglecting the offset, Equation 5-27 can be simplified as


Vo(k) 1 2 -1 z Vi(k). (5-28)
CI + C2 2









The compensator is followed by the comparator. Therefore, the term 02 is not important.

Comparing Equation 5-28 with Equation 5-22, we obtain


C,
a C2


(5-29)


5.3.3 Comparator Design

The comparator used in the AE modulator consists of a open-loop amplifier followed

by a positive feedback stage. The net feedback of the comparator is determined by the Latch

signal. The comparator has two phases of operation namely "Track" and "Latch". During the

track phase (Latch = 0), the negative feedback in the comparator is greater than the positive

feedback, and the comparator functions as an open-loop amplifier. During the latch phase

(Latch = 1), the positive feedback is greater than the negative feedback, and the comparator

output goes high or low corresponding to whether the the input signal is greater or less than the

reference voltage. In this system, the reference voltage is 2.5V. The schematic of the comparator

is shown in Figure 5-8. The latch mode time constant is given in terms of the load capacitance


Figure 5-8. Transistor level schematic of comparator


CL, transconductance gm, change in input signal AVo, and the change in the output logic signal

AVlogc by the equation [79]


CL AVo
Tcompar =- In
gm AVlogic


(5-30)









The time constant of the comparator stage is much lower than the amplifier and compensator

stages and can be neglected.

5.3.4 Bias Network Design

The currents and voltages required to set the operating points of the amplifier, comparator,

and compensator is generated via the on-chip bias generation network shown in Figure 5-11.

This network designs the bias voltages (Vbn, V;,, V4p, p) and the bias currents.





M8 M7 M11
M14 Vo, M18




M16
M1 M4 Mi M1














5.4 System Level Timing
M2 M3 --


M12





Figure 5-9. Schematic of bias network.


5.4 System Level Timing

In the previous section, the various building blocks of the A. interface technique was

explained. For the proper operation of the circuit, the various building blocks must operate in a

correct sequence. This section describes the sequence of operation of the various blocks and the

timing consideration required for the sequencing. The various time intervals discussed in this

section are related to the settling times discussed in Section 5.3.

The overall sequence of operation of the microphone system is shown in Figure 5-10.

The clock signals involved in each phase of operation are shown inside the block. The various

components of the AZ modulator, namely switched capacitor amplifier, compensator, and

comparator have two phases of operation. The two phases of operation of the amplifier are the

124










reset phase and the amplify phase and is controlled by clocks 01 and 02 respectively. The two

phases of operation of the compensator are sample and evaluate and is controlled by three clock

signals namely Sampl, Eval, and Chop. The two phases of operation of the comparator are

track and latch and is controlled by one clock signal, namely Latch. The various phases of the

building blocks of the AZ modulator and the clock signals used to control them are listed in

Table 5-1. The complete interface circuit is shown in Figure 5-11.

Table 5-1. Clock signals used in the various components of the AZ modulator.
Component Phases Clock
Amplifier Reset, Amplify 1, 2
Compensator Sample, Eval Sampl, Eval, Eval
Comparator Track, Latch Latch


Input
Pressure




Sense
(Reset, Amplify)



Feedback Compensate
(TP, BP) (Sampl, Eval, Chop)



Compare
(Track, Latch)





Digital Output

Figure 5-10. Flowchart showing the sequence of operation of the overall AZ modulation based
interface circuit.


The timing diagram begins at the reset phase of the amplifier during the sensing operation.

At instant denoted by 1, the reset phase ends and the amplifier is prepared to enter the amplify

phase. The time from 1 to 2 is the non overlap period of the switched capacitor amplifier. At the

instant 2, the amplifier comes into the amplify phase. The time from 2 to 3 is the time required

for the amplifier to settle before starting a new measurement. At instant 3, the top (TP) and

125











bottom (BP) plates voltages make a step equal in magnitude but opposite in sign to perform

the measurement. During this period, the Sampl clock is high and the compensator is in the

Sampling phase. The sampling capacitors form the load capacitance of the amplifier. The time

from 1 to 4 is the time required for the amplifier to settle, Tamp, discussed earlier.

Once the amplifier has settled and the output of the amplifier has been sampled on to the

compensator capacitors, the compensator enters the evaluation mode. The period from 4 to 5 is

the non-overlap time for this to occur. Once the amplifier enters the evaluation mode, it takes

some time to settle and this is the time from 5 to 6. This is the settling period of the compensator

and is denoted by rcompen. Once the comparator settles, the comparator needs to make a decision.

At instant 6, the comparator goes from the track mode to the latch mode. The output of the

comparator goes high or low depending on the compensator input as compared to the reference

voltage. Based on the output of the comparator, an external circuitry changes the top and bottom

plate voltages to null the diaphragm deflection.


-------- ----
C,

SComparator
Compensator

0 V S p om VE|
C +

VV
Switched Capacitor
Amplifier Sampl*Chop S Ea Latc



C,
SamplEvl SamplChop

Vm V|m
C ------------------------

Eval*Chop SamplChop

Vo Vom


Figure 5-11. Schematic of AZ modulator circuit.


The time period from 6 to 8 is the latch mode time constant, Tcompar. The period from 6 to

7 is the non overlap period during which the amplifier goes to the reset phase and prepares for

another measurement. Finally, the time from the end of the feedback phase to the time instant 1

is adjusted so that the top and bottom plate voltages during the sense phase is spent equally high

and low. The function performed by the system during the various time intervals are given in

Table 5.4.













PHI_ 1

PHI 2

Sampl I _

Eval

Latch

Chop

TP

BP F

SENSE -- FEEDBACK
S$Ts

Figure 5-12. System level timing diagram over one complete cycle.

Table 5-2. Function performed during each time interval.
Time Interval Function
ToI Time for equalization
T12 Non-overlap time of amplifier
T23 Amplifier settling before measurement
T34 Amplifier settling time
T45 Non overlap time of compensator
T56 Settling time of compensator
T67 Non-overlap time of amplifier
Trs Latching time of comparator


5.5 System Level Simulation

The system level simulation of the closed-loop system was carried out using the Simulink

tool in Matlab 7.0. Figure 5-13 shows the Simulink model of the microphone system. The

microphone was modelled using a "s-domain transfer function" block with a second order

transfer function given by


H(s) 2 + 2 f + f.) (5-31)
s2 + 2((27f.)s + (27f,)2

The input to the microphone system block is the "summing" block which computes the

difference between the force corresponding to the input pressure signal modeled using a "signal

























Figure 5-13. Simulink Model of the AY modulator based interface circuit.


generator" block and the electrostatic feedback force. This error signal forms the input to the

microphone which is represented by a "s-domain transfer function" block. The output of the

microphone block is inputted to the compensator which is modeled using the "z-domain transfer

function" block. The comparator is modeled as a "zero order hold" and "one-bit quantizer"

block. The magnitude of the feedback force is computed using the "equation" block whose inputs

are the current displacement and a predetermined feedback voltage. The sign of the force is

determined by a "2 to 1 multiplexer" block which is driven by the output of the comparator. The

output of the multiplexer forms the electrostatic feedback signal. The simulation parameters

are as listed in Table 5.5. Preliminary simulation results of the voice grade microphone are

Table 5-3. Parameters used in closed loop simulation.
Parameter Magnitude
Resonant Frequency fres 30 kHz
Damping Coefficient ( 0.15
Sampling Frequency f, 2.56Mhz
Input frequency 5kHz
Oversampling Ratio M 256
Solver Type Fixed point (ODE45)
Step Size 1/fs
Length of FFT 5120
Frequency Resolution 500Hz


presented here. The voice grade microphone was excited with a 5kHz size wave pressure source

with an amplitude corresponding to 2000Pa. The power spectral density of the output signal A.

modulated bitstream is shown in Figure 5-14. From Figure 5-14, it can be seen that the output










bitstream has a signal spike at the input frequency of 5kHz. At higher frequencies, the rising

noise shape of the output bitstream can be seen.

S100



-0 2
C 10


E
2 10-4
10



09
Q-)
/0)
0 -8
a 108

1012 104 106 108
Frequency

Figure 5-14. Power spectral density of output AE modulated bitstream.


5.6 Circuit Implementation

Circuit requirements for the implementation of AE modulation based controller, as a

function of resonant frequency and dynamic range, presented in Chapter 3 is repeated below in

Table 5-4 and Table 5-5 respectively.

Table 5-4. Frequency requirements of AE interface circuits with increasing resonant frequency.
Property Accelerometer Audio microphone Aeroacoustic
microphone
frs 1kHz[8] 20kHz 200kHz
f, 1MHz[8] 20MHz 200MHz
f, 4.5MHz 90MHz 900Mhz


Table 5-5. Feedback voltage requirements of AE interface circuits with increasing dynamic
range.
Property Audio microphone Aeroacoustic microphone
Maximum input 120dB 160dB
Feedback voltage 6V 45V


From Table 5-4 and Table 5-5 it can be seen that to implement a digital closed loop

controller for an aeroacoustic microphone the process technology must support voltages in excess

129









of 45V with a gain bandwidth of 900MHz. This is not feasible with current IC technology.

Hence, in this work, only a closed loop analog controller was implemented.

5.7 Summary

In this chapter, a design procedure for the development of a A. modulator for a capacitive

microphone based on the system level specification was developed. A system level simulation

was performed to verify the functionality of the approach. The requirements of the IC fabrication

technology to implement the controller was discussed. Based on this discussion, it was concluded

that a digital closed loop controller was not feasible with current process technology.

In the next chapter, the experimental setup and experimental results to characterize an

analog closed loop controller are presented.










CHAPTER 6
EXPERIMENTAL SETUP AND RESULTS

6.1 Introduction

In this chapter, the experimental setup and experimental results for characterizing the

capacitive microphone using an analog controller is presented. The chapter begins with a

literature review of test apparatus that can perform simultaneous acoustic, electrical and optical

measurement. This is followed by the design and characterization of the acoustic pressure

coupler. Experimental results of open loop and closed loop measurements of the microphone

are presented next. The experimental results are compared to simulation results. The simulation

results presented include modeling of circuit phenomena such as saturation of amplifiers and

thermal noise. This chapter concludes with a summary of the various results. A graphical

overview of this chapter is provided in Figure 6-1.


Introduction



Pressure coupler Experimental setup



Literature review Experimental Pressure coupler
results characterization


Design Open loop
characterization

SClosed loop
Characterization
Summary



Figure 6-1. Overview of Chapter 6.



6.2 Acoustic Pressure Coupler

The acoustic characterization of the microphone involves performing experiments to

determine the sensitivity, dynamic range, noise floor, and frequency response of the microphone.

As part of the characterization, the performance of the microphone with and without a feedback










controller is evaluated. To study the effect of feedback, a test setup that allows measurement of

the electrical output of the microphone and the input acoustic signal is required.

6.2.1 Design of Acoustic Pressure Coupler

The pressure coupler consists of a rigid-walled duct with a square cross section (8.5mm x

8.5mm). On one end, a BMS4590P compression driver is used to excite the duct using a Techron

7450 power amplifier controlled by the PULSE multi-analyzer system. A cut away section of the

pressure coupler is shown in Figure 6-2.


(a) Photograph of cutaway section of pressure coupler showing
rectangular duct with 8.5mm square cross section.


(b) Photograph of assembled pressure coupler showing PCB,
plug for reference microphone and optical port.

Figure 6-2. Photograph of acoustic pressure coupler.


On the opposite face of the duct, a Bruel and Kjaer Type 4138 microphone is mounted at

normal incidence to the acoustic waves and serves as the reference microphone. The device

under test (DUT) is mounted at grazing incidence to the acoustic waves close to the reference

microphone. The distance between the DUT and reference microphone d is 2.4mm. This









determines the bandwidth over which the pressure coupler can be used. The frequency f at

which the pressure seen by the DUT and reference microphone are 90 degrees out of phase is

related to the velocity of sound in air c by

d/4
f d (6-1)
C

An inlet and exit port is machined in the duct to allow the chamber to be filled with helium.

Up to the maximum frequency of operation, the DUT and the reference microphone are

exposed to the same pressure. This is ensured by making the distance from the DUT and the

reference microphone less than a quarter wavelength at the maximum frequency of operation.

Also to ensure that the higher order modes are sufficiently attenuated at the end of the test

section, the cross section dimensions are established such that the cut on frequency is higher

than the maximum frequency of interest. To increase the operating frequency range, the duct is

filled with helium as the isentropic speed of sound in helium is 2.7 times as that in air. Since the

distance between the DUT and the reference microphone is fixed, this will increase the maximum

frequency by approximately 2.7X. The minimum frequency of operation of the pressure coupler

is 300Hz and is determined by the speaker. The maximum frequency of operation of the pressure

coupler filled with helium is 18.3kHz and is determined by the duct dimensions. Without

helium, the pressure coupler can be operated up to a frequency of 6.7kHz.

6.3 Experimental Results

6.3.1 Characterization of Pressure Coupler

The pressure coupler was characterized by mounting a Briiel and KjerType 4138

microphone at the location where the DUT is to be mounted. The pressure coupler was excited

with a periodic random bandlimited white noise signal from 100Hz to 25.6kHz with 4 periods

using the PULSE system. Since the speaker has a built in bandpass filter with a low frequency

cut in of 300Hz, all the data plotted is shown from 300Hz. The spectrum of the microphone

output was measured using an FFT analyzer with 800 lines. A plot of the input pressure signal as

a function of frequency is shown in Figure 6-3.

The frequency response between the reference microphone and the second Bruel and

Kjaer Type 4138 microphone is shown in Figure 6-4 with air and helium as the medium in the










duct respectively. It can be seen that magnitude and phase frequency response between the

microphones is flat within 3dB from 300Hz to 5.6kHz in air and from 300Hz to 14.2kHz in

helium respectively. The frequency response shows that the two microphones are exposed to the

same acoustic field. Furthermore, it can be seen that the bandwidth of the system with helium is

approximately 2.7 times as that of the system with air.

140

135

130






115 -

110

105
102 103 104
Frequency


Figure 6-3. Plot of input pressure signal as a function of frequency.



The roll-off of the magnitude frequency response plot is because of the relative position of

the reference microphone and DUT. In the acoustic pressure coupler, the reference microphone

is placed at normal incidence to the acoustic source and the DUT is placed at grazing incidence

to the acoustic source. At low frequencies (large wavelength), both the reference microphone

and the DUT are subjected to the same acoustic pressure amplitude. With increasing frequency

(smaller wavelength), the pressure amplitude on the reference microphone remains constant

whereas the pressure amplitude on the DUT keeps on decreasing and reaches a minimum based

on the duct dimension and then increases. In the open and closed loop frequency response plots

shown later in this chapter, this effect has been mitigated by multiplying the magnitude frequency

response between the DUT and the reference microphone by the inverse of the magnitude

frequency response shown in Figure 6-4.

Once the pressure coupler was characterized, open loop and closed loop characterization

of the microphone are performed. Figure 6-5 shows a block diagram of the characterization












Air
2 -: : Helium

1.5


0.5


102 103 104 105
Frequency (Hz)

300
Air
200 Helium
0)
100
0)
a) .




100
102 103 104 105
Frequency (Hz)

Figure 6-4. Comparison of frequency response of pressure coupler with air as medium and with
helium as medium.



procedure used. The first experiment performed on the microphone is to obtain the frequency


Frequency response
with dc bias


Open loop
characterization


Compensator
characterization


Closed loop
characterization


Figure 6-5. Block diagram describing characterization procedure used.



response of the microphone with only a dc bias and a unity gain follower. This experiment is

performed to characterize the microphone and ensure that a working microphone is available

for further tests. The next set of experiments consists of characterizing the microphone with

a dc and ac bias. A synchronous modulation/demodulation interface circuit is used in these

experiments. These experiments provide the frequency response, linearity and noise floor of the

microphone in open loop configuration. To facilitate stable closed loop operation, a compensator

is required. The compensator, independent of the rest of the system, is characterized next. This is










accomplished by exciting the compensator with a swept sine signal and measuring the frequency

response. Finally, the closed loop configuration is characterized. The closed frequency response,

linearity, and noise floor is measured. The results of the experiments described are presented

next.

6.3.2 Open Loop Frequency Response of Microphone with Voltage Amplifier (dc bias)

The first test performed on the microphone is to obtain the frequency response of the

microphone with only a dc bias and a unity gain follower. For this test, the PCB consisting of

the microphone and buffer amplifier, is mounted in the acoustic pressure coupler. A Briiel and

KjerType 4138 microphone is used as the reference microphone. The microphone is excited

with a periodic random bandlimited white noise signal going from 100Hz to 25.6kHz with 4

periods using the PULSE system. The interface circuit used for this test is a voltage buffer.

102
Corrected Data
Raw Data


0
100

10
U)

102 103 104


300 .
& 200-
100-


-100

102 103 104
Frequency (Hz)

Figure 6-6. Frequency response with dc bias (5V) with air medium.


The top and bottom plate of the microphone is biased with a regulated 5V dc bias. Since the

speaker has a built in bandpass filter with a low frequency cut in of 300Hz, all the data plotted is

shown from 300Hz. The spectrum of the DUT and reference microphone output was measured

using an FFT analyzer with a 4Hz bin width and 1000 averages and is shown in Figure 6-6. The

roll-off of the microphone magnitude frequency response is due to the relative placement of the

reference microphone and DUT which was discussed earlier in this chapter in Section 6.3.1. The

frequency response corrected for the placement of DUT and reference microphone is also shown.

136














o
c 10

0)
I,


N


100


(n
0
-100
_ -100


103
Frequency (Hz)


Figure 6-7. Frequency response with dc bias (5V) with helium medium.


The magnitude and phase frequency response is shown in Figure 6-6. It can be seen that the

microphone has a flat frequency response from 300Hz to 5.6kHz with a sensitivity of 3.8uV/Pa

at 1kHz. Over this frequency range the phase varies from 39 deg to 9 deg.The acoustic pressure

coupler is then filled with helium and the experiment is repeated. The magnitude and phase

frequency response of the microphone in helium medium is shown in Figure 6-7. The system has

a flat frequency response from 300Hz to 13.4kHz and a sensitivity of 3.9uV/Pa.

6.3.3 Open Loop Frequency Response of Microphone with Synchronous Modulator /
Demodulator (dc + ac bias)

In this section, the open loop characterization of the microphone is presented. The block

diagram of the open loop system described in Chapter 2 is repeated in Figure 6-8. The top


Vdc+ V, sin(wt)


SMultiplier Lowpass Filter A
Charge Amp (Demod)


-Vdc-Vac sin(wct)

Figure 6-8. Schematic of capacitive microphone with a synchronous modulation and
demodulation technique using a voltage amplifier based circuit.












and bottom plate 1MHz sine wave voltages are generated using two Agilent 33014C function

generator that are 180 deg out of phase. Two different ac voltage amplitudes are used, 6Vpp and

9Vpp. The reference input to the demodulator is obtained from another Agilent 33014C function

generator that is phase locked to the top plate voltage with a 5Vpp signal. The test setup used is

similar to the previous experiment.


102L


0-






102

200

S100
0)
10

-100-


-200
102


Figure 6-9.


Tr. r i.: I


Tr, ,:,r^l,,: fll ,i- ,
Tr, :,r~,,: ll, G.


103
Frequency (Hz)


Open loop frequency response measured at the output of low pass filter (Node A,
Figure 6-8) with dc (5V) and ac bias voltages (6Vpp and 9Vpp and demodulator at
5Vpp) after compensating for magnitude frequency response slope introduced by the
acoustic pressure coupler.


IU
0-


10
0)

o 100
102

200

100

0-

1 -100

-200
102


--9V
--pp


103
Frequency (Hz)


Figure 6-10.


Open loop frequency response normalized to the peak ac voltage measured at the
output of low pass filter (Node A, Figure 6-8) with dc (5V) and ac bias voltages
(6Vpp and 9Vpp and demodulator at 5Vpp) after compensating for magnitude
frequency response slope introduced by the acoustic pressure coupler.


_ _










3500
---9V PP
3000_ ,p

2500

S2000

5 1500 ,
0
1000

500


0 50 100 150 200 250 300 350
Input pressure (Pa)

Figure 6-11. Open loop linearity of the open loop system as measured at the output of the
demodulator (Node A,Figure 6-8 ) for two different ac bias voltage at an input
frequency of 2kHz. The straight line fit has a regression value, R2, of 0.98


A dc bias of 5V is used. The microphone is excited with a periodic random bandlimited

white noise signal going from 100Hz to 6.4kHz with 4 periods using the PULSE system. The

spectrum of the DUT and reference microphone output was measured using an FFT analyzer

with a 4Hz bin width and 1000 averages. The open loop frequency response of the system after

accounting for the acoustic pressure coupler is shown in Figure 6-9 along with the theoretical

sensitivity. At 2kHz, the microphone has a sensitivity of 7.3uV/Pa for a 6Vpp ac bias and

10.47uV/Pa for a 9Vpp ac bias. Table 6-1 shows a comparison between the theoretical and

simulated values for the sensitivity. A plot of the sensitivity of the system normalized to the peak

ac bias voltage is shown in Figure 6-10. It can be seen that both the curves coincide with each

other. This shows that the sensitivity is linearly related to the peak ac voltage.

Table 6-1. Comparison of simulated and experimental sensitivity.
Biasing Bias Voltages Simulation Experimental
Technique
dc dc voltage : 5V 4tV/Pa 3.9ptV/Pa
dc + ac dc voltage : 5V, Top and bottom plate ac 7tIV/Pa 7.3[tV/Pa
voltage 6Vpp, Demodulator : 5Vpp



6.3.4 Linearity of Open Loop Microphone

The linear range of the open loop system with ac bias is presented next. The microphone is

excited with a 2kHz acoustic wave of increasing amplitude. A plot of the rms output voltage as a












PP 1o
350 -- Vp

300

250

200

S150
0
100

50

0 50 100 150 200 250 300 350
Input pressure (Pa)

Figure 6-12. Linearity of the open loop system normalized to peak ac bias voltage as measured at
the output of the demodulator (Node A,Figure 6-8 ) for two different ac bias voltage
at an input frequency of 2kHz. The straight line fit has a regression value, R2, of
0.98


function of increasing input pressure is shown in Figure 6-11. It can be seen that the microphone

has a linear response up to 315Pa(143dB) for 6Vpp and 9Vpp ac bias. The maximum applied

pressure was stopped at 143dB to prevent possible damage to the device. The slope of the line

in the linearity plot is a measure of the sensitivity of the system at 2kHz. The microphone

itself has been shown to be linear up to 166dB [64]. The circuit operates with a 15 power

supply and hence will be linear up to this pressure input. The slope of the input pressure versus

output voltage curve is 7.29uV/Pa for a 6Vp ac bias and 10.3uV/Pa for a 9Vp ac bias. This

corresponds to the sensitivity of the microphone at 2kHz discussed in Section 6.3.3. A plot of

the open loop linearity of the microphone normalized to the peak ac bias voltage is shown in

Figure 6-12. It can be seen that the curves for ac bias voltages of 6Vpp and 9Vpp approximately

coincide with each other.

6.3.5 Noise Floor of Open Loop Microphone

The noise floor of the system is obtained by measuring the output for no acoustic input.

In this experiment, the input to the amplifier driving the speaker was disconnected. A plot

of the power spectral density at the output of the demodulator for two different ac bias

voltages (6Vpp and 9Vpp is shown in Figure 6-13. From Figure 6-13, the noise floor at 2kHz

is 1.76uV. Dividing the noise floor with the sensitivity, the minimum detectable signal (MDS) is











9V
6V
-- -Theory



0
> 101
'0
E -------------- --------------





100
101 102 103 104
Frequency (Hz)

Figure 6-13. Noise floor of microphone for two different ac bias voltages measured at the output
of the demodulator.


obtained. The MDS at 2kHz for 6Vpp and 9Vp ac bias is 0.24Pa(81.6dB) and 0.17Pa(78.5dB)

respectively. The theoretical noise floor derived in Chapter 4 is also superimposed on the plot.

It is seen that the experimental noise is lower than the theoretical noise. One possible reason for

this could because the amplifier noise specification used are the typical values. The actual noise

values could be lower than the typical values provided in the data sheet.

6.3.6 Compensator Characterization

For stable closed loop operation, a compensator is required to provide adequate gain and

phase margin. In this section, the results of the compensator characterization independent of

the system is presented. For this experiment, a luV swept sine signal was used as the input

to the compensator circuitry. The output of the compensator was measured, and a magnitude

and phase frequency response plot of the compensator is shown in Figure 6-14. A plot of the

theoretical compensator response discussed in Chapter 4 is overlaid on the experimental plot in

Figure 6-14. From Figure 6-14, it can be seen that the compensator has a dc gain of 77.8dB. It

can also be seen that the compensator has a dominant pole at 1kHz and a non-dominant pole at

30kHz. The zero of the system is at 230kHz and is not captured in Figure 6-14. This is because

the spectrum analyzer (SRS785) has a bandwidth of 101.3kHz. It can be observed from the

phase frequency plot that the phase does not continue reducing after the second pole but starts to










increase indicating the presence of a high frequency pole. It can be seen that the location of the

pole and zero positions are the same between simulation and experiment.

100
Theory
E 50 Experiment

0)
-50
-100
100 102 104 106 108


100
0)

-100

100 102 104 106 108
Frequency (Hz)

Figure 6-14. Theoretical and experimental frequency response of compensator.


6.3.7 Closed Loop Frequency Response

In this section, the results of closed loop operation of the system is presented. The block

diagram of the closed loop system described in Chapter 4 is repeated in Figure 6-15. From


A) (B)
rea ra foce Intface Comp n
Pressure I N nteface I.pu Compen !
force rce circuit sator
Sensor Displacement
(Force to to voltage
displacement)


Electrostatic Voltage to
force Force

Figure 6-15. Block diagram of analog closed-loop system showing microphone and interface
circuit.


linear control theory [126], during closed loop operation, if the forward path of the system has

a high gain, the signal at any point in the forward path of a closed loop system is attenuated by

the loop gain. Thus, to demonstrate closed loop operation, the frequency response of the system

measured at the output of the interface circuit (Node A in Figure 6-15) is compared to the open

loop response shown in Figure 6-9. The frequency response of the closed loop system measured

142











at the output of the interface circuit (Node A in Figure 6-15) is shown in Figure 6-16 along with

the theoretical curve for ac bias voltages of 6Vp and 9Vpp. This frequency response is compared

to the frequency response of the open loop system shown in Figure 6-17 and Figure 6-18 for 6Vp

and 9Vpp ac bias. The magnitude frequency response at 2kHz is 1.27uV/Pa with a 6Vpp ac bias

102
9V
S 01 Theoretlcal 6Vp .
10
s Theoretical Vpp ..o 9V? : .

10 10T' 104






10


-100 ,,

-150 -
-200
10 10 10
Frequency

Figure 6-16. Closed loop frequency response measured at the output of demodulator with dc
bias of 5V, ac bias of 9Vp, and demodulator at 5Vp (Node A in Figure 6-15).
Frequency response corrected for reference microphone and DUT location.



voltage and 1.89uV/Pa with a 9Vp ac bias voltage. An increase in phase was observed in closed

loop response at the output of the demodulator which was not observed in open loop. The cause

of this phase increase is not clear but may be due to the coupling capacitors used between the

stages. Compared with the open loop magnitude frequency response measured at 2kHz, with the

same bias voltage, this represents a 5 x reduction in diaphragm displacement. Thus the system

operates in a stable fashion in closed loop. The experimental loop gain of the system, T evaluated

at 2kHz, is given by the product of the system sensitivity, H1, expressed in uV/N evaluated

at 2kHz, compensator gain H2, evaluated at 2kHz, and voltage to force transfer function, H3,

evaluated at 2kHz i.e.

T = H1 x H2 x H3 (6-2)


It should be noted that in this work, the voltage to force transfer function was not experimentally

evaluated. Therefore, H3 is calculated using Equation 4-10. Also, the sensitivity of the system

in uV/Pa obtained from Figure 6-9 is converted to uV/N by dividing the sensitivity by the











Open Loop
S-- Closed Loop


E 100
-10 :
10

102 103 104

200

100
0-


a- -100-

-200 2 3 4
10 10 10
Frequency

Figure 6-17. Closed loop frequency response measured at the output of demodulator (Node A in
Figure 6-15) with dc bias of 5V, ac bias of 6Vpp, and demodulator at 5Vpp in open
and closed loop operation. Frequency response corrected for reference microphone
and DUT location.


effective area of the diaphragm. Using Equation 6-2, the experimental loop gain of the system

is 5.28. From linear control theory, the diaphragm displacement is attenuated by 1 + T. This

implies that the diaphragm displacement should be attenuated by 6.28 x. The actual diaphragm

displacement is only 5 x. This discrepancy can be attributed to the modeling errors used in

the voltage to force transfer function. Figure 6-17 and Figure 6-18 show a comparison of the

magnitude and phase frequency response in open and closed loop operation with an ac bias of

6Vpp and 9Vpp respectively compensated for the reference microphone and DUT location.

The frequency response of the closed loop system at the output of the compensator (Node

B, Figure 6-15) after compensating for the transfer function of the pressure coupler is shown

in Figure 6-19. The magnitude frequency response at 2kHz is 2.63mV/Pa with a 6Vpp ac bias

voltage and 3.56mV/Pa with a 9Vpp ac bias voltage.

6.3.8 Closed Loop Linearity

A plot of the closed loop linearity measured at the output of the compensator (Node B,

Figure 6-15) for 6Vpp and 9Vpp ac bias voltages ac bias voltages is shown in Figure 6-20. It can

be seen that the system has a linear response upto an input pressure of 315Pa(144dB). Further











Open Loop
> Closed Loop
10' Lop


100

10-

102 103 104

200

100
0-
o, 0

a- -100-

-2002 3 -4
10 10 10
Frequency

Figure 6-18. Closed loop frequency response measured at the output of demodulator (Node A in
Figure 6-15) with dc bias of 5V, ac bias of 9Vpp, and demodulator at 5Vpp in open
and closed loop operation. Frequency response corrected for reference microphone
and DUT location.


increase in input pressure was not carried out to prevent possible damage to the microphone

diaphragm. The slope of the input pressure versus output voltage curve is 2.63mV/Pa for

a 6Vpp ac bias and 3.56mV/Pa for a 9Vpp ac bias. This corresponds to the sensitivity of the

microphone at 2kHz in Figure 6-19. A plot of the open loop linearity of the microphone

normalized to the peak ac bias voltage is shown in Figure 6-21. It can be seen the plots for for ac

bias voltages of 6Vpp and 9Vpp approximately coincide with each other.

6.3.9 Closed Loop Noise Floor

The noise floor of the system is obtained by measuring the output of the compensator

for no acoustic input. In this experiment, the input to the amplifier driving the speaker was

disconnected. A plot of the power spectral density at the output of the demodulator for 6Vpp and

9Vpp ac bias voltages is shown in Figure 6-22. From Figure 6-22, the noise floor at 2kHz is

1.11mV. Dividing the noise floor with the sensitivity, the minimum detectable signal (MDS) is

obtained. The MDS at 2kHz for 6Vpp and 9Vpp ac bias is 106.5dB and 103.8dB respectively.

The large noise floor is discussed in the next section.












--6V
pp
9V

100

10-1..
102 103 104

200

100
0)
0 /_ .. ._ _. ,,_--

a -100-

-200
102 103 104
Frequency

Figure 6-19. Closed loop frequency response measured at the output of compensator (Node
B in Figure 6-15) with dc bias of 5V ac bias of 9Vpp, and demodulator at 5Vpp.
Frequency response corrected for reference microphone and DUT location.


1200
9V
6V
1000 -


800-


600


0 400-


200


0
0 50 100 150 200 250 300 350
Input pressure (Pa)

Figure 6-20. Linear range of system in closed loop configuration as measured at the output of
the compensator (Node B in Figure 6-15) with 6Vpp and 9Vpp ac bias voltages for an
input frequency of 2kHz. The straight line fit has a regression value, R2, of 0.97



6.4 Discussion

A closed loop analog controller for a MEMS dual backplate capacitive microphone

has been characterized in terms of frequency response, noise floor, and linear range. The

performance characteristics of the system in open and closed loop operation is listed in Table 6.4.

The reduction in the output of the demodulator is approximately 5 x which is approximately

consistent with that predicted to the reduction of loop gain.











140
9V
pp
120 6Vpp

100

E
80

S60

0
40


0

0 50 100 150 200 250 300 350
Input pressure (Pa)

Figure 6-21. Linearity of the closed loop system normalized to peak ac bias voltage as measured
at the output of the compensator (Node B in Figure 6-15 ) for two different ac bias
voltage at an input frequency of 2kHz. The straight line fit has a regression value,
R2, of 0.98




10 pp




104



103




102
10 102 103 104
Frequency (Hz)

Figure 6-22. Noise floor of the circuit in closed loop operation as measured at the output of the
compensator (Node B in Figure 6-15) with 6Vpp and 9Vpp ac bias voltages.


The performance metrics of the open loop microphone using synchronous modulation

and demodulation interface circuit is compared to the performance metrics of the microphone

using a low noise voltage amplifier and a charge amplifier [64]. Martin reports a sensitivity

of 166uV/Pa and:1 .; iV/Pa with a 9.3V bias using a voltage amplifier and charge amplifier

respectively. The reported interface circuits have a noise floor of 22.7dB and 41.5dB with a

voltage and charge amplifier respectively. The linear range of the interface circuit extends up

to an input pressure of 166dB. Based on this information, the dynamic range reported in this









Table 6-2. Characterization results of microphone in open and closed loop mode of operation.
Bias condition Sensitivity at Noise floor Linear range
2kHz
Open loop (ac 6Vpp) 7.3uV/Pa 81.6 81.6dB- 143dB
Open loop (ac 9Vpp) 10.47uV/Pa 78.5 78.5dB- 143dB
Closed loop (ac 6Vpp) 2.63mV/Pa 106.5 106.5dB -
143dB
Closed loop (ac 9Vpp) 3.56mV/Pa 103.8 103.8dB -
143dB


work is considerably lower. The reasons are as follows. In this work, the amplifiers are chosen

to have a large bandwidth to be used in synchronous modulation and demodulation and a high

operating voltage and slew rate to support the feedback voltage requirements whereas in the work

published by Martin only open loop operation was investigated. Amplifiers which simultaneously

satisfy the high voltage, bandwidth and slew rate conditions have higher noise than amplifiers

with lower bandwidth and lower power supply voltages. This is because the transistors used to

implement the amplifiers must have larger size to support the increased current required for the

high frequency operation and the high voltage requirement. The closed loop interface circuit

also has increased component count due to the need for a high gain compensation network.

All these factors increase the noise floor of the closed loop microphone system. In closed loop

accelerometers, the bandwidth of the sensor is lower (approximately 1kHz). This reduces the

bandwidth requirement of the amplifiers used to implement closed loop control. Hence closed

loop accelerometers can potentially be implemented with a lower noise floor.

Closed loop capacitive interface circuits have primarily been used for low bandwidth

(1 2kHz) application such as accelerometers [68, 67, 11, 2, 112]. The main advantage

of a closed loop interface circuit is the possibility of increased bandwidth and linear range.

Also closed loop interface circuits have the potential to address pull in issues as the diaphragm

deflection is reduced. In this work, a proof of concept closed loop interface circuit has been

presented for a comparatively higher bandwidth ( 20kHz) application compared to potential

closed loop accelerometer applications. Closed loop control was demonstrated by showing that

the diaphragm deflection was reduced during closed loop operation. This was verified for two

different voltage bias conditions. The disadvantages of closed loop interface circuits is a higher









noise floor because of the increased number of components needed and the requirement of a high

gain stage due to the small voltage-to-force factor.

6.5 Summary

In this section, the experimental results were presented. The design and characterization of

the test setup was first presented. This was followed by open and closed loop characterization

of the microphone. A comparison of the interface circuit performance was compared to other

interface circuits for the same device and with interface circuits for other capacitive transducers.









CHAPTER 7
CONCLUSION AND FUTURE WORK

In this chapter, the key goals of the work and the results are summarized. Key contributions

of this research are highlighted. This chapter concludes with a discussion on the avenues for

further research in this area.

7.1 Conclusion

The goal of this work is to design and characterize interface circuits that are optimized

for miniature capacitive transducers. As transducer geometry is scaled down, improved circuit

techniques are required to measure the sensor output. This is because at small geometries sensor

capacitance is comparable to unwanted parasitic capacitance which reduces the transducer

sensitivity. Also, at reduced sensor geometry, the voltage required to bias the microphone could

cause the microphone plates to pull in.

In this work, an extensive study of open loop and closed interface techniques for capacitive

sensors has been presented along with theoretical formulations for key performance metrics.

An analog closed loop controller for a dual backplate MEMS capacitive microphone was

designed. A detailed MATLAB model which captures key dynamics of the sensor and controller

was developed. Noise models of the closed loop system was also developed. Open and closed

loop simulations were performed for step and sinusoidal excitations. From these simulations, it

can be seen that in closed loop operation, the diaphragm displacement is attenuated which can

potentially address pull in issues.

A proof of concept closed-loop analog controller for a MEMS capacitive microphone was

developed and characterized. A test apparatus was developed to characterize the system over the

audio range by operating the microphone in a helium medium which increases the bandwidth of

the test apparatus. Characterization of the microphone in open and closed loop mode of operation

is presented. Results show that stable closed loop operation of the microphone is feasible with

increased sensitivity and the potential to address pull-in issues.

7.2 Future Work

The future areas of research in this project can be partitioned into simulation related and

hardware implementation related future work. These two topics are discussed next.









7.2.1 Simulation

In the open and closed loop Matlab simulations presented in this work, a 2nd order linear

model of the microphone was used. This can be expanded to include a detailed non-linear model

of the microphone which incorporates pull-in effects and mechanical non-linearities of the

microphone.

Another area of research is detailed transistor level simulations using SPICE. Transistor

level simulations allows noise analysis of the system which is not possible with MATLAB

simulations. To successfully perform these simulations, the microphone must also be modeled in

SPICE using analog modeling blocks. Furthermore noise models for the circuit elements must be

available.

7.2.2 Hardware Implementation

A number of improvements can be suggested for the next version of the closed loop

controller. In the current work, a two PCB approach wherein the microphone and buffer amplifier

is mounted on one PCB and the controller is mounted on a different PCB and connected using

ribbon connectors is used. This increased the 60Hz pick-up of the circuit. In future versions

of the controller, a single PCB which incorporates both the microphone and the controller can

be used. In the current work, dual in-line package (DIP) components were used to allow easy

testing and debugging. The downside of using DIP components is increased parasitics and noise.

Future versions of the controller may use surface mount components in order to reduce parasitic

capacitance and electromagnetic coupling.

It was also noticed that 60Hz noise in both the input channels of the demodulator caused

120Hz noise signals at the output of the demodulator. One of the sources of 60Hz pick up is the

cables connecting the function generator output to the controller PCB. This source of noise pick

up can be reduced by generating the clock signal on the PCB itself using oscillator ICs.

Another avenue for research could be to monolithically integrate the interface circuit and

controller using a high voltage, high frequency, and low noise process. This will greatly improve

circuit performance because of the reduced parasitics which will improve the sensitivity. Also

monolithic integration will eliminate the need for cables that are used for clock generation as the

clock signal can be generated on chip.









APPENDIX A

A.1 Derivation of Sensitivity for a Voltage Follower

In this technique the top and bottom backplate of the microphone with top and bottom

capacitances given by C1 and C2 are biased at a dc potential of +Vdc and -Vdc volts. A unity

gain voltage amplifier is connected to the middle plate as shown in Figure A-1 This results

+Vdc


C /
V7
S+ V'ut


> C

-Vdc 0 1

Figure A-1. Noise model of dual backplate microphone connected to voltage buffer.


in fixed charge of +Qda and -Qdc on the top and bottom capacitor respectively. C, and Ci

represent the interconnect parasitic capacitance and the input capacitance of the amplifier

respectively. The top and bottom plate capacitance is given by


(A-l)


coA
xo X
coA
oC2 +
Xo +x'


The nominal capacitance Co is given by the expression


coA


(A-2)


Equation A-2 can be expressed in terms of the nominal capacitance Co and a varying capacitance

AC using the relation


Co + ACi

Co AC2


(A-3)










Using Equation A-1 the voltage across the top and bottom capacitor Va and V2 respectively is

given by


xo X'


(A-4)


xo + x'
V2
coA


This shows that the voltage varies linearly with gap distance. The voltage across each capacitor

can also be expressed in terms of the fixed charge Qdc as


(A-5)


dc
Co + AC

V2 -
Co AC

Substituting Equation A-l1, Equation A-2 in Equation A-6 we obtain


cA
x0 x.
Xo X'
cA
X0 + x'


(A-6)


X'
ACi + -

X'
xo
AC2--
Xo


Solving Equation A-6 for AC1 and AC2


ACi

AC2


x' cA
Xo X0 X'
x' cA
Xo Xo + X'


(A-7)


To obtain the output voltage we solve for the voltage at the inverting node using the

principle of superposition. Initially we consider only the output due to the top capacitance. This

implies that the bottom plate is grounded. The initial charge on the top capacitor Qli system is

given by


Qli CoVdc


(A-8)


The final charge is given by


Qlf = CIVI C2Voutc,


(A-9)


Equating Equation A-8 and Equation A-9 we obtain


CoVdc CIVI C2 VoutC1


(A-10)









Substituting for C1 as Co + AC and Vi as Vdc Voutc, in Equation A-10 we obtain


CVd = (Co + AC) V1 C2Vo c,. (A-11)

Solving for Voutc, we obtain
AC1
Voutc1 Vdc. (A-12)
C1 + C2
Similarly for Voutc, we obtain
AC2
Vo.utc Vdc. (A-13)
C1 + C2
The total output of the amplifier Vot is the sum of outputs given by Equation A-12 and

Equation A-13,
ACI + AC2
Vout = c (A-14)
C1 + C2
If the capacitance change AC1 and AC2 are equal, i.e,


ACI = AC2 = AC (A-15)

and using Equation A-1, in Equation A-14 we obtain

AC
Vout = CVd (A-16)
Co

From the input output relation given by Equation A-16, the electrical sensitivity of the circuit is

given by the equation

Se = (A-17)
Co

A.2 Noise Analysis for a Capacitive Microphone with a Voltage Follower

The complete noise model of the circuit with all the noise generators of the interface circuit

is shown in Figure A-2. The current and voltage noise spectral density (PSD) of the amplifier are

denoted by Sia and Sa respectively. The current and voltage noise PSD Si(f) and Se(f) can be

expressed as a function of the current and voltage noise corner frequency fei and fe by


Se(f) = 2(f) -= (i+ ) (A-18)

s (f) = 2f (f) _= 1+f
\i i f














C2 a
TT SVR
f ^I



Figure A-2. Noise model of dual backplate microphone connected to voltage buffer.

The power spectral density of the noise of the resistor is denoted by SR. It is given by

SR 4kTRdc. (A-19)

The impedance of the parasitic capacitances and sense capacitor denoted by Zi is given by

1
zi = (2Co +C ) (A-20)
jJ (2Co + C/ + C)

We now use superposition to calculate the output voltage noise power spectral density. We
first consider only the noise generator Sa. With only the Sa noise generator, no current flow in
Zi or Rdc. The noise at the output S,,o is given by

Sol = Sa. (A-21)

Next we consider only the noise generator Sia. It can be seen that the total impedance seen by
this noise generator is is the parallel combination of Z, and Rdc. Thus the noise generated at the
output So2 is given by
So2 = S z + Rd 2 (A-22)
Zi + Rdc
Finally we consider the noise due to the resistor. It can be seen from the circuit that noise due to
the resistor is shaped by the voltage divider consisting of the of Rdc and Zi. Hence the noise at
the output Svo3 due to the resistor is given by

Z 2
So,3 = S zR + Rd (A-23)
Zi + Rdc









Using superposition the power spectral density of the output noise voltage v-. of the circuit is

given by the sum of Equation A-21, Equation A-22, Equation A-23

ZiRdc 2 Z 2
S vo = a + Sia + Sn (A-24)
SZ + Rdac + + Z +Rdc

It should be noted that the noise analysis presented here considers only the noise sources

due to the electrical interface circuit. The thermo-mechanical sources of noise associated with the

sensor itself are not considered in this analysis.

A.3 Derivation of Sensitivity for a Charge Amplifier

The top and bottom plate capacitor is given by


C1 Co 1 + (A-25)

C2 co f-
C2 = CO 1 -X

The charge on each capacitor is given by


Qi = ClVdc (A-26)

and

Q2 -C2Vd

The net charge transferred is given by


AQ = Q1 + Q2 (A-27)

The output voltage for a charge amplifier is given by


Vot = CitAQ (A-28)

Substituting for Qi and Q2 in Equation A-28 and simplifying


V.t Vdc] AC. (A-29)

From the input-output relationship given by Equation A-29 the electrical sensitivity of the circuit

is given by the equation
2Va,
Se = (A-30)
156
156









A.4 Noise Analysis for a Capacitive Microphone with a Charge Amplifier

The complete noise model of the circuit with all the noise generators is shown in

Figure A-3. The current and voltage noise power spectral density of the amplifier are denoted by

S,, and Sia respectively. The power spectral density of the current noise of the resistor is denoted

by SiR. It is given by


1
S,R = 4kT Af.
Rdc


(A-31)


The impedance of the parasitic capacitances and sense capacitor which is denoted by Zi is given
by ma


(A-32)


-1 ja (2Co + C + C) *
The impedance of the feedback network denoted by Zf is given by

1
Zf =- 1II
j3^ Cint


now use superposition to calculate to the output voltage noise power spectral density. We first

SiR


RDC







C2 CL
-- _




Figure A-3. Noise model of dual backplate microphone connected to charge amplifier.


consider only the noise generator Sa. The noise at the output S,,o due to S,, noise generator is

given by


( 1+ )
( Zy2


(A-33)


Svol = Sva


(A-34)









Next we consider only the noise generator Sia. It can be seen that the total impedance seen by

this noise generator is Zf. the noise at the output So2 due to Sia noise generator is given by


So2 SaZ. (A-35)

Similarly the output voltage noise due to the resistor current noise, Svo3, is given by the relation


So3 = S PaZ. (A-36)

Using superposition the power spectral density of the output noise voltage So is given by the

sum of Equation A-34, Equation A-35, and Equation A-36 and is given by the expression,


Svl = Sa (Z 2 + S,7 + SiRZ2 (A-37)

Since we are interested in the noise PSD after the cut in frequency i.e.,

1
>> CitR, (A-38)
Cint Rdc '

Simplifying

SvO = SCa [ + + (Si. + SR) wC 2 (A-39)

Equation A-39 shows that both the resistor current noise and the amplifier current noise are 1/f2

shaped.

A.5 Sensitivity for Synchronous Modulation and Demodulation Interface Circuit

In this section the linearized sensitivity of a synchronous modulation and demodulation

based interface circuit is derived. In this technique, the top plate is biased with a a square wave

with 5(' duty cycle of amplitude Vac at a frequency of w, on a dc voltage Vdc. The bottom plate

is biased with signal with the same magnitude as the top plate but 180deg out of phase.

The carrier wave input r(t) to the demodulator also consists of a square wave at a frequency

of w~ but with an amplitude of V. The fourier series representation x(t) of a square wave of

amplitude A with 5(' duty cycle is given by

A 2A 12A 12A
x(t) = + -cos(wt) + --cos(3wt) + --cos(5wt)... (A-40)
2 7 3 7r 5 7










To reduce the movement of microphone diaphragm due to the applied AC voltage, Vac is

kept small. Since we are doing a linearized analysis, the higher order harmonics are neglected.

The top plate voltage vtp is given by


Vac, 2Va
vtp(t) Vdc + 2. + Cos(ocJt).
2 Ti


(A-41)


The carrier wave applied to the demodulator is given by


2V
r(t) -+ cos (wt).
2 T7


(A-42)


The output of the buffer amplifier Vmod(t) is given by Equation 3-13 as


X'
Vmod(t) = Vt(t)
Xo


(A-43)


If the input acoustic signal is at a frequency Wm, and the maximum displacement of the

diaphragm is x' then


x'= I,..-( ).


xo


Substituting for vtp


[x' -V 2V
Vmod (t)= COs (mt) X Vdc + V + os (ct) ] (A-4


metric identity cosAcosB = 1/2(cos(A + B) cos(A b)), Vmod(t) can be


expressed as


Vmod(t) =
X(

+

+


6)


(A-47)


( Vdc + V. ) COs (Umt)

x'o 2V COS(Um + wc)t
Xo 7
X'. 2V,- )t
XO iT


The output of the buffer amplifier Vdemod(t) is given by as


Vdemod(t) = Vmod(t) X r(t).


Therefore


(A-44)


(A-45)


Using the trigonom


(A-48)









Substituting for Vmod(t) and r(t)

'/ V V
Vdemod(t) a= o0 + C- COS(mt) X + -COS(U)
xO V 2 2 7
+ cos(Wm + Wc)t x + -cos(ct)

+x 2 VS( c)t x + 2-- os(wt)
Xo 7" 2 7"
XO 7 2 7


(A-49)


(A-50)


Expanding the above equation


xo x (A-51)
XO

[Vdc + V) COS(,mt) + Vdc + COS( Umt)-- os(ct) +

2Vac 2Vc 2,
-2-COS( Wm + ct)t- + -2 COS(Wm + c)t2cos(wcJt)
7T 2 77 7
2Voc K 2Vc 2V,
- COS(Wm cLU)t + COS(Wm c)t-) cos(wcJt)
iT 2 7 7


Using trigonometric identity Vdemod(t) can be expressed as


Vdemod(t) X 0 x (A-52)
xo
[(Vdc + V,) V] cos(Ut) + (VdC + ) 2 C OW+c t

+c t 2 2V [c 2X [COS(WLm)t + COS m-
w, t, r +@ 2V_ 2Vcl h [cos0W()t t COS( m-)- ;
.....r-----1--- CO L t2
...._ar7T -7 r r


The low pass filtered signal Vout(t) can be expressed as


Vot(t) = o x
xo
[(Vdc + V') ] COS(Wmt)
2V_ 2"c 1I ( )
77 2 [cos( -t]

77 2 [cos( -t]


Vdemod (t)


(A-53)









which reduces to


V., (t) x
xo


I(c + V-)
(c2


V +
cos(Wmt) +
2


Substituting for ,.- (uWm)t as x' we obtain


vout(t)
xo


The linearized sensitivity of this interface technique is


S X(t) 1
x' xo 2


+ 2V,, 2V,
+ 7


(A-56)


2KV 2Vc (
T T


(A-54)


X >(Vd, +


Vac
2



V,
2


2 K 2V,
xx ~


(A-55)









APPENDIX B

B.1 Bandwidth Considerations for Amplifiers in Switched Capacitor Amplifiers

This derivation is based on the work presented in Wongkomet's doctoral dissertation [108].

The time constant 7 of a single pole amplifier is related to its closed loop bandwidth f, by the

relation
1
f (B-l)
27tT

Let the number of time constants required for the amplifier to settle be denoted by n,, sampling

frequency by f, and the fraction of time spent in the amplify phase by m(< 1). The time
1
available for the switched capacitor amplifier to settle is thus m x The time required for the

amplifier to settle within a given error criteria is n, x r. For example, for 0.1 settling, n, must

be equal to seven. Equating the time required and time available for settling, we obtain

1
m x 1 = nr x T (B-2)


Substituting for 7 and rearranging the equation, we obtain
[ -1

S1 1 (B-3)
27 nfs

B.2 Analysis of Noise in a Switched Capacitor Amplifier

This derivation is based on the work presented in Wongkomet's doctoral dissertation [108].

The noise model of the amplifier during the amplification phase is shown in Figure B-1

V+ /

Crest
C,=C1 4kTF


vC out







Figure B-1. Noise model of switched capacitor amplifier during amplify phase.









The noise of the amplifier is modeled as a wide band white noise vT at the inverting terminal

of the amplifier. The total capacitance at the input of the amplifier is given by


Cr = 2Co + Cp + Cgain (B-4)


We shall now derive the total noise at the output due to kT/C noise and due to the wideband

amplifier noise.

B.2.1 Switch Resistance Noise

In this section the total output noise due to the on resistance of the switches is derived.

When the switches are closed, the noise due to on resistance of the switches is low pass filtered

by the total capacitance CT
kT
v2 (B-5)
VkT/Cinput T-

Referring this noise to the output by multiplying the previous equation by the square of the gain

we obtain the total output kT/C noise The total noise at the output of the amplifier is given by

2 Cr 12 2kT
VkT/C C (B-6)


B.2.2 Wideband Amplifier Noise

In this section the total output noise due to the wideband amplifier noise is presented. The

noise at the input of the amplifier is

-2 -2 (B-7)
Vopampinput n (7)


The total output noise voltage is


Vaonmp j [ ndf. (B-8)
0 Cgain

For a single pole system, the equivalent noise bandwidth is 7r/2 times the bandwidth[79].

Therefore the total output noise for an amplifier with a closed loop bandwidth of fu


v opamp C 1 2 U2 (B-9)
op p gain 2 2









REFERENCES


[1] Marc Madou, Fundamentals of Microfabrication, CRC Press, Boca Raton, 1997.

[2] Analog Devices, "Data sheet adxl50, monolithic accelerometer with signal conditioning,"
Tech. Rep., One Technology Way, Norwood, MA, 1993.

[3] Texas Instruments, "Digital light processing," World Wide Web, http: / /www. dip.
com/index.html, 1996.

[4] Freescale Semiconductors, "Data sheet mpxy8040a : Tire pressure monitor sensor," Tech.
Rep., Phoenix, AZ, 1997.

[5] Knowles Acoustics, "Sisonic microphones data sheet," Tech. Rep., 1111 Maple Drive,
Ithasca, IL 60143, 2002.

[6] Weijei Yun, Roget T. Howe, and Paul R. Gray, "Surface micromachined, digitally force
balanced accelerometer," in Solid-State Sensor and Actuator Workshop. 1992, pp. 126 -
131, IEEE.

[7] R.P. van Kampen, M.J.Vellekoop, P.M.Sarro, and R.F.Wolffenbuttel, "Application
of electrostatic feedback to critical damping of an integrated silicon capacitive
accelerometer," Sensors andActuatorA, vol. 43, pp. 100-106, 1994.

[8] Mark Lemkin and Bernhard E. Boser, "A three-axis micromachined accelerometer with a
cmos position-sense interface and digital offset-trim electronics," IEEE Journal of Solid
State Circuits, vol. 34, no. 4, pp. 456-468, April 1999.

[9] Y. Matsumoto, M. Nishimura, M. Matsura, and M. Ishida, "Three-axis soi capacitive
accelerometer with pll cv converter," Sensors and Actuator A, vol. 75, no. 1, pp. 77-85,
May 1999.

[10] Tetsuya Kajita, Un-Ku Moon, and Gabor C. Temes, "A noise shaping accelerometer
interface circuit for two chip implementation," in IEEE Instrument and Measurement
T, lhmrl,, Cti, r, t c. Budapest, Hungary. IEEE, 2001, pp. 1581-1585, IEEE.

[11] Bernhard E. Boser and Roger T. Howe, "Surface micromachined accelerometers," in
Custom Integrated Circuits Cif, rT, c ,. 1995, pp. 337-344, IEEE.

[12] J. M. Huang, A. Q. Liu, Z. L. Deng, J. Ahn Q. X. Zhang, and A. Asundi, "An approach to
the coupling effect between torsion and bending for electrostatic torsional micromirrors,"
Sensors and Actuator A, vol. 115, no. 1, pp. 159-167, September 2004.

[13] John Bergqvist, Rudolf F, Masiano J, and Rossi M, "A silicon condenser microphone with
a highly perforated backplate," 1991.

[14] W.H. Hsieh, Tseng-Yang Hsu, and Yu-Chong Tai, "A micromachined thin-film teflon
electret microphone," in International COif, rr,h c on Solid State Sensors and Actuators,
TRANSDUCERS '97, Chicago, 1997, vol. 1, pp. 425-428.

[15] W. Kronast, B. Muller, W. Siedel, and A. Stoffel, "Single-chip condenser microphone
using porous silicon as sacrificial layer for the air gap," in The El, i', '/h Annual
International Workshop on Micro Electro Mechanical Systems, MEMS 98, Heidelberg,
Germany, 1998, pp. 591-596, IEEE.









[16] Siebe Bouwstra, Torben Storgaard, Patrick Scheeper, Jens Ole Gullov, Jesper Bay,
Matthias Mtillenborg, and Pirmin Rombach, "Silicon microphones a Danish
perspective," Journal of Micromechanical Microengineering, vol. 8, pp. 64-68, 1998.

[17] Altti Torkkelia, Outi Rusanenb, Jaakko Saarilahtia, Heikki Seppc, Hannu Sipolac, and
Jarmo Hietanen, "Capacitive microphone with low-stress polysilicon membrane and
high-stress polysilicon backplate," Sensors and Actuator A, vol. 85, pp. 116-123, 2000.

[18] Mullenborn M., Rombach P., Klein U., Rasmussen K., Kuhnmann J.F., Heschel M.,
Amskov Gravad M., Janting J., Branebjerg J., Hoogerwerf A.C., and Bouwstra S.,
"Chip-size packaged silicon microphone," Sensors and Actuator A, vol. 92, pp. 23-29,
2001.

[19] Rombach P., Mii llenbom, Klein U., and Rasmussen K., "The first low voltage, low noise
differential condenser silicon microphone, technology development and measurement
results," Sensors and Actuator A, vol. 95, pp. 196-201, 2002.

[20] Jing Chen, Litian Liu, Zhijian Li, Zhimin Tan, Yang Xu, and Jun Ma, "Single-chip
condenser miniature microphone with a high sensitive circular corrugated diaphragm," in
The Fifteenth IEEE International C 'nf, rr, on Micro Electro Mechanical Systems, 2002,
pp. 284 287.

[21] Borge Nordstrand Patrick Scheeper, Jens Gullev, Bin Liu, Thomas Clausen, Lise Midjord,
and Torben Storgaard-Larsen, "A new measurement microphone based on mems
technology," IEEE Journal of Microelectromechanical Systems, vol. 12, no. 6, pp.
880-891, December 2003.

[22] S. T. Moe, K. Schjlberg-Henriksen, D. T. Wang, E. Lund, J. Nysth, L. Furuberg, M. Visser,
T. Fallet, and R. W. Bernstein, "Capacitive differential pressure sensor for harsh
environments," Sensors and Actuator A, vol. 83, no. 1-3, pp. 30-33, May 2000.

[23] Shuwen Guo, Jun Guo, and Wen H. Ko, "A monolithically integrated surface
micromachined touch mode capacitive pressure sensor," Sensors andActuatorA, vol.
80, no. 3, pp. 224-232, March 2003.

[24] Klaus Kasten, Norbert Kordas, Holger Kappert, and Wilfried Mokwa, "Capacitive
pressure sensor with monolithically integrated cmos readout circuit for high temperature
applications," Sensors andActuatorA, vol. 97-98, no. 1, pp. 83-87, April 2002.

[25] Min-Xin Zhou, Qing-An Huang, and Ming Qin, "Modeling, design and fabrication of a
triple-layered capacitive pressure sensor," Sensors andActuatorA, vol. 117, no. 1, pp.
71-81, January 2005.

[26] Y C. Loke and K. M. Liew, "VHF enhancement of micromechanical resonator via
structural modification," Sensors and Actuator A, vol. 96, no. 1, January 2002.

[27] F. F. Faheem and Y C. Lee, "Tether- and post-enabled flip-chip assembly for
manufacturable rf-mems," Sensors andActuatorA, vol. 114, no. 2-3, pp. 486-495,
September 2004.

[28] R. Fritschi, S. Frdrico, "High tuning range capacitors fabricated with sacrificial amorphous
silicon surface micromachining," Microelectronics Engineering, vol. 73-74, pp. 447-451,
June 2004.









[29] Siavash Pourkamali and Farrokh Ayazi, "Electrically coupled MEMS bandpass filters:
Part i: With coupling element," Sensors and Actuator A, vol. 122, no. 2, pp. 307-316,
August 2005.

[30] Siavash Pourkamali and Farrokh Ayazi, "Electrically coupled MEMS bandpass filters:
Part ii: Without coupling element," Sensors and Actuator A, vol. 122, no. 2, pp. 317-325,
August 2005.

[31] Motoaki Hara, Jan Kuypers, Takashi Abe, and Masayoshi Esashi, "Surface micromachined
A1N thin film 2 ghz resonator for cmos integration," Sensors and Actuator A, vol. 117, no.
2, pp. 211-216, January 2005.

[32] Bin Xiong, Lufeng Che, and Yuelin Wang, "A novel bulk micromachined gyroscope with
slots structure working at atmosphere," Sensors and Actuator A, vol. 7, no. 2, pp. 137-145,
October 2003.

[33] W. Geiger, T. Link, J. Merz, S. Steigmajer, A. Hauser, H. Sandmaier, W. Lang, and
N. Niklasch, "New digital readout electronics for capacitive sensors by the example of
micro-machined gyroscopes," Sensors andActuatorA, vol. 97-98, pp. 557-562, April
2002.

[34] Huikai Xie and Gary K. Fedder, "Vertical comb-finger capacitive actuation and sensing for
CMOS-MEMS," Sensors andActuatorA, vol. 95, no. 2-3, pp. 212-221, January 2002.

[35] Farrokh Ayazi and Khalil Najafi, "High aspect-ratio polysilicon micromachining
technology," Sensors and Actuator A, vol. 87, no. 1-2, pp. 46-51, December 2002.

[36] Larry K. Baxter, Capacitive Sensors, Design and Application, Electronics Technology.
1997.

[37] George S.K.Wong and Tony F.W. Embleton, AIP Handbook of Condensor Microphones,
Modem Acoustics and Signal Processing. American Institute of Physics, New York, 1995.

[38] Bruel and Kjaer, Microphone Handbook Technical Documentation, vol. I, Bruel and
Kjaer, 1996.

[39] Knowles Acoustics, "Data sheet, EM series high performance microphone," Tech. Rep.,
Itasca, IL, 2003.

[40] National Semiconductors, "Data sheet LMV1012 pre-amplified ICs for high gain 2-wire
microphones," Tech. Rep., Santa Clara, CA, 2004.

[41] Naiyavudhi Wongkomet and Bernhard E. Boser, "Correlated double sampling in
capacitive position sensing circuits for micromachined applications," in IEEE Asia-Pacific
C, rirc rt on Circuits and Systems. 1998, pp. 723-726, IEEE.

[42] Widge Henrion, Len DiSanza, Matthew Ip, Stephen Terry, and Hai Jerman, "Wide
dynamic range direct digital accelerometer," in IEEE Solid-State Sensor and Actuator
Workshop, Hilton Head Island, South Carolina, June, 1990, pp. 153-157.

[43] Kouji Mochizuki and Kenzo Watanabe, "A relaxation oscillator based interface IEEE
Transactions on Instrumentation and Measurement, vol. 47, no. 1, pp. 11-15, February
1998.









[44] Andrzej Clchocki and Rolf Unbehauen, "A switched-capacitor interface for capacitive
sensors based on relaxation oscillators," IEEE Transactions on Instrumentation and
Measurement, vol. 39, no. 5, pp. 797-799, October 1990.

[45] Frank M. L. van der Goes and Gerard C. M. Meijer, "A novel low-cost capacitive-sensor
interface," IEEE Transactions on Instrumentation and Measurement, vol. 45, no. 2, pp.
536-540, April 1996.

[46] Stephen D. Senturia, Microsystem Design, Kluwer Academic Publishers, 1st edition,
2001.

[47] Sonion Technologies, "Digsimic microphones tcl00z21a data sheet," Tech. Rep., Sonion
Horsens A/S Fuglevangsvej 45 DK-8700 Horsens Denmark, 2005.

[48] George Brasseur, "Design rules for robust capacitive sensor," IEEE Transactions on
Instrumentation and Measurement, vol. 52, no. 4, pp. 1261-1265, 2003.

[49] George Brasseur, "State of the art of robust capacitive sensors," in 1st International
Workshop on Robotic Sensing, 2003.

[50] Christianne Thierlemann and Gisela Hess, "A micro-machined capacitive electret
microphone," in Symposium on Design, Test, and Microfabrication of MEMS and
MOEMS, Paris, France, 1999, SPIE, vol. 3680, pp. 748-756.

[51] Quanbo Zou, Zhimin Tan, Zhenfeng Wang, Jiangtao Pang, Xin Qian, Qingxin Zhang,
Rongming Lin, Sung Yi, Haiqing Gong, Litian Liu, and Zhijian Li;, "A novel integrated
silicon capacitive microphone-floating electrode electret microphone (feem)," IEEE
Journal of Microelectromechanical Systems, vol. 7, no. 2, pp. 224-234, June 1998.

[52] Hohm D. and Gerhard Multhaupt R., "Silicon-dioxide electret transducer," Journal of the
Acoustical Society of America, vol. 75, no. 4, pp. 1297-1298, April 1984.

[53] Bruel and Kjaer, "Measuring microphones," World Wide Web, http: / /www. bksv.
com/pdf/Measuring_Microphones.pdf, 1998.

[54] George S.K.Wong and Tony F.W. Embleton, AIP Handbook of Condensor Microphones,
Modem Acoustics and Signal Processing. American Institute of Physics, New York, 1995.

[55] Patrick Scheeper, Wouter Olthuis, and Piet Bergveld, "Fabrication of a subminiature
silicon condenser microphone using the sacrificial layer technique," in International
C,,if r, ccr on Solid-State Sensors and Actuators, TRANSDUCERS '91, San Francisco,
CA USA, 1991, pp. 408-411, IEEE.

[56] Patrick Scheeper, A.G.H van der Donk, Wouter Olthuis, and Piet Bergveld, "Fabrication
of silicon condenser microphones using single wafer technology," IEEE Journal of
Microelectromechanical Systems, vol. 1, no. 3, pp. 147-154, September 1992.

[57] Wolfgang Kiihnel and Gisela Hess, "A silicon condenser microphone with structure
backplate and silicon nitride membrane," Sensors andActuatorA, vol. 30, pp. 251-258,
1992.









[58] Johan Bergqvist and Jean Gobet, "Capacitive microphone with a surface micromachined
backplate using electroplating technology," IEEE Journal of Microelectromechanical
Systems, vol. 3, no. 2, pp. 69-75, June 1994.

[59] Michael Pederson, Wouter Olthuis, and Piet Bergveld, "High performance condenser
microphone with fully integrated cmos amplifier and dc-dc voltage converter," IEEE
Journal of Microelectromechanical Systems, vol. 7, no. 4, pp. 387-394, December 1998.

[60] Michael Pedersen, Wouter Olthuis, and Piet Bergveld, "An integrated silicon capacitive
microphone with frequency-modulated digital output," Sensors and Actuator A, vol. 69,
pp. 267-275, 1998.

[61] Jesper Bay, Ole Hansen, and Siebe Bouwstra, "Design of a silicon microphone with
differential read-out of a sealed double parallel-plate capacitor," Sensors and Actuator A,
vol. 53, pp. 232-236, 1996.

[62] Romach P., Mii llenbom, Klein U., and Rasmussen K., "The first low voltage, low
noise differential condenser silicon microphone," in The 14th European Cf r', r ci c
on Solid-State Transducers, EUROSENSORS XIV, Copenhagen, Denmark, 2000, pp.
213-216.

[63] Mi llenbom, P. Rombach, U.Klein, K.Rasmussen, J.F.Kuhmann, M.Heschel, M.Amskov
Gravad, J.Janting, J.Branebjerg, A.C.Hoogerwer, and S.Bouwstra, "Chip-size packaged
silicon microphones," Sensors and Actuator A, vol. 92, pp. 23-29, 2001.

[64] David T. Martin, Design, Fabrication Characterization of a MEMS Dual Backplate
Capacitive Microphone, Ph.D. dissertation, University of Florida, Gainesville, FL, 2007.

[65] Joost C. Litters, Wouter Olthuis, Peter H. Veltnik, and Piet Bergveld, "A sensitive
differential capacitance to voltage converter for sensor applications," IEEE Transactions
on Instrumentation and Measurement, vol. 48, no. 1, pp. 89-96, February 1999.

[66] Satomi Ogawa, Yukata Oisugi, Kouiki Mochizuki, and Kenzo Watanabe, "A switched
capacitor interface for differential capacitance transducers," IEEE Transactions on
Instrumentation and Measurement, vol. 50, no. 5, pp. 1296-1301, October 2001.

[67] van der Donk A.G.H, Scheeper PR., Olthuis W., and Bergveld, "Amplitude modulated
electro mechanical feedback system for silicon condenser microphones," Journal of
Micromechanical Microengineering, vol. 2, pp. 211-214, 1992.

[68] A.G.H. van der Donk, A. Sprenkels, Wouter Olthuis, and Piet Bergveld, "Preliminary
results of a silicon condenser microphone with internal feedback," in 1991 International
Cri f r, c on Solid-State Sensors and Actuators TRANSDUCERS '91, New York, 1991,
pp. 262-265, IEEE.

[69] Michael Kraft, Closed Loop Digital Accelerometer Employing Oversampling Conversion,
Ph.D. dissertation, School of Engineering, Coventry University, UK, July, 1997.

[70] Krylov S. and Maimon R., "Pull-in dynamics of electrostatic beam actuated by
continuously distributed electrostatic force," Journal of the Acoustical Society of America,
vol. 126, pp. 332-342, 1993.









[71] Hunt F., Electroacoustics : the analysis of transduction, and its historical background,
Harvard University Press, Cambdridge, MA, 1954.

[72] Scheeper P.R., van der Donk A.G.H., Olthuis W., and Bergveld P., "A review of silicon
microphone," Sensors and Actuator A, vol. 44, pp. 1-11, February 1994.

[73] David T. Blackstock, Fundamentals of Physical Acoustics, Wiley Interscience, New York,
2000.

[74] M. Rossi, Acoustics and Electroacoustics, Artech House, Norwood, MA, 1988.

[75] Z. Skvor, "On the acoustical resistance to viscous losses in the air gap of electrostatic
transducers," Acoustica, vol. 19, pp. 295-299, 1996.

[76] Liu J., Martin D., Kadirvel K. Nishida T., Cattafesta L., Sheplak M., and Mann B.,
"Nonlinear model and system identification of a capacitive dual-backplate mems
microphone," Journal of Sound and Vibration, 2006.

[77] Bruel and Kjaer, "Measuring sound," World Wide Web, http: / /www. bksv. corn/
pdf/Measuring_Sound.pdf, 1998.

[78] Timosheko S. and Krieger S., Theory of plates and shells, McGraw Hill Book Company,
New York, 1959.

[79] David A. Johns and Ken Martin, Analog Integrated Circuit Design, John Wiley and Sons,
Inc, New York, 1997.

[80] A. van der Ziel, Noise in Solid State Devices and Circuits, John Wiley and Sons, New
York, 1986.

[81] Fredrico Alimenti, Paolo Mezzanotte, Luca Roselli, and Roberto Sorrentino, "Modeling
and characterization of bonding wire interconnection," IEEE Transactions on Microwave
Theory and Techqniques, vol. 49, no. 1, pp. 142-151, January 2001.

[82] Standard Wire and Cable Company, "Data sheet rg164 coaxial cable," Tech. Rep., Rancho
Domniguez, CA, 2000.

[83] Mark Serridge and Torben R. Licht, Piezoelectric Accelerometers and Vibration
Preamplifiers Theory and Application Handbook, Bruel and Kjaer, New York, 1987.

[84] Furst C. and Marczak T, "Preamplifier for two terminal electret condenser microphones,"
US Patent 6888408, 2005.

[85] National Semiconductors, "LmvlOxx analog pre-amplified ic's for high gain
microphones," Tech. Rep., 2900 Semiconductor Drive, Santa Clara, California USA
95050,2003.

[86] Killion M. and Carlson E., "Audio frequency amplification circuit," US Patent 3512100,
1977.

[87] Murphy P., Hubschi K., Rooji N., and Racine, "Subminiature silicon integrated electret
capacitor microphone," IEEE Transactions on Electrical Insulation, vol. 24, no. 3, pp.
495-499, June 1989.









[88] Y.B. Ning, A.W. Mitchell, and R.N.Tait, "Fabrication of a silicon micromachined
capacitive microphone using a dry-etch process," Sensors andActuatorA, vol. 53, pp.
237-242, 1996.

[89] Voorthuyzen J., Sprenkels A., van der Donk, Scheeper P, and Bergveld P, "Optimization of
capacitive microphone and pressure sensor performance by capacitor electrode shaping,"
Sensors and Actuator A, vol. 25, pp. 331-336, 1991.

[90] Kuhnel W and Hess G., "A silicon condenser microphone with structured back plate and
silicon nitride membrane," Sensors and Actuator A, vol. 30, pp. 251-258, 1992.

[91] Bergqvist J. and Gobet J., "Capacitive microphone using electroplating technology," IEEE
Journal of Microelectromechanical Systems, vol. 3, no. 2, pp. 69-75, June 1994.

[92] Baker M. and Sarpeshkar R., "A low-power high-psrr current-mode microphone
preamplifier," vol. 38, no. 10, pp. 1671-1679, October 2003.

[93] Otmar Kern, "Amplifying circuit," US Patent 6812788, 2004.

[94] Loeppert P. and Lee S., "Sisonic the first commercialized mems microphone," in Solid
State Sensors and Actuators Microsystem Workshop, Hilton Head South Carolina, 2006,
pp. 27-31, IEEE.

[95] Hsu P., Klastrangelo C., and Wise K, "A high sensitivity polysilicon diaphragm condenser
microphone," in MEMS 1998, Anchorage AK, USA, 1998, pp. 580-585.

[96] G. Amendola, Y. Blanchard, A. Exertier, S. Spirkovitch, G. Lu, and G. Alquie, "A CMOS,
self-biased charge amplifier," in Third International Workshop on Design of Mixed-Mode
Integrated Circuits and Applications, Puerto Vallarta, Mexico, 1999, IEEE, vol. 3680, pp.
5-8.

[97] D.T. Martin, K. Kadirvel, R.M. Fox, T. Nishider, J. Liu, and M. Sheplak, "Surface and
bulk micromachined dual back-plate condenser microphone," in International C f, ;r,c c
on Micro Electro Mechanical Systems, Miami,FL, 2005, IEEE.

[98] Thermoptics, "Data sheet dn620, charge amplifier," Tech. Rep., Carson City, NV, 1995.

[99] Texas Instruments, "Data sheet opal28," Tech. Rep., One TI Blvd., Dallas, TX, 1993.

[100] Texas Instruments, "Data sheet opal 11," Tech. Rep., One TI Blvd., Dallas, TX, 1993.

[101] Analog Devices, "Data sheet ad745," Tech. Rep., One Technology Way, Norwood, MA,
1993.

[102] Analog Devices, "Data sheet ad734, 10 mhz, 4-quadrant multiplier/divider," Tech. Rep.,
One Technology Way, Norwood, MA, 1993.

[103] Baoqing Li, Deren Lu, and Weiyuan Wang, "Open-loop operating mode of micromachined
capacitive accelerometer," Sensors and Actuator A, vol. 79, pp. 219-223, 2000.

[104] Michael Kraft, C.P.Lewis, and T.G.Hesketh, "Closed loop silicon accelerometers," IEE
Proceedings of Circuits Devices and Systems, vol. 145, no. 5, pp. 325-331, October 1998.









[105] Yakabe et al., "Electrostatic capacitance detection circuit and microphone device," US
Patent 7034551, 2006.

[106] Christian C. Enz and Gabor C. Temes, "Circuit techniques for reducing the effects
of op-amp imperfections: Autozeroing, correlated double sampling, and chopper
stabilization," Proceedings of the IEEE, vol. 84, no. 11, pp. 1584-1614, November
1996.

[107] Cluade-Alain Gobet and Alexander Knob, "Noise analysis of switched capacitor
networks," IEEE Transactions of Circuits and Systems, vol. 30, no. 1, pp. 37-43, 1983.

[108] Naiyavudhi Wongkomet, Position Sensing for Electrostatic Micropositioners, Ph.D.
dissertation, University of California, Berkeley, 199.

[109] Potter R and Boor S, "Switched microphone buffer," US Patent 7092538, 1977.

[110] S.A. Jawed, M. Gottardi, and A. Baschirotto, "A switched capacitor interface for a
capacitive microphone," Research in Microelectronics and Electronics 2006, pp. 384-388,
2006.

[111] Frederick V. Hunt, Electroacoustics The Anial %i\ of Transduction, and Its Historical
Background, American Institute of Physics, 2 edition, 1982.

[112] K.Y.Park, C.W.Lee, H.S.Jang, Y.S.Oh, and B.J.Ha, "Capacitive sensing typesurface
micromachined silicon accelerometer with a stiffness tuning capability," IEEE, pp.
637-642, 1998.

[113] Navid Yazdi, Farrokh Ayazi, and Khalil Najafi, "Micromachined inertial sensors,"
Proceedings of the IEEE, vol. 86, no. 8, pp. 1640-1659, August 1998.

[114] Hall N.A., Bicen B., Jeelani M.K., Lee W, Quereshi S., Degertekin EL, and Okandan
Murat, "Micromachined microphones with diffraction-based optical displacement
detection," Journal of the Acoustical Society of America, vol. 118, no. 5, pp. 3000-3010,
2005.

[115] van der Donk A.G.H., Scheeper PR.,, Olthuis W., and Bergveld P., "Amplitude modulated
electro-mechanical feedback system for silicon condenser microphones," 1992.

[116] S. Hadjiloucas, L.S. Karatzas, D.A. Keating, and M.J. Usher, "Optical sensors for
monitoring water uptake in plants," Journal of Lightwave Technology, vol. 13, no. 7, pp.
1421-1428, 1995.

[117] Hao Luo, Gang Zhang, Richard Carley, and Garry Fedder, "A post-cmos micromachiend
lateral accelerometer," vol. 2, no. 3, pp. 188-195, 2002.

[118] Stephen A. Joyce and J.E.Houston adn Bradley K. Smith, "Interfacial force sensor with
force-feedback control," in IEEE Electron Devices Meeting. 1990, pp. 26.3.1-26.3.4,
IEEE.

[119] Mark A. Lemkin, Micro Accelerometer Design with Digital Feedback Control, Ph.D.
dissertation, University of California, Berkeley, 1997.

[120] Gelb A and van der Velde, Multiple Input Describing Functions, Mcgraw Hill, London,
1968.









[121] Wu N., Miles R., and Aydin O., "A digital feedback damping scheme for a micromachined
directional microphone," in 1994 Proceeding of the American Controls CO',f, ;r, ,.
Boston MA, 1994, pp. 3315 3321, IEEE.

[122] Mark Lemkin and Bernhard E. Boser, "A three-axis micromachined accelerometer with a
cmos position-sense interface and digital offset-trim electronics," IEEE Journal of Solid
State Circuits, vol. 34, no. 4, pp. 456-468, April 1999.

[123] Tetsuya Kajita, Un-Ku Moon, and Gabor C. Temes, "A noise shaping accelerometer
interface circuit for two chip implementation," in IEEE International Symposium on
Circuits and Systems, ISCAS 2000, Geneva Switzerland. IEEE, 2000, pp. 337-340, IEEE.

[124] Edelson J. and Ward N., "Force balance microphone," US Patent 6285769 B1, 2001.

[125] C.C. Enz and G.C. Temes, "Circuit techniques for reducing the effects of
op-ampimperfections: autozeroing, correlated double sampling, and chopperstabilization,"
Proceedings of IEEE, vol. 84, no. 11, pp. 326-337, March 1996.

[126] K. Ogata, Modem Control Engineering, Prentice Hall, Norwood, MA, 2001.

[127] Analog Devices, "Data sheet op470, very low noise quad operational amplifier," Tech.
Rep., One Technology Way, Norwood, MA, 1993.

[128] Altera Semiconductors, "Data sheet epm7128 cpld," Tech. Rep., 101 Innovation Drive,
San Jose, CA,, 2003.









BIOGRAPHICAL SKETCH

Karthik Kadirvel was in Chennai, Tamil Nadu, India. He moved to the United Arab

Emirates in 1989 and graduated from Our Own English High School in Dubai, United Arab

Emirates in 1996. He went back to India for his Bachelors degree and obtained his Bachelor of

Engineering degree specializing in Electronics and Instrumentation from Annamalai University,

Tamil Nadu, India in May 2000. He obtained his Master's degree at the University of Florida in

2002. His thesis was on the design and characterization of a MEMS optical microphone. He is

currently pursuing his doctoral degree at Interdisciplinary Microsystem Group at the University

of Florida.





PAGE 1

1

PAGE 2

2

PAGE 3

3

PAGE 4

IwouldliketothankmyadvisorandmentorDr.ToshikazuNishidaforhisconstantmotivationandencouragementtobringoutthebestinme.Thisresearchwouldnothavebeenpossiblewithouthislogicalapproachtoproblemsolvingandemphasisonfundamentalprinciples.Dr.MarkSheplakandDr.RobertFoxprovidedindispensablehelponMEMSprocessingandcircuitdesignrespectivelyatvariouspointsduringthisresearch.ThemanydiscussionswithDr.LouisN.Cattafestaoncontroltheoryprovidedvaluableinformationatcrucialjuncturesofthisresearch.IwouldliketothankmyprojectcolleaguesDavidMartinforprovidingmewithaMEMSmicrophonetouseasatestvehicleforthecircuitdevelopment,JianLiuforthehelpinmodelingusingMatlabandJoelFusterforthevariousboardlevelcircuitdesignideas.IwouldliketothankallmyfriendsattheInterdisciplinaryMicrosystemsGroupfortheirfriendshipandsupportduringthecourseofmygraduateprogram.SpecialmentiongoestoRobertTaylor,ErinPatrick,andStephenHorowitzforalltheirhelp.Mr.T.KSundaram,myhighschoolmathematicsandphysicsinstructor,deservesspecialmentionhere.Iwouldliketoextendmyheartfeltthanksandgratitudeforprovidingmewiththestrongfoundationonwhichallmyachievementsstand.ThehelpprovidedbyAlOgden,IvanKravachenkoandBillLewisoftheUFNanofabricationfacilitiesandKenReedofTMREngineeringisgratefullyacknowledged.Finally,andmostimportantIwouldliketothankmyfamilyandmywifeforbelievinginallmygoalsandaspirationsand,fortheirlove,encouragement,andconstantsupportinallmyendeavors. 4

PAGE 5

page ACKNOWLEDGMENTS .................................... 4 LISTOFTABLES ....................................... 8 LISTOFFIGURES ....................................... 10 ABSTRACT ........................................... 14 CHAPTER 1INTRODUCTION .................................... 16 1.1Background ..................................... 18 1.2ResearchObjectives ................................. 20 1.3DissertationOverview ................................ 21 2CAPACITANCETRANSDUCERS ........................... 23 2.1Introduction ..................................... 23 2.2CanonicalCapacitiveTransducers ......................... 23 2.2.1SingleBackplateCapacitiveTransducers .................. 24 2.2.2CanonicalDualBackplateCapacitiveTransducer ............. 33 2.3StructureofMicrophone .............................. 39 2.4LumpedElementModel .............................. 41 2.5Summary ...................................... 44 3INTERFACECIRCUITS ................................. 46 3.1Introduction ..................................... 46 3.2PerformanceMetricsofCapacitiveTransducers .................. 46 3.3AnalogOpenLoopSenseTechniques ....................... 51 3.3.1SensingUsingAVoltageAmplier ..................... 52 3.3.2SensingUsingAChargeAmplier ..................... 65 3.3.3SensingUsingSynchronousModulation/Demodulation .......... 72 3.4DigitalOpenLoopSenseTechniques ........................ 80 3.5Closed-LoopSenseTechniques ........................... 85 3.5.1SensingUsingClosed-LoopAnalogInterfaceCircuit ........... 86 3.5.2SensingUsingDigitalclosed-loopInterfaceTechnique .......... 91 3.6ScalingAnalysisofInterfaceTechniques ................... 93 3.7Summary ...................................... 96 4DESIGNOFANALOGINTERFACECIRCUIT .................... 97 4.1Introduction ..................................... 97 4.2Simulation ...................................... 97 4.2.1SimulinkModel ............................... 98 4.2.2SimulationResults ............................. 102 4.3DesignofAnalogInterfaceCircuit ......................... 103 4.3.1DesignofAmplierforTopandBottomPlateVoltageGeneration .... 103 4.3.2InputBufferAmplierDesign ....................... 105 5

PAGE 6

............................ 106 4.3.4CompensatorDesign ............................ 106 4.4IntegrationofInterfaceCircuitComponentsonaPCB ............... 108 4.5Noisemodel ..................................... 109 4.6LimitationsoftheCurrentImplementationoftheInterfaceCircuit ........ 112 4.7Summary ...................................... 112 5DESIGNOFDIGITALCLOSEDLOOPINTERFACECIRCUIT ........... 113 5.1Introduction ..................................... 113 5.2DesignofInterfaceCircuit ........................... 113 5.3ComponentDesign ................................. 116 5.3.1SwitchedCapacitorAmplier ........................ 116 5.3.2CompensatorDesign ............................ 120 5.3.3ComparatorDesign ............................. 123 5.3.4BiasNetworkDesign ............................ 124 5.4SystemLevelTiming ................................ 124 5.5SystemLevelSimulation .............................. 127 5.6CircuitImplementation ............................... 129 5.7Summary ...................................... 130 6EXPERIMENTALSETUPANDRESULTS ....................... 131 6.1Introduction ..................................... 131 6.2AcousticPressureCoupler ............................. 131 6.2.1DesignofAcousticPressureCoupler .................... 132 6.3ExperimentalResults ................................ 133 6.3.1CharacterizationofPressureCoupler .................... 133 6.3.2OpenLoopFrequencyResponseofMicrophonewithdcBias ....... 136 6.3.3OpenLoopFrequencyResponseofMicrophonewithdc+acBias ..... 137 6.3.4LinearityofOpenLoopMicrophone .................... 139 6.3.5NoiseFloorofOpenLoopMicrophone ................... 140 6.3.6CompensatorCharacterization ....................... 141 6.3.7ClosedLoopFrequencyResponse ..................... 142 6.3.8ClosedLoopLinearity ............................ 144 6.3.9ClosedLoopNoiseFloor .......................... 145 6.4Discussion ...................................... 146 6.5Summary ...................................... 149 7CONCLUSIONANDFUTUREWORK ......................... 150 7.1Conclusion ..................................... 150 7.2FutureWork ..................................... 150 7.2.1Simulation .................................. 151 7.2.2HardwareImplementation .......................... 151 APERFORMANCEMETRICSFORMULATION ..................... 152 A.1DerivationofSensitivityforaVoltageFollower .................. 152 A.2NoiseAnalysisforaCapacitiveMicrophonewithaVoltageFollower ....... 154 A.3DerivationofSensitivityforaChargeAmplier .................. 156 6

PAGE 7

...... 157 A.5SensitivityforSynchronousModulationandDemodulationInterfaceCircuit .. 158 BFORMULATIONUSEDINSWITCHEDCAPACITORCIRCUITS .......... 162 B.1BandwidthConsiderationsforAmpliersinSwitchedCapacitorAmpliers ... 162 B.2AnalysisofNoiseinaSwitchedCapacitorAmplier ............... 162 B.2.1SwitchResistanceNoise .......................... 163 B.2.2WidebandAmplierNoise ......................... 163 REFERENCES ......................................... 164 BIOGRAPHICALSKETCH .................................. 173 7

PAGE 8

Table page 1-1Comparisonofaudioandaeroacousticmicrophonespecications ............ 21 2-1Dimensionsofaeroacousticandaudiomicrophone[ 64 ]. ................. 40 2-2Materialpropertiesofmicrophonestructure. ....................... 40 2-3Theoreticalformulationforlumpedelementmodelparameters. ............. 43 2-4Lumpedelementmodelparametersofaeroacousticandaudiomicrophone. ....... 43 2-5Secondordersystemparametersofmicrophones. .................... 44 3-1Parasiticcapacitanceforvariouspackagingtechnologies. ................ 54 3-2SpecicationofVoltageAmpliersforBruelandKjaerMicrophones. ......... 61 3-3ReviewofinterfacecircuitsusedinMEMSmicrophonespublishedinliterature. .... 63 3-4Specicationofarepresentativesampleofoperationalampliersrecommendedbymanufacturersforuseasvoltagefollowersforcapacitivetransducers. .......... 65 3-5Reviewofchargeamplierinterfacecircuitsusedincapacitivemicrophones. ..... 72 3-6Specicationofarepresentativesampleofgeneralpurposeoperationalampliersrecommendedbythemanufacturerforuseaschargeampliersusingafeedbackresistorandcapacitor. ....................................... 72 3-7Summaryofprevioussynchronousmodulation/demodulationbasedopenloopcapacitivetransducerinterfacecircuit. ................................ 79 3-8Summaryofprevioussynchronousmodulation/demodulationbasedclosed-loopcapacitivetransducerinterfacecircuit. ................................ 89 3-9Frequencyrequirementsofinterfacecircuitswithincreasingresonantfrequency. 94 3-10Feedbackvoltagerequirementsofinterfacecircuitswithincreasingdynamicrange. 94 3-11Summaryofpreviousbasedclosed-loopcapacitivetransducerinterfacecircuit. .. 95 4-1Feedbackvoltagerequirementsofaeroacousticandaudiomicrophoneforforcebalancingmaximuminputpressure. ................................. 105 4-2Elementvaluesforresistorsandcapacitorsusedincompensator. ............ 107 4-3Magnitudeofparametersusedinnoiseanalysis ..................... 110 5-1Clocksignalsusedinthevariouscomponentsofthemodulator. .......... 125 5-2Functionperformedduringeachtimeinterval. ...................... 127 5-3Parametersusedinclosedloopsimulation. ........................ 128 5-4Frequencyrequirementsofinterfacecircuitswithincreasingresonantfrequency. 129 8

PAGE 9

129 6-1Comparisonofsimulatedandexperimentalsensitivity. ................. 139 6-2Characterizationresultsofmicrophoneinopenandclosedloopmodeofoperation. .. 148 9

PAGE 10

Figure page 1-1Microphoneclassicationbasedonsensorstructure. ................... 19 1-2Commoncondensermicrophonestructures. ....................... 20 1-3Microphoneclassicationbasedonsensetechnique. ................... 21 2-1OverviewofChapter2. .................................. 23 2-2Schematicofasinglebackplatecapacitivetransducer. .................. 24 2-3Schematicofasinglebackplatecapacitivetransducerconnectedtoavoltagebuffer. .. 26 2-4Schematicofasinglebackplatecapacitivetransducerconnectedtoachargeamplier. 28 2-5Schematicofasinglebackplatecapacitivetransducershowingelectrostaticforceandrestoringforce. ...................................... 31 2-6Schematicofadifferentialcapacitivetransducer. ..................... 33 2-7Schematicofadualbackplatecapacitivetransducerconnectedtoavoltageamplier. 36 2-8Schematicofadualbackplatecapacitivetransducerconnectedtoachargeamplier. .. 36 2-9Schematicofadualbackplatecapacitivetransducershowingforcesonthemovableplate. ............................................ 37 2-10Crosssectionviewofdualbackplatemicrophone[ 64 ]. ................. 40 2-11LowfrequencyLEMofdualbackplatemicrophone. ................... 43 2-12Theoreticalnormalizedfrequencyresponseofaudiomicrophone. ............ 44 2-13Theoreticalnormalizedfrequencyresponseofaeroacousticmicrophone. ........ 45 3-1OverviewofChapter3. .................................. 46 3-2Characteristicfrequencyresponseofacapacitivetransducer. .............. 48 3-3Characteristicidealandactualresponseofacapacitivetransducer. ........... 48 3-4Characteristicnoiseoorofacapacitivetransducerusingavoltageamplierinterfacecircuit. ........................................... 50 3-5Noisemodelofresistor. .................................. 51 3-6Noisemodelofopampshowingvoltageandcurrentnoisesource. ............ 51 3-7PlotofnoisePSDofamplier. .............................. 52 3-8OrganizationofeachinterfacecircuitdiscussedinChapter3. .............. 53 3-9Schematicofadualbackplatemicrophoneconnectedtovoltagebuffer. ......... 53 3-10Noisemodelofdualbackplatemicrophoneconnectedtovoltagebuffer. ........ 55 10

PAGE 11

........................................ 57 3-12Powerspectraldensityofindividualnoisecontributorsandtotaloutputnoiseinavoltageamplierbasedinterfacecircuit .............................. 58 3-13Plotofoutputnoiseforvoltageamplierwithaxedbiasresistor(10M) 59 3-14Plotofoutputnoiseforvoltageamplierwithaxedparasiticcapacitanceof(10pF). 60 3-15Schematicofdualbackplatemicrophoneconnectedtoachargeamplier. ....... 66 3-16Noisemodelofdualbackplatemicrophoneconnectedtochargeamplier. ....... 66 3-17Powerspectraldensityofindividualnoisecontributorsandtotaloutputnoiseinachargeamplierbasedinterfacecircuit. ............................. 68 3-18Plotofoutputnoiseforchargeamplierwithaxedbiasresistor(10M) 70 3-19Plotofoutputnoiseforchargeamplierwithaxedparasiticcapacitanceof(10pF). 71 3-20Schematicofcapacitivemicrophonewithasynchronousmodulationanddemodulationtechniqueusingavoltageamplierbasedcircuit. .................... 73 3-21Schematicofacapacitivemicrophonewithasynchronousmodulationanddemodulationtechniqueusingachargeamplierbasedcircuit. ..................... 73 3-22PlotofspectrumatvariousnodesofFigure 3-21 ..................... 74 3-23Noisemodelofsynchronousmodulatoranddemodulatorusingvoltageamplier. ... 77 3-24Noisemodelofsynchronousmodulatoranddemodulatorusingchargeamplier. .... 77 3-25Switchedcapacitorimplementationofcorrelateddoublesampling. ........... 81 3-26Equivalentcircuitofswitchedcapacitoramplierduringamplifyandresetphase. ... 82 3-27Noisemodelofswitchedcapacitoramplierduringamplifyphase. ........... 82 3-28Schematicofanalogclosed-loopforcefeedbackscheme. ................ 87 3-29Schematicofrstordermodulator. .......................... 92 3-30Schematicofsecondordermechanicalmodulator. ................. 92 4-1OverviewofChapter4. .................................. 97 4-2Blockdiagramofanalogclosed-loopsystemshowingmicrophoneandinterfacecircuit. 98 4-3Bodeplotofcompensator. ................................ 101 4-4Bodeplotofloopgainwithandwithoutcompensator. .................. 102 4-5Simulinkmodelofanalogclosed-loopcontrolsystem. .................. 102 4-6SubsystemsusedinSimulinkmodel. ........................... 103 11

PAGE 12

.................. 104 4-8Diaphragmdisplacementwithandwithoutfeedback. .................. 104 4-9Schematicofsummingamplierthatgeneratestopandbottomplatevoltages. ..... 105 4-10Schematicofvoltageamplier. .............................. 106 4-11Schematicofanalogmultiplier(AD835)showingrelevantpins. ............. 106 4-12Schematicoftwostagecompensator. ........................... 107 4-13AnnotatedphotographofPCB1showingbufferamplier,breakoutheadersandopticalport. ............................................ 108 4-14AnnotatedphotographofPCB2. ............................. 108 4-15Schematicofforwardpathoftheclosedloopsystem .................. 109 4-16Plotoftheoreticalnoiseooroftheopenloopsystemandforwardpathsystem .... 111 5-1OverviewofChapter5. .................................. 113 5-2Flowchartfordesignofmodulatorcomponentparametersfromsystemlevelspecication. ....................................... 114 5-3Schematicofsecondordermechanicalmodulator. ................. 115 5-4Transistorlevelschematicofamplier. .......................... 117 5-5Schematicofamplierduringamplifyphase. ...................... 118 5-6Transistorlevelschematicofcompensator. ........................ 121 5-7Compensatorschematicduringthetwophasesofoperation. ............... 122 5-8Transistorlevelschematicofcomparator ......................... 123 5-9Schematicofbiasnetwork. ................................ 124 5-10Flowchartshowingthesequenceofoperationoftheoverallmodulationbasedinterfacecircuit. ...................................... 125 5-11Schematicofmodulatorcircuit. ........................... 126 5-12Systemleveltimingdiagramoveronecompletecycle. .................. 127 5-13SimulinkModelofthemodulatorbasedinterfacecircuit. .............. 128 5-14Powerspectraldensityofoutputmodulatedbitstream. ............... 129 6-1OverviewofChapter6. .................................. 131 6-2Photographofacousticpressurecoupler. ......................... 132 6-3Plotofinputpressuresignalasafunctionoffrequency. ................. 134 12

PAGE 13

.................................... 135 6-5Blockdiagramdescribingcharacterizationprocedureused. ............... 135 6-6Frequencyresponsewithdcbias(5V)withairmedium. ................. 136 6-7Frequencyresponsewithdcbias(5V)withheliummedium. .............. 137 6-8Schematicofcapacitivemicrophonewithasynchronousmodulationanddemodulationtechniqueusingavoltageamplierbasedcircuit. .................... 137 6-9Openloopfrequencyresponsemeasuredattheoutputoflowpasslter. ........ 138 6-10Openloopfrequencyresponsenormalizedtothepeakacvoltagemeasuredattheoutputoflowpasslter. ..................................... 138 6-11Openlooplinearityoftheopenloopsystem. ...................... 139 6-12Linearityoftheopenloopsystemnormalizedtopeakacbiasvoltagemeasuredattheoutputofthedemodulator. ................................ 140 6-13Noiseoorofmicrophonefortwodifferentacbiasvoltagesmeasuredattheoutputofthedemodulator. .................................... 141 6-14Theoreticalandexperimentalfrequencyresponseofcompensator. ........... 142 6-15Blockdiagramofanalogclosed-loopsystemshowingmicrophoneandinterfacecircuit. 142 6-16Closedloopfrequencyresponsemeasuredattheoutputofdemodulator ........ 143 6-17Closedloopfrequencyresponsemeasuredattheoutputofdemodulatorwithacbiasof6Vpp. .......................................... 144 6-18Closedloopfrequencyresponsemeasuredattheoutputofdemodulatorwithacbiasof9Vpp. .......................................... 145 6-19Closedloopfrequencyresponsemeasuredattheoutputofcompensator. ........ 146 6-20Linearrangeofsysteminclosedloopcongurationasmeasuredattheoutputofthecompensator. ....................................... 146 6-21Linearityoftheclosedloopsystemnormalizedtopeakacbiasvoltagemeasuredattheoutputofthecompensator. .............................. 147 6-22Closedloopnoiseoor. .................................. 147 A-1Noisemodelofdualbackplatemicrophoneconnectedtovoltagebuffer. ........ 152 A-2Noisemodelofdualbackplatemicrophoneconnectedtovoltagebuffer. ........ 155 A-3Noisemodelofdualbackplatemicrophoneconnectedtochargeamplier. ....... 157 B-1Noisemodelofswitchedcapacitoramplierduringamplifyphase. ........... 162 13

PAGE 14

Assensorgeometryisscaleddown,improvedcircuittechniquesarerequiredtomeasurethesensoroutput.Thisisbecauseatsmallgeometriessensorcapacitanceiscomparabletounwantedparasiticcapacitancewhichreducesthetransducersensitivity.Also,atreducedsensorgeometry,thevoltagerequiredtobiasthemicrophonecouldcausethemicrophoneplatestopullin.Thegoalofthisworkistodesignandcharacterizeinterfacecircuitsthataresuitableforminiaturecapacitivetransducers.Toachievethisgoal,theperformanceofexistingopenandclosed-loopinterfacecircuitsareinvestigated.Scalingoftheperformancemetricsofthemicrophoneandinterfacecircuitassensorgeometrydecreasesisalsoinvestigated. Aproofofconceptclosed-loopanalogcontrollerforaMEMScapacitivemicrophoneisdesigned.Atestapparatusisdevelopedtocharacterizethesystemovertheaudiorangebyoperatingthemicrophoneinaheliummediumwhichincreasesthebandwidthofthetestapparatus.Characterizationofthemicrophoneinopenandclosedloopmodeofoperationis 14

PAGE 15

15

PAGE 16

Therehasabeenamajortrendtowardsminiaturizationandbatchfabricationoftransducersinspiredbyasimilartrendintheelectronicsindustry.Withtheadventofmicromachining[ 1 ]techniquesusedinmicroelectromechanicalsystem,(MEMS)manufacturing,fabricationoftransducerswithverysmallsizeshavebeenmadepossible.Micromachiningalsofacilitatesthebatchfabricationofdeviceswhichpossessnearlymatchedstructuralandelectricalproperties.Batchfabricationofdevicesalsocanreducethecostofmanufacturingifthevolumeislargeenough. Onekeytransducerleadingthewayinthistrendarecapacitivetransducers.Capacitivetransducersaredevicesthatrespondtophysicalphenomenonviaachangeincapacitancethatismeasuredusingvariousinterfacecircuits.CommerciallysuccessfulapplicationsthatutilizecapacitivetransducersincludeaccelerometersusedinairbagdeploymentsystemforcarsandlaptopharddrivesbyAnalogDevices[ 2 ],projectiontechnologyforlargescreenTVsusingTexasInstrumentsDigitalLightProcessing(DLP)processors[ 3 ],pressuresensorsbyFreescaleSemiconductors[ 4 ],andaudiomicrophonesforcellphonesbyKnowlesAcoustics[ 5 ].Othercapacitivetransducerspublishedintheliteratureincludeaccelerometers[ 2 ],[ 6 ][ 11 ],micro-mirrors[ 12 ],microphones[ 13 ][ 21 ],pressuresensors[ 22 ][ 25 ],oscillators[ 26 ][ 31 ],andgyroscopes[ 32 ][ 35 ].Inalltheseapplications,thecapacitancechangeofthesensorisconvertedintoausablevoltageusinginterfacecircuits[ 36 ]suchasthevoltagebuffer[ 37 ][ 40 ],chargeamplier[ 41 ],acbridge[ 6 ],forcebalancingcircuits[ 6 8 42 ],andoscillatorbasedcircuits[ 43 ][ 45 ]. Theminiaturizationofcapacitivetransducersresultsinnewchallengesforthedesignoftheinterfacecircuits.Someoftheissuesthatareexacerbatedatsmallsensorgeometriesthatmakethedesignofinterfacecircuitsmorechallengingincludeaparasiticcapacitancecomparabletothesensorcapacitance,reducedsensitivity,andpullin[ 46 ].Allofthesefactorscandegradetheperformanceofmicromachinedcapacitivetransducers.AswescaledownaggressivelyfrommicroscaledevicesinMEMSsystemstonanoscaledevices,anunderstandingofthetradeoffsofinterfacecircuitsisimportanttoobtainthefullperformancebenetsoftheminiaturecapacitivetransducer. 16

PAGE 17

38 ],useavoltagebufferamplierastheinterfacecircuit.Capacitivemicrophonesusedovertheaudiobandwidth(20Hz20kHz)suchastheBruelandKjrType4135microphonealsouseavoltagebufferastheinterfacecircuit.CommercialMEMSmicrophonesthatusevoltagebuffersincludetheSiSonicmicrophones[ 5 ]manufacturedbyKnowlesacousticsandDigiSiMicmicrophones[ 47 ]bySonionTechnologies. Thoughvoltageamplierinterfacecircuitsarecommonlyused,thereexistotherinterfacetechniquesthatmaypossessadvantagesforhighperformancecapacitivetransducersincertainapplications.Forexample,sincecapacitiveaccelerometersaresimilarinoperatingprincipleandstructureascomparedtocapacitivemicrophonescircuitsusedforaccelerometersmaybeadaptedforuseincapacitivemicrophones.Aclosed-loopinterfaceinterfacecircuittechniqueusedinsomeaccelerometers,[ 2 ],mayprovidebenetssuchasimprovedpull-inperformance,highersensitivity,anddirectdigitaloutput.Thisprovidesthemotivationforinvestigatingthepotentialadvantagesanddisadvantagesforclosed-loopinterfacecircuitsforcapacitivemicrophones. Toachievethesegoals,athoroughunderstandingofthetheoreticalperformanceofexistingcapacitiveinterfacetechniquesisrequired.Althoughpreviouswork[ 48 ],[ 49 ],reviewinterfacetechniquessuchasvoltagebuffer,chargeamplier,synchronousmodulationanddemodulation,theydonotprovidedetailedtheoreticalformulationstocomparethevarioustechniques.Inaddition,closedloopinterfacecircuitsarenotdiscussedindetail. Thegoalofmyworkistodesignandevaluatecapacitiveinterfacecircuitswithanemphasisonclosed-loopinterfacecircuitsforMEMScapacitivemicrophones.Toachievethisgoal,theperformanceofexistingopenandclosed-loopinterfacecircuitsareinvestigated.Theoreticalformulationsfortheperformancemetricsoftheinterfacecircuitsarederivedandcompiledintoacommonframeworkforcomparison.Usingtheinformationobtained,aproofofconceptclosed-loopanalogcontrollerforaMEMScapacitivemicrophoneisdesigned. 17

PAGE 18

Amicrophoneisatransducerthatconvertsinputacousticalenergyintocorrespondingelectricalenergy.Capacitivemicrophonesconsistofaexiblediaphragmthatisexposedtothepressureuctuationsoftheacousticeldofinterestandoneormorexedbackplates.Thepressureinducedstructuraldeectionofthediaphragmcausesavariationinthecapacitancebetweenthevibratingdiaphragmandthexedbackplateswhichisdetectedandmeasured. Therearetwomainwaysofmeasuringthecapacitancechange.ThetwoapproachesarebasedonthefundamentalequationthatrelatesthechargeQstoredonthecapacitorplatestothecapacitanceCandthevoltageVacrosstheplatesofthecapacitor, Inthexedchargescheme[ 14 ],[ 50 ][ 52 ]the,chargeisheldconstantandthevoltagechangeassociatedwiththecapacitancechangeismeasured.Thechargeisheldconstantbyusingadcvoltagesourcetobiasthecapacitororbyusingaspecialdielectricmaterial(electret)suchaspolypropylene,mylar,orsilicondioxidethatcanholdchargeforlongperiodsoftimemeasuredinyears.Theelectretmaterialisbondedeithertothemovablediaphragmortothexedbackplatewhichischargedwithahighpotentialofhundredsofvoltsusinganexternaldcsource.Themainadvantageofelectretmicrophonesisthattheydonotrequireanexternalpowersupplyandarethususefulforportableapplications[ 53 ].Thexedvoltagethatcanbegeneratedbythechargeontheelectretismuchhigherthanwhatcanbeachievedviaabatterywhichleadstoahighersensitivity.Themaindisadvantagesofmicromachinedelectretsarethatthemicromachinedelectretmaterialslosetheirchargeovertimewhenexposedtohumidity[ 16 ].Inthexedvoltagescheme,axedvoltageismaintainedusinganexternalvoltagesourcewhile 18

PAGE 19

Capacitivemicrophonescanbebroadlyclassiedbasedonthestructureandonthedetectionmechanism.Figure 1-1 showsaclassicationtreeofthecommoncapacitivemicrophonestructure.Basedonthestructure,theycanbeclassiedassinglebackplateanddualbackplatemicrophones.Variousauthorshavereportedsinglebackplatemicrophones[ 17 21 ],[ 54 ][ 60 ]anddualbackplatemicrophones[ 19 ],[ 61 ][ 64 ]intheliterature.AschematicofthecommonmicrophonestructuresisshowninFigure 1-2 MostcommercialMEMS[ 5 47 ]andnon-MEMS[ 38 ]microphonesuseasinglebackplatedesign.Inthesemicrophones,asinglevariablecapacitorisformedbetweenthexedbackplateandthemovableplatethatismeasuredbytheinterfacecircuit.Indualbackplateanddualdiaphragmmicrophones,twocapacitorsthatvarydifferentiallyareformed.Dualbackplateanddualdiaphragmmicrophoneshavecertainadvantagessuchasincreasedsensitivityascomparedtosinglebackplatemicrophones,ahigherbiasvoltagewhichincreasesthesensitivity,improvedpullinperformance,andincreasedlinearitybutaremorecomplextofabricateandhavehighercost. Figure1-1. Microphoneclassicationbasedonsensorstructure. Figure 1-3 showsaclassicationtreeofthevariouscapacitivemicrophoneinterfacetechniques.Basedonthesensingscheme,microphonescanbebroadlydividedintoopenloop[ 41 59 65 66 ]andclosedloop[ 6 ][ 8 ],[ 10 60 ],[ 67 ][ 69 ]categories.Bothofthesecategoriescanbefurtherdividedintoanaloganddigitaltechniquesbasedonthepresenceor 19

PAGE 20

Schematicofsinglebackplatesinglediaphragmmicrophone[ 55 ]. (b) Schematicofsinglebackplatedualdiaphragmmicrophone[ 55 ]. (c) Schematicofdualbackplatesinglediaphragmmicrophone[ 55 ]. Commoncondensermicrophonestructures. absenceofaclocksignal,respectively.Adetailedliteraturereview,principleofoperation,andperformancemetricsofeachofthesecircuitsisdiscussedinSection2.2toSection2.5. 1-1 .Thesespecicationsprovidetypicalvaluesforthevariousparameters.Theactualvaluesdependupontheapplication.Forexample,telephonyapplicationsrequireaudiomicrophoneswith 20

PAGE 21

Microphoneclassicationbasedonsensetechnique. abandwidthof300Hz3:4kHzwhilehearingaidmicrophoneshaveabandwidthof20Hz-20kHz.Also,inthetable,itisassumedthatthedynamicrangeis100dB.Toaidin Table1-1. Comparisonofaudioandaeroacousticmicrophonespecications MaxPressure120dB160dBBandwidth20Hz-20kHz45Hz-160kHzNoiseFloor20dB60dB 21

PAGE 22

Chapter3discussesvariousopenandclosedloopinterfacecircuitsforcapacitivetransducerswithanemphasisoncapacitivemicrophones.Theoreticalformulationsfortheperformancemetricsoftheinterfacecircuitsaredeveloped. Chapter4andChapter5detailthedesignandsimulationoftheanaloganddigitalclosed-loopinterfacecircuitsrespectively.Thetwointerfacecircuitsarecomparedintermsofperformancespecications,designcomplexity,easeofimplementation,andcost.Chapter5concludesbypresentingthereasonforimplementingananalogclosed-loopinterfacecircuitoveradigitalclosed-loopinterfacecircuit. InChapter6,thetestapparatususedtocharacterizetheperformanceoftheinterfacecircuitisexplained.Resultsofthecharacterizationofthetestapparatusispresentedrstfollowedbythecharacterizationresultsofthemicrophonewiththeanalogopenandclosedloopinterfacecircuit.Chapter7discussestheresultsobtainedalongwiththefuturework. 22

PAGE 23

Thischapterbeginswithadiscussionofacanonicalsinglebackplateanddualbackplatecapacitivetransducer.Theoreticformulationofthevoltage,charge,capacitance,andforcerelationsforthetwocanonicaltransducersarederived.Linearizationoftheformulationisalsopresented.Adiscussionoftheperformancemetricsofcapacitivetransducersisthenpresented.Thestructuralandmaterialpropertiesofthemicrophoneunderconsiderationarediscussednext.Alumpedelementmodel(LEM)isthendevelopedforthemicrophone.UsingtheLEM,areducedsecondordersystemisdevelopedwhichisusedtoextractthesensorresonantfrequencyanddampingcoefcient.ThischapterconcludeswithMATLABsimulationsofthereducedordermodels.AgraphicaloverviewofthischapterisshowninFigure 2-1 Figure2-1. OverviewofChapter2. 23

PAGE 24

StructureandPrincipleofOperation. 2-2 showsasimplesinglebackplatecapacitivetransducer.Itconsistsoftwoplatesthatareatanominalgapx0betweenthem.Oneplateisxedandtheotherplateisfreetomoveinresponsetoaninputphysicalphenomenon. Figure2-2. Schematicofasinglebackplatecapacitivetransducer. TheplatesaresoarrangedsuchthatonlythedistancexbetweentheplatescanvaryandnottheareaofoverlapAbetweenthetwoplates.Undertheapplicationofanexternalpressuresignal,thediaphragmmovesfromitsequilibriumposition. Thedistancebetweentheplatescanbeexpressedintermsofthenominalgapdistance,x0,andtheinstantaneousdisplacementfromthenominalposition,x0,bytheequation TheinstantaneouscapacitanceC,betweentheplatesisgivenintermsofthepermitivityoffreespace0,relativepermitivityr,areaofoveralapA,andthedistancebetweentheplatesx,by x:(2) Forairgap,therelativepermitivityisapproximatelyunity.SubstitutingEquation 2 intothepreviousequation,weobtain x0"1 1x0 DeningthenominalcapacitanceC0as x0;(2) 24

PAGE 25

1x0 i.e., 1x0 FromEquation 2 ,itcanbeobservedthatthereexistsanon-linearrelationshipbetweenthecapacitanceandthedistancebetweentheplates.Forsmalldisplacements(x0<
PAGE 26

2 equationcanbelinearizedbyusingeitheraxedchargeoraxedvoltagetomeasurethedisplacement.Thesetwolinearizationmethodsdeterminethechoiceofinterfacecircuit.Forthexedchargetechnique,avoltageamplierisused.Thisisbecauseforthexedchargetechnique Forthexedvoltagetechnique,achargeamplierisused.Thisisbecauseforthexedvoltagetechnique,theoutputischargewhichwillbeshowntobelinearwithdisplacementafterintegrationbyachargeamplier.Apreliminarydiscussionofthetwointerfacetechniquesispresentednext.AdetailedreviewofthesetwotechniquesalongwithdetailedperformancemetricsandaliteraturereviewisgiveninSection 3.3 andSection 3.3.2 2-3 showsasinglebackplatecapacitivetransducerlinearizedusingaxedchargetechniquebyconnectingtoadcvoltagesourceVdc.Vdcisconnectedtooneendofthecapacitivetransducer.AbiasresistorRdc Schematicofasinglebackplatecapacitivetransducerconnectedtoavoltagebuffer. isconnectedbetweentheotherendofthecapacitivetransducerandgroundandisusedtosetthedcoperatingpoint.Aunitygainvoltageamplierisconnectedtothecommonpointbetweenthe 26

PAGE 27

2-3 .Thexedchargeonthetransducerisgivenbytherelation TheRCnetworkconsistingofRdcandCsetsupahighpasslterwithacutinfrequencyof1 2RdcC.Forallfrequenciesabovethecutinfrequency,thevoltageontheplateisfreetochangewhichisbufferedbythevoltageamplierandcanbemeasured.SubstitutingQwithQdcinEquation 2 ,thevoltageacrosstheplatesofthecapacitorisgivenby ThesmallsignaloutputvoltageVoutisgivenby FromEquation 2 ,itcanbeseenthat whichisalinearrelationbetweenthevoltageandthechangeingapdistance.Itshouldbenotedthatinthisdiscussion,theparasiticcapacitancesareneglected.TheeffectofparasiticcapacitancesarediscussedinChapter 3 2-4 showsasinglebackplatecapacitivetransducerconnectedtoachargeamplier.Inthistechnique,thexedvoltageisprovidedbyaDCvoltagesourceVdcwhichisconnectedtooneendofthecapacitivetransducer.Theotherendisconnectedtotheinvertinginputofaninvertingamplier.Theinvertingterminaloftheamplierisheldatvirtualgroundduetonegativefeedback.AfeedbackcapacitorCintisconnectedbetweentheoutputoftheamplierandtheinvertinginput.Thiscircuitintegratesthechargeowingoutofthetransducer.AbiasresistorRdcisconnectedinparalleltothefeedbackcapacitorandisusedtosettheDCoperatingpoint.TheRCnetworkconsistingofRdcandCintsetsupahighpasslterwithacutinfrequencyof1 2RdcCint.Forallfrequenciesabovethecutinfrequency,thechargeowingfromthetransducerisintegratedbythefeedbackcapacitor.Therelationbetweentheinputandoutputofachargeamplierabovethe 27

PAGE 28

Schematicofasinglebackplatecapacitivetransducerconnectedtoachargeamplier. cutinfrequencyisgivenby CintVdc(2) SubstitutingforCfromEquation 2 andEquation 2 ,Voutcanbeexpressedas FromEquation 2 itcanbeseenthat whichisalinearrelationbetweentheoutputvoltageandthechangeingapdistance.TheeffectofparasiticcapacitanceontheoutputvoltageisdiscussedinChapter3. TheincrementalchangeinmechanicalpotentialenergyisgivenintermsoftheforcebetweenthetwoplatesFandtheincrementalchangeingapdistancedx0by 28

PAGE 29

2 ,weobtain SubstitutingEquation 2 inEquation 2 ,weobtain TheelectricalpotentialenergystoredacrosstheplatesofthecapacitorEe,isdependentonthebiasingtechniqueused.IfthecapacitorisbiasedwithaxedvoltageVdc,thentheenergystoredisgivenbytherelation 2CV2dc:(2) SubstitutingforCfromEquation 2 andusingEquation 2 inEquation 2 ,weobtainanexpressionforthemagnitudeoftheelectrostaticforceactingonthemovableplatebiasedatVdc 2V2dcA i.e., (x0x0)2:(2) FromEquation 2 ,itcanbeseenthattheforcebetweentheplatesofacapacitorisnon-linearlyrelatedtothegap.Thedirectionoftheelectrostaticforceistowardsthexedplate.Themovableplatecomesclosertothexedplate,theforcewillincreaseandapproachinnityasgapdistancedecreases.Theforcerelationcanbelinearizedforsmalldisplacements(x0<
PAGE 30

2V2dcA x20(2) Fromthisequation,itcanbeseenthattheforceisindependentofthedisplacementofthemovableplateforthexedchargetechnique. 46 ].Thereasonforthisisasfollows.Whenavoltageisappliedtotheplates,anelectrostaticforceattractsthemovableplatetowardsthexedplate.Atsmallvoltages,thisisbalancedbytherestoringforceofthediaphragm.Asthevoltageisincreased,theelectrostaticforcebetweentheplateskeepsincreasing.Atsomecriticalvoltage,knownasthepull-involtageVPI,thetwoplateswillsnapin.Thestudyofthepull-incharacteristicsifimportanttodeterminethemaximumvoltageVdcthatcanbeappliedonthetransducerplates.Theactualvoltageusedismuchlowerthanthepullinlimittoensurestableoperationofthetransducer.Therearetwomaintypesofpullin. 1. QuasistaticPull-in:Inthistypeofpull-inthemovableplateisinitiallyatrest.Notimevaryingexternalforceisappliedonthemovableplate. 2. DynamicPull-in:Inthistypeofpull-in,atimevaryingexternalforceisactingonthemovableplateand/oratimevaryingbiasvoltage[ 70 ]. Inthisdiscussion,onlythequasistaticcaseisconsidered.Thepull-involtageandthecriticalgapatwhichpull-inoccursisderivednextfortheconstantvoltageandconstantchargetechniques. 2-5 showsthexedplateandmovableplatealongwiththedirectionofforces.ItcanbeseenthattheforceduetoelectrostaticattractionFelectrostaticandtherestoringspringforceFspringopposeeachother.Theelectrostaticforceisalwaysattractiveandtriestobringbothplatestogether.Therestoringforcealwaystriestobringthemovableplatebacktoitsequilibriumposition.Basedonthesignconventionshowninthegureanincreaseintheelectrostaticforcecausesadecreaseingapdistance. 30

PAGE 31

Schematicofasinglebackplatecapacitivetransducershowingelectrostaticforceandrestoringforce. TherestoringforceinthediaphragmFrisgivenintermsofthemechanicalcomplianceCm;dofthediaphragmby UsingEquation 2 TheFelectrostaticduetotheelectrostaticattractionbetweentheplatesisgivenbyEquation 2 whichisrepeatedbelow. 2Vdc2A x2(^x):(2) ThenetforceactingonthediaphragmFnetisthusgivenbythesumoftherestoringspringforcegivenbyEquation 2 andtheelectrostaticforcegivenbyEquation 2 2Vdc20A x2+1 Initially,whenthebiasvoltageissmall,themovableplatewillbedisplacedbyasmallamountfromitsequilibriumposition.Asthevoltageincreases,thegapdistancedecreases.Atanygivenvoltage,thestabilityisdeterminedbyapplyingasmallperturbationtothediaphragmandanalyzingthesubsequentdiaphragmmotion.Ifasmallperturbationcausesthediaphragmtostartmovingtowardsthexedbackplatetheequilibriumisunstable,elsetheequilibriumisstable.Todeterminethecriticalgapdistanceandcriticalvoltageatwhichthesystemtransformsfromstabletounstable,wepartiallydifferentiatethenetforceFnetwithrespecttothegap 31

PAGE 32

Ifthediaphragmisperturbedtowardsthebackplate,i.e.,@x<0,thentoensurestableoperationthequantityinparenthesismustbepositivesothat@Fnetisalsonegative.Thus, DenotingthecriticalgapdistancebyxPIandthecriticalvoltagebyVPI,Equation 2 becomes BysubstitutingEquation 2 intoEquation 2 andevaluatingatxPIandVPI,thecriticalgapdistanceis 3x0;(2) andthepull-involtageis 2 whichisrepeatedbelow 2V2dcA x20(^(x)):(2) Itshouldbenotedthattheelectrostaticforceduetoaxedchargedoesnotdependonthepositionofthemovableplate.Thenetforceisgivenby 2Vdc2A x02+1 Differentiatingthenetforce, 32

PAGE 33

2Cm;dVdc2A x02:(2) Iftoolargeavoltageischosen,thenxeq<0whichimpliesthatthetwoplatesaretouchingeachother.FurthermoreifVdcistoolargethentheinitialdeectionofthemovableplatewillbetoolargeandthetransducermightoperateinthenon-linearregion.Todeterminethepull-involtagewesetxeq=0andsolveforthevoltage.Thisisgivenby Comparingthepull-involtagebetweentheconstantchargeandconstantvoltagebiasingtechniques,itcanbeseenthatthepull-involtagelimitishigherfortheconstantchargetechnique. 71 ]asshowninFigure 2-6 .Theneedforadifferentialstructureismotivatedbyaneedtolinearize Figure2-6. Schematicofadifferentialcapacitivetransducer. thesystemusingstructuralmodication.Alsothedifferentialstructurehastwicethechangeincapacitanceofthesinglebackplatestructure.Theimprovementinsensitivityandlinearityisdiscussedlaterinthissection. Thisstructureconsistsofamovableplatesandwichedbetweentwoxedplates.TwocapacitancesC1andC2areformed,onebetweenthetopbackplateandthemovableplateandtheotherbetweenthebottombackplateandthemovableplate.Undertheapplicationofaninput 33

PAGE 34

and Theeffectsofcapacitancemismatchwillbediscussedlaterinthischapter.Fromthegeometry,C1andC2isgivenby x0x0 x0+x0 SubstitutingforC1andC2,weobtain x0x00rA x0+x0:(2) Thiscanbesimpliedto x20x02x0:(2) Forsmalldisplacementsi.e[x0x0],x02canbeneglectedandtheaboveequationcanbesimpliedas x20x0:(2) 34

PAGE 35

2 ,Equation 2 canbesimpliedto FromEquation 2 ,itcanbeobservedthatthecapacitancechangeisalinearfunctionofthechangeingapdistanceforsmalldisplacements.ComparingEquation 2 withEquation 2 itcanbeobservedthatthechangeincapacitanceforthedualbackplatestructureistwicethatofthestructurewithasinglebackplate. Wenowconsiderthevoltageacrosstheplatesofthetwocapacitors.IfthevoltageacrosscapacitorsC1andC2isdenotedbyV1andV2,respectively,andthechargestoredisQ1andQ2,respectively,then 2 showthatthevoltagesacrosstheplatesofacapacitorareanon-linearfunctionofthedisplacementandchargestoredintheplatesofthecapacitor.Similartothecasewithasinglebackplate,theaboveequationscanbelinearizedusingeitheraxedchargeoraxedvoltage.Thisisdiscussednext.i.e, 2-7 .Aunitygainvoltageamplierisconnectedtothemiddleplate.RdcisabiasresistorthatsetstheDCoperatingpoint.Thelinearizedrelationbetweentheinputandoutputofachargeamplierabovethecut-infrequencyisgivenby ThedetailedderivationofthisequationisprovidedinAppendixB. 35

PAGE 36

Schematicofadualbackplatecapacitivetransducerconnectedtoavoltageamplier. 2-8 showsadualbackplatecapacitivetransducerconnectedtoachargeamplier.AfeedbackcapacitorCintisconnected Figure2-8. Schematicofadualbackplatecapacitivetransducerconnectedtoachargeamplier. betweentheoutputoftheamplierandtheinvertinginput.Thiscapacitorintegratesthechargeowingoutofthetransducer.AbiasresistorRdcisconnectedinparalleltothefeedbackcapacitorandisusedtosettheDCoperatingpoint.TheRCnetworkconsistingofRdcandCintsetsupahighpasslterwithacut-infrequency,of1 2RdcCint.Forallfrequenciesabovethecutinfrequencythechargeowingfromthetransducerisintegratedbythefeedbackcapacitor.Thelinearizedrelationbetweentheinputandoutputofachargeamplierabovethecutinfrequencyisgivenby 36

PAGE 37

2-9 showstheschematicofthedualbackplatecapacitivetransducerandtheforceactingonthemovableplate.Theforceonthemovableplatedependsonthetypeofbiasing Figure2-9. Schematicofadualbackplatecapacitivetransducershowingforcesonthemovableplate. techniqueused.Werstconsiderthexedvoltagebiasingtechnique.Thenetelectrostaticforce,Felectrostatic,onthemovableplatecanbeexpressedintermsoftheindividualforces~F1and~F2by UsingEquation 2 ,~F1and~F2aregivenby 2V2dcA 2V2dcA Thetwoforcesoperatingonthemovableplateareactinginopposition.Thisisbecausetheforcebetweenthemovableplateandeachofthexedplatesisattractive.~F1isactinginthenegativexdirection.Thetotalforce~Felectrostaticisgivenbythesumof~F1and~F2.SubstitutingEquation 2 andEquation 2 inEquation 2 andsimplifying,Felectrostaticisgivenby ComparingEquation 2 withEquation 2 ,itcanbeseenthatthenetelectrostaticforcewithaconstantvoltagebiasislowerforadualbackplatecapacitivetransducerascomparedtoasingle 37

PAGE 38

Wenowconsiderthecasewhenthetwoplatesarebiasedwithaxedcharge.Inthiscasetherearetwocomponentstotheforce.Theyaretheforcebetweenthetopbackplateandthemovableplateandtheforcebetweenthebottombackplateandthemovableplate.UsingEquation 2 ,thetwoforcesonthediaphragmare 2 i.e whichiszeroforconstantchargecaseformatcheddualbackplates. 2 andtherestoringforceisgivenbyEquation 2 .Thenetforceisgivenby Inthecaseofadualbackplatestructure,therearetwogapdistances.Hence,weanalyzethenetforcebydifferentiatingitwithrespecttothedistancemovedfromthenominalgapx0insteadofthegapdistance.DifferentiatingFnetwithrespecttox0,weobtain Forthesystemtobestable,asmallperturbationinthechangeingapdistance,i.e@x0>0,mustcauseadecreaseinnetforcei.e.@Fnet<0. 38

PAGE 39

DeningthevoltageatthetransitionpointfromstabletounstableregionofoperationbyVPIandthedisplacementfromequilibriumpositionbyx0PI,wehave SubstitutingEquation 2 inEquation 2 andsolvingforx0PI,wehave Thisshowsthattheequilibriumpositionforadualbackplatestructurebiasedwithaxedvoltageisattherestpositionofthediaphragm.Thepullinvoltageisgivenby 2 iszero.Hencetheequilibriumpositionisx0=0.Basedonthemodelused,thereisnoupperboundonthechargethatcanbeplacedonthetransducerplates.Thisisnotpracticallytrueasthemodeldidnotaccountfornon-idealitiessuchascurvatureoftheplates,orotherdynamicconditions.Basedonthephysicalrealizationofthetransducer,therewillbelimitsonthemaximumchargethatcanbeappliedontheplates. 39

PAGE 40

2-10 .ThemicrophoneswerefabricatedusingtheSandiaSUMMiTVprocess[ 64 ].Thisve-layerpolyprocessallowsthefabricationofultra-planarpolylayerswithlowin-planestress.Twomicrophonestructuresweredesignedforuseinaeroacousticapplicationandforaudioapplication.ThematerialpropertiesanddimensionsofthetwomicrophonestructuresarelistedinTable 2-1 Figure2-10. Crosssectionviewofdualbackplatemicrophone[ 64 ]. Table2-1. Dimensionsofaeroacousticandaudiomicrophone[ 64 ]. PropertyAeroacousticAudio RadiusofTopBackplate256m756mThicknessofTopBackplate2:25m2:25mRadiusofBottomBackplate213m713mThicknessofBottomBackplate2:5m2:5mRadiusofDiaphragm230m730mThicknessofDiaphragm2:25m2:25mThicknessofTopGap2m2mThicknessofBottomGap2m2mRadiusofHoles5m5mNo.ofholesinTopBackplate5575030No.ofholesinBottomBackplate3674361DepthofCavity650m650mRadiusofCavity186m686m Materialpropertiesofmicrophonestructure. PropertyMagnitude Young'smodulusofpolysilicon(e)1:61011PaPoisson'sratioofpolysilicon()0:22Densityofpolysilicon(p)2:2103kg=m3

PAGE 41

46 ].Inthiscase,thelengthscaleofinterestisthediaphragmdiameterandthecharacteristicwavelengthistheacousticwavelength,.UsingLEM,atheoreticalformulationfortheresonantfrequencyanddampingcoefcientisdeveloped.ThegeneraltechniquefordevelopingtheLEMforatransducerwillbediscussednextfollowedbytheapplicationofthetechniquetothemicrophoneunderconsideration. TodeveloptheLEM,weconsiderthepowerowbetweenthevariousmechanicalelementsusingtheprincipleofconjugatepowervariables.Inthistechnique,powerisdenotedastheproductoftwogeneralizedquantities,effortandow.Intheelectricaldomain,theeffortvariableistheappliedvoltageandtheowvariableistheelectriccurrent.Theproductofthesetwoquantitiesisthepower.Intheacousticdomain,theeffortvariableispressure,andtheowvariableisvolumevelocity.Usingtheeffortandowvariablesinthesystem,wenextdiscusstheenergystorageanddissipativeelementsinthesystem. Theratiooftheeffortvariabletotheowvariablerepresentsthegeneralizedimpedanceinthesystem.Intheelectricaldomain,theratiooftheappliedvoltagetothecurrentisequaltotheimpedancewhichifinphaserepresentstheelectricallossesintheelectricaldomain.Similarly,theratiooftheinputpressuretothevolumevelocityistheacousticimpedance.Arealacousticimpedanceisdenotedbyaresistoranditrepresentstheenergylossmechanismsintheacousticdomain.Thephysicalcauseoftheenergylossintheacousticdomainisviscousdampingandradiationresistance. Theenergystorageelementsarediscussednext.Intheelectricaldomain,thecapacitorstorespotentialenergy(energyassociatedwiththeeffortvariable)andtheinductorstoreskineticenergy(energyassociatedwiththeowvariable).Intheacousticdomain,thekineticenergyandpotentialarestoredbythemotionanddeectionoftheelasticdiaphragmrespectively.The 41

PAGE 42

Finally,wediscusstherulesthatgoverntheinterconnectionofthevariouselements.Intheimpedanceanalogy,elementsthatsharethesameowanddisplacementvariableareconnectedinserieswhileelementsthatsharethesameeffortvariableareconnectedinparallel.Also,fortheimpedanceanalogy,byKirchoff'slaws,thesumofalltheeffortsaroundaclosedloopiszero,andthesumofalltheowsenteringanodeiszero. TheLEMofthedualbackplatemicrophonestructureisdiscussednext.AlltheLEMparametersdiscussedareintheacousticdomain.Ideally,thetopandbottombackplatesarerigid.However,physicallytheydodeectbyasmallamount.Hence,theyaremodelledascapacitorsCbp1;Cbp2whichrepresentsthestorageofpotentialenergy.Themainlossmechanisminthebackplatesistheviscousdampingloss[ 72 ].ThesearerepresentedbyRbp1;Rbp2whichareconnectedinparalleltothecorrespondingcomplianceastheybothsharethesameeffortvariable.ThediaphragmrespondstotheinputpressurePinanddeectsstoringpotentialenergyandismodelledasacapacitorCd.ThediaphragmmotionstoreskineticenergyandismodelledasaninductorMd.Theseelementssharethesamevolumedisplacement(owvariable)andarethusconnectedinseries.Viscouslossesassociatedwiththediaphragmareneglected.ThecavityimpedesthemotionofthediaphragmbystoringpotentialenergyandismodelledasacomplianceCcav.Thecavitybetweenthetopandbottombackplatesandthediaphragmisneglected.TheresistanceoftheventismodelledbyasaresistorRv.TheoreticalformulationforeachofthemodelparametersislistedinTable 2-3 [ 73 74 75 ].UsingtheformulationinTable 2-3 ,andthestructuralandmaterialpropertieslistedinTable 2-1 andTable 2-2 respectively,themodelparametersforthetwomicrophonesarecalculated.TheresultsarelistedinTable 2-4 TheLEMofthemicrophonewithapproximationsdiscussedisshowninFigure??.Inthefrequencyrangeofinterest,thetopandbottomplatecompliancesandtheventresistancescanbe 42

PAGE 43

Theoreticalformulationforlumpedelementmodelparameters. ParameterSymbolExpression DiaphragmMassMma9ph 16Eh3CavityComplianceCcavV airc2BackplateHoleResistanceRa12airB(A) Lumpedelementmodelparametersofaeroacousticandaudiomicrophone. Topbackplatecompliance(m4s2=kg)Cbp12:810171:91014Bottombackplatecompliance(m4s2=kg)Cbp26:910189:81015DiaphragmCompliance(m4s2=kg)Cd1:510171:61016CavityCompliance(m4s2=kg)Ccav5:010166:81015 2-11 .TheLCRsystemcanbedescribedbyasecondorder Figure2-11. LowfrequencyLEMofdualbackplatemicrophone. displacementtopressuretransferfunctionH(s)[ 76 ] wheretheresonantfrequencyfresofthesystemisgivenbytheexpression 2s 43

PAGE 44

CcavCd 2 andEquation 2 ,weobtainthe Table2-5. Secondordersystemparametersofmicrophones. QuantityAeroacoustic(Theoretical)Audio(Theoretical)Aeroacoustic(Experimental)[ 76 ] Figure2-12. Theoreticalnormalizedfrequencyresponseofaudiomicrophone. resonantfrequencyanddampingcoefcientforthetwomicrophonedesigns.TheresultsarelistedinTable 2-5 .AplotofthenormalizedtheoreticaltransferfunctionoftheaudioandaeroacousticmicrophoneisshowninFigure 2-13 andFigure 2-13 [ 76 ]. 44

PAGE 45

Theoreticalnormalizedfrequencyresponseofaeroacousticmicrophone. Todesigntheclosedloopcircuit,amodeloftheinterfacecircuitisrequired.Chapter3presentsanoverviewofthevariouscapacitivetransducerinterfacecircuitsalongwithmodelsthatwillbeusedinChapter4andChapter5forthedevelopmentoftheoverallclosedloopsystem. 45

PAGE 46

Figure3-1. OverviewofChapter3. inthepublishedliteratureusingtheseinterfacecircuitsispresented.Thischapterconcludeswithasummaryofthevarioustechniques.AgraphicaloverviewofthischapterisprovidedinFigure 3-1 77 ]ofcapacitivetransducerswithanemphasisoncapacitivemicrophonesisintroduced.Thesensitivity,S,ofacapacitivetransducerisdenedasthechangeintheoutputvoltageforaninputphysicalphenomenon-derivedchange 46

PAGE 47

Theemphasisofthischapterisoninterfacecircuitswhichmeasurecapacitance.Hence,theelectricalsensitivityoftheinterfacecircuitisexpressedinV=F.Itshouldbenotedthatthesensitivityisdenedasthemagnitudeoftheatregionofthefrequencydependenttransferfunction.Foralineartransducer,thetransferfunctionH(!)isdenedastheratiooftheFouriertransformoftheoutputoftransducerY(!)totheFouriertransformoftheinputofthetransducerX(!),i.e. Thetransferfunctionisingeneralacomplexquantity.Hence,thecompletedescriptionincludesboththemagnitudejH(!)jandphase\H(!)jasafunctionoffrequency.Themagnitudeofthetransferfunction,asdenedinEquation 3 ,isthusaplotofthesensitivityofthetransducerasafunctionoffrequency.Foranidealtransducer,thephasemustbeconstantandzero. ThetypicalfrequencyresponseofacapacitivetransducerisshowninFigure 3-2 .Itcanbeseenthatoveracertainfrequencyrange,thefrequencyresponsemagnitudeisatwithinagiventolerance,say3dB.Foratransducer,thisistherangeoffrequenciesoverwhichitcanbeusedandisdenedasthebandwidthofthesensor.Thelowerandhigherrolloffofthetransferfunctioniscausedbyboththetransducerstructureandtheinterfaceelectronics.Foramicrophone,thelowerendoffrequencyrange(denotedbyfl),alsoknownasthecutonfrequency,isdeterminedbyeitherthebiasingresistoroftheinterfacecircuitortheacousticresistanceoftheventchannelandcavitycompliance[ 53 ].Similarly,thehigherendofthefrequencyrange(denotedbyfh),alsoknownasthecutofffrequency,isdeterminedbythe 47

PAGE 48

Characteristicfrequencyresponseofacapacitivetransducer. bandwidthoftheinterfacecircuitorthemechanicalresonanceofthesensorstructurewhichinturnisdeterminedbythediaphragmmassandcompliance. Thenextperformancemetrictobeconsideredisthelinearityofthesensor.Thelinearityofthesensorisameasureofhowclosetheoutputversusinputcalibrationcurveapproximatesastraightlineatagivenfrequency.AplotoftheidealandactuallinearityofatransducerisshowninFigure 3-3 .Theslopeofthestraightlineprovidesthesensitivityofthetransducerat Figure3-3. Characteristicidealandactualresponseofacapacitivetransducer. thatfrequency.Itcanbeseenthatathighinputamplitudes,theoutputofthetransducerdeviatesfromtheidealstraightlinecurve.Thelowerandhigherendsofthelinearrangearedeterminedbyboththesensorinterfacecircuitandthesensor.Thelowerendislimitedbysystemnoisesuchasthermalnoise,1=fnoise,andmechanicalnoise.Thehigherendofthelinearrangeisdeterminedbystructuralnon-linearitiessuchasspringstiffening[ 78 ]orbycircuitnon-linearities 48

PAGE 49

79 ].Thedynamicrangeofacapacitivetransducerisdenedastheratioofthemaximumandminimuminputsignalofthelinearrange. Thedeviationoftheoutputfromtheideallinearcurvecausesdistortioninthemicrophoneoutput.Distortioncanbedenedintwodifferentwaysbasedontheinputsignal.Whenthesystemisexcitedatasinglefrequency,distortionisdenedastheminimuminputamplitudethatcausestheoutputtodeviatefromlinearitybyaxedpercentage.Forameasurementmicrophone,thedeviationisusuallychosentobe3%.Whenthetransducerischaracterizedoverafrequencyrange,thedistortionisdenedintermsofthetotalharmonicdistortion(THD).THDisdenedastheratioofthesumofthepowerinalltheharmonicfrequencies(!n)tothefundamentalfrequency(!0). Theoutputnoiseoorofatransduceristhenoisemechanismgeneratedoutputforzeroinputsignal.Referredtotheinput,theinputnoiseoorrepresentstheminimumdetectablesignal(MDS).Theoutputreferrednoisevoltagepowerspectraldensity(PSD)oftheinterfacecircuitdenotedbySVoutisexpressedinV2=Hz.Theelectricalnoiseattheoutputoftheinterfacecircuitcanbereferredtotheinterfacecircuitinputbydividingtheoutputnoisebythesquareoftheelectricalsensitivity.TheinputsignalPSDoftheinterfacecircuitisexpressedinF2=Hz.Thus,theoutputnoisepowerinasystemisdependentonthebandwidthofoperation.ThetotaloutputnoisepowerormeansquareoutputnoisevoltageisobtainedbyintegratingtheoutputnoisevoltagePSDoverthemeasurementbandwidthi.e. andthermsoutputnoisevoltageisvn;rms=p 3-4 .Theoverallsystemnoiseincludesnoisecontributionsfromboththesensorandtheinterfacecircuit.Atlowfrequencies,thenoisePSDisdominatedbytheickernoise[ 80 ]oftheinterfacecircuitwhichhasa1=fspectralshape.Itshouldbenotedthatickernoiseispresentonlywhenadccurrentows.Athighfrequencies,thenoiseisdominatedbythethermalnoise[ 80 ]aboveacornerfrequency.Thecornerfrequencyisthepointatwhichtheasymptoteofthe1=fnoiseregionmeetsthethermalnoiseoor.Thermalnoiseispresentinallsystemsthat 49

PAGE 50

Characteristicnoiseoorofacapacitivetransducerusingavoltageamplierinterfacecircuit. areinthermodynamicequilibrium.ThemagnitudeofthethermalnoisePSDisproportionaltothedissipationinthesystem. Tostudytheeffectoftheinterfacecircuitnoiseonoverallsystemperformance,thenoisemodelsdevelopedinthischapterconsideronlytheelectronicsnoiseoftheinterfacecircuitelementsandthespectralshapingofthesenoisesourcesbytheinterfacecircuitelements.Thecommonlyusedelementsintheinterfacecircuitareresistors,capacitorsandoperationalampliers(opamps). Thenoisemodeloftheresistorisconsideredrst.The1=fnoiseofaresistorisdependentonthefabricationmethodusedtomanufacturetheresistor.Metallmresistorshavethelowest1=fnoiseandhaveaatfrequencyresponseacrosstheentirespectrum.Figure 3-22 showsthevoltageandcurrentnoiserepresentationsofaresistor.ThePSDofresistorvoltageandcurrentnoiserepresentationsisgivenby and R(3) respectively. Ideally,capacitorsdonothaveanynoiseassociatedwiththem.Inpracticalapplications,capacitorshaveanequivalentseriesresistor(ESR)whichhasanoiseassociatedwithit.TheESRofacapacitorissmallincomparisonwithotherphysicalresistorspresentinthesystem.HencethenoiseofthephysicalresistordominatesoverthenoiseduetotheESR.Hence,noiseincapacitorsarenotconsideredinthediscussionsinthischapter. 50

PAGE 51

Voltagenoisemodelofresistor. (b) Currentnoisemodelofresistor. Noisemodelofresistor. Thenoisemodelfortheopampisdiscussednext.Figure 3-6 showsthenoisemodelofanopampconsistingofavoltageandcurrentnoisevoltagesourceatthenon-invertinginputoftheamplier.Thefundamentalphysicalphenomenawhichcausenoiseintheopamparethethermalnoise,shotnoise,and1=fnoiseinthetransistorsandresistorsinsidetheopamp.ThetypicalnoisevoltagePSD(Sv)andcurrentnoisePSD(Si)isshowninFigure 3-7 .ItcanbeseenthatatlowfrequencythePSDhasa1=fshapeandatlargefrequency,thePSDisconstant.Thepointatwhichtheasymptoteofthe1=fnoisemeetsthethermalnoiseooristhecornerfrequencyandisdenotedbyfcvandfciforthevoltageandcurrentnoisePSDsrespectively.Atanygivenfrequency,thevoltageandcurrentnoisePSD,Sv(f)andSi(f)respectively,canbeapproximatelyexpressedintermsofthecorrespondingvoltageandcurrentthermalnoiselimits,SvoandSioby and Figure3-6. Noisemodelofopampshowingvoltageandcurrentnoisesource. 51

PAGE 52

PlotofvoltagenoisePSDofamplier. (b) PlotofcurrentnoisePSDofamplier. PlotofnoisePSDofamplier. Theperformancemetricsdiscussedsofarwillbederivedforthevariousopenandclosed-loopinterfacetechniques.Theanalogopen-loopsenseschemesarecharacterizedbytheabsenceofaclocksignalascomparedtothedigitalopenloopsensetechniques.Thecommonlyusedanalogopen-looptechniquessensethevoltagechangeusingavoltageamplierunderaconstantchargebias,sensethechargeusingachargeamplierunderaconstantvoltagebias,andsenseusingsynchronousmodulation/demodulation.Eachofthesetechniquesisdiscussedindetail.Foreachoftheseinterfacetechniques,theprincipleofoperation,formulationforperformancemetrics,tradeoffs,andaliteraturereviewispresented.Figure 3-8 showsaowdiagramforeachsection.ThederivationoftheequationsforeachoftheinterfacetechniquesispresentedinAppendixB. OperatingPrinciple. 3-9 showsaschematicofadualbackplatecapacitivemicrophoneconnectedtoavoltagebufferincludingparasitics.Cprepresentsthetotalparasiticcapacitanceattheinputtotheamplierduetothepackaging.TheparasiticcapacitanceofcommonlyusedpackagingtechniquesispresentedinTable 3-1 .Forasinglechip 52

PAGE 53

OrganizationofeachinterfacecircuitdiscussedinChapter3. Figure3-9. Schematicofadualbackplatemicrophoneconnectedtovoltagebuffer. solutioninwhichboththemicrophoneandinterfacecircuitryareonthesamedie,thesensorandinterfacecircuitryareelectricallyconnectedusingmetalorpolylinesandtheparasiticorinputcapacitanceis<1pF.Fortwochipsolutions,theparasiticcapacitanceisontheorderofafewpFdependingontheinterconnectusedsuchasbondwiresorPCBtracks.Cirepresentsthetotalinputcapacitanceoftheamplier.Foranidealinterfacecircuitwithnoparasiticcapacitance,theoutputofthecircuit,foragivenchangeincapacitance,C,isgivenby, 53

PAGE 54

Parasiticcapacitanceforvariouspackagingtechnologies. PolysiliconormetallinesOndieinterconnect<10fF=um(Layoutdependent)Wirebond[ 81 ]Interconnectbetweenunpackageddies30fF=mmPCBTrace(FR4PCB)Interconnectbetweenpackageddies12pF=cm(Layoutdependent)BNCcable(RG164[ 82 ]])Interconnectbetweendifferentboards20:6pF=ft Equation 3 canberewrittenas DeningtheoverallvoltageattenuationfactorHcas Equation 3 canbewrittenas ComparingEquation 3 andEquation 3 ,itcanbeseenthattheoutputvoltageisattenuatedbythefactorHcthatisdependentontheparasiticcapacitance. 3.2 andusingEquation 3 ,theelectricalsensitivitySeofthecircuitisgivenby Fromthisequation,itcanbeseenthatthesensitivityofthecircuitisreducedwithincreasedparasiticcapacitance. Todeterminethetotaloutputnoise,thenoisemodelshowninFigure 3-10 isused.ThePSD 54

PAGE 55

Noisemodelofdualbackplatemicrophoneconnectedtovoltagebuffer. ofthecircuitSVoutisdeterminedbytheampliernoiseandthethermalnoiseofthebiasresistorandisgivenby whereZi=1 Toobtaintheminimumdetectablecapacitanceperp FromtheequationforCmin,itcanbeseenthattheminimumdetectablecapacitancecanbedecreasedbyincreasingthebiasvoltage.Itcanalsobeseenthattheparasiticcapacitanceincreasestheminimumdetectablecapacitance. Theupperendofthecircuitdynamicrangeislimitedbythemaximumoutputvoltageoftheamplierandisdependentonthetopologyoftheamplierused.ForagivenpowersupplyVdd,themaximumoutputvoltageVomaxisdeterminedbyacircuitdependentvoltagedropVdropbelowthesupplyrailsVddandisgivenbytheexpression Forexample,foranactivelyloadedsinglestageamplier,theminimumVdropisgivenbytheoverdrivevoltageacrosstheloadtransistor.Dividingtheaboveequationbytheelectrical 55

PAGE 56

Thisshowsthattheupperendofthedynamicrangecanbeincreasedbydecreasingthetopandbottomplatevoltageandincreasingthesensecapacitance. Thenextperformancemetricthatwederiveisthecircuitbandwidth.Thelowerlimitofthefrequencyresponseissetbythetotalcapacitanceattheinvertingnodeandthebiasresistor.[ 83 ].Itisgivenbytheexpression 2Rdc(Cp+Ci+2C0):(3) Theupperlimitofthebandwidthfhofthiscircuitisdeterminedbytheamplier.Forafoldedcascodeamplier,thebandwidthisdeterminedbytheunitygainfrequency,andisexpressedintermsoftheloadcapacitanceCLandtransconductancegmbytheexpression[ 79 ] 2gm 3-11 fordifferentsensorcapacitances.Itcanbeseenthatassensorcapacitanceisscaleddown,theeffectofparasiticcapacitanceisexacerbated.ThenoiseooroftheinterfacecircuitgivenbyEquation 3 isdependentonthelteringoftheamplierandresistornoisebythesenseandparasiticcapacitance.Furthermore,thechoiceofampliersalsomakesadifferenceasampliersareoptimizedforeitherlowcurrentnoiseorlowvoltagenoiseandnotboth.Figure 3-12 showsaplotofthenoisecontributionsbythevariousnoisesourcesandthetotaloutputnoise. Tounderstandtheeffectofthevariouscircuitelements,ontheoutputnoisePSD,Figure 3-13 andFigure 3-14 areplotted.Figure 3-13 showsasetofplotsoftheoutputnoisePSDasthesensorcapacitanceisscaledfrom100pFto1pFforvariousparasiticcapacitancesassumingaxedbiasresistorof10M. 56

PAGE 57

Variationofsensitivityasafunctionoftotalparasiticcapacitanceforvarioussensecapacitances Figure 3-13(a) ,Figure 3-13(c) ,andFigure 3-13(e) showtheseplotsforanamplierwithlowvoltagenoiseandFigure 3-13(b) ,Figure 3-13(d) ,andFigure 3-13(f) showtheseplotsforalowcurrentnoiseamplier.Itcanbeseenthatforaxedbiasresistor,anamplieroptimizedforlowcurrentnoisehasalowernoiseoorascomparedtoanamplierwithlowvoltagenoise.Itcanalsobeseenthatthisistrueforallcombinationsofsensorandparasiticcapacitance. Figure 3-14 showsasetofplotsoftheoutputnoisePSDcapacitancesasthesensorcapacitanceisscaledfrom1pFto100pFforvariousbiasresistorsassumingaxedparasiticcapacitorof10pF.InFigure 3-14(a) ,Figure 3-14(c) ,Figure 3-14(f) showtheseplotsforanamplierwithlowvoltagenoiseandFigure 3-14(b) ,Figure 3-14(d) ,Figure 3-14(e) showtheplotsforlowcurrentnoise.Itcanbeseenthatforaxedparasiticcapacitor,anamplieroptimizedforlowcurrentnoisehasalowernoiseoorascomparedtoanamplierwithlowvoltagenoise.Itcanalsobeseenthatthisistrueforallcombinationsofsensorcapacitanceandbiasresistor. 57

PAGE 58

Powerspectraldensityofindividualnoisecontributorsandtotaloutputnoiseinavoltageamplierbasedinterfacecircuit.(LowVoltageNoiseAmplier(OP471)Sv0=10nV=p applications.Table 3-2 ,Table 3-3 ,andTable 3-4 listthepropertiesofcommercialandresearchprototypevoltageampliersusedforcapacitivemicrophones.OneofthecommonlyusedcommercialmeasurementmicrophoneforaeroacousticapplicationsistheBruelandKjaerType4138microphone[ 38 ].Thenominalcapacitanceofthismicrophoneis6:2pF.TheinterfacecircuitusedistheType2670voltageamplier. OneofthecommonlyusedcommercialmeasurementmicrophoneforaeroacousticapplicationsistheBruelandKjaerType4138microphone[ 38 ].Thenominalcapacitanceofthismicrophoneis6:2pF.TheinterfacecircuitusedistheType2670voltageamplier.TheType2670amplierhasaatfrequencyresponsefrom15Hz200kHzandanoiselevelof14Vintegratedovera30Hzto300kHz.Theinputcapacitanceofthepreamplieris0:5pF.Thesemicrophonearebiasedwith200VusingaBruelandKjaerType2804microphonepowersupplyunit. 58

PAGE 59

LowVoltageNoise(Cp+Ci=1pF) (b) LowCurrentNoise(Cp+Ci=1pF) (c) LowVoltageNoise(Cp+Ci=10pF) (d) LowCurrentNoise(Cp+Ci=10pF) (e) LowVoltageNoise(Cp+Ci=100pF) (f) LowCurrentNoise(Cp+Ci=100pF) Plotofoutputnoiseforvoltageamplierwithaxedbiasresistor(10M).(LowVoltageNoiseAmplier(OP471)Sv0=10nV=p 59

PAGE 60

LowVoltageNoise(Rdc=100k) LowCurrentNoise(Rdc=100k) LowVoltageNoise(Rdc=10M) LowCurrentNoise(Rdc=10M) LowVoltageNoise(Rdc=1G) LowCurrentNoise(Rdc=1G) Plotofoutputnoiseforvoltageamplierwithaxedparasiticcapacitanceof(10pF).(LowVoltageNoiseAmplier(OP471)Sv0=10nV=p 60

PAGE 61

3-2 showsalistofcommonlyusedBruelandKjramplierssuitableforaeroacousticmeasurementandtheircharacteristics. Table3-2. SpecicationofVoltageAmpliersforBruelandKjaerMicrophones. Type26600:3pF0:8V20Hz-200kHzType2669B0:3pF1:9V20Hz-200kHzType26700:3pF4:0V30Hz-300kHz KnowlesAcousticssellscommercialMEMSbasedmicrophonesforaudioapplicationsnamely,theSiSonicseriesofmicrophones[ 5 ].TheyconsistofaMEMSmicrophonewirebondedtoavoltageamplier.Theinterfacecircuitryoperatesona1.5Vsource.Thehighvoltage(11V)requiredforbiasingthemicrophoneisobtainedbyanon-chipchargepump.Thecircuithasabandwidthinexcessof10kHz.Theperformancemetricsoftheamplieritselfarenotprovided. SonionTechnologiessellsMEMSmicrophonesforcellphoneapplicationsnamelytheTC100seriesmicrophones[ 47 84 ].ItconsistsofaPMOSbufferampliercascadedwithaNMOSgainstage.Thechipispoweredbya1:8Vsupply.TheoutputoftheamplierisprocessedviaaADCandthusthemicrophonehasadigitaloutput. TheNationalSemiconductorsLV10XXseriesofampliers[ 85 ]isspeciallydesignedforelectretcondensermicrophonesandconsistsofahighgain,lowinputcapacitancestagefollowedbyanoptionalsigmadeltabasedanalogtodigitalconverterthusprovidingbothanaloganddigitaloutputs.Thecircuitrunsona1:8to5Vsupply.Theinputcapacitanceis2pFovertheentirepowersupplyrange.Thegainvariationis8V=Vto25V=Vovertheentireproductfamilyandpowersupplyoptions.Theseampliershaveaatfrequencyresponsefrom100Hzto110kHz.Theinputnoiseoftheseampliersisapproximately500nV=p 61

PAGE 62

3-3 presentsanoverviewofvoltageampliercircuitsusedinresearchprototypecapacitivemicrophones.Ithasbeenobservedthatinthesepublications,theoverallmicrophonesensitivityinV=Paisreportedandnottheelectricalsensitivity(V=F).Hence,thegainreportedinTable 3-3 isthesensitivityoftheamplieraloneandisexpressedinV/V.ThegeneraltrendintheinputstageofvoltageamplierimplementationusingcustomICsistheuseofunitygainsingletransistorCMOSstages.Thisisdonetominimizeinputparasiticcapacitanceandinputcurrentnoise. Inthepublishedliterature,oneoftheearliestsemiconductorvoltagefollowerbasedcapacitivemicrophoneinterfacecircuitwasreportedinapatentbyKillionetal.in1970[ 86 ].ThecircuitconsistedofaJFETvoltagefollower.Back-to-backdiodeswereusedforbiasingthemicrophonecapacitance. In1989,Murphyetal,[ 87 ]presentedasiliconelectretmicrophonethatusesaCMOSpreamplifer.Theamplierhadareportedgainof0:9V=V.Theoveralldevicehadabandwidthof15kHz.Thebiasvoltageusedwasnotreported.In1991,Bergqvistetalpresentedasinglebackplatemicrophonewithaperforatedbackplate.Themicrophonecapacitancewas5pFandwasbiasedwitha5Vdcbattery.AmodiedBruelandKjrType2619wasusedastheinterfacecircuit.Theinputparasiticcapacitanceoftheinterfacecircuitwas2pF,andthebandwidthwasfrom20Hzto20kHz.In1995,Ningetal.[ 88 ],presentedavoltagebufferbasedinterfacecircuitforamicromachinedcapacitivemicrophonetobeusedincellphones. In1995,Ningetal.[ 88 ],presentedavoltagebufferbasedinterfacecircuitforamicromachinedcapacitivemicrophonetobeusedincellphones.Thenominalcapacitanceofthesensorwas9:1pF.Theinputparasiticcapacitanceofthecablesandtheinterfacecircuitwas12pFwiththeampliercontributing2pF.Themicrophonewasbiasedatadcpotentialof6V. In1996,BernsteinandBorensteinpresentedtherstMEMSmicrophonewithonchipinterfacecircuitry[ 18 ]in1996.ItconsistedofaJFETsourcefollowerwithaninputparasiticcapacitanceof0:5pF. In1998,PedersenandBergveldreportedaMEMSamplier[ 59 ]withintegratedCMOSamplieranddc-dcconverter.Theamplierinputconsistedofap-MOSFETbasedvoltagefollower.Ap-MOSFETwasusedbecauseofitslowernoiseoorascomparedtoann-MOSFET. 62

PAGE 63

ReviewofinterfacecircuitsusedinMEMSmicrophonespublishedinliterature. Killion,1970[ 86 ] NR,NR NR NR Murphyetal,1989[ 87 ] NR,NR 15kHz Sprenkelsetal.,1991[ 89 ] Kuhneletal.,1992[ 90 ] 20kHz Bergqvistetal,1994[ 91 ] 20Hz20kHz Ningetal,1994[ 88 ] Bernsteinetal,1996[?] NR NR Hsual,1997[ 14 ] Kabiral,1999[ 14 ] NR,NR <1 100Hz9kHz Bakeretal,2003[ 92 ] NR,NR NR Kernetal,2004[ 93 ] NR,NR NR NR Furstetal,2005[ 84 ] <10pF,NR NR NR Loeppertetal,2006[ 94 ] <1pF,0:25pF 95 ].Thedevicehadazerobiascapacitanceof16:2pF.TheinterfacecircuitusedwasaHPACOP4012preamplier.Ithadaninputimpendanceof2:5G.Theinputcapacitanceisnotreported.Themicrophonewaspolarizedat10V. In1999,Kabiretal.reportedaMEMSmicrophonethatusedadopedp+polyasthemembraneandgoldbackplates.Thedevicehadazerobiascapacitanceof2:4pF.During 63

PAGE 64

In2003,BakeretalreportedaJFETinputpreamplierthatusedthecurrentoutputofasourcefollowerasopposedtothetraditionaloutputvoltage[ 92 ].Theamplierwasdesignedformicrophonesusedincochlearimplants.Theamplierwasoptimizedforhighpowersupplyrejectionratio(PSRR)whichisarequirementforlownoisedesign.Thecircuitoperatedona3Vpowersupplyandhadaatfrequencyresponseupto10kHz. In2004,Kernetalobtainedapatentforamicrophoneinterfacecircuit.Thecircuitbiasesthemicrophoneusinglargeresistorsfromback-to-backdiodes.Otherperformancemetricsofthecircuitwerenotdisclosedinthepatent. In2005,Furstetal[ 84 ]ofSonionTechnologyobtainedapatentforatwostagepreamplierforanelectretcondensermicrophone.ItconsistsofaPMOSunitygainsourcefollowerdrivingaNMOSgainstage.Thecircuithasaninputcapacitanceoflessthan10pF.Usingatwostagetopologyallowstheindependenttuningoftheampliergainandtheinputcapacitance.Thepreamplierisdesignedforuseintelephonyapplications. In2006,Loeppertetal[ 94 ]presentedtherstcommercialMEMSmicrophone.Theinterfacecircuitconsistsofavoltagebufferwitha0:25pFinputcapacitancefollowedbyagainstage.Themicrophoneisbiasedbya11Vdcsignalthatisgeneratedbyachargepumpthatisonthesamedieastheamplier. GeneralpurposeoperationalamplierscanalsobeusedasinterfacecircuitsforcapacitivemicrophonesusinganyofthepackagingtechniqueslistedinTable 3-1 .ComparedtocustomICs,theinputparasiticcapacitanceoftheseampliersarecomparativelyhigherandisfurtherincreasedbythepackagingtechnologyusedtoconnecttheampliertothemicrophone.Theseampliersareusuallyprovidedinopenloopconguration.Thus,carefulcircuitdesign,whichmayresultinaddingmorecomponentssuchasresistorsandcapacitorsisrequiredfortheproperoperationoftheseampliers.Theseextracomponentsmayintroduceparasiticcapacitancethathavetobetakenintoconsiderationduringtheinterfacecircuitdesign.Asetoperational 64

PAGE 65

3-4 alongwiththeirperformancespecications. Table3-4. Specicationofarepresentativesampleofoperationalampliersrecommendedbymanufacturersforuseasvoltagefollowersforcapacitivetransducers. NR NR OperatingPrinciple. 3-15 .Theinvertinginputisconnectedtothemiddleplateofthemicrophoneandthenon-invertinginputisgrounded.RDCisabiasresistorthatsetsthedcoperatingpointoftheamplierinput.Themiddleplateisbiasedat0Vbytheprincipleofvirtualshortbetweenthenon-invertingterminalandinvertingterminaloftheamplier.ThechargeowingthroughthemiddleplateisintegratedbythefeedbackcapacitorCint.Thetopandbottomplatesarebiasedat+VdcandVdcrespectively.Theinput-outputrelationofthecircuitforacapacitancechangeisgivenbytheequation 3 ,theelectricalsensitivityofthecircuitisgivenbytheequation whichistwicethatofasinglebackplatemicrophone.Fromthesensitivityequation,itcanbeinferredthatthesensitivitycanbeincreasedbyincreasingthebiasvoltageordecreasingthe 65

PAGE 66

Schematicofdualbackplatemicrophoneconnectedtoachargeamplier. integratingcapacitor.TodeterminethetotaloutputvoltagenoisePSDSVout,thenoisemodel Figure3-16. Noisemodelofdualbackplatemicrophoneconnectedtochargeamplier. showninFigure 3-16 isused.DeningCtot=2C0+Cp+CiZi=1 Forfrequenciesabovethecutonfrequencyi.e.!>1 (!Cint)2:(3) 66

PAGE 67

(!Cint)2:(3) FromtheequationforCmin,itcanbeseenthatthenoiseoorisincreasedbytheparasiticcapacitance.Furthermore,thecurrentnoisePSDduetothebiasresistorandthecurrentnoiseoftheamplierisshapedas1=!2whichincreasesthenoiseatlowfrequenciesconsiderably. Theupperendofthedynamicrangeislimitedbythemaximumoutputvoltageoftheamplierandisdependentonthetopologyoftheamplierused.ForagivenpowersupplyVdd,themaximumoutputvoltageVomaxisdeterminedbyacircuitdependentvoltagedropVdropbelowthesupplyrailsVddandisgivenbytheexpression Dividingtheaboveequationbytheelectricalsensitivity,themaximuminputcapacitancethatcanbemeasuredis Thisshowsthattheupperendofthecircuitdynamicrangecanbeincreasedbydecreasingthetopandbottomplatevoltageandincreasingtheintegratingcapacitance. Thelowerlimitofthefrequencyresponseissetbythefeedbackcapacitanceandthefeedbackresistor[ 83 ].Itisgivenbytheexpression 2RdcCint:(3) Theupperendoffrequencyresponseofthiscircuitisdeterminedbythebandwidthoftheamplierandisdependentontheampliertopology.Forexample,forasinglestagefoldedcascodeamplierwithatransconductancegm,thebandwidthisgivenbytheexpression 2gm Fromtheequationforfh,itcanbeinferredthatthebandwidthisalsoincreasedwhentheCintisdecreased. 67

PAGE 68

ThemaindisadvantageofthiscircuitistheneedforalargedcbiasresistorRdcbecausethenoisePSDduetothisresistoris1=f2shapedwhichconsiderablyincreasesthenoiseatlowfrequency.ChoosingRdcisatradeoffbetweennoiseandthelowerendofthefrequencyresponsefl.ImplementingalargeRdcresistorisnon-trivialinanintegratedcircuitprocess.Optionsincludethicklmresistors,back-to-backdiodes,orlowresistancepolyresistors.Eachofthesetechniqueshasitsownadvantagesanddisadvantages[ 8 ].Figure 3-17 showsaplot Figure3-17. Powerspectraldensityofindividualnoisecontributorsandtotaloutputnoiseinachargeamplierbasedinterfacecircuit.(LowVoltageNoiseAmplier(OP471)Sv0=10nV=p ofthePSDofvariousnoisecontributorsinachargeamplierandthetotaloutputreferrednoise.Itcanbeinferredthatatlowfrequenciestheresistorcurrentnoiseisthedominantnoisemechanismandathigherfrequenciestheampliervoltagenoiseisthedominantnoisemechanism.Figure 3-18 andFigure 3-19 plottheeffectofscalingvariouscircuitelements,on 68

PAGE 69

3-18 showsasetofplotsoftheoutputnoisePSDasthesensorcapacitanceisscaledfrom100pFto1pFforvariousparasiticcapacitanceassumingaxedbiasresistorof10M.Figure 3-18(a) ,Figure 3-18(c) ,andFigure 3-18(e) showtheseplotsforalowvoltagenoiseamplierandFigure 3-18(b) ,Figure 3-18(d) ,andFigure 3-18(f) showtheseplotsforalowcurrentnoiseamplier.Itcanbeseenthat,foralargexedbiasresistor(10M),anamplieroptimizedforlowcurrentnoisehasalowernoiseoorascomparedtoanamplierwithlowvoltagenoise.Itcanalsobeseenthatthisistrueforallcombinationsofsensorandparasiticcapacitances.Figure 3-19 showsasetofplotsoftheoutputnoisePSDasthesensorcapacitanceisscaledfrom100pFto1pFforvariousbiasresistorsassumingaxedparasiticcapacitorof10pF.Figure 3-19(a) ,Figure 3-19(c) ,andFigure 3-19(f) showtheseplotsforanamplierwithlowvoltagenoiseandFigure 3-19(b) ,Figure 3-19(d) ,andFigure 3-19(e) showtheplotsforalowcurrentnoiseamplier.Itcanbeseenthat,foraxedparasiticcapacitance,anamplieroptimizedforlowcurrentnoisehasalowernoiseoorascomparedtoanamplierwithlowvoltagenoisewherethedifferenceinlargeforalargebiasresistor. 3-5 In1999,Amendolaetal.[ 96 ]reportedachargeamplier-basedinterfacecircuitforacapacitivemicrophone. Thetargetapplicationsforthemicrophonewasmeasurementsinsidearesonantcavitywhichhadanarrowbandwidthcenteredat25kHz.Theamplierhadasimulatedbandwidthfrom19Hzto1:4MHz.Thesimulatedsensitivityofthecircuitwas2V=pCandameasuredsensitivityof0:3V=pC. 69

PAGE 70

LowVoltageNoise(Cp+Ci=1pF) (b) LowCurrentNoise(Cp+Ci=1pF) (c) LowVoltageNoise(Cp+Ci=10pF) (d) LowCurrentNoise(Cp+Ci=10pF) (e) LowVoltageNoise(Cp+Ci=100pF) (f) LowCurrentNoise(Cp+Ci=100pF) Plotofoutputnoiseforchargeamplierwithaxedbiasresistor(10M).(LowVoltageNoiseAmplier(OP471)Sv0=10nV=p 70

PAGE 71

LowVoltageNoise(Rdc=100k) LowCurrentNoise(Rdc=100k) LowVoltageNoise(Rdc=10M) LowCurrentNoise(Rdc=10M) LowVoltageNoise(Rdc=1G) LowCurrentNoise(Rdc=1G) Plotofoutputnoiseforchargeamplierwithaxedparasiticcapacitanceof(10pF).(LowVoltageNoiseAmplier(OP471)Sv0=10nV=p 71

PAGE 72

97 ]reportedadualbackplateMEMScapacitivemicrophoneforuseinaeroacousticapplicationsthatusedacommercial-off-the-shelf(COTS)ThermopticsDN620chargeamplierwhichhasasensitivityof1V=pCandabandwidthof100kHz.Thefeedbackelementsconsistedofa1pFcapacitorinparallelwitha1Gresistor. Table3-5. Reviewofchargeamplierinterfacecircuitsusedincapacitivemicrophones. Amendolaetal.1999[ 96 ] Microphone0:7pF Martinetal.2005[ 97 ] Microphonedifferentialcapacitorwith1pF,0:7pFnominalcapacitance 98 ]chargeamplier 3-6 Table3-6. Specicationofarepresentativesampleofgeneralpurposeoperationalampliersrecommendedbythemanufacturerforuseaschargeampliersusingafeedbackresistorandcapacitor. OPA128(TexasInstruments)[ 99 ]DiFET27nV=p 100 ]DiFET8nV=p 101 ]FET3:2nV=p OperatingPrinciple. 65 46 ]ofthecircuitbymodulatingthecapacitivesignaltoahighcarrierfrequencyabovethelow1=fcornerfrequencyoftheamplier.Thishighfrequencysignalisthenampliedanddemodulated.Duringthedemodulationandsubsequentlowpasslteringprocess,theoriginalsignalisrecoveredandthenoiseislteredout. 72

PAGE 73

Schematicofcapacitivemicrophonewithasynchronousmodulationanddemodulationtechniqueusingavoltageamplierbasedcircuit. Inthistechnique,thetopandbottomplatesareexcitedbyasinusoidalsignalofequalmagnitude(Vac)but180degreesoutofphase.Theexcitationfrequency(!c)ischosentobemuchhigher(100)thantheresonantfrequencyofthesensor.Thisisdonesoastopreventtheelectrostaticexcitationofthesensorbytheacwaveform.Theinterfacecircuitusedcanbeeitheravoltageamplierorachargeamplier.AschematicofthesystemusingavoltageamplierandchargeamplierinterfacecircuitisshowninFigure 3-20 andFigure 3-21 respectively. Figure3-21. Schematicofacapacitivemicrophonewithasynchronousmodulationanddemodulationtechniqueusingachargeamplierbasedcircuit. Theamplitudemodulated(AM)signalisconvertedbacktoabasebandsignalusingasynchronousdemodulator.Thecarrierfrequencyusedtodemodulatetheoutputofthechargeamplierisphaselockedtothemodulatingasignalappliedtothetopandbottombackplate.The 73

PAGE 74

Spectrumofvarioussignalsinvolvedinsynchronousmodulation/demodulationtechnique (b) Spectrumofsignalsatthemiddleplateofthemicrophone. (c) Spectrumofsignalatoutputofdemodulator (d) Spectrumofsignalatoutputoflowpasslter. PlotofspectrumatvariousnodesofFigure 3-21 synchronousdemodulatorconsistsofananalogmultiplierfollowedbyalowpasslter.Thiscircuitovercomesthenoiseduetoamplier1=fnoiseandthe1=f2shapedcurrentnoiseofthebiasresistorandcurrentnoiseoftheamplierbymodulatinghighfrequencycarrierfrequencywiththelowfrequencyacousticsignal.Thissignalisthenamplied,demodulated,andlowpasslteredtoobtaintheoriginalsignal.Duringthedemodulationprocessthelowfrequency1=fand1=f2noisearemodulatedtothecarrierfrequencywhichisthenlteredoutwhiletheacousticsignalatthecarrierfrequencyistranslatedtothebasebandandpassesthrough.Therelationbetweenthecapacitanceinputandthelowpasslteredoutputofthistechniqueisgivenby[ 46 ] foravoltageamplierandby forachargeamplierwhereU=1Visascalefactor.ThedimensionofEquation 3 andEquation 3 needsconsideration.Inpractice,thedemodulator[ 102 ]isimplementedasanan 74

PAGE 75

U;(3) whereZistheoutputofthemultipliermeasuredinVolts,XandYaretheinputstothemultipliermeasuredinVoltsandUisascalefactoralsomeasuredinVolts.Thus,theoutputofthedemodulatorhasthedimensionofVoltsandEquations 3 and 3 aredimensionallyconsistent. 3 andEquation 3 ,theelectricalsensitivityofthecircuitisgivenbytheequation foravoltageamplierandby forachargeamplier. Fromtheaboveequation,itcanbeinferredthatthesensitivitycanbeincreasedforbothinterfacecircuitsbyincreasingthepeakacbiasvoltage,Vac.Thebiasvoltagecannotbeincreasedwithoutbound.Therearetwomainfactorsthatdeterminethemaximumvalueoftheacsignal.Theyaretheelectrostaticpull-inlimitwhichwasdiscussedinChapter2andtheactuationofthemicrophoneplatesbytheelectricalacvoltageinsteadoftheinputacousticsignal.Althoughthefrequencyofexcitationis100higherthantheresonanttheresonantfrequency,iftheinputamplitudeislarge,thediaphragmcanvibrateatthehighcarrierfrequency. Forthechargeamplierbasedinterfacecircuit,thesensitivitycanalsobeincreasedbydecreasingthevalueoftheintegratingcapacitor.Theminimumvalueoftheintegratingcapacitorisdeterminedbythelimitationsofthefabricationprocessandtheoverallbandwidthofthecircuit. Forillustration,thesensitivityofavoltageamplierbasedinterfacecircuitusingsquarewavesformodulationanddemodulationisderived.Letthetopplatevoltagevtpandbottomplate 75

PAGE 76

where 22 !cvac(t)=01 22 !c
PAGE 77

Forthereminderofthissection,weusethesimpliedsensitivity.Todeterminethetotaloutputnoise,thenoisemodelshowninFigure 3-23 isusedforavoltageamplierandFigure 3-24 isusedforachargeamplier.Forthevoltageamplierinterfacecircuit,thetotal Figure3-23. Noisemodelofsynchronousmodulatoranddemodulatorusingvoltageamplier. Figure3-24. Noisemodelofsynchronousmodulatoranddemodulatorusingchargeamplier. inputnoiseisgivenby whereZi=1 77

PAGE 78

3 andEquation 3 ,bythecorrespondingelectricalsensitivity,weobtaintheminimumdetectablecapacitanceandisgivenbytheequation V2acs foravoltageamplierand V2acs forachargeamplier.Theupperendofthedynamicrangeisthesameforbothinterfacecircuitsandisgivenby Dividingthemaximumoutputvoltageequationbythecorrespondingelectricalsensitivity,themaximuminputcapacitancethatcanbemeasuredisgivenby V2ac[VddVdrop]:(3) foravoltageamplierand V2ac[VddVdrop]:(3) forachargeamplier.Thisshowsthattheupperendofthedynamicrangecanbeincreasedbydecreasingthepeakacbiasvoltage,Vacforbothinterfacetechniques. Thelowerlimitofthefrequencyresponseisthesameasthedcvoltageamplierandchargeampliertechniquesconsideredpreviously.Itisgivenbytheexpression 2Rdc(2C0+Cp+Ci):(3) foravoltageamplierand 2RdcCint:(3) forachargeamplier.Theupperlimitofthefrequencyresponseofthiscircuitisdeterminedbythecutofffrequencyflpfofthelowpasslterandisgivenbytheexpression 78

PAGE 79

Therearetwomaindisadvantageswiththistechnique.Alargebiasresistorisneededfortheoperationofthecircuit.Implementingthisresistorisanon-trivialproblem.Forexample,intheADXL-50,a3Mresistorwasusedandwasfabricatedusingaspecialresistorlayer.Anothersolutionistouseback-to-backdiodes.Theissuesinusingbacktobackdiodesareleakagecurrentsandpoortransientperformance.Thesecondmajordisadvantageofthistechniqueistheuseofthelowpasslter.Thelowpasslterwillcausephasedelaywhichneedstobecompensatedwhenusedinclosed-loopoperationaswillbeshowninSection 3.5.1 .Furthermore,themultipliernoise,increasestheoutputnoise. 3-7 Table3-7. Summaryofprevioussynchronousmodulation/demodulationbasedopenloopcapacitivetransducerinterfacecircuit. Kraftetal.1998[ 104 ](accelerometer) 7:2pF Lietal.2000[ 103 ] NR NR Yakabeetal.2006[ 105 ](patent) NR NR NR 103 ]reportedanopenloopaccelerometerusingsynchronousmodulation/demodulationastheinterfacecircuit.Theyreportedasensitivityof75V=gwhentheinputwasin0:1grange.Thecharacteristicsofthecarriersignalandthebandwidthofthedevicewasnotreported. 79

PAGE 80

104 ]reportedananalogopenloopaccelerometerthatused2Vpp1MHzcarriersignal.Thedevicehadareportedsensitivityof0:7V=gandabandwidthof300Hz.Theopenloopbandwidthofthedevicewas56Hz. Yakabeetal.[ 105 ]obtainedapatentforamodiedformofavoltageampliersynchronousmodulationbasedinterfacecircuitthatallowsforreducedsensitivitytoparasiticcapacitance.Thisisaccomplishedbyplacingthesensoroutputsignal,anextrafeedbackcapacitor,andanauxiliaryopampinthefeedbackpathofaprimaryamplier.Nodetailsoftheperformancespecicationsarepresented. 79 ].Inthesecircuits,aresistorisreplacedwithappropriatelyconnectedMOSFETswitchesandcapacitorsanddrivenbynon-overlappingclockwaveformsthatsimulatearesistor.Switchedcapacitorsthuseliminatetheneedforlargeresistors.Switchedcapacitorsusingthetechniqueofcorrelateddoublesampling(CDS)[ 106 ]canattenuatedcoffsetsand1=fnoisefromtheinterfacecircuit. ThefundamentalideaintheCDSschemeistomeasurethenoisedcoffsetvoltageatsomeinstantoftimeandsubtractitfromthemeasuredsignalduringthenextinstant.Thisisperformedintwophaseswhicharetimemultiplexed.Duringthesamplingphase,thenoiseandoffsetsaremeasuredandstored.Afterwards,duringthesignalprocessingphasetheoffsetfreeamplierisavailableforoperation.Theoperationperformedbythesystemcanbeexpressedintermsoftheoutputsignaly[n],theinputsignalx[n],andthe1=2periodpreviousinputsignalx[n1 2]bytheequation Usingthez-transform,thetransferfunctionH(z)oftheabovesystemcanbeexpressedas 2:(3) 80

PAGE 81

ItcanbeseenfromthistransferfunctionthattheCDSoperationputsazeroatdc(i.e!=0)andhasahighpassresponseforincreasingfrequenciesuptotheNyquistfrequency.Thus,theoretically,CDSeliminatesdcoffsetsandhighpasslterstheinputsignal. 3-25 showsoneimplementationofaswitchedcapacitoramplierthatimplementsCDS.Theswitchedcapacitoramplier Figure3-25. Switchedcapacitorimplementationofcorrelateddoublesampling. consistsoftwocapacitorsCgainandCresetandfourswitchesarrangedasshowninFigure 3-25 .Theoutputcommonmodevoltage,Vcm,ischosensoastomaximizetheoutputrange.Foracircuitwith05Vsupplyrail,Vcmischosentobe2.5V.Theswitchesaredrivenwithtwononoverlappingclocks1and2. TheequivalentcircuitduringeachclockphaseisshowninFigure 3-26 .Duringtheresetphase(1),thenoisevoltagesaremeasuredandstoredontheintegratingcapacitor.Duringtheamplifyphase(2),thecircuitperformstheactualmeasurement.Toperformthemeasurement,stepvoltagesofequalmagnitudeVbutoppositeinsignareappliedtothetopandbottomplatesrespectively.Theinputoutputrelationduringtheamplifyphaseisgivenby[ 79 ], Cgain2C:(3) 81

PAGE 82

ResetPhase1. (b) AmplifyPhase2. Equivalentcircuitofswitchedcapacitoramplierduringamplifyandresetphase. Cgain:(3) Fromtheaboveequation,itcanbeinferredthatthesensitivitycanbeincreasedbyvaryingtheheightofthesamplingpulseordecreasingtheintegratingcapacitor. ThenoiseanalysisoftheswitchedcapacitornetworkisdoneusingasimpliedmodelfortheampliernoisebasedontheworkbyGobetetal[ 107 ].ThecompletenoisemodelisshowninFigure 3-27 .ThedetailedderivationofthenoisesourcesisinAppendixB.Thenoiseanalysis Figure3-27. Noisemodelofswitchedcapacitoramplierduringamplifyphase. 82

PAGE 83

108 ].Theswitchedcapacitorinterfacecircuitisanoversampledsystem.ThemainsourcesofnoisearethekT=Cnoiseassociatedwiththeswitchesandthewidebandwhitenoiseoftheamplierthatisaliasedtothebaseband.ThetotalcapacitanceCTattheinputoftheamplierthatissampledbytheswitchesis ThetotalkT=Cnoiseattheoutputcanbeexpressedintermsofthesamplingfrequency,fs,thetotalcapacitance,CT,andthegaincapacitanceCgain,andisgivenby CT:(3) Thetotalsampledwidebandnoiseattheoutput AddingEquation 3 andEquation 3 weobtainthetotaloutputnoiseas CT+CT Equation 3 canbedividedbythesensitivitytodeterminetheminimumdetectablecapacitanceperp CT+CT Theupperendofthecircuitdynamicrangeislimitedbythemaximumoutputvoltageoftheamplierandisdependentonthetopologyoftheamplierused.Itisgivenbytheexpression Dividingtheaboveequationbytheelectricalsensitivitythemaximuminputcapacitancethatcanbemeasuredisgivenby 83

PAGE 84

Thebandwidthofthiscircuitisdeterminedbytheclosed-loopbandwidthoftheamplier.BasedonthediscussioninAppendixB,foraswitchedcapacitoramplier,theclosed-loopbandwidthfucanbeexpressedintermsofthedutycycleofthesensephasem(<1),numberofsettlingtimeconstantsn,andthesamplingfrequencyfs,[ 41 ] 2m nfs1:(3) Theampliercoreismadeofafoldedcascodeamplierwhoseopen-loopbandwidthfampisgivenintermsofthetotaloutputcapacitanceCLT, bytheexpression 2gm Theclosed-loopbandwidthfuisrelatedtotheopen-loopbandwidthfampbythefeedbackfactor, bytheexpression Themaindisadvantagesareasfollows.Beingasampleddatasystem,thewidebandnoiseoftheamplierisaliasedintothebasebandthusincreasingthenoiseoorinthebandwidthofinterest.Thisproblemisnotpresentinthesynchronousmodulation/demodulationtechnique.Tomitigatethiseffect,theminimumallowableclosed-loopbandwidthmustbechosenfortheamplier.BasedonthediscussioninAppendixB,theminimumclosed-loopbandwidthisgiven 84

PAGE 85

2m nfs1:(3) 66 ]reportedaswitchedcapacitorinterfacecircuitforcapacitivepressuresensorswithasensecapacitorof1nF..Theinterfacecircuitwasimplementedusingdiscretecomponents.Theclockfrequencyusedwas10kHzwitharesetcapacitorof3:9nFandaresetcapacitorof101:2pF. In2004,Potteretal.ofKnowlesAcousticsreceivedapatent[ 109 ]foraswitchedcapacitorbasedinterfacecircuitforacapacitivemicrophone.Itisdesignedsoastoreducetheinputoffsetvoltageand1=fnoise.Performancespecicationswerenotspeciedinthepatent. In2006,Jawedetal[ 110 ]presentedthesimulationresultsofaswitchedcapacitorbasedinterfacecircuitforasinglebackplatecapacitivemicrophone.Thenominalcapacitanceofthemicrophonewas4:5pFandthepullinvoltagewasequalto3:5V.TheparasiticcapacitancewasontheorderofafewpF.Themicrophonehadabandwidthextendingfrom50Hzto20kHz.ThecircuitisasimplerversionthanthestructureshowninFigure 3-25 .Ithasonlyonefeedbackcapacitorandusesaxedon-chippoly-polydummycapacitortoformapseudo-differentialstructure. 111 ]. Inclosed-loopschemes,therearetwophasesofoperation,namelysenseandfeedback.Duringthesensephase,thecapacitancechangeismeasuredandduringthefeedbackphase,anelectrostaticforceisappliedusingavoltagetoeitherthebackplateorthediaphragmtonullthedeection.Thesetwophasesareseparatedintimeinthecaseofdigitalclosed-loopschemesandinfrequencyinthecaseofanalogclosed-loopschemes.Thisseparationisessentialandisdonetopreventtheexcitationsignalusedtothemeasurethecapacitancechangefrominterfering 85

PAGE 86

69 ].Toimplementtheforcefeedbackscheme,therearetwocommonapproaches.Inoneapproach,aseparatesetofplatesforsenseandforcefeedback[ 68 67 11 ]areusedwhileforanotherapproach,thesameplatescanbemultiplexedforthesenseandfeedbackphase[ 2 112 ]. Themainadvantagesoftheclosed-loopapproachisincreasedlinearityanddynamicrange[ 11 42 113 ].Bandwidthimprovementbeyondtheresonantfrequencyisalsopredictedfortheanalogclosed-loopinterfacecircuitapproach.Sincethefeedbackelectrostaticforcekeepstheplateatthecenterposition,thediaphragmbecomesmuchstifferthanthesimpleopen-loopcasewhichresultsinincreasedbandwidth.Forcefeedbackalsoallowsonetousehighervoltagesbyreducinginstabilityduetopull-involtageandtherebyincreasingthesensitivity. Theadvantagesprovidedbytheclosed-loopschemescomewiththeirownlimitationsandconstraints.Sincethesensorisafeedbackloop,thestabilityoftheloopmustbecarefullyconsideredduringthedesignoftheelectronics.Largevoltages,whichmightnotbeeasilyavailableincurrentCMOStechnologies,mayberequiredtobalancelargeinputsoundpressurelevels.Forexample,toforcebalancea160dBacousticsignal,approximately45VisrequiredwhichisnotacommonlyavailablevoltageinstandardICprocesses. Thedetailedoperatingprincipleofthevariousclosed-looptechniquesandtheirperformancemetricsarediscussednext. Thedesignofthedcbiasvoltageisdiscussedlaterinthissection.Thus,threedifferentvoltagesaresummedandappliedtothetopandbottomplates.Thetopandbottomplatevoltages 86

PAGE 87

3-28 showsaschematicoftheanalogclosed-looptechnique. Schematicofanalogclosed-loopforcefeedbackscheme. ThenetforceactingonthemovableplateFcanbeexpressedasthedifferencebetweentheforce,Ftp,actingbetweenthetopplateandthemiddleplateandtheforce,Fbp,actingbetweenthebottomplateandthemiddleplateandisgivenbytheexpression AssumingthatthemiddleplateiskeptatgroundpotentialbecauseofnegativefeedbackoftheamplierandusingEquation 2 ,FtpandFtparegivenbytheexpressions 2Av2tp 2Av2bp 3 andEquation 3 inEquation 3 ,neglectinghighfrequencytermsastheywillbelteredbythelowpasslter,andassumingx0<
PAGE 88

SubstitutingEquation 3 inEquation 3 ,forsmalldisplacementsweobtain Equation 3 showsthatforsmalldisplacements(i.e.x0<
PAGE 89

114 ]anduseelectrostaticsonlyforthefeedbackphase.Commerciallythesynchronousmodulaiton/demodulationtechniquehasbeenusedintheAnalogDevices(ADXL-50)accelerometer. Table3-8. Summaryofprevioussynchronousmodulation/demodulationbasedclosed-loopcapacitivetransducerinterfacecircuit. vanderDonketal1992[ 115 ](singlebackplatemicrophone)(Design1) NR NR1MHz vanderDonketal1992[ 115 ](singlebackplatemicrophone)(Design2) NR NR1MHz Hadjiloucasetal.,1995[ 116 ](singlebackplatemicrophone) NRNR OpticalsensingusingFabry-Perotinterferometer. Halletal.,2005[ 114 ](singlebackplatemicrophone) Diaphragmresonanceat44:8kHz(Theoretical) NRNR OpticalsensingusingMichaelsoninterferometer. AnalogDevices1993[ 2 ](accelerometer) NR Commerciallyavailable Kampenetal1994[ 7 ](accelerometer) NR Parketal.1998[ 112 ](accelerometer) Sensitivityis39mV=g 104 ](accelerometer) Sensitivityis700mV=g 117 ](accelerometer) Sensitivityis2:2mV=gNoiseFlooris1mg=p 115 ]reportedanalogclosedloopcontrolforasinglebackplatemicrophone.Themicrophoneconsistedofinterdigitatedelectrodeswithacousticholesonthe 89

PAGE 90

In1995,Hadjiloucasetal.[ 116 ]reportedaforcefeedbackmicrophonethatusesopticalsensingandelectrostaticforcefeedback.ThemovementofthemicrophonediaphragmisdetectedusingaFabry-Perotinterferometer.Sincethediaphragmpositionisdetectedoptically,noacsignalisrequiredfortheoperationofthissensor. In2005,Halletal.[ 114 ]reportedameasurementmicrophonethatusesopticalsensingwithelectrostaticforcefeedback.ThesensingisdoneusingaMichaelsoninterferometer.Thelaserrequiredforsensingisintegratedonthesensoritselfusingaverticalcavitysurfaceemittinglaser.Thebackplateofthemicrophoneisperforatedandintegratesthediffractiongratings.Thepaperdoesnotreportanyclosed-loopresults. Next,closedloopaccelerometersarediscussed.In1993,AnalogDevicesADXL-50[ 2 ]producedacommercialaccelerometerthatusesclosed-loopcontrolwitha2Vpp,1MHzcarriersignal.Itusesthesameplatesforbothsenseandfeedback. In1994,Kampenetal.[ 7 ]reportedaclosed-loopaccelerometerthatusesa2Vpp1MHzcarriersignal.Inthisdesign,thesamecapacitorwasusedforbothsenseandfeedback.Noperformancemetricsoftheaccelerometerisprovided. In1998,Kraftetal.[ 104 ]reportedaanalogclosed-loopaccelerometerthatusesa2Vpp1MHzcarriersignal.Thedevicehasareportedsensitivityof0:7V=gandabandwidthof300Hz.Theopenloopbandwidthofthedeviceis56Hz. In1998,Parketal.[ 112 ]reportedaclosedloopaccelerometerthatusesa1Vpp50kHzcarriersignal.Thesensitivityoftheaccelerometeris39mV=gandtheclosed-loopbandwidthis350Hz.Thissensorusesthesameplatesforsensingandfeedback. In2002,Luoetal.[ 117 ]reportedalateralaccelerometerwithalinearrangeof13gandanoiseoorof1mg=p 90

PAGE 91

In1990Joyceetal.[ 118 ]reportedacapacitiveforceforatomicforcemicroscopythatusesanalogclosed-loopforcefeedback.Thenominalcapacitanceofthesensoris0:4pFandtheresonantfrequencyis30kHz.Thesensorwasexcitedwitha2:5MHzcarriersignal.Theyreportaminimumdetectableforceof108Nwhichcorrespondstoadisplacementof0:5nmforabandwidthof1kHz. 3-29 .Itconsistsofadiscretetimeintegratorwhoseoutputisquantizedusingaone-bitquantizer.Theinputtotheintegratoristheerrorbetweentheanaloginputsignalandthequantizeroutput.Thesystemisoperatedatasamplingfrequency,fs,thatismuchhigherthantheNyquistfrequencyfNoftheinputsignal.Overmanyperiods,theaverageerrorsignaleisdriventozerobythefeedbackloopassumingthattheloopisproperlycompensated.Theinputinformationcanberecoveredfromtheoutputbitstreambylowpassprincipleisusedasaninterfacetechnique,thediscretetimeintegratorisreplacedbythemicrophonewhichismodelledasasecondordersystem.Figure 3-30 showsaschematicofthemechanicalmodulator.Theforwardpathisrepresentedviathreephysicalblocksnamely,themicrophone,achargeamplier,adiscretetimecompensatorandacomparator.Thefeedbackpathisrepresentedviatwofunctionalblocks(digitaltoanalog(D/A)converterandvoltagetoforce(V/F)block)whichrepresenttheelectrostaticfeedbackforce 91

PAGE 92

Schematicofrstordermodulator. appliedontheplatesataninstantoftime.Thesummingnodecomputesthedifferencebetweentheforcecorrespondingtotheinputacousticpressuresignalandtheelectrostaticfeedbackforce.Afeedbackforceisappliedovertimethatnullsthiserrorsignal. Schematicofsecondordermechanicalmodulator. 92

PAGE 93

79 ] Themodulatorisanon-linearclosed-loopsystem.Linearsystemstheorycannotbeappliedfortheanalysisofthesesystems.Someoftheexistingworkintheliterature[ 119 6 ]donotdiscussthebandwidthvariationwithclosed-looptechniques.Kraft[ 69 ]discussesananalyticmethodusingthedescribingfunctiontechnique[ 120 ]toqualitativelypredictthebandwidthimprovement.Theconclusionbasedonhisworkisthatthefeedbackvoltageincreasesthesensorbandwidthaslongastheinputsignalfrequencyisbelowthelimitcyclefrequencypredictedbythedescribingfunctionanalysis. Themaindisadvantageofthistechniqueisthecomplexityofthecircuit.Theotherdisadvantagesofthistechniqueaswescaledevicegeometriesdownarediscussedinthenextsectiononscaling. 93

PAGE 94

3-9 showsacomparisonofthesamplingfrequenciesrequiredforsensorswithincreasingresonantfrequencies.Thetargetdynamicrangeis120dB.Toachievethetargetdynamicrange,asecondordermodulatorneedsanoversamplingratioofapproximately725.Thisisroundedupto1000Xtoprovideamarginofsafety.Todeterminetheclosed-loopbandwidthfu,weuseEquation 3 withn=7(i.e.0:1%settling),andm=0:25(i.e.25%oftimespentinsensephase).Thenextissuetoconsider Table3-9. Frequencyrequirementsofinterfacecircuitswithincreasingresonantfrequency. PropertyAccelerometerAudiomicrophoneAeroacousticmicrophone 8 ]20kHz200kHzfs1MHz[ 8 ]20MHz200MHzfu4:5MHz90MHz900Mhz 3-10 showsthevoltagerequiredtobalancethemaximumpressureforaudiogradeandaeroacousticmicrophones.FromTable 3-9 Table3-10. Feedbackvoltagerequirementsofinterfacecircuitswithincreasingdynamicrange. PropertyAudiomicrophoneAeroacousticmicrophone Maximuminput120dB160dBFeedbackvoltage6V45V 3-10 ,itcanbeseenthatdesigningamodulationbasedcontrollerforaeroacousticapplicationisanon-trivialproblem.Thisisbecausecurrentsemiconductormanufacturingprocesstechnologiesdonothave45Vprocesseswhichsupportamplierswitha900Mhzgainbandwidthproduct.Onemethodtoovercomethisproblemistouseseparatevoltagesforsensingandfeedbackusingoffchipdrivers. 3.6 providesareviewofcapacitivetransducersusingbasedclosed-looptechniques.Thistechniquehasbeenreportedextensivelyforcapacitiveaccelerometers.Theuseofthistechniqueforclosedloopforcefeedbackinhighbandwidthoperationsisverychallengingduetothelargesimultaneousgainbandwidthandvoltagerequirementsfortheseapplications. Therstdigitalclosed-loopaccelerometerwasreportedbyHenrionetal.[ 42 ]in1990.Thesystemoperatesataclockfrequencyof512kHz.Thereportedsystemsensitivityis1V=g.The 94

PAGE 95

Summaryofpreviousbasedclosed-loopcapacitivetransducerinterfacecircuit. Henrionetal.,1990[ 42 ](accelerometer) NR Sensitivityis1V=g 6 ](accelerometer) Theoreticalresonantfrequencyofsensor15kHz NR Onlyopenloopmodeofoperationwastestedthoughtthechiphastheelectronicsforclosed-loop. Wuetal.,1992[ 121 ](microphone)theoreticalpaper NR Kraftetal.,1998[ 104 ](accelerometer) Lemkinetal.,1998[ 122 ](accelerometer) Theoreticalresonantfrequencyofsensor3:4kHz Sensitivityis1V=g,Noiseooris110g=p Moonetal.,2000[ 123 10 ](accelerometer) Theoreticalresonantfrequencyofsensor15kHz Edelson2001[ 124 ](microphone) NR NR NR NR In1992,Yunelal.[ 6 ]reportedthedesignofadigitalclosed-loopaccelerometer.Inthisdesignseparatecapacitorswasusedforsenseandfeedback.Thefabricateddevicewastestedonlyintheselftestopenloopmodeusingelectricalexcitation. In1994,Wuetal.[ 121 ]presentedatheoreticalpaperonthedesignofabasedclosed-loopcontroller.Theproposedinterfacecircuitusesinterferometricopticalsensingtodeterminethediaphragmpositionofadualbackplatemicrophoneandelectrostaticfeedback.Thecircuitisdesignedtohaveaoverallsystembandwidthof25kHz.Theoversamplingfrequencyusedis5MHz:

PAGE 96

104 ]reportedadigitalclosed-loopaccelerometerthatusesa10kHzclocksignal.Theopenloopbandwidthofthedeviceis56Hzandtheclosed-loopbandwidthis300Hz. In1998,Lemkinetal.[ 122 ]presentedathreeaxisMEMSaccelerometer.Itusesaswitchcapacitoramplierfordetectingthecapacitancechange.Thecircuitwascharacterizedataclockfrequencyof500kHz.Thex-axisaccelerometerhasasensitivityof1V=gwithanoiseoorof110g=p In2001,Edelsonetal[ 124 ]ofBorealisTechnicalLimitedobtainedapatentforamodulationbasedforcebalancedmicrophone.Themicrophonedescribedisoftheelectrodynamictypebutthepatentcoverselectrostaticdevicesalso.Inthispatent,thesameplatesareusedforbothsensingandforfeedback. 96

PAGE 97

4-1 Figure4-1. OverviewofChapter4. 4-2 .Theinputtothesystemisa 97

PAGE 98

Blockdiagramofanalogclosed-loopsystemshowingmicrophoneandinterfacecircuit. timevaryingpressuresignalgeneratedbytheacousticsource.Theinputpressureactingontheconstantareaofthemicrophonediaphragmproducesaninputforce.Thenetforceactingonthediaphragmisthedifferencebetweentheinputforceandtheelectrostaticfeedbackforce.Thediaphragmdisplacementiselectrostaticallytransducedasachangeincapacitancebyappropriatelybiasingthemicrophoneplates[ 37 ].Synchronousmodulation/demodulation[ 125 ]isusedastheinterfacecircuittodetectthecapacitancechange.Theoutputofthecircuitaftercompensationprovidesthefeedbackvoltage.Thisvoltagecausesanelectrostaticforcewhichopposestheinputforceandactstoreducethediaphragmdisplacement. Themicrophoneismodeledasasecondordersystemgivenbytheequation s2+2(2fres)s+(2fres)2;(4) developedinChapter2. TheinterfacecircuitmeasuresthechangeincapacitancesC1andC2inducedbytheinputpressureactingonthediaphragm.Thetopandbottomplatesofthemicrophoneareexcitedbyadcvoltage(Vdc)superimposedonasquarewaveaccarrier(vac)ofoppositepolarity.Theexcitationfrequency(!c)ischosentobemuchhigherthantheresonantfrequencyofthesensor 98

PAGE 99

wherex0isthenominalgapdistancebetweentheplatesofthemicrophone,x0isthedisplacementofthediaphragmfromitsequilibriumposition,andC10andC20aretheequilibriummicrophonecapacitances.InEquation 4 ,theratioC10+C20 andhassignalsatthebaseband,carrierfrequency,andmultiplesofthecarrierfrequency.Theoutputofthedemodulatorislowpasslteredandisgivenby UsingEquation 4 ,thetransferfunctionoftheinterfacecircuit,Hintisgivenby Thevoltage-to-forceblockmodelsthetransductionofthefeedbackvoltagevftoanelectrostaticforceactingonthediaphragm.Theforcesbetweenthetopplateandthediaphragm,Ftp,andbetweenthebottomplateandthediaphragm,Fbp,aregivenby 2Av2tp 2Av2bp 99

PAGE 100

Thefeedbackvoltagevfisgeneratedbythecompensator.Thedesignofthecompensatorisdiscussedinthenextsection.ThenetelectrostaticfeedbackforceonthediaphragmFeisgivenbythedifferenceofFtpandFbpi.e., Forsmalldisplacements(x00),Fecanbelinearizedto Equation 4 representsthelinearizedinput-outputrelationbetweenthefeedbackvoltageandtheforceonthemicrophonediaphragm.UsingEquation 4 ,thefeedbacktransferfunction,Hfbisgivenby Thefunctionofthecompensatoristolterthehighfrequencymodulatingsignalandprovideappropriategainandphasesothattheresultantclosedloopsystemisstable.Thecompensatorconsistsoftwostages.Therststageisahigh-gainlow-passlterthatsetsadominantpoleatlowfrequency.Thesecondstageconsistsofalead-lagsectionthatprovidesprimarilyphaseleadtoobtainsufcientphasemargin.Theloopgainofthesystemwithoutthecompensator,Tnc(s),isgivenby FromtheuncompensatedsystemBodeplotinFigure 4-4 ,itcanbeseenthatthesystemhasverylowloopgainatdcandahighQresonantpeakduetothecomplexconjugatepolesofthemicrophone.Toimprovetheloopgainandprovidesufcientphasemargin,atwostagecompensatorgivenbythetransferfunctionH(s) 100

PAGE 101

4-3 .Itcanbeseenthatthe Figure4-3. Bodeplotofcompensator. compensatorhasahighgainlowfrequencypoleat1kHz.Ahighfrequencyzeroat230kHzisusedtoprovidephasemargin.Theloopgainofthesystemwithcompensator,T(s),isgivenby s2+2(2fres)+(2fres)2 4-4 .Itcanbeseenthatthesystemhasaphasemarginof42andadcloopgainof87. ThecompleteclosedloopSimulinkmodelisshowninFigure 4-5 .Themicrophoneandcompensatoraremodeledusingas-domaintransferfunctionblockwhichimplementsEquation 4 andEquation 4 respectively.Thesynchronousmodulator/demodulatorinterfacecircuitmodelandvoltage-to-forcesubsystemmodelimplementEquation 4 andEquation 4 respectivelyusingmultipliersandgainblocks.Thus,atimedomainmodeloftheclosedloopsystemhasbeendeveloped.Openandclosedloopoperationforstepandsinusoidalinputsweresimulatedandispresentednext. 101

PAGE 102

Bodeplotofloopgainwithandwithoutcompensator. Simulinkmodelofanalogclosed-loopcontrolsystem. 126 ],duringclosedloopoperation,thediaphragmdisplacementatagivenfrequencywillbeattenuatedbytheloopgainevaluatedatthatfrequency.Fromthetimedomainsimulation(Figure 4-7 ),itcanbeseenthattheoutputdisplacementfora2kHzinputisreducedbyafactorof40whichmatchestheattenuationindiaphragmdisplacementpredictedbytheloopgain(Figure 4-4 ).Tofurtherexamineclosedloopoperation,a1mdroptestwassimulated.Thepressurecorrespondingtoa1mdropactingonthediaphragmissimulatedbya1ms,0:029Pa(63dB)pulse.Thediaphragmdisplacementis 102

PAGE 103

Simulinkmodeloftopplatevoltagesubsystem. Simulinkmodelofdemodulator. Simulinkmodelofvoltage-to-forcesubsystem SubsystemsusedinSimulinkmodel. attenuatedfrom6:2pmto7:1102pmduringthepulse(Figure 4-8 ).Thismatchesthedcloopgainof87(38dB)obtainedfromtheloopgainBodeplot(Figure 4-4 ). ThevoltagerequiredforthetwomicrophonestructuresdiscussedinChapter2isrepeatedinTable 4-1 .Basedonthetable,itcanbeseenthatweneedanamplieroperatingatavoltage 103

PAGE 104

Diaphragmdisplacementwithandwithoutfeedback. Figure4-8. Diaphragmdisplacementwithandwithoutfeedback. greaterthan45Vwithaunitygainbandwidthofatleast2MHz.Suchampliersarenotavailablecommercially.Themaximumoperatingvoltageofhighbandwidthcommerciallyavailableampliersis15V.AnOP275dualoperationalamplierwhichhastwomatchedamplierswithaunitygainbandwidthof11MHzatanoperatingvoltageof15Visused.Topreventpowersupplynoisefromenteringthesystem,thetopandbottomplatedcvoltagesareprovidedbybatteriesregulatedusingalinearregulator.Themaximumpressurethatcanbeforcebalancedisapproximately134dB. TheschematicofthesummingampliercircuitthatgeneratesthetopandbottomplatevoltagesisshowninFigure 4-9 .InthisgureVdcandVdcrepresentthedcvoltagesused 104

PAGE 105

Feedbackvoltagerequirementsofaeroacousticandaudiomicrophoneforforcebalancingmaximuminputpressure. Property Audiomicrophone Aeroacousticmicrophone Maximuminput 160dB(2000Pa)Feedbackvoltage Figure4-9. Schematicofsummingamplierthatgeneratestopandbottomplatevoltages. 4-10 .Theunitygainamplierwasimplementedusingalownoise,highbandwidthamplier,TLE2074[ 127 ]conguredforunitygain.Aquadopampwaschosensothatthefouraeroacousticmicrophonesoneachsensordiecouldbeinterfacedusingasinglechip.Thisamplieroperatesona15Vpowersupply,hasaninputcapacitanceof10:6pF,aninputreferrednoisevoltageof11nV=(Hz)1 2,andaninputreferrednoisecurrentof1:6fA=(Hz)1 2.Themiddleplateofthemicrophoneisbiasedusingtwosurfacemount22Mresistorsina 105

PAGE 106

ThetotalparasiticcapacitanceattheinputnodeisgivenbythesumofthePCBtracecapacitanceandtheinputcapacitanceoftheamplier.Thetotalparasiticcapacitanceisapproximately20pF. 102 ].ThisICispoweredwitha5powersupplyandhasa200MHzfullpowerbandwidth.IfthetwoinputsignalstothemultiplierareXandY Schematicofvoltageamplier. respectivelyandtheoutputsignalisW,thenthetransferfunctionofthemultiplierisgivenby AschematicofthedemodulatorICisshowinFigure 4-11 Figure4-11. Schematicofanalogmultiplier(AD835)showingrelevantpins. 4-12 .Thetransferfunctionofthecircuit 106

PAGE 107

4-12(a) ,Hc1,is 1+Rf1Cf1s;(4) andthetransferfunctionoftheschematicshowninFigure 4-12(b) ,Hc2,is ForthetransferfunctiongiveninEquation 4 andEquation 4 ,thecomponentvaluesusedarelistedinTable 4-2 .ThevariouscircuitryusedintheclosedloopoperationaresolderedontwodifferentPCBs.Thisisdiscussedinthenextsection. Table4-2. Elementvaluesforresistorsandcapacitorsusedincompensator. ElementValue ElementValue Schematicofrststageofcompensator. (b) Schematicofsecondstageofcompensator. Schematicoftwostagecompensator. 107

PAGE 108

AnnotatedphotographofPCB1showingbufferamplier,breakoutheadersandopticalport. Figure4-14. AnnotatedphotographofPCB2. Thecompletesystemconsistingofthemicrophoneandinterfacecircuitischaracterizedinanacousticpressurecoupler(APC).ThedetailsoftheAPCareprovidedinChapter6.TofacilitatemountingofthemicrophoneintheAPC,thecircuitryisdividedintotwoseparateprintedcircuitboards(PCBs).ThemicrophoneandbufferamplieraresolderedontothemicrophonePCB(denotedasPCB1)whichismountedontotheAPCwhiletheremainderofthecircuitryconsistingofthedemodulator,compensator,andtopandbottomplatevoltagegenerationcircuitryPCBaremountedonanotherPCB(denotedasPCB2).Thetwoboardsareelectricallyconnectedwitheachotherusingaribboncable.ThiscablecouplesthetopandbottomplatevoltagesgeneratedbythePCB2tothemicrophoneplatesonPCB1.Thiscablealso 108

PAGE 109

4-13 andFigure 4-14 showsanannotatedpictureofthePCB1andPCB2.Theheadersfortheribboncableareshowninthegures. 4-16 .Thetotaloutputreferrednoiseoftheforwardpathisdeterminedby Figure4-15. Schematicofforwardpathoftheclosedloopsystem summingthenoiseattheoutputofeachstageoftheforwardpathwiththeappropriategain.Thisnoisemodelisvalidinthebandwidthofoperationoftheclosedloopsystemwhichextendsfrom300Hzto20kHz.Inthebandwidthofinterest,thenoisespectrumisdeterminedythethermalnoiseofthevariouscomponentsandisatoverthespectrum.Henceinthisanalysis,theshapingofthenoisebythevariouscomponentsisnotconsidered. 109

PAGE 110

Magnitudeofparametersusedinnoiseanalysis Voltagenoiseofinputamplier(Sva1)18nV=p 110

PAGE 111

4-3 .SubstitutingthevaluesintheEquation 4 weobtain Figure4-16. Plotoftheoreticalnoiseooroftheopenloopsystemandforwardpathsystem 111

PAGE 112

and Thus,thenoiseattheoutputofforwardpathofthedemodulatorishighduetothecomponentsthatarerequiredtogeneratethelargegainintheforwardpath.Itcanbeseenthattheforwardpathnoiseisveryhigh.Thereasonforthehighnoiseooristheneedforlargegainintheforwardpathtoachievealargeloopgain.Onesolutiontoreducethenoiseandstillmaintainthelargegainwouldbetouseamplierswithlowernoise.Inthiswork,alltheICsutilizethesame15VpowersupplywhereaslowernoiseICsusuallyaresingleendedandoperateat3:3Vand1:8V.Hence,tousetheselowernoisecomponents,multiplepowersupplyrailsarerequired.Also,astheICsaresingleended,appropriatelevelshiftingneedstobeperformedtoobtainstableoperation. Thevoltagerequiredtoforcebalancethemaximuminputpressureof160dBasshowninTable 4-1 is45V.Thiscannotbemetduetopowersupplylimitationsof15forthecommercialofftheshelfoperationalampliersthatareusedinthedesign.Thebandwidthoftheoverallsystemisdeterminedbythecompensator.Thecompensatordesignwaslimitedtoanoverallbandwidthofonly20kHzbecauseofthegainbandwidthconstraintsoftheampliersused. 112

PAGE 113

5-1 Figure5-1. OverviewofChapter5. 5.3 .ThesequenceofoperationofthevariousblocksandtheoverallsystemleveltimingisexplainedinSection 5.4 .Section 5.5 describesthesystemlevelsimulationofthemicrophoneandsystemusingMatlab. 113

PAGE 114

5-3 .Thedesignoftheinterfacecircuitryinvolvesthedesignofthecomponentsinsidetheboxlabeledelectricaldomain.Themodulatorbasedinterfacetechniqueisasampleddatasystem.Hence,alltheblocksneedaclocksignalfortheiroperation.Therefore,therststepinthedesignprocedureoftheinterfacecircuitistodeterminethesamplingfrequencyfs.Thesamplingfrequencyisafunctionoftheresonantfrequencyofthemechanicalsystemandthedynamicrangespecicationoftheoverallsystem.ThisisillustratedinFigure 5-2 .Theblocksinellipsoidalboxesrepresentthesystemlevelinputspecicationandrectangularboxesrepresenttheassumptionsusedforthedesignandtheparametersderivedfromthedesignprocedure.Thedynamicrangeisdetermined Flowchartfordesignofmodulatorcomponentparametersfromsystemlevelspecication. bythenoisesourcesinthesystem.Themainsourcesofnoisearethesampledthermalnoiseoftheamplier,kT=Cnoiseoftheswitches,andquantizationnoise.Foranidealmodulator,thenoiseoorisdominatedbythequantizationnoise.ThesampledthermalnoiseandkT=Cnoiseisassumedtobebelowthequantizationnoiseoor.Assumingthatthenoiseoorisdominatedbythequantizationnoise,thedynamicrangeDR,indB,ofaone-bitsecondordermodulatorcanbeexpressedintermsoftheoversamplingratioM(=fs=2fres)bytherelation[ 79 ] 114

PAGE 115

Schematicofsecondordermechanicalmodulator. Oncethesamplingfrequencyisdetermined,thenextstepisthepartitioningofthetimeavailableduringoneclockcycleT(=1=fs)intothesenseandfeedbackphase.Tomaximizethedynamicrangeandbandwidth,thetimespentinthesensephasemustbeminimizedandthetimespentinthefeedbackphasemustbemaximized.ThetimespentinfeedbackisdeterminedbythemaximumpressurethatmustbeforcebalancedandthemaximumvoltageVmaxavailableinthesystem.ThevoltagerequiredtoforcebalancethemaximuminputpressurePmaxcanbeexpressedintermsofthenominalcapacitanceC0,thecapacitanceareaA,andthenominalgapdistanceofthecapacitorx0bytheexpression Thisvoltagemustbeavailableduringthefractionoftimefbspentinfeedback.IfthemaximumvoltageavailableinthesystemisVmax,then Usingtheaboverelation,thetimespentinfeedbackcanbedetermined.Then,thetimespentinsensesenseisgivenby Next,weconsiderthepartitioningofthetimespentinthesensephase.Thetimespentduringsensesenseisdeterminedbythesettlingtimerequirementsoftheamplieramp,compensator 115

PAGE 116

Theindividualblocksmusttosatisfytwoconictingcriteria,namelyhighspeed(settlingtimerequirements)andlownoise.Toachievethis,thecircuitisrstdesignedforlownoiseandthentestedforsettlingtimerequirements.Ifthelownoisedesigndoesnotsatisfythespeedrequirements,thecircuitdesignisreiteratedbychangingthegaincapacitortomeetthesettlingtimerequirements,andthenoiseperformanceisreevaluated.TheoreticalformulationforthenoiseandsettlingtimeforeachofthecomponentsaredescribedinSection 5.3 Oncetheindividualcomponentsaredesigned,thesequenceofoperationoftheindividualpartsofthesenseandfeedbackphaseisdesigned.ThisisexplainedinSection 5.4 .Forthiswork,thetimingsignalsaregeneratedoffchipusingaAlterEPM7128[ 128 ]complexprogrammablelogicdevice(CPLD).Thisallowsonetohavebettercontrolofthesenseandfeedbackphaseduringtheinitialtestingphase.Basedonthisinformationthesystemlevelsimulations(Subsection 5.4 )areperformed. 116

PAGE 117

5-11 Switchedcapacitoramplier. Ampliercore. Transistorlevelschematicofamplier. ThedesignoftheampliercloselyfollowstheworkbyWongkomet[ 108 ].Thefollowingassumptionsaremade: TheamplierschematicduringtheamplifyphaseusedfortherestoftheanalysisisshowninFigure 5-5 .Thethermalnoiseoftheampliercanbeexpressedintermsoftheampliertransconductancegmandtheratioofthetotaltransistornoisetotheinputtransistornoisenfbytheexpression[ 79 ] 3gmnf:(5) 117

PAGE 118

Schematicofamplierduringamplifyphase. Forawelldesignedamplier,onlytheinputtransistorscontributenoise,andnfisequaltotwo.Inpracticeitcanbegreater.Fromthediscussioninthepreviouschapter,thetotaloutputthermalnoiseoftheamplieris Intheaboveequation,CTisgivenby Theparasiticcapacitance,Cp,canbesplitintotheinputcapacitanceoftheamplier,Cp;ampandotherexternalparasiticsCp;ext.Therefore,CTcanbeexpressedas Foranytransistor,thetransconductancegm,canbeexpressedintermsoftheinputcapacitanceCp;amp,andtheunitygainfrequency,fT,as Substituting,Equation 5 andEquation 5 inEquation 5 ,weobtain 2fTCp;ampfu 118

PAGE 119

Fromthepreviousequation,itcanbeinferredthattominimizethenoiseoftheamplier,theinputcapacitancemustbeequaltothetotalcapacitanceattheampliersummingnode.SubstitutingEquation 5 inEquation 5 ,gmcanbedesigned.Fromthetransconductance,theW=Lofthetransistorsoftheampliercanbecalculated. Oncethetransconductanceisdesigned,theloadcapacitorisdesignedtomeetthebandwidthrequirement.Asdiscussedinthepreviouschapter,theclosed-loopbandwidthoftheswitchedcapacitoramplierisgivenby 2m nfs1:(5) Theclosed-loopbandwidthfuisrelatedtotheopen-loopbandwidthfampbythefeedbackfactor, andtheopen-loopgain, 2gm bytheexpression ThetotalloadcapacitanceCLT,isgivenbytheexpression SubstitutingEquation 5 inEquation 5 ,weobtain 2gm 2gm CL+Creset+Cgain[2C0+Cp;amp+Cp;ext]

PAGE 120

SubstitutingforCp;ampfromEquation 5 andfufromEquation 5 intothepreviousequation,weobtain 2nfs Thiscanbesimpliedto Thisistheloadcapacitancerequiredtosatisfythelownoiserequirementsandsettlingtimeconstraints.Iftheactualloadcapacitancewhichisformedbythecompensatorislarger,thenalargerintegratingcapacitorcanbeusedandtheprocedureisreiterated.Thus,theamplierhasbeendesignedtomeetthesettlingtimeconstraintsandminimizenoise. 119 ]forcompensatingamodulationbasedcapacitivesensoristouseatwotapdiscretetimeniteimpulseresponse(FIR)lter.Theltertransferfunctioncanbeexpressedintermsofadesignparameterbythegeneralz-domaintransferfunction, Thedesignparameterisdesignedsoastoreducethetotalsystemnoiseandprovidestablecompensation.Fortheinitialdesign,waschosentobe2using[ 119 ]andveriedviasimulationandnotviaanalyticalmethods.ThesettlingtimeofthecompensatorisdeterminedbytheRCtimeconstantofthecompensatorcapacitanceandtheon-resistanceoftheswitches. TheactualtransistorlevelimplementationofthecompensatorisshowninFigure 5-6 .Thecircuitconsistsofonlyswitchesandcapacitors,andthetransferfunctionisdeterminedbytheprincipleofchargesharing.Sincethiscircuitdoesnothaveanyampliers,thesettlingtimeisconsiderablyreduced.ItconsistsofthreecapacitorsC1,C2,andC3connectedtoswitchesthat 120

PAGE 121

Transistorlevelschematicofcompensator. phase.DuringtheSamplephase,theoutputoftheamplierissampledontocapacitorsC1andoneofthetwocapacitorsC2andC3asdeterminedbythechopsignal.ThederivationofthetransferfunctionofthecircuitisdonebyanalyzingthetwohalfcyclesofthiscircuitasshowninFigure 5-7 .Figure 5-7(a) showsthecasewhencapacitorC1andC3arebeingchargedwiththeinformationfromthecurrentcyclewhileC2holdstheamplieroutputfromthepreviouscycle.Thechargeonthethreecapacitorsduringthesamplingphaseatsomeinstantoftimekisgivenby 121

PAGE 122

Samplephase. Evaluationphase. Compensatorschematicduringthetwophasesofoperation. DuringtheEvaluationphaseofthesamecycle,showninFigure 5-7(b) ,theplatesofcapacitorC2arereversedandconnectedtotheoutput.Applyingchargeconservation,weequatethenalchargeonthecapacitorsconnectedtotheoutputnodetotheirrespectiveinitialcharges.Thiscanbeexpressedas Takingthez-transformofthepreviousequationandrearrangingtheterms,weobtain, FromEquation 5 ,itcanbeseenthatthetransferfunctionofthecompensatorhasaDCoffset.Whenthiscompensatorisusedinclosed-loop,thisoffsetwillcausethecenterplatetobedisplacedfromthecenterposition.Thisdoesnotaffecttheoperationofthecircuitoritsperformancemetrics.Neglectingtheoffset,Equation 5 canbesimpliedas 122

PAGE 123

5 withEquation 5 ,weobtain 5-8 .Thelatchmodetimeconstantisgivenintermsoftheloadcapacitance Transistorlevelschematicofcomparator 79 ] 123

PAGE 124

5-11 .Thisnetworkdesignsthebiasvoltages(Vbn;Vcn;Vbp;Vcp)andthebiascurrents. Schematicofbiasnetwork. 5.3 TheoverallsequenceofoperationofthemicrophonesystemisshowninFigure 5-10 .Theclocksignalsinvolvedineachphaseofoperationareshowninsidetheblock.Thevariouscomponentsofthemodulator,namelyswitchedcapacitoramplier,compensator,andcomparatorhavetwophasesofoperation.Thetwophasesofoperationoftheamplierarethe 124

PAGE 125

5-1 .ThecompleteinterfacecircuitisshowninFigure 5-11 Table5-1. Clocksignalsusedinthevariouscomponentsofthemodulator. ComponentPhasesClock AmplierReset,Amplify1,2CompensatorSample,EvalSampl,Eval,EvalComparatorTrack,LatchLatch Flowchartshowingthesequenceofoperationoftheoverallmodulationbasedinterfacecircuit. Thetimingdiagrambeginsattheresetphaseoftheamplierduringthesensingoperation.Atinstantdenotedby1,theresetphaseendsandtheamplierispreparedtoentertheamplifyphase.Thetimefrom1to2isthenonoverlapperiodoftheswitchedcapacitoramplier.Attheinstant2,theampliercomesintotheamplifyphase.Thetimefrom2to3isthetimerequiredfortheampliertosettlebeforestartinganewmeasurement.Atinstant3,thetop(TP)and 125

PAGE 126

Oncetheamplierhassettledandtheoutputoftheamplierhasbeensampledontothecompensatorcapacitors,thecompensatorenterstheevaluationmode.Theperiodfrom4to5isthenon-overlaptimeforthistooccur.Oncetheamplierenterstheevaluationmode,ittakessometimetosettleandthisisthetimefrom5to6.Thisisthesettlingperiodofthecompensatorandisdenotedbycompen.Oncethecomparatorsettles,thecomparatorneedstomakeadecision.Atinstant6,thecomparatorgoesfromthetrackmodetothelatchmode.Theoutputofthecomparatorgoeshighorlowdependingonthecompensatorinputascomparedtothereferencevoltage.Basedontheoutputofthecomparator,anexternalcircuitrychangesthetopandbottomplatevoltagestonullthediaphragmdeection. Figure5-11. Schematicofmodulatorcircuit. Thetimeperiodfrom6to8isthelatchmodetimeconstant,compar.Theperiodfrom6to7isthenonoverlapperiodduringwhichtheampliergoestotheresetphaseandpreparesforanothermeasurement.Finally,thetimefromtheendofthefeedbackphasetothetimeinstant1isadjustedsothatthetopandbottomplatevoltagesduringthesensephaseisspentequallyhighandlow.ThefunctionperformedbythesystemduringthevarioustimeintervalsaregiveninTable 5.4 126

PAGE 127

Systemleveltimingdiagramoveronecompletecycle. Table5-2. Functionperformedduringeachtimeinterval. TimeIntervalFunction 5-13 showstheSimulinkmodelofthemicrophonesystem.Themicrophonewasmodelledusingas-domaintransferfunctionblockwithasecondordertransferfunctiongivenby Theinputtothemicrophonesystemblockisthesummingblockwhichcomputesthedifferencebetweentheforcecorrespondingtotheinputpressuresignalmodeledusingasignal 127

PAGE 128

SimulinkModelofthemodulatorbasedinterfacecircuit. generatorblockandtheelectrostaticfeedbackforce.Thiserrorsignalformstheinputtothemicrophonewhichisrepresentedbyas-domaintransferfunctionblock.Theoutputofthemicrophoneblockisinputtedtothecompensatorwhichismodeledusingthez-domaintransferfunctionblock.Thecomparatorismodeledasazeroorderholdandone-bitquantizerblock.Themagnitudeofthefeedbackforceiscomputedusingtheequationblockwhoseinputsarethecurrentdisplacementandapredeterminedfeedbackvoltage.Thesignoftheforceisdeterminedbyato1multiplexerblockwhichisdrivenbytheoutputofthecomparator.Theoutputofthemultiplexerformstheelectrostaticfeedbacksignal.ThesimulationparametersareaslistedinTable 5.5 .Preliminarysimulationresultsofthevoicegrademicrophoneare Table5-3. Parametersusedinclosedloopsimulation. ParameterMagnitude ResonantFrequencyfres30kHzDampingCoefcient0:15SamplingFrequencyfs2:56MhzInputfrequency5kHzOversamplingRatioM256SolverTypeFixedpoint(ODE45)StepSize1=fsLengthofFFT5120FrequencyResolution500Hz 5-14 .FromFigure 5-14 ,itcanbeseenthattheoutput 128

PAGE 129

Figure5-14. Powerspectraldensityofoutputmodulatedbitstream. 5-4 andTable 5-5 respectively. Table5-4. Frequencyrequirementsofinterfacecircuitswithincreasingresonantfrequency. PropertyAccelerometerAudiomicrophoneAeroacousticmicrophone 8 ]20kHz200kHzfs1MHz[ 8 ]20MHz200MHzfu4:5MHz90MHz900Mhz Feedbackvoltagerequirementsofinterfacecircuitswithincreasingdynamicrange. PropertyAudiomicrophoneAeroacousticmicrophone Maximuminput120dB160dBFeedbackvoltage6V45V 5-4 andTable 5-5 itcanbeseenthattoimplementadigitalclosedloopcontrollerforanaeroacousticmicrophonetheprocesstechnologymustsupportvoltagesinexcess 129

PAGE 130

Inthenextchapter,theexperimentalsetupandexperimentalresultstocharacterizeananalogclosedloopcontrollerarepresented. 130

PAGE 131

6-1 Figure6-1. OverviewofChapter6. 131

PAGE 132

6-2 Photographofcutawaysectionofpressurecouplershowingrectangularductwith8:5mmsquarecrosssection. (b) PhotographofassembledpressurecouplershowingPCB,plugforreferencemicrophoneandopticalport. Photographofacousticpressurecoupler. Ontheoppositefaceoftheduct,aBruelandKjaerType4138microphoneismountedatnormalincidencetotheacousticwavesandservesasthereferencemicrophone.Thedeviceundertest(DUT)ismountedatgrazingincidencetotheacousticwavesclosetothereferencemicrophone.ThedistancebetweentheDUTandreferencemicrophonedis2:4mm.This 132

PAGE 133

Aninletandexitportismachinedintheducttoallowthechambertobelledwithhelium. Uptothemaximumfrequencyofoperation,theDUTandthereferencemicrophoneareexposedtothesamepressure.ThisisensuredbymakingthedistancefromtheDUTandthereferencemicrophonelessthanaquarterwavelengthatthemaximumfrequencyofoperation.Alsotoensurethatthehigherordermodesaresufcientlyattenuatedattheendofthetestsection,thecrosssectiondimensionsareestablishedsuchthatthecutonfrequencyishigherthanthemaximumfrequencyofinterest.Toincreasetheoperatingfrequencyrange,theductislledwithheliumastheisentropicspeedofsoundinheliumis2:7timesasthatinair.SincethedistancebetweentheDUTandthereferencemicrophoneisxed,thiswillincreasethemaximumfrequencybyapproximately2:7X.Theminimumfrequencyofoperationofthepressurecoupleris300Hzandisdeterminedbythespeaker.Themaximumfrequencyofoperationofthepressurecouplerlledwithheliumis18:3kHzandisdeterminedbytheductdimensions.Withouthelium,thepressurecouplercanbeoperateduptoafrequencyof6:7kHz. 6.3.1CharacterizationofPressureCoupler 6-3 ThefrequencyresponsebetweenthereferencemicrophoneandthesecondBruelandKjaerType4138microphoneisshowninFigure 6-4 withairandheliumasthemediuminthe 133

PAGE 134

Figure6-3. Plotofinputpressuresignalasafunctionoffrequency. Theroll-offofthemagnitudefrequencyresponseplotisbecauseoftherelativepositionofthereferencemicrophoneandDUT.Intheacousticpressurecoupler,thereferencemicrophoneisplacedatnormalincidencetotheacousticsourceandtheDUTisplacedatgrazingincidencetotheacousticsource.Atlowfrequencies(largewavelength),boththereferencemicrophoneandtheDUTaresubjectedtothesameacousticpressureamplitude.Withincreasingfrequency(smallerwavelength),thepressureamplitudeonthereferencemicrophoneremainsconstantwhereasthepressureamplitudeontheDUTkeepsondecreasingandreachesaminimumbasedontheductdimensionandthenincreases.Intheopenandclosedloopfrequencyresponseplotsshownlaterinthischapter,thiseffecthasbeenmitigatedbymultiplyingthemagnitudefrequencyresponsebetweentheDUTandthereferencemicrophonebytheinverseofthemagnitudefrequencyresponseshowninFigure 6-4 Oncethepressurecouplerwascharacterized,openloopandclosedloopcharacterizationofthemicrophoneareperformed.Figure 6-5 showsablockdiagramofthecharacterization 134

PAGE 135

Comparisonoffrequencyresponseofpressurecouplerwithairasmediumandwithheliumasmedium. procedureused.Therstexperimentperformedonthemicrophoneistoobtainthefrequency Figure6-5. Blockdiagramdescribingcharacterizationprocedureused. responseofthemicrophonewithonlyadcbiasandaunitygainfollower.Thisexperimentisperformedtocharacterizethemicrophoneandensurethataworkingmicrophoneisavailableforfurthertests.Thenextsetofexperimentsconsistsofcharacterizingthemicrophonewithadcandacbias.Asynchronousmodulation/demodulationinterfacecircuitisusedintheseexperiments.Theseexperimentsprovidethefrequencyresponse,linearityandnoiseoorofthemicrophoneinopenloopconguration.Tofacilitatestableclosedloopoperation,acompensatorisrequired.Thecompensator,independentoftherestofthesystem,ischaracterizednext.Thisis 135

PAGE 136

Figure6-6. Frequencyresponsewithdcbias(5V)withairmedium. Thetopandbottomplateofthemicrophoneisbiasedwitharegulated5Vdcbias.Sincethespeakerhasabuiltinbandpasslterwithalowfrequencycutinof300Hz,allthedataplottedisshownfrom300Hz.ThespectrumoftheDUTandreferencemicrophoneoutputwasmeasuredusinganFFTanalyzerwitha4Hzbinwidthand1000averagesandisshowninFigure 6-6 .Theroll-offofthemicrophonemagnitudefrequencyresponseisduetotherelativeplacementofthereferencemicrophoneandDUTwhichwasdiscussedearlierinthischapterinSection 6.3.1 .ThefrequencyresponsecorrectedfortheplacementofDUTandreferencemicrophoneisalsoshown. 136

PAGE 137

Frequencyresponsewithdcbias(5V)withheliummedium. ThemagnitudeandphasefrequencyresponseisshowninFigure 6-6 .Itcanbeseenthatthemicrophonehasaatfrequencyresponsefrom300Hzto5:6kHzwithasensitivityof3:8uV=Paat1kHz.Overthisfrequencyrangethephasevariesfrom39degto9deg.Theacousticpressurecoupleristhenlledwithheliumandtheexperimentisrepeated.ThemagnitudeandphasefrequencyresponseofthemicrophoneinheliummediumisshowninFigure 6-7 .Thesystemhasaatfrequencyresponsefrom300Hzto13:4kHzandasensitivityof3:9uV=Pa. 6-8 .Thetop Figure6-8. Schematicofcapacitivemicrophonewithasynchronousmodulationanddemodulationtechniqueusingavoltageamplierbasedcircuit. 137

PAGE 138

Figure6-9. Openloopfrequencyresponsemeasuredattheoutputoflowpasslter(NodeA,Figure 6-8 )withdc(5V)andacbiasvoltages(6Vppand9Vppanddemodulatorat5Vpp)aftercompensatingformagnitudefrequencyresponseslopeintroducedbytheacousticpressurecoupler. Figure6-10. Openloopfrequencyresponsenormalizedtothepeakacvoltagemeasuredattheoutputoflowpasslter(NodeA,Figure 6-8 )withdc(5V)andacbiasvoltages(6Vppand9Vppanddemodulatorat5Vpp)aftercompensatingformagnitudefrequencyresponseslopeintroducedbytheacousticpressurecoupler. 138

PAGE 139

Openlooplinearityoftheopenloopsystemasmeasuredattheoutputofthedemodulator(NodeA,Figure 6-8 )fortwodifferentacbiasvoltageataninputfrequencyof2kHz.Thestraightlinethasaregressionvalue,R2,of0.98 Adcbiasof5Visused.Themicrophoneisexcitedwithaperiodicrandombandlimitedwhitenoisesignalgoingfrom100Hzto6:4kHzwith4periodsusingthePULSEsystem.ThespectrumoftheDUTandreferencemicrophoneoutputwasmeasuredusinganFFTanalyzerwitha4Hzbinwidthand1000averages.TheopenloopfrequencyresponseofthesystemafteraccountingfortheacousticpressurecouplerisshowninFigure 6-9 alongwiththetheoreticalsensitivity.At2kHz,themicrophonehasasensitivityof7:3uV=Pafora6Vppacbiasand10:47uV=Pafora9Vppacbias.Table 6-1 showsacomparisonbetweenthetheoreticalandsimulatedvaluesforthesensitivity.AplotofthesensitivityofthesystemnormalizedtothepeakacbiasvoltageisshowninFigure 6-10 .Itcanbeseenthatboththecurvescoincidewitheachother.Thisshowsthatthesensitivityislinearlyrelatedtothepeakacvoltage. Table6-1. Comparisonofsimulatedandexperimentalsensitivity. dcdcvoltage:5V4V=Pa3:9V=Padc+acdcvoltage:5V,Topandbottomplateacvoltage6Vpp,Demodulator:5Vpp7V=Pa7:3V=Pa 139

PAGE 140

Linearityoftheopenloopsystemnormalizedtopeakacbiasvoltageasmeasuredattheoutputofthedemodulator(NodeA,Figure 6-8 )fortwodifferentacbiasvoltageataninputfrequencyof2kHz.Thestraightlinethasaregressionvalue,R2,of0.98 functionofincreasinginputpressureisshowninFigure 6-11 .Itcanbeseenthatthemicrophonehasalinearresponseupto315Pa(143dB)for6Vppand9Vppacbias.Themaximumappliedpressurewasstoppedat143dBtopreventpossibledamagetothedevice.Theslopeofthelineinthelinearityplotisameasureofthesensitivityofthesystemat2kHz.Themicrophoneitselfhasbeenshowntobelinearupto166dB[ 64 ].Thecircuitoperateswitha15powersupplyandhencewillbelinearuptothispressureinput.Theslopeoftheinputpressureversusoutputvoltagecurveis7:29uV=Pafora6Vppacbiasand10:3uV=Pafora9Vppacbias.Thiscorrespondstothesensitivityofthemicrophoneat2kHzdiscussedinSection 6.3.3 .AplotoftheopenlooplinearityofthemicrophonenormalizedtothepeakacbiasvoltageisshowninFigure 6-12 .Itcanbeseenthatthecurvesforacbiasvoltagesof6Vppand9Vppapproximatelycoincidewitheachother. 6-13 .FromFigure 6-13 ,thenoiseoorat2kHzis1:76uV.Dividingthenoiseoorwiththesensitivity,theminimumdetectablesignal(MDS)is 140

PAGE 141

Noiseoorofmicrophonefortwodifferentacbiasvoltagesmeasuredattheoutputofthedemodulator. obtained.TheMDSat2kHzfor6Vppand9Vppacbiasis0:24Pa(81:6dB)and0:17Pa(78:5dB)respectively.ThetheoreticalnoiseoorderivedinChapter4isalsosuperimposedontheplot.Itisseenthattheexperimentalnoiseislowerthanthetheoreticalnoise.Onepossiblereasonforthiscouldbecausetheampliernoisespecicationusedarethetypicalvalues.Theactualnoisevaluescouldbelowerthanthetypicalvaluesprovidedinthedatasheet. 6-14 .AplotofthetheoreticalcompensatorresponsediscussedinChapter4isoverlaidontheexperimentalplotinFigure 6-14 .FromFigure 6-14 ,itcanbeseenthatthecompensatorhasadcgainof77:8dB.Itcanalsobeseenthatthecompensatorhasadominantpoleat1kHzandanon-dominantpoleat30kHz.Thezeroofthesystemisat230kHzandisnotcapturedinFigure 6-14 .Thisisbecausethespectrumanalyzer(SRS785)hasabandwidthof101:3kHz.Itcanbeobservedfromthephasefrequencyplotthatthephasedoesnotcontinuereducingafterthesecondpolebutstartsto 141

PAGE 142

Figure6-14. Theoreticalandexperimentalfrequencyresponseofcompensator. 6-15 .From Figure6-15. Blockdiagramofanalogclosed-loopsystemshowingmicrophoneandinterfacecircuit. linearcontroltheory[ 126 ],duringclosedloopoperation,iftheforwardpathofthesystemhasahighgain,thesignalatanypointintheforwardpathofaclosedloopsystemisattenuatedbytheloopgain.Thus,todemonstrateclosedloopoperation,thefrequencyresponseofthesystemmeasuredattheoutputoftheinterfacecircuit(NodeAinFigure 6-15 )iscomparedtotheopenloopresponseshowninFigure 6-9 .Thefrequencyresponseoftheclosedloopsystemmeasured 142

PAGE 143

6-15 )isshowninFigure 6-16 alongwiththetheoreticalcurveforacbiasvoltagesof6Vppand9Vpp.ThisfrequencyresponseiscomparedtothefrequencyresponseoftheopenloopsystemshowninFigure 6-17 andFigure 6-18 for6Vppand9Vppacbias.Themagnitudefrequencyresponseat2kHzis1:27uV=Pawitha6Vppacbias Figure6-16. Closedloopfrequencyresponsemeasuredattheoutputofdemodulatorwithdcbiasof5V,acbiasof9Vpp,anddemodulatorat5Vpp(NodeAinFigure 6-15 ).FrequencyresponsecorrectedforreferencemicrophoneandDUTlocation. voltageand1:89uV=Pawitha9Vppacbiasvoltage.Anincreaseinphasewasobservedinclosedloopresponseattheoutputofthedemodulatorwhichwasnotobservedinopenloop.Thecauseofthisphaseincreaseisnotclearbutmaybeduetothecouplingcapacitorsusedbetweenthestages.Comparedwiththeopenloopmagnitudefrequencyresponsemeasuredat2kHz,withthesamebiasvoltage,thisrepresentsa5reductionindiaphragmdisplacement.Thusthesystemoperatesinastablefashioninclosedloop.Theexperimentalloopgainofthesystem,Tevaluatedat2kHz,isgivenbytheproductofthesystemsensitivity,H1,expressedinuV=Nevaluatedat2kHz,compensatorgainH2,evaluatedat2kHz,andvoltagetoforcetransferfunction,H3,evaluatedat2kHz,i.e. 4 .Also,thesensitivityofthesysteminuV=PaobtainedfromFigure 6-9 isconvertedtouV=Nbydividingthesensitivitybythe 143

PAGE 144

Closedloopfrequencyresponsemeasuredattheoutputofdemodulator(NodeAinFigure 6-15 )withdcbiasof5V,acbiasof6Vpp,anddemodulatorat5Vppinopenandclosedloopoperation.FrequencyresponsecorrectedforreferencemicrophoneandDUTlocation. effectiveareaofthediaphragm.UsingEquation 6 ,theexperimentalloopgainofthesystemis5:28.Fromlinearcontroltheory,thediaphragmdisplacementisattenuatedby1+T.Thisimpliesthatthediaphragmdisplacementshouldbeattenuatedby6:28.Theactualdiaphragmdisplacementisonly5.Thisdiscrepancycanbeattributedtothemodelingerrorsusedinthevoltagetoforcetransferfunction.Figure 6-17 andFigure 6-18 showacomparisonofthemagnitudeandphasefrequencyresponseinopenandclosedloopoperationwithanacbiasof6Vppand9VpprespectivelycompensatedforthereferencemicrophoneandDUTlocation. Thefrequencyresponseoftheclosedloopsystemattheoutputofthecompensator(NodeB,Figure 6-15 )aftercompensatingforthetransferfunctionofthepressurecouplerisshowninFigure 6-19 .Themagnitudefrequencyresponseat2kHzis2:63mV=Pawitha6Vppacbiasvoltageand3:56mV=Pawitha9Vppacbiasvoltage. 6-15 )for6Vppand9VppacbiasvoltagesacbiasvoltagesisshowninFigure 6-20 .Itcanbeseenthatthesystemhasalinearresponseuptoaninputpressureof315Pa(144dB).Further 144

PAGE 145

Closedloopfrequencyresponsemeasuredattheoutputofdemodulator(NodeAinFigure 6-15 )withdcbiasof5V,acbiasof9Vpp,anddemodulatorat5Vppinopenandclosedloopoperation.FrequencyresponsecorrectedforreferencemicrophoneandDUTlocation. increaseininputpressurewasnotcarriedouttopreventpossibledamagetothemicrophonediaphragm.Theslopeoftheinputpressureversusoutputvoltagecurveis2:63mV=Pafora6Vppacbiasand3:56mV=Pafora9Vppacbias.Thiscorrespondstothesensitivityofthemicrophoneat2kHzinFigure 6-19 .AplotoftheopenlooplinearityofthemicrophonenormalizedtothepeakacbiasvoltageisshowninFigure 6-21 .Itcanbeseentheplotsforforacbiasvoltagesof6Vppand9Vppapproximatelycoincidewitheachother. 6-22 .FromFigure 6-22 ,thenoiseoorat2kHzis1:11mV.Dividingthenoiseoorwiththesensitivity,theminimumdetectablesignal(MDS)isobtained.TheMDSat2kHzfor6Vppand9Vppacbiasis106:5dBand103:8dBrespectively.Thelargenoiseoorisdiscussedinthenextsection. 145

PAGE 146

Closedloopfrequencyresponsemeasuredattheoutputofcompensator(NodeBinFigure 6-15 )withdcbiasof5Vacbiasof9Vpp,anddemodulatorat5Vpp.FrequencyresponsecorrectedforreferencemicrophoneandDUTlocation. Figure6-20. Linearrangeofsysteminclosedloopcongurationasmeasuredattheoutputofthecompensator(NodeBinFigure 6-15 )with6Vppand9Vppacbiasvoltagesforaninputfrequencyof2kHz.Thestraightlinethasaregressionvalue,R2,of0.97 6.4 .Thereductionintheoutputofthedemodulatorisapproximately5whichisapproximatelyconsistentwiththatpredictedtothereductionofloopgain. 146

PAGE 147

Linearityoftheclosedloopsystemnormalizedtopeakacbiasvoltageasmeasuredattheoutputofthecompensator(NodeBinFigure 6-15 )fortwodifferentacbiasvoltageataninputfrequencyof2kHz.Thestraightlinethasaregressionvalue,R2,of0.98 Figure6-22. Noiseoorofthecircuitinclosedloopoperationasmeasuredattheoutputofthecompensator(NodeBinFigure 6-15 )with6Vppand9Vppacbiasvoltages. Theperformancemetricsoftheopenloopmicrophoneusingsynchronousmodulationanddemodulationinterfacecircuitiscomparedtotheperformancemetricsofthemicrophoneusingalownoisevoltageamplierandachargeamplier[ 64 ].Martinreportsasensitivityof166uV=Paand368uV=Pawitha9:3Vbiasusingavoltageamplierandchargeamplierrespectively.Thereportedinterfacecircuitshaveanoiseoorof22:7dBand41:5dBwithavoltageandchargeamplierrespectively.Thelinearrangeoftheinterfacecircuitextendsuptoaninputpressureof166dB.Basedonthisinformation,thedynamicrangereportedinthis 147

PAGE 148

Characterizationresultsofmicrophoneinopenandclosedloopmodeofoperation. Openloop(ac6Vpp)7:3uV=Pa81:681:6dB143dBOpenloop(ac9Vpp)10:47uV=Pa78:578:5dB143dB Closedloopcapacitiveinterfacecircuitshaveprimarilybeenusedforlowbandwidth(12kHz)applicationsuchasaccelerometers[ 68 67 11 2 112 ].Themainadvantageofaclosedloopinterfacecircuitisthepossibilityofincreasedbandwidthandlinearrange.Alsoclosedloopinterfacecircuitshavethepotentialtoaddresspullinissuesasthediaphragmdeectionisreduced.Inthiswork,aproofofconceptclosedloopinterfacecircuithasbeenpresentedforacomparativelyhigherbandwidth(20kHz)applicationcomparedtopotentialclosedloopaccelerometerapplications.Closedloopcontrolwasdemonstratedbyshowingthatthediaphragmdeectionwasreducedduringclosedloopoperation.Thiswasveriedfortwodifferentvoltagebiasconditions.Thedisadvantagesofclosedloopinterfacecircuitsisahigher 148

PAGE 149

149

PAGE 150

Inthischapter,thekeygoalsoftheworkandtheresultsaresummarized.Keycontributionsofthisresearcharehighlighted.Thischapterconcludeswithadiscussionontheavenuesforfurtherresearchinthisarea. Inthiswork,anextensivestudyofopenloopandclosedinterfacetechniquesforcapacitivesensorshasbeenpresentedalongwiththeoreticalformulationsforkeyperformancemetrics. AnanalogclosedloopcontrollerforadualbackplateMEMScapacitivemicrophonewasdesigned.AdetailedMATLABmodelwhichcaptureskeydynamicsofthesensorandcontrollerwasdeveloped.Noisemodelsoftheclosedloopsystemwasalsodeveloped.Openandclosedloopsimulationswereperformedforstepandsinusoidalexcitations.Fromthesesimulations,itcanbeseenthatinclosedloopoperation,thediaphragmdisplacementisattenuatedwhichcanpotentiallyaddresspullinissues. Aproofofconceptclosed-loopanalogcontrollerforaMEMScapacitivemicrophonewasdevelopedandcharacterized.Atestapparatuswasdevelopedtocharacterizethesystemovertheaudiorangebyoperatingthemicrophoneinaheliummediumwhichincreasesthebandwidthofthetestapparatus.Characterizationofthemicrophoneinopenandclosedloopmodeofoperationispresented.Resultsshowthatstableclosedloopoperationofthemicrophoneisfeasiblewithincreasedsensitivityandthepotentialtoaddresspull-inissues. 150

PAGE 151

AnotherareaofresearchisdetailedtransistorlevelsimulationsusingSPICE.TransistorlevelsimulationsallowsnoiseanalysisofthesystemwhichisnotpossiblewithMATLABsimulations.Tosuccessfullyperformthesesimulations,themicrophonemustalsobemodeledinSPICEusinganalogmodelingblocks.Furthermorenoisemodelsforthecircuitelementsmustbeavailable. Itwasalsonoticedthat60Hznoiseinboththeinputchannelsofthedemodulatorcaused120Hznoisesignalsattheoutputofthedemodulator.Oneofthesourcesof60HzpickupisthecablesconnectingthefunctiongeneratoroutputtothecontrollerPCB.ThissourceofnoisepickupcanbereducedbygeneratingtheclocksignalonthePCBitselfusingoscillatorICs. Anotheravenueforresearchcouldbetomonolithicallyintegratetheinterfacecircuitandcontrollerusingahighvoltage,highfrequency,andlownoiseprocess.Thiswillgreatlyimprovecircuitperformancebecauseofthereducedparasiticswhichwillimprovethesensitivity.Alsomonolithicintegrationwilleliminatetheneedforcablesthatareusedforclockgenerationastheclocksignalcanbegeneratedonchip. 151

PAGE 152

A-1 Thisresults FigureA-1. Noisemodelofdualbackplatemicrophoneconnectedtovoltagebuffer. inxedchargeof+QdcandQdconthetopandbottomcapacitorrespectively.CpandCirepresenttheinterconnectparasiticcapacitanceandtheinputcapacitanceoftheamplierrespectively.Thetopandbottomplatecapacitanceisgivenby x0x0 x0+x0 x0 Equation A canbeexpressedintermsofthenominalcapacitanceC0andavaryingcapacitanceCusingtherelation

PAGE 153

A thevoltageacrossthetopandbottomcapacitorV1andV2respectivelyisgivenby A ,Equation A inEquation A weobtain x0x0=C1+x0 x0+x0=C2x0 A forC1andC2 x0x0 x0+x0 Thenalchargeisgivenby EquatingEquation A andEquation A weobtain 153

PAGE 154

A weobtain SolvingforVoutC1weobtain SimilarlyforVoutC2weobtain ThetotaloutputoftheamplierVoutisthesumofoutputsgivenbyEquation A andEquation A IfthecapacitancechangeC1andC2areequal,i.e, andusingEquation A ,inEquation A weobtain C0Vdc:(A) FromtheinputoutputrelationgivenbyEquation A ,theelectricalsensitivityofthecircuitisgivenbytheequation A-2 .Thecurrentandvoltagenoisespectraldensity(PSD)oftheamplieraredenotedbySiaandSvarespectively.ThecurrentandvoltagenoisePSDSi(f)andSe(f)canbeexpressedasafunctionofthecurrentandvoltagenoisecornerfrequencyfciandfceby

PAGE 155

Noisemodelofdualbackplatemicrophoneconnectedtovoltagebuffer. ThepowerspectraldensityofthenoiseoftheresistorisdenotedbySvR.Itisgivenby TheimpedanceoftheparasiticcapacitancesandsensecapacitordenotedbyZiisgivenby Wenowusesuperpositiontocalculatetheoutputvoltagenoisepowerspectraldensity.WerstconsideronlythenoisegeneratorSva.WithonlytheSvanoisegenerator,nocurrentowinZiorRdc.ThenoiseattheoutputSvo1isgivenby NextweconsideronlythenoisegeneratorSia.ItcanbeseenthatthetotalimpedanceseenbythisnoisegeneratorisistheparallelcombinationofZiandRdc.ThusthenoisegeneratedattheoutputSvo2isgivenby Finallyweconsiderthenoiseduetotheresistor.ItcanbeseenfromthecircuitthatnoiseduetotheresistorisshapedbythevoltagedividerconsistingoftheofRdcandZi.HencethenoiseattheoutputSvo3duetotheresistorisgivenby 155

PAGE 156

A ,Equation A ,Equation A Itshouldbenotedthatthenoiseanalysispresentedhereconsidersonlythenoisesourcesduetotheelectricalinterfacecircuit.Thethermo-mechanicalsourcesofnoiseassociatedwiththesensoritselfarenotconsideredinthisanalysis. and Theoutputvoltageforachargeamplierisgivenby SubstitutingforQ1andQ2inEquation A andsimplifying Fromtheinput-outputrelationshipgivenbyEquation A theelectricalsensitivityofthecircuitisgivenbytheequation 156

PAGE 157

A-3 .ThecurrentandvoltagenoisepowerspectraldensityoftheamplieraredenotedbySvaandSiarespectively.ThepowerspectraldensityofthecurrentnoiseoftheresistorisdenotedbySiR.Itisgivenby TheimpedanceoftheparasiticcapacitancesandsensecapacitorwhichisdenotedbyZiisgivenbyma TheimpedanceofthefeedbacknetworkdenotedbyZfisgivenby FigureA-3. Noisemodelofdualbackplatemicrophoneconnectedtochargeamplier. consideronlythenoisegeneratorSva.ThenoiseattheoutputSvo1duetoSvanoisegeneratorisgivenby 157

PAGE 158

Similarlytheoutputvoltagenoiseduetotheresistorcurrentnoise,Svo3,isgivenbytherelation UsingsuperpositionthepowerspectraldensityoftheoutputnoisevoltageSvoisgivenbythesumofEquation A ,Equation A ,andEquation A andisgivenbytheexpression, SinceweareinterestedinthenoisePSDafterthecutinfrequencyi.e., Simplifying (!Cint)2:(A) Equation A showsthatboththeresistorcurrentnoiseandtheampliercurrentnoiseare1=f2shaped. Thecarrierwaveinputr(t)tothedemodulatoralsoconsistsofasquarewaveatafrequencyof!cbutwithanamplitudeofVc.Thefourierseriesrepresentationx(t)ofasquarewaveofamplitudeAwith50%dutycycleisgivenby cos(!t)+1 32A cos(3!t)+1 52A cos(5!t):::(A) 158

PAGE 159

Thecarrierwaveappliedtothedemodulatorisgivenby Theoutputofthebufferampliervmod(t)isgivenbyEquation 3 as Iftheinputacousticsignalisatafrequency!m,andthemaximumdisplacementofthediaphragmisx00then Therefore Substitutingforvtp UsingthetrigonometricidentitycosAcosB=1=2(cos(A+B)cos(Ab)),vmod(t)canbeexpressedas 159

PAGE 160

Expandingtheaboveequation 2[cos(!m)t+((((((((cos(!m+2!c)t]+(((((((((((2Vac 2[((((((((cos(!m2!c)t+cos(!m)t] 2[cos(!m)t]+2Vac 2[cos(!m)t]

PAGE 161

Substitutingforx00cos(!m)tasx0weobtain Thelinearizedsensitivityofthisinterfacetechniqueis 161

PAGE 162

108 ].Thetimeconstantofasinglepoleamplierisrelatedtoitsclosedloopbandwidthfubytherelation 2:(B) Letthenumberoftimeconstantsrequiredfortheampliertosettlebedenotedbyn,samplingfrequencybyfsandthefractionoftimespentintheamplifyphasebym(<1).Thetimeavailablefortheswitchedcapacitorampliertosettleisthusm1 Substitutingforandrearrangingtheequation,weobtain 2m nfs1(B) 108 ].ThenoisemodeloftheamplierduringtheamplicationphaseisshowninFigure B-1 FigureB-1. Noisemodelofswitchedcapacitoramplierduringamplifyphase. 162

PAGE 163

WeshallnowderivethetotalnoiseattheoutputduetokT=Cnoiseandduetothewidebandampliernoise. CT:(B) ReferringthisnoisetotheoutputbymultiplyingthepreviousequationbythesquareofthegainweobtainthetotaloutputkT=CnoiseThetotalnoiseattheoutputoftheamplierisgivenby CT:(B) Thetotaloutputnoisevoltageis Forasinglepolesystem,theequivalentnoisebandwidthis=2timesthebandwidth[ 79 ].Thereforethetotaloutputnoiseforanamplierwithaclosedloopbandwidthoffu 163

PAGE 164

[1] MarcMadou,FundamentalsofMicrofabrication,CRCPress,BocaRaton,1997. [2] AnalogDevices,Datasheetadxl50,monolithicaccelerometerwithsignalconditioning,Tech.Rep.,OneTechnologyWay,Norwood,MA,1993. [3] TexasInstruments,Digitallightprocessing,WorldWideWeb, [4] FreescaleSemiconductors,Datasheetmpxy8040a:Tirepressuremonitorsensor,Tech.Rep.,Phoenix,AZ,1997. [5] KnowlesAcoustics,Sisonicmicrophonesdatasheet,Tech.Rep.,1111MapleDrive,Ithasca,IL60143,2002. [6] WeijeiYun,RogetT.Howe,andPaulR.Gray,Surfacemicromachined,digitallyforcebalancedaccelerometer,inSolid-StateSensorandActuatorWorkshop.1992,pp.126131,IEEE. [7] R.P.vanKampen,M.J.Vellekoop,P.M.Sarro,andR.F.Wolffenbuttel,Applicationofelectrostaticfeedbacktocriticaldampingofanintegratedsiliconcapacitiveaccelerometer,SensorsandActuatorA,vol.43,pp.100,1994. [8] MarkLemkinandBernhardE.Boser,Athree-axismicromachinedaccelerometerwithacmosposition-senseinterfaceanddigitaloffset-trimelectronics,IEEEJournalofSolidStateCircuits,vol.34,no.4,pp.456,April1999. [9] Y.Matsumoto,M.Nishimura,M.Matsura,andM.Ishida,Three-axissoicapacitiveaccelerometerwithpllcvconverter,SensorsandActuatorA,vol.75,no.1,pp.77,May1999. [10] TetsuyaKajita,Un-KuMoon,andGaborC.Temes,Anoiseshapingaccelerometerinterfacecircuitfortwochipimplementation,inIEEEInstrumentandMeasurementTechnologyConference,Budapest,Hungary.IEEE,2001,pp.1581,IEEE. [11] BernhardE.BoserandRogerT.Howe,Surfacemicromachinedaccelerometers,inCustomIntegratedCircuitsConference.1995,pp.337,IEEE. [12] J.M.Huang,A.Q.Liu,Z.L.Deng,J.AhnQ.X.Zhang,andA.Asundi,Anapproachtothecouplingeffectbetweentorsionandbendingforelectrostatictorsionalmicromirrors,SensorsandActuatorA,vol.115,no.1,pp.159,September2004. [13] JohnBergqvist,RudolfF,MasianoJ,andRossiM,Asiliconcondensermicrophonewithahighlyperforatedbackplate,1991. [14] W.H.Hsieh,Tseng-YangHsu,andYu-ChongTai,Amicromachinedthin-lmteonelectretmicrophone,inInternationalConferenceonSolidStateSensorsandActuators,TRANSDUCERS'97,Chicago,1997,vol.1,pp.425. [15] W.Kronast,B.Muller,W.Siedel,andA.Stoffel,Single-chipcondensermicrophoneusingporoussiliconassacriciallayerfortheairgap,inTheEleventhAnnualInternationalWorkshoponMicroElectroMechanicalSystems,MEMS98,Heidelberg,Germany,1998,pp.591,IEEE. 164

PAGE 165

SiebeBouwstra,TorbenStorgaard,PatrickScheeper,JensOleGullov,JesperBay,MatthiasMullenborg,andPirminRombach,Siliconmicrophones-aDanishperspective,JournalofMicromechanicalMicroengineering,vol.8,pp.64,1998. [17] AlttiTorkkelia,OutiRusanenb,JaakkoSaarilahtia,HeikkiSeppc,HannuSipolac,andJarmoHietanen,Capacitivemicrophonewithlow-stresspolysiliconmembraneandhigh-stresspolysiliconbackplate,SensorsandActuatorA,vol.85,pp.116,2000. [18] MullenbornM.,RombachP.,KleinU.,RasmussenK.,KuhnmannJ.F.,HeschelM.,AmskovGravadM.,JantingJ.,BranebjergJ.,HoogerwerfA.C.,andBouwstraS.,Chip-sizepackagedsiliconmicrophone,SensorsandActuatorA,vol.92,pp.23,2001. [19] RombachP.,Mullenborn,KleinU.,andRasmussenK.,Therstlowvoltage,lownoisedifferentialcondensersiliconmicrophone,technologydevelopmentandmeasurementresults,SensorsandActuatorA,vol.95,pp.196,2002. [20] JingChen,LitianLiu,ZhijianLi,ZhiminTan,YangXu,andJunMa,Single-chipcondenserminiaturemicrophonewithahighsensitivecircularcorrugateddiaphragm,inTheFifteenthIEEEInternationalConferenceonMicroElectroMechanicalSystems,2002,pp.284287. [21] BorgeNordstrandPatrickScheeper,JensGullev,BinLiu,ThomasClausen,LiseMidjord,andTorbenStorgaard-Larsen,Anewmeasurementmicrophonebasedonmemstechnology,IEEEJournalofMicroelectromechanicalSystems,vol.12,no.6,pp.880,December2003. [22] S.T.Moe,K.Schjlberg-Henriksen,D.T.Wang,E.Lund,J.Nysth,L.Furuberg,M.Visser,T.Fallet,andR.W.Bernstein,Capacitivedifferentialpressuresensorforharshenvironments,SensorsandActuatorA,vol.83,no.1-3,pp.30,May2000. [23] ShuwenGuo,JunGuo,andWenH.Ko,Amonolithicallyintegratedsurfacemicromachinedtouchmodecapacitivepressuresensor,SensorsandActuatorA,vol.80,no.3,pp.224,March2003. [24] KlausKasten,NorbertKordas,HolgerKappert,andWilfriedMokwa,Capacitivepressuresensorwithmonolithicallyintegratedcmosreadoutcircuitforhightemperatureapplications,SensorsandActuatorA,vol.97-98,no.1,pp.83,April2002. [25] Min-XinZhou,Qing-AnHuang,andMingQin,Modeling,designandfabricationofatriple-layeredcapacitivepressuresensor,SensorsandActuatorA,vol.117,no.1,pp.71,January2005. [26] Y.C.LokeandK.M.Liew,VHFenhancementofmicromechanicalresonatorviastructuralmodication,SensorsandActuatorA,vol.96,no.1,January2002. [27] F.F.FaheemandY.C.Lee,Tether-andpost-enabledip-chipassemblyformanufacturablerf-mems,SensorsandActuatorA,vol.114,no.2-3,pp.486,September2004. [28] R.Fritschi,S.Frdrico,Hightuningrangecapacitorsfabricatedwithsacricialamorphoussiliconsurfacemicromachining,MicroelectronicsEngineering,vol.73-74,pp.447,June2004. 165

PAGE 166

SiavashPourkamaliandFarrokhAyazi,ElectricallycoupledMEMSbandpasslters:Parti:Withcouplingelement,SensorsandActuatorA,vol.122,no.2,pp.307,August2005. [30] SiavashPourkamaliandFarrokhAyazi,ElectricallycoupledMEMSbandpasslters:Partii:Withoutcouplingelement,SensorsandActuatorA,vol.122,no.2,pp.317,August2005. [31] MotoakiHara,JanKuypers,TakashiAbe,andMasayoshiEsashi,SurfacemicromachinedAlNthinlm2ghzresonatorforcmosintegration,SensorsandActuatorA,vol.117,no.2,pp.211,January2005. [32] BinXiong,LufengChe,andYuelinWang,Anovelbulkmicromachinedgyroscopewithslotsstructureworkingatatmosphere,SensorsandActuatorA,vol.7,no.2,pp.137,October2003. [33] W.Geiger,T.Link,J.Merz,S.Steigmajer,A.Hauser,H.Sandmaier,W.Lang,andN.Niklasch,Newdigitalreadoutelectronicsforcapacitivesensorsbytheexampleofmicro-machinedgyroscopes,SensorsandActuatorA,vol.97-98,pp.557,April2002. [34] HuikaiXieandGaryK.Fedder,Verticalcomb-ngercapacitiveactuationandsensingforCMOS-MEMS,SensorsandActuatorA,vol.95,no.2-3,pp.212,January2002. [35] FarrokhAyaziandKhalilNaja,Highaspect-ratiopolysiliconmicromachiningtechnology,SensorsandActuatorA,vol.87,no.1-2,pp.46,December2002. [36] LarryK.Baxter,CapacitiveSensors,DesignandApplication,ElectronicsTechnology.1997. [37] GeorgeS.K.WongandTonyF.W.Embleton,AIPHandbookofCondensorMicrophones,ModernAcousticsandSignalProcessing.AmericanInstituteofPhysics,NewYork,1995. [38] BruelandKjr,MicrophoneHandbook-TechnicalDocumentation,vol.I,BruelandKjr,1996. [39] KnowlesAcoustics,Datasheet,EMserieshighperformancemicrophone,Tech.Rep.,Itasca,IL,2003. [40] NationalSemiconductors,DatasheetLMV1012pre-ampliedICsforhighgain2-wiremicrophones,Tech.Rep.,SantaClara,CA,2004. [41] NaiyavudhiWongkometandBernhardE.Boser,Correlateddoublesamplingincapacitivepositionsensingcircuitsformicromachinedapplications,inIEEEAsia-PacicConferenceonCircuitsandSystems.1998,pp.723,IEEE. [42] WidgeHenrion,LenDiSanza,MatthewIp,StephenTerry,andHaiJerman,Widedynamicrangedirectdigitalaccelerometer,inIEEESolid-StateSensorandActuatorWorkshop,HiltonHeadIsland,SouthCarolina,June,1990,pp.153. [43] KoujiMochizukiandKenzoWatanabe,Arelaxationoscillatorbasedinterface,IEEETransactionsonInstrumentationandMeasurement,vol.47,no.1,pp.11,February1998. 166

PAGE 167

AndrzejClchockiandRolfUnbehauen,Aswitched-capacitorinterfaceforcapacitivesensorsbasedonrelaxationoscillators,IEEETransactionsonInstrumentationandMeasurement,vol.39,no.5,pp.797,October1990. [45] FrankM.L.vanderGoesandGerardC.M.Meijer,Anovellow-costcapacitive-sensorinterface,IEEETransactionsonInstrumentationandMeasurement,vol.45,no.2,pp.536,April1996. [46] StephenD.Senturia,MicrosystemDesign,KluwerAcademicPublishers,1stedition,2001. [47] SonionTechnologies,Digsimicmicrophonestc100z21adatasheet,Tech.Rep.,SonionHorsensA/SFuglevangsvej45DK-8700HorsensDenmark,2005. [48] GeorgeBrasseur,Designrulesforrobustcapacitivesensor,IEEETransactionsonInstrumentationandMeasurement,vol.52,no.4,pp.1261,2003. [49] GeorgeBrasseur,Stateoftheartofrobustcapacitivesensors,in1stInternationalWorkshoponRoboticSensing,2003. [50] ChristianneThierlemannandGiselaHess,Amicro-machinedcapacitiveelectretmicrophone,inSymposiumonDesign,Test,andMicrofabricationofMEMSandMOEMS,Paris,France,1999,SPIE,vol.3680,pp.748. [51] QuanboZou,ZhiminTan,ZhenfengWang,JiangtaoPang,XinQian,QingxinZhang,RongmingLin,SungYi,HaiqingGong,LitianLiu,andZhijianLi;,Anovelintegratedsiliconcapacitivemicrophone-oatingelectrodeelectretmicrophone(feem),IEEEJournalofMicroelectromechanicalSystems,vol.7,no.2,pp.224,June1998. [52] HohmD.andGerhardMulthauptR.,Silicon-dioxideelectrettransducer,JournaloftheAcousticalSocietyofAmerica,vol.75,no.4,pp.1297,April1984. [53] BruelandKjr,Measuringmicrophones,WorldWideWeb, [54] GeorgeS.K.WongandTonyF.W.Embleton,AIPHandbookofCondensorMicrophones,ModernAcousticsandSignalProcessing.AmericanInstituteofPhysics,NewYork,1995. [55] PatrickScheeper,WouterOlthuis,andPietBergveld,Fabricationofasubminiaturesiliconcondensermicrophoneusingthesacriciallayertechnique,inInternationalConferenceonSolid-StateSensorsandActuators,TRANSDUCERS'91,SanFrancisco,CAUSA,1991,pp.408,IEEE. [56] PatrickScheeper,A.G.HvanderDonk,WouterOlthuis,andPietBergveld,Fabricationofsiliconcondensermicrophonesusingsinglewafertechnology,IEEEJournalofMicroelectromechanicalSystems,vol.1,no.3,pp.147,September1992. [57] WolfgangKuhnelandGiselaHess,Asiliconcondensermicrophonewithstructurebackplateandsiliconnitridemembrane,SensorsandActuatorA,vol.30,pp.251,1992. 167

PAGE 168

JohanBergqvistandJeanGobet,Capacitivemicrophonewithasurfacemicromachinedbackplateusingelectroplatingtechnology,IEEEJournalofMicroelectromechanicalSystems,vol.3,no.2,pp.69,June1994. [59] MichaelPederson,WouterOlthuis,andPietBergveld,Highperformancecondensermicrophonewithfullyintegratedcmosamplieranddc-dcvoltageconverter,IEEEJournalofMicroelectromechanicalSystems,vol.7,no.4,pp.387,December1998. [60] MichaelPedersen,WouterOlthuis,andPietBergveld,Anintegratedsiliconcapacitivemicrophonewithfrequency-modulateddigitaloutput,SensorsandActuatorA,vol.69,pp.267,1998. [61] JesperBay,OleHansen,andSiebeBouwstra,Designofasiliconmicrophonewithdifferentialread-outofasealeddoubleparallel-platecapacitor,SensorsandActuatorA,vol.53,pp.232,1996. [62] RomachP.,Mullenborn,KleinU.,andRasmussenK.,Therstlowvoltage,lownoisedifferentialcondensersiliconmicrophone,inThe14thEuropeanConferenceonSolid-StateTransducers,EUROSENSORSXIV,Copenhagen,Denmark,2000,pp.213. [63] Mullenborn,P.Rombach,U.Klein,K.Rasmussen,J.F.Kuhmann,M.Heschel,M.AmskovGravad,J.Janting,J.Branebjerg,A.C.Hoogerwer,andS.Bouwstra,Chip-sizepackagedsiliconmicrophones,SensorsandActuatorA,vol.92,pp.23,2001. [64] DavidT.Martin,Design,FabricationCharacterizationofaMEMSDualBackplateCapacitiveMicrophone,Ph.D.dissertation,UniversityofFlorida,Gainesville,FL,2007. [65] JoostC.Lotters,WouterOlthuis,PeterH.Veltnik,andPietBergveld,Asensitivedifferentialcapacitancetovoltageconverterforsensorapplications,IEEETransactionsonInstrumentationandMeasurement,vol.48,no.1,pp.89,February1999. [66] SatomiOgawa,YukataOisugi,KouikiMochizuki,andKenzoWatanabe,Aswitchedcapacitorinterfacefordifferentialcapacitancetransducers,IEEETransactionsonInstrumentationandMeasurement,vol.50,no.5,pp.1296,October2001. [67] vanderDonkA.G.H,ScheeperP.R.,OlthuisW.,andBergveld,Amplitudemodulatedelectromechanicalfeedbacksystemforsiliconcondensermicrophones,JournalofMicromechanicalMicroengineering,vol.2,pp.211,1992. [68] A.G.H.vanderDonk,A.Sprenkels,WouterOlthuis,andPietBergveld,Preliminaryresultsofasiliconcondensermicrophonewithinternalfeedback,in1991InternationalConferenceonSolid-StateSensorsandActuatorsTRANSDUCERS'91,NewYork,1991,pp.262,IEEE. [69] MichaelKraft,ClosedLoopDigitalAccelerometerEmployingOversamplingConversion,Ph.D.dissertation,SchoolofEngineering,CoventryUniversity,UK,July,1997. [70] KrylovS.andMaimonR.,Pull-indynamicsofelectrostaticbeamactuatedbycontinuouslydistributedelectrostaticforce,JournaloftheAcousticalSocietyofAmerica,vol.126,pp.332,1993. 168

PAGE 169

HuntF.,Electroacoustics:theanalysisoftransduction,anditshistoricalbackground,HarvardUniversityPress,Cambdridge,MA,1954. [72] ScheeperP.R.,vanderDonkA.G.H.,OlthuisW.,andBergveldP.,Areviewofsiliconmicrophone,SensorsandActuatorA,vol.44,pp.1,February1994. [73] DavidT.Blackstock,FundamentalsofPhysicalAcoustics,WileyInterscience,NewYork,2000. [74] M.Rossi,AcousticsandElectroacoustics,ArtechHouse,Norwood,MA,1988. [75] Z.Skvor,Ontheacousticalresistancetoviscouslossesintheairgapofelectrostatictransducers,Acoustica,vol.19,pp.295,1996. [76] LiuJ.,MartinD.,KadirvelK.NishidaT.,CattafestaL.,SheplakM.,andMannB.,Nonlinearmodelandsystemidenticationofacapacitivedual-backplatememsmicrophone,JournalofSoundandVibration,2006. [77] BruelandKjr,Measuringsound,WorldWideWeb, [78] TimoshekoS.andKriegerS.,Theoryofplatesandshells,McGrawHillBookCompany,NewYork,1959. [79] DavidA.JohnsandKenMartin,AnalogIntegratedCircuitDesign,JohnWileyandSons,Inc,NewYork,1997. [80] A.vanderZiel,NoiseinSolidStateDevicesandCircuits,JohnWileyandSons,NewYork,1986. [81] FredricoAlimenti,PaoloMezzanotte,LucaRoselli,andRobertoSorrentino,Modelingandcharacterizationofbondingwireinterconnection,IEEETransactionsonMicrowaveTheoryandTechqniques,vol.49,no.1,pp.142,January2001. [82] StandardWireandCableCompany,Datasheetrg164coaxialcable,Tech.Rep.,RanchoDomniguez,CA,2000. [83] MarkSerridgeandTorbenR.Licht,PiezoelectricAccelerometersandVibrationPreampliers-TheoryandApplicationHandbook,BruelandKjr,NewYork,1987. [84] FurstC.andMarczakT,Preamplierfortwoterminalelectretcondensermicrophones,USPatent6888408,2005. [85] NationalSemiconductors,Lmv10xxanalogpre-ampliedic'sforhighgainmicrophones,Tech.Rep.,2900SemiconductorDrive,SantaClara,CaliforniaUSA95050,2003. [86] KillionM.andCarlsonE.,Audiofrequencyamplicationcircuit,USPatent3512100,1977. [87] MurphyP.,HubschiK.,RoojiN.,andRacine,Subminiaturesiliconintegratedelectretcapacitormicrophone,IEEETransactionsonElectricalInsulation,vol.24,no.3,pp.495,June1989. 169

PAGE 170

Y.B.Ning,A.W.Mitchell,andR.N.Tait,Fabricationofasiliconmicromachinedcapacitivemicrophoneusingadry-etchprocess,SensorsandActuatorA,vol.53,pp.237,1996. [89] VoorthuyzenJ.,SprenkelsA.,vanderDonk,ScheeperP,andBergveldP,Optimizationofcapacitivemicrophoneandpressuresensorperformancebycapacitorelectrodeshaping,SensorsandActuatorA,vol.25,pp.331,1991. [90] KuhnelWandHessG.,Asiliconcondensermicrophonewithstructuredbackplateandsiliconnitridemembrane,SensorsandActuatorA,vol.30,pp.251,1992. [91] BergqvistJ.andGobetJ.,Capacitivemicrophoneusingelectroplatingtechnology,IEEEJournalofMicroelectromechanicalSystems,vol.3,no.2,pp.69,June1994. [92] BakerM.andSarpeshkarR.,Alow-powerhigh-psrrcurrent-modemicrophonepreamplier,vol.38,no.10,pp.1671,October2003. [93] OtmarKern,Amplifyingcircuit,USPatent6812788,2004. [94] LoeppertP.andLeeS.,Sisonic-therstcommercializedmemsmicrophone,inSolidStateSensorsandActuatorsMicrosystemWorkshop,HiltonHeadSouthCarolina,2006,pp.27,IEEE. [95] HsuP.,KlastrangeloC.,andWiseK,Ahighsensitivitypolysilicondiaphragmcondensermicrophone,inMEMS1998,AnchorageAK,USA,1998,pp.580. [96] G.Amendola,Y.Blanchard,A.Exertier,S.Spirkovitch,G.Lu,andG.Alquie,ACMOS,self-biasedchargeamplier,inThirdInternationalWorkshoponDesignofMixed-ModeIntegratedCircuitsandApplications,PuertoVallarta,Mexico,1999,IEEE,vol.3680,pp.5. [97] D.T.Martin,K.Kadirvel,R.M.Fox,T.Nishider,J.Liu,andM.Sheplak,Surfaceandbulkmicromachineddualback-platecondensermicrophone,inInternationalConferenceonMicroElectroMechanicalSystems,Miami,FL,2005,IEEE. [98] Thermoptics,Datasheetdn620,chargeamplier,Tech.Rep.,CarsonCity,NV,1995. [99] TexasInstruments,Datasheetopa128,Tech.Rep.,OneTIBlvd.,Dallas,TX,1993. [100] TexasInstruments,Datasheetopa111,Tech.Rep.,OneTIBlvd.,Dallas,TX,1993. [101] AnalogDevices,Datasheetad745,Tech.Rep.,OneTechnologyWay,Norwood,MA,1993. [102] AnalogDevices,Datasheetad734,10mhz,4-quadrantmultiplier/divider,Tech.Rep.,OneTechnologyWay,Norwood,MA,1993. [103] BaoqingLi,DerenLu,andWeiyuanWang,Open-loopoperatingmodeofmicromachinedcapacitiveaccelerometer,SensorsandActuatorA,vol.79,pp.219,2000. [104] MichaelKraft,C.P.Lewis,andT.G.Hesketh,Closedloopsiliconaccelerometers,IEEProceedingsofCircuitsDevicesandSystems,vol.145,no.5,pp.325,October1998. 170

PAGE 171

Yakabeetal.,Electrostaticcapacitancedetectioncircuitandmicrophonedevice,USPatent7034551,2006. [106] ChristianC.EnzandGaborC.Temes,Circuittechniquesforreducingtheeffectsofop-ampimperfections:Autozeroing,correlateddoublesampling,andchopperstabilization,ProceedingsoftheIEEE,vol.84,no.11,pp.1584,November1996. [107] Cluade-AlainGobetandAlexanderKnob,Noiseanalysisofswitchedcapacitornetworks,IEEETransactionsofCircuitsandSystems,vol.30,no.1,pp.37,1983. [108] NaiyavudhiWongkomet,PositionSensingforElectrostaticMicropositioners,Ph.D.dissertation,UniversityofCalifornia,Berkeley,199. [109] PotterRandBoorS,Switchedmicrophonebuffer,USPatent7092538,1977. [110] S.A.Jawed,M.Gottardi,andA.Baschirotto,Aswitchedcapacitorinterfaceforacapacitivemicrophone,ResearchinMicroelectronicsandElectronics2006,pp.384,2006. [111] FrederickV.Hunt,ElectroacousticsTheAnalysisofTransduction,andItsHistoricalBackground,AmericanInstituteofPhysics,2edition,1982. [112] K.Y.Park,C.W.Lee,H.S.Jang,Y.S.Oh,andB.J.Ha,Capacitivesensingtypesurfacemicromachinedsiliconaccelerometerwithastiffnesstuningcapability,IEEE,pp.637,1998. [113] NavidYazdi,FarrokhAyazi,andKhalilNaja,Micromachinedinertialsensors,ProceedingsoftheIEEE,vol.86,no.8,pp.1640,August1998. [114] HallN.A.,BicenB.,JeelaniM.K.,LeeW,QuereshiS.,DegertekinF.L,andOkandanMurat,Micromachinedmicrophoneswithdiffraction-basedopticaldisplacementdetection,JournaloftheAcousticalSocietyofAmerica,vol.118,no.5,pp.3000,2005. [115] vanderDonkA.G.H.,ScheeperP.R.,,OlthuisW.,andBergveldP.,Amplitudemodulatedelectro-mechanicalfeedbacksystemforsiliconcondensermicrophones,1992. [116] S.Hadjiloucas,L.S.Karatzas,D.A.Keating,andM.J.Usher,Opticalsensorsformonitoringwateruptakeinplants,JournalofLightwaveTechnology,vol.13,no.7,pp.1421,1995. [117] HaoLuo,GangZhang,RichardCarley,andGarryFedder,Apost-cmosmicromachiendlateralaccelerometer,vol.2,no.3,pp.188,2002. [118] StephenA.JoyceandJ.E.HoustonadnBradleyK.Smith,Interfacialforcesensorwithforce-feedbackcontrol,inIEEEElectronDevicesMeeting.1990,pp.26.3.1.3.4,IEEE. [119] MarkA.Lemkin,MicroAccelerometerDesignwithDigitalFeedbackControl,Ph.D.dissertation,UniversityofCalifornia,Berkeley,1997. [120] GelbAandvanderVelde,MultipleInputDescribingFunctions,McgrawHill,London,1968. 171

PAGE 172

WuN.,MilesR.,andAydinO.,Adigitalfeedbackdampingschemeforamicromachineddirectionalmicrophone,in1994ProceedingoftheAmericanControlsConference,BostonMA,1994,pp.33153321,IEEE. [122] MarkLemkinandBernhardE.Boser,Athree-axismicromachinedaccelerometerwithacmosposition-senseinterfaceanddigitaloffset-trimelectronics,IEEEJournalofSolidStateCircuits,vol.34,no.4,pp.456,April1999. [123] TetsuyaKajita,Un-KuMoon,andGaborC.Temes,Anoiseshapingaccelerometerinterfacecircuitfortwochipimplementation,inIEEEInternationalSymposiumonCircuitsandSystems,ISCAS2000,GenevaSwitzerland.IEEE,2000,pp.337,IEEE. [124] EdelsonJ.andWardN.,Forcebalancemicrophone,USPatent6285769B1,2001. [125] C.C.EnzandG.C.Temes,Circuittechniquesforreducingtheeffectsofop-ampimperfections:autozeroing,correlateddoublesampling,andchopperstabilization,ProceedingsofIEEE,vol.84,no.11,pp.326,March1996. [126] K.Ogata,ModernControlEngineering,PrenticeHall,Norwood,MA,2001. [127] AnalogDevices,Datasheetop470,verylownoisequadoperationalamplier,Tech.Rep.,OneTechnologyWay,Norwood,MA,1993. [128] AlteraSemiconductors,Datasheetepm7128cpld,Tech.Rep.,101InnovationDrive,SanJose,CA,,2003. 172

PAGE 173

KarthikKadirvelwasinChennai,TamilNadu,India.HemovedtotheUnitedArabEmiratesin1989andgraduatedfromOurOwnEnglishHighSchoolinDubai,UnitedArabEmiratesin1996.HewentbacktoIndiaforhisBachelorsdegreeandobtainedhisBachelorofEngineeringdegreespecializinginElectronicsandInstrumentationfromAnnamalaiUniversity,TamilNadu,IndiainMay2000.HeobtainedhisMaster'sdegreeattheUniversityofFloridain2002.HisthesiswasonthedesignandcharacterizationofaMEMSopticalmicrophone.HeiscurrentlypursuinghisdoctoraldegreeatInterdisciplinaryMicrosystemGroupattheUniversityofFlorida. 173