|Table of Contents|
Summary and conclusions
Introduction, description of material and specimens, and testing
Presentation of data and analysis of data
Description of adhesive
7 o .." ,a< Ap J' u0a r
SIFATICUIE r 0 SAN WIC CCNSTIIUGIN$
IouI AIICIPAIT kT LANT A.
(Aluminum Facing and Aluminumm Ulencycemb Core Sandwich
Material Tested in Shear)
This Report is Une of a Series
Issued In Cooperation with the
AII IFCICCE-NAVY-CIVIL SUICCMMITT[E
AIVClAIFT SIGN CIPITIEIA
Under the Supervision of the----- ----
AIICIPAIT COMMITTlt .'.'_. ,_
of the F I
**~~~' No. 1559-IH DPOST'
UNITED STATES DEPARTMENT OF AGRICULTURE
FOREST PRODUCTS LABORATORY
Madison 5, Wisconsin
In Cooperation with the University of Wisconsin
FATIGUE OF SAXITDICH CONSTRUCTIONS FOR AIRCRAFT1
(Aluminum Facing and Aluminum Honeycomb Core Sandwich Material
Tested in Shear)2
By FRED WERREN, Engineer
Summary and. Conclusions
A limited number of tests have been made at the Forest Products
Laboratory to determine the shear fatigue properties of an assembled sandwich
panel with aluminum facings and perforated-aluminum-foil honeycomb core
materiz-.l. Repeated tests were made at a ratio of minimum to maximnumn loading
of O.1. The results of the tests and the S-N curves obtained from them are
The shear and shear fatigue properties in the LT plane (fig. l) are
different from those in LR plane, and the results of tests in each direction
are given. The shear strength in the LT plane is almost twice as great as
the shear strength in the LR plane, but the fatigue properties are proportion-
ately better in the LR plane. If equal repeated shear stresses were applied
to o. specimen in each plane, however, the specimen with deformation in the
LR plane would be expected to fail first. The results of the series of tests
in the LR plane indicate a fatigue strength at 30 million cycles of approxi-
mately 36 percent of the static strength for the condition of loading used.
-This progress report is one of a series prepared and distributed by the
Forest Products Laboratory under U. S. Navy, Bureau of Aeronautics 1T).
17BA-PC-17TAor 00619, Amendment No. 1, and U. S. Air Force No. USAF-PO-
(33o-03)48-41E. Results here reported are preliminary and may be revised
as additional data become available.
2Tis is the eighth of a eries of reports intended to offer a comrson of
-!-This is the eighth of a series of reports intended to offer a comparison of
the shear fatigue properties of different sandwich materials. The follow-
ing FPL reports discuss the shear fatigue properties of:
1559 "Cellular Cellulose Acetate Core Material"
1559-A "Aluminum Face and Paper Honeycomb Core Sandwich Material"
1559-? "Aluminum Face and End-grain Balsa Core Sandwich Material"
1559-0 "Aluminum or Fiberglas-laminate Face and Fiberglas Honeycomb Core
1559-D "Fiberglas-laminate Face and End-grain Balsa Core Sandwich Material"
1559-R "Aluminum or Fiberglas-laminate Face and Cellular-hard-rubber Core
1559-F "Cellular Cellulose Acetate Core Material with Aluminum or Fiberglas-
1559-G "Fiberglas-laminate Facing and Paper Honeycomb Core Sandwich Material"
Rept. No. 1559-H
In the IT plano, tihe fatl-ue strength at 3) million cycles is about 23 percent
of the static strength.
If plates of sandwich construction are designed so that their facings
are elastically stable, the most critical stress to which the core is subjected
is ohp::r. The consideration of the effect of repeated shear stresses on the
material of the cores and on the bands between the core and facings is, there-
The general testing procedures and nomenclature applied to these tests
are 3irilanr to those used by the Forest Products Laboratory in testing aluminum
facing and paper honeycomb core sandwich material.2
Description of Material and Srccimens
Three panels of the sandwich material were furnished to the Laboratory
by the manufacturerr2 The honeycomb core material consisted of O.C'..K-inch
perforated aluminum foil formed into 3/8-inch cells of hexagonal shape, and
weiF.ed alout 5-1/2 pounds per cubic foot. The core was cut to a thicktness
of 0.500 + 0.C05 inch and was bonded to the 0.C20-inch alurinun facings with
r.n adhesive especially formulated for bending metals. The core and facings
weec assembled and cured in a press at 10 pounds per square inch at 3CC F.
for 20 Ainutes.
lpec'l-.enq were cut from the panels with a metal-cutting band saw to a
width ?nd length of 2.00 and 5.67 inches, respectively. The specimens were
cut in two directions with respect to the core orientation to permit applica-
tion of shear deformation in (1) the LR plane and (2) the LT plane (fii. I).
The specimens were glued to the 1/2-inch shear plates with Adhesive Y,1 vith
a final cure at l0 pounds per square inch and 300 F. for 1 hour.
The results of 33 fatigue tests and 24 control tests are presented in
All tests were made in accordance with the methods described in
Forest Products Laroratory Report No. 1559-A.-2
The failure of fatigue specimens tested to produce shear deformation
in the LR plane was a combination of buckling of the cell walls and of
dial-na.l tension (fig. 2). The diagonal-tension cracks originated relatively
-A.' "itional information on the panels and on the adhesives used in these
'oLts is given in Appendix 1.
F-,--t. "-o. 1559-H
early in the test and always had their inception at one of the perforations
-in the cell wall. These early cracks almost always occurred in the cell wall
of single thickness. Slight buckling of some of the cell walls was also
evident almost from the beginning of the test. In spite of the early frac-
tures and buckling, the load dropped off only slightly until shortly before
the final failure. Usually the final failure was a progressive buckling and
diagonal-tension failure of the cell walls, and the specimen would not hold
the load. There was no sudden failure or drop off in load such as has been
experienced with several other core materials.
The failure of the specimens that were tested to produce shear defor-
mation in the LT plane was considerably different from that mentioned above.
The final failure was a combination of (1) buckling of the cell walls, (2)
diagonal-tension failure of the core originating at the perforations, and
(3) glue-line failure between core and facings. It appeared that the glue-
line failure occurred after the other two types, and the specimen failed more
rapidly once the bond had begun to fail. As with the LR specimens, however,
the failure was progressive until the specimen would no longer hold the load.
It was noted that the initial buckling appeared limited to the single cell
walls but that the diagonal-tension cracks originated at the perforations
in both the single and double walls.
The failure of the control specimens tested in the LR plane was due
to buckling of the cell walls. If the load was carried on beyond the maximum,
the result was a farther collapse of the cell walls. In specimens tested in
the LT plane, the failure was also due to buckling of the cell walls. The
double cell walls buckled noticeably at the maximum load, and the buckling
was followed by progressive buckling and a slow drop in load. In each case,
the load appeared to increase at a uniform rate until the maximum load was
Presentation of Data
The results of the individual fatigue and control tests are presented
in table 1. Values are calculated as in Forest Products Laboratory Report
No# 1559-A. It is evident that the shear strength in the LT plane is almost
double that in the LR plane.
The results of the fatigue tests are plotted in figure 3, and an S-N
curve is plotted through the points representing the two planes tested..
Analysis of Data
From an examination of the construction of the core material (fig. 1),
it can be seen that the cell walls of double thickness are in an LT plane.
Therefore it appears that the strength in the LT plane would be greater than
that in the LR plane, provided the glue bond between core and facings is
satisfactory. This was confirmed in static tests wherein the bond did not
Rept. No. 1559-H
fail. Since the fatigue characteristics in the two planes might be completely
different in much a construction, however, tests in both the weak and the
str'nf direction seemed advisable,
The 3-7' curve associated with the weaker direction (LR plane) is an
indication of the shear fatigue properties of the core in that direction, and
t-. I:--. rcet-.'cen core a-nd facing appeared to be satisfactory. Failure was
&u to Q1i.-onal-tension cracks originating at the perforations in the cell
I.."als r..; to buckling of the cell walls (fig. 2). AlthouWh no diagona'-
tensioi. fnilures were visible in the static tests, the effect of repeated
strc3sses resulted in the tension cracks at the perforations, whore there is
a zone of stress concentration. In specimen A6-2-13, one such diao-nal-
teonsisn crack was observed at about 3 million cycles, and additional onos at
a77ut 13 -.il` ion cycles, Nlevertheless, the specimen withstood more than-
3 -.!llion cycles without complete failure.
For specimens tested in the LT plane, the failure was different than
that atove, and the resultant S-1T curve reflects a combination of failure of
the core material and failure of the bond between core and facin.s.
A comparison of the data and the two S-1T curves indicates that the
LR .l'.ne, within the ran,--e tested, is still the critical plane as far as
fatigue is concerned. Even though the curve for the LT plane is lower, these
percentage values are based on a higher control strength. As e:ca:-ple, a
speci-en from uanel 2 subjected to stresses in the I. lane at a raxi..r=
rr',eat-ed shear stress of 100 pounds per square inch would be expected to
withstand about 200,OCO cycles before final failure. A similar Zeci-men from
the sane par.el but subjected to the same repeated shear stress in the LT plane
would be expected to withstand almost 2 millionn cycles. The above cm-oarison
is of course limited to the type of loading used in t-.ese tests.
It is important to repeat here that the perforations in te.o cell walls
are a point of stress concentration and that when the core rnateril is
subjected to repeated shear stresses, a few dia,:-:n:i!-t -nsion craci:s boco-e
evident long before the final failure takes place.
Prior to testing, it was agreed to discontinue testing, an;z f-tiu,.e
-,ecimen +-r-t withstood 30 million cycles without cor.plete failure, Four
suc:-. -oecimens were removelI from the r'ac-hi.r.e. It car. be seen fro- the
plotted points and curves of figure 3 that the endurance limit c:-r_-.Ot be
accurately determined from these tests. For specimens tested in the LP.
plane, it appears that the endurance limit ,ri-.ht te about 35 percent of
.. s2tic strenri.th for the condition of loaqinj used,; but for I-c'er.
tes-tol in the LT Ilane, there does not appear to be a definite Inicc-.tion
":.t the -nfurrn.e limit is being approacholi, even at 30 -.illion cycles.
Description of Sandwich Panels
The following description of the sandwich panels was submitted to the
Forest Products Laboratory by their manufacturer.
Molded panel thickness
Number of panels submitted
Preheat time: in press
Weight of panels
24 ST clad aluminum alloy
3/8-0.004 PR (3/8-inch cells,
perforated aluminum foil 0.004
0.500 + 0.002 inch
0.541 0.002 inch
10 pounds per square inch
3 minutes at zero pressure
20 minutes in steam-heated
5.55 pounds per cubic foot
FM-45, co-position unknovm
0.92 pound per square foot
Description of Adhesive
Adhesive Y, a modified polyvinyl butyral adhesive.
Rept. ITo. 1559-H -5-
-ie .-- .r fi. -trcni:th Df sand ic. c.nst-'ucticns of
"a."! _1 a.... _d ih o ycorn cores.=
S Control tests
S ..- ... : Control:Patlo of : Cy.cles :Sneci".en: S]e-.r
-*. 2repeated:2 stc'r~Ct}:T'.ai.Lt' : to : :Tu. : strene-h
s' c-: : :resneated : failure : I
* stress :
: s.ear a
S:sre;T to: :
:* :strcn-th :
(.) (2) (3) (4) (5) (6) (7)
-. * T.S.,- _
-- --- .i- - L- -. -. : r- - ---- -- -- --1 -
'.,-'.' 1 - 1 :'." ::' i u - i .." .. ... . ..... ..
',' .8 l .l 2 ] ?0 97'+ A6 1 -.
55 1V .1: <. 26, 4 :
: 63.2 : 154.1 41.0 : ,. : 6
5?. 1 : I' .! : ?K.0 :13, ,, : I :
5- 6 l '..l 3," 0 1--,:*5, '> 10
: 134.1 : 154.1 : 87.0 6,n 12
* U S
l. t -5
Av. . 1!'
S.... 1 : -, T--,,- i-. -" ... . - '
-.. .--- a'. -_ . .. --- _-- -. -
: 1,650, :-*
*" *" 1'
: 1 C ," 1
Sheet of 2
I /' ro "
Table l.--Shear fatiaue strength of sandwich constructions of
aluminum facings and aluminum honeycomb corslA (Cont'd)
: Control tests
Specinen:!Iaximum : Control Ratio of :
No. :repeated:strength:maxium :
: shear : :repeated :
Stress : : shear :
to : No:.
S: :strength :
(* ) (2) : (3) : (4.) (5) : (6) : (7)
SP.s.1. : P.s.i. : Percent : : : P*s.i.
2 1 b
: 0 .
PANEL 2 -- 18 by 28 INCHES -- TESTED IN LT DI-ECTION
AMTEL 3 --
18 by 28
43.0 : 356,700 : A6-
61.5 : 45,100 :
80.0 : 4,400 :
49.2 : 107,800 :
73.7 : 9,000 :
38.4 : 1,108,000 :
33.8 : 3,126,900 : Av.
67.0 : 27,000 :
36.1 : 1,339,300 :
40.7 : 455,000 :
55.3 : 78,100 :
INCHES TESTED IN LT DI:
28.8 : 9,381,500 :A6-
25.0 :21,801,900 :
27.0 : 5,190,900 :
30.8 : 4,065,900
23.0 :31,242,500+: Av.
loaded at a rate of 900 cycles per minute in
direct-stress fatigue machine. Ratio of minimum to maximum load
was 0.10. Control specimens tested at a head speed of 0.01 inch
Sheet 2 of 2
Rept. No. 1559-H
Figure 1.--Section of a block of aluinun h'-neycomb
core material made from perforated alurinm- foil.
Directional orientation referred to as L (longi-
tVdinal), R (radial), and T (tangential).
9 .r.. p
- 9 ,
"ii '. I'"
Figure 2.--Sandwich material with alutr.irum facirgs and
aluminum honeycomb core after failure in shear
fatigue test. Specimen was tested to produce .hear
deformation in the LR plane,
ZM 82098 F
Figure 53.--"S-N" curve for sandwich material with aluminum facings
and perforate-d-aluminum-foil honeycomb core, tested in shear.
Ratio of minimum to maximum loading (range ratio) was 0.10.
ZN 93329 F
-IU i 1 1 i- 11i 1 1i r I--- I tIll
-- ^ K'. -- --
? < 6 tO 2
t I -,
vtt... ', H
_- l, \
6 /0' 2 4
- iI I -I I! I I ""
II I I"
ii-4 .1: I-i..
i i \' ,-- i ( i
.. . .I + I L t
i. I i
6 i '
l ,i ,i,
6 iO5 2 4 6 ,'06 2
I I iI iI,
' 0 > I '
4. 4 -4-- -. ---.-.---- -- I
4- i0 2- 4 6 i;
4 6 /Q? 2 4 6 ,/08
CYCLES TO FAILURE (NUMBER)
z 8329 P
-I.. R DIRECTION
- L. R DIRECTIONV
I I I I &
! : ! ,
i 1iE2 SIT Y F FLORIDA
3 1262 08927 4277
-, j^ '
L .*.'. :'
,: ... .../ : .; -^
* .*. -.. .', '
* .. . l
S. ... *...
'- *: ; :
1. 44 ^
... .- *^ "*
* 4 "
,' 4. .
4. 44 '