EFFECT CF DEFECTS ON STRENGTH CF
AIRCRAFT-TYPE SANDWICU PANELS
Information Reviewed and Reaffirmed
No. 1809
March Me
FOREST PRODUCTS LABORATORY
MADISON 5 , WISCONSIN UNITED STATES DEPARTMENT OF AGRICULTURE
•
FOREST SERVICE
In Cooperation with the University of Wisconsin
EFFECT OF DEFECTS ON STRENGTH OF AIRCRAFT-TYPE
SANDWICH PANEL
By
B. G. HEEBINK, Engineer
and
A. A. MOHAUPT, Technologist
Summary
This report presents results of tests requested by the ANC-23 Panel
on Sandwich Construction to determine the effect of certain specific defects
on the strength of a sandwich panel. Flat, short-column specimens and
larger curved panels were tested in edgewise compression.
It was found that poor bonds that drastically reduced the tensile
strength between the facing arid core also reduced the compressive strength
of a sandwich panel. Insufficient data were obtained, however, to establish
any exact relationship between the tensile strength and the edgewise compressive strength. Unbonded circular areas between the facing and the core
in most cases reduced the edgewise compressive strength. Wrinkles 0.015 to
0.020 inch deep in glass- . cloth facings lowered the eOzewithe compressive
strength from 1/3 to 1/2 of the control strength; however, sh'&1Jow wrinkles
(less than 0.003 inch) caused by a fold in a cellophane parting film
appeared to have no weakening effect. The effect of a 1/16-inch core gap
under 0.012-inch glass-cloth facia:.: was questionable, but core gaps of
1/8 inch and larger reduced the compressive strength 40 percent'or more.
Three fabric lap arrangements in glass-cloth-faced sandwich panels were
tested, but no adverse effect on strength was apparent. Areas of delamina
tion within the facing produced results similar to those at unbonded areas
between the facing and the core. In general, as the resin content of glasscloth facings was decreased, the edgewise compressive strength decreased.
A fold in the glass-cloth facing of a sandwich panel reduced compressive
strength very little. Neither gradual nor abrupt increase in the number of
plies of glass cloth in the facing had any substantial effect on the edgewise compressive strength of the test panels. An unglued joint in the core
was difficult to obtain in a sandwich panel with glass-cloth facings.
Although the core sections were not glued together initially, they became
bonded during the fabrication of the panel by resin from the facings, provided they were reasonably close together. The compressive strength
developed by such a panel was equivalent to that of normal construction.
1This progress report is one of a series prepared and distributed by the
Forest Products Laboratory under U. S. Navy, Bureau of Aeronautics No.
NBA-PO-NAer 00619, Amendment No. 2 and U. S. Air Force No. USAF-PO(33-038)48-41E. Results here reported are preliminary and may be revised
as additional data become available.
Rept. No. 1809
-1-
Agriculture-Madison
Introduction
The purpose of this study was to make a preliminary investigation
of the effect of defects on the strength of typical sandwich panels of the
aircraft type. The ANC-23 Subcommittee on Sandwich Construction in 1947
initiated a project at the Forest Products Laboratory to determine the
effect of unbonded and poorly bonded areas. Later other defects in glasscloth facings were included in the study.
Types of Defects
A preliminary study of the effect of unbonded and poorly bonded
areas on the strength of typical aircraft sandwich panels was first made,
using the, following five panel constructions:
1. Aluminum facings 0.012 inch thick glued to an end-grain balsa
core 0.240 inch thick with adhesive A2 and adhesive B.
2. Aluminum facings 0.020 inch thick glued to a cotton-cloth ,
honeycomb core of 7/16-inch cell size, 0.640 inch thick with adhesive C.E
3. Glass-cloth facings, four plies of cloth 112-114, laminated
with resin DE to a cellular cellulose acetate core 0.430 inch thick.
4. Glass-cloth facings, four plies of cloth 112-114, laminated
resin
D to glass-cloth honeycomb core of 1/4-inch cell size, 0.400
with
inch thick.
5. Glass-cloth facings, four plies of cloth 112-114, laminated
with the liquid component of adhesive. C to an end-grain balsa core 0.480
inch thick.
In addition to the: .tanbonded and poorly bonded construction,
several other defects in sandwich panels of construction 3 and 4 were
later investigated. These defects were:
(a) Wrinkles in one facing.
(b) Gaps in the core.
-See Appendix I.
Rept. No. 1809
-2-
(c)Fabric laps in one facing.
(d) Unbonded areas repaired by hypodermic injection of resin
and cured.
(e) Local areas of delamination within one facing.
(f) Low resin content in one facing.
(g) Fold in one facing.
(h) Variation of the number of glass-cloth plies in both facings.
(i) Unglued butt joint in the core.
Test Methods
Two tests were used to evaluate the effect of defects in sandwich construction, the edgewise compression test and the flatwise tension test. The
defects were included in flat sandwich panels 12 inches square, and were
positioned so that 12 to 15 test specimens with or without defects were
obtained from each panel. Edgewise compression tests were made on flat and
curved panels. The flat edgewise compression specimen was 2 by 3 inches,
with the load applied parallel to the 3-inch dimension. Details regarding
the specimen are given in Forest Products Laboratory Report No. 1556,
"Methods for Conducting Mechanical Tests of Sandwich Construction at Normal
Temperatures," revised October 1948. Test methods used were essentially
those described in that report; however, instead of using a clamp-type
arrangement at the ends of the specimen, as shown in Report No. 1556, a
slotted round steel bar was used.
The curved edgewise-compression specimens were approximately 12 by
18 inches, all formed to a 50-inch radius except the aluminum-on-balsa panels,
which were formed to a 24-inch radius. The panels were loaded in the axial
direction (18-inch dimension), and loading rails of 0.020-inch aluminum were
bonded to the ends of the aluminum-faced panels, and 1/16-inch birch plywood
to the panels faced with glass cloth to prevent end failures.
Tension tests were taken from flat panels made by identical fabrication techniques as used for making the panels for edgewise compression tests.
Tension specimens 1 inch square were made and tested, in accordance with
Forest Products Laboratory Report No. 1556, to determine the quality of the
bond between the facing and the core for comparison between poorly bonded
end well-bonded constructions.
Fabrication, of Sandwich
Panels
Controls
One 12-inch square flat panel of each of the five constructions listed
above was fabricated for tension tests. Fabrication procedures for constructions 1 through 4 were typical, and similar to the methods presented in
Forest Products Laboratory Report No. 1574, 'Fabrication of Light Weight
Sandwich Panels of the Aircraft Type," dated June 1947. In addition
Rept. No.
1809
-3-
construction 5, a combination of glass-cloth facings laminated with adhesive
C liquid to a balsa core, was included. The facingp.of this combination
were prepared by impregnating the individual sheets of glass cloth with the
liquid and building up the facing on a clean aluminum or glass surface to
the desired thickness. The facings were dried at room temperature for 2
hours or more to allow the volatile solvents in the adhesive to escape. The
resin content of the dried facings was from 40 to 45 percent. The balsa core
was sprayed with adhesive C liquid, thinned with 50 percent by volume of
'ethyl alcohol, at the rate of 5 to 7 grams per square foot and immediately
dusted with adhesive C powder, the excess being removed by inverting the
core and tapping it gently. After drying overnight, the core and facings
were assembled and placed between cellophane-covered aluminum cauls. 'A
fluffed wool blanket was used between each caul and press platen as a
cushion to take up slight irregularities. *
The panel was pressed at 260° F. and 70 pounds per square inch for
1 hour and cooled under pressure.
Details of fabrication t'or the I2-inch square panels of each combination to be used for flat edgewise compression tests were identical to those
used for the tension panels. The 12- by 18-inch curved panels were bag
molded under curing conditions corresponding to those required for the
adhesive and the core material.
Poorly Bonded Construction
Poorly bonded constructions were produced in the various panels by
means of starved glue joints or by inserting a foreign material between the
facing and core. A poor bond was produced on only one side of the panel,
the other side having a normal bond. Two flat panels of each combination
were made, one for edgewise_ , compression tests and the other for tension tests
to determine the reduction in , strength as compared with the values obtained
from a control panel. Curved panels, 12 by 18 inches, were fabricated with
the poorly bonded construction on the concave side of the panel. The methods
used to attain a poor bond in the various panel combinations are as follows:
1. Aluminum facings on balsa core.--For poor bonds in this construction, the aluminum facings were etched as usual in a sulphuric acid-sodium
dichromate bath, sprayed with the primary adhesive A, and cured in an oven
at 325° F. for 30 minutes. The very light spread of the secondary
adhesive B, 1 gram per square foot, was next applied. After drying for 1 to
2 hours, the panel was assembled and pressed between blanket-backed cauls in
a press at 230° F. and 75 pounds per square inch for 1 hour, and cooled to
room temperature while under pressure.
2. Aluminum facings on cotton-cloth honeycomb core.--Poor bonds in
these panels were made by reducing the spread of adhesive C. For one facing
the normal spread, 35 grams per square foot of wet adhesive C liquid, was
-used and dusted with adhesive C powder, but the poorly bonded facing was
coated with a wet spread of 15 grams per square foot of the liquid with no
Rept. No. 1809
-4-
powder. The application of adhesive C liquid to the core was from 5 to 10
grams per square foot for both facings. The panels were pressed between
blanket-backed cauls and cooled in the press.
3. •lass-cloth facings on cellular cellulose acetate core.--All construction details for fabricating poorly bonded panels of this combination
were normal except that a sheet of sanded polyvinyl alcohol film was placed
between the glass-cloth facing and the core on one side to produce the poor
bond. Since no resin was applied to the cellulose acetate core, the side of
the polyvinyl alcohol film applied to the core was spread with resin. Sanding or roughing the surface of the film was necessary to produce the desired
strength level. The panels were pressed between cellophane-covered blanket
cauls and cooled underAaressure.
4. Glass-cloth facinjs on glass-cloth honeycomb core.--Fabrication
details were similar to those used with the cellular cellulose acetate cores.
Since the normal procedure required an application of 10 grams per square
foot of resin D to each side of the honeycomb core, no additional resin was
applied to the polyvinyl alcohol film.
5. Glass-cloth facings laminated with adhesive C to balsa core.--The
fabrication techniques used for these poorly bonded panels were the sane as
those used to make the controls, except that a parting film of cellophane
was used between one of the glass-cloth facings and the balsa core. The
surface of the cellophane was roughened with 3/0 or 120 garnet finishing paper,
applying 10 strokes of the finishing paper in two directions, perpendicular
to each other.
Constructions with Unbonded Areas
The unbonded areas in the flat panels for edgewise compression tests
were 1/2, 3/4, or 1 inch in diameter and were incorporated only in one side
of the panel between the core and the facing. In the panels with aluminum
facings, the unbonded areas were produced by applying a tape patch of the
desired size and shape to the aluminum facings before applying the adhesive.
After the adhesive was ;3pread on the facings the patches were removed,
leaving the area void of adhesive. Unbonded areas in panels with glasscloth facings were produced by inserting two cellophane discs of the desired
size and shape between the glass-cloth facing and the core. A double layer
of cellophane was used because a bond of low strength is often obtained
between the cellophane and the facing or core if coated with adhesive. Other
fabrication details were the same as those used to make the other sandwiches
except that in all cases the panels were cooled to room temperature while
under pressure in the press. The cutting diagram for these panels is shown
in figure 1.
The unbonded areas in the 12- by 18-inch curved panels were 2 inches
in diameter and were located in the center of the concave side of the panel
between the facing and the core.
Rept. No. 1809
-5-
Wrinkles in One Facing
In the sandwich constructions consisting of glass-cloth facings on
cellular cellulose acetate or glass- cloth honeycomb cores, wrinkles were
simulated in one facing of 12-inch square panels by placing a 0.010-inch
diameter wire and a 0.024-inch diameter string under the cellophane of one
cellophane-covered aluminum caul. Both sizes of wrinkles were included in
'the same panel but were spaced so that only one appeared in each test
'specimen. When the panels were pressed, the imprint of the wire or string
was plainly left as a straight line across the center of the glass cloth
facing. The technique of fabrication was otherwise the same as for the
other sandwiches. Pressing conditions were 225° F. and 15 pounds per square
inch of pressure for 1 hour.
Gaps in the Core
Gaps in the core were made in three nominal sizes, 1/16, 1/8, and
1/4 inch in width. The cellulose acetate core, being quite rigid, was
easily spaced by inserting small wood separators at the edges of the panel.
The glass-cloth honeycomb was more difficult to handle because of its
flexibility, and therefore the gaps were measured and graded for size for
each specimen after fabrication.
Fabric Laps in One Facing
One 1/2-inch lap, two 1/2-inch laps on top of each other, and two
1/2-inch laps spaced one inch apart were included in one glass-cloth facing
of a 12-inch square panel of the two constructions.
Unbonded Areas Repaired by Hypodermic Injection
of Resin and Cured
No procedure was devised which was considered satisfactory for
simulating this defect, and therefore no test specimens were made.
Local Areas of Delamination Within One Facing
Local areas of delamination were made in the' same way as the unbonded
construction, except that the two cellophane discs were placed between the
second and third ply of one of the four-ply facings. These areas were 1/2,
3/4, or 1 inch in diameter, and positioned in the panel so that the areas
of delamination would fall in the center of the test specimens. These
panels were cured and cooled under pressure in the press.
Rept. No. 1809
-6-
Low Resin Content in One Facing
The resin content of one entire facing of each panel was varied to
determine the effect of resin content on the edgewise compressive strength
of the sandwich panels. A series of test specimens were made with 45 percent of resin in both facings as controls, and two other series were
fabricated so that one facing of each specimen of one series had 35 and the
other 25 percent of resin. One facing of each test specimen contained 45
percent of resin. The panels were pressed under usual conditions and
cooled under pressure.
Fold in One Facing
tested
folded
on top
normal
The effect of a fold in an entire four-ply glass-cloth facing was
in only one of the two facings. The facing was laid up and then
to produce a fold about 1/8 to 1/4 inch wide. The core was placed
of this facing and the other four-ply facing laid on the core in the
manner. The panels were cured and cooled under pressure in a press.
Variation of the Number of Glass-cloth
Plies in Both Facings
The number of glass-cloth plies in the facings was varied abruptly
or gradually. Starting with a four-ply facing, the thickness was gradually
increased to eight plies over a distance of 1-1/2 inches (fig. 2A). In the
same panel the number of plies was also increased abruptly from four to
eight, as shown in figure 2B. One panel with each core material (acetate
and glass-cloth honeycomb) was fabricated so that both methods of varying
facing thickness were used in one facing, and another panel with each core
material was fabricated with the two variations in both facings.
Unglued Butt Joint in Core
The panels for these tests were fabricated in accordance with
established techniques for panels faced with glass cloth, except that
adjacent pieces of acetate and glass-cloth honeycomb core were laid
together as tightly as possible, but were not edge glued to make a single
piece.
Results and Discussions
The following discussion deals with the effect of the various defects
described heretofore on the results of edgewise compression tests only. The
effect of these defects on the results of other tests, such as longitudinal
tension or bending, will require additional work.
Rept. No. 1809
-7-
Controls
Results of tests on controls tested in accordance with standard
procedures are recorded in tables 1, 2, and 3.
Poorly Bonded Construction
The reduction in strength of the poorly bonded sandwich panels, as
indicated by the flat edgewise compression tests (table 1) was not consistent in all the sandwich combinations tested. Results of tension tests made
on controls and poorly bonded panels to determine the relative quality of
the poor bonds are also given in table 1. The combinations of aluminum
facings on balsa cores, glass-cloth facings on cellulose acetate cores, and
on glass-cloth honeycomb cores showed no substantial
glass-cloth:,,
reduction in compressive strength. Although the poorly bonded compression
specimens with aluminum facings on balsa cores gave values approximately
9 percent less than the controls, this difference was considered to be
within the normal variation produced by inconsistencies in materials,
fabrication techniques, and testing procedures. The panels made of
aluminum facings on cotton-cloth honeycomb cores and of glass-cloth facings
on balsa cores showed a marked decrease in compressive strength. The types
of failures that occurred in these specimens were glue failures between the
facing and the core or shear failure in the core, but in the other panel
combinations failures were produced primarily by end buckling of the facings.
If the tensile strength of the bond between the facings and the core is
reduced sufficiently to produce glue , failures, it is probable that the
edgewise compressive strength of the. sandwich panel will be reduced as compared to that of the controls. •The strength values are also probably
affected by the degree of smoothness of the surface. A considerable number
of additional tests would be required, therefore, to establish a correlation between the flatwise tensile and compressive strength of poorly bonded
panels.
The.effect of poorly bonded construction on the compressive strength
'of the curved sandwich panels, as shown in table 2, followed the same trend
as with the flat sandwich'panels, but the percentage reduction in strength
was not so great.
Unbonded
Construction
In general, the edgewise compressive strength of all small, flat
sandwich combinations tested decreased as the size of the unbonded area
increased (table 1).' In a preliminary study of this type, where only a
few specimens were tested for each condition, irregularities in results
were often apparent. Unbonded areas of 3/4-inch diameter and larger in
most instances reduced the edgewise compressive strength in excess of 10
percent of the control strength. In the aluminum-on-balsa sandwich no
strength reduction was recorded for unbonded areas of 1/2- and 3/4-inch
diameter. This lack of reduction in strength could have been caused by
Rept. No. 1809
-8-
accidental formation of a bond between the facing and the core, by creeping
of the adhesive into the starved area during the curing process, or by a
better than average flatness on the facings of the panels.
In most of the combinations, only one or two of the three failures
occurred in the unbonded areas of 1/2-inch diameter. The best example of
this unexpected lack of effect was the sandwich of glass cloth on cellulose
acetate, in which failures occurred near the ends on all three specimens,
and the average compressive strength was 12.4 percent higher than that of
the controls. Close examination showed that these areas were poorly bonded
by resin that crept between the cellophane discs. Reductions were obtained,
however, when specimens with unbonded areas of 3A- and 1-inch diameter
were tested, but it is very likely that these unbonded areas were also
reduced in size during the curing cycle.
The unbonded area of 2-inch diameter between the facing and core of
a 12- by 18-inch curved panel also caused a substantial reduction in the
edgewise compressive strength of all the sandwich combinations. The
results are summarized in table 2. These reductions were greater than
those obtained from corresponding curved sandwich panels with poorly bonded
construction.
Wrinkles in One Facing
The wrinkles produced in the glass-cloth facings of the sandwich
panels by a 0.010-inch diameter wire and a 0.024-inch diameter string reduced the edgewise compressive strength by as much as 62 percent of the
control strength. The reduction in strength was greater in the panels
with glass-cloth facings on cellulose acetate cores than in the panels
with glass-cloth facings on glass-cloth honeycomb cores (table 3).
Although a substantial reduction was indicated by these tests,
wrinkles produced by these methods were deeper than those usually formed
by a fold in a cellophane parting film in normal production. The depth of
several wrinkles caused by a fold in the cellophane was measured and compared to those fabricated in the defective panels for study. The wrinkles
formed by the string were about 0.020 inch deep and those formed by the
wire were 0.015 inch deep, while the wrinkles caused by the cellophane
would not exceed 0.003 inch in depth. Five flat edge-compression specimens
having depressions caused by a cellophane wrinkle in one facing on a
cellulose acetate core were made and tested. The results, shown in table 3,
revealed strength greater than that of the controls; none failed at the
depression caused by the wrinkle.
Gaps in the Core
Gaps in the core appeared to be detrimental to the compressive
strength of a sandwich panel, with the strength generally decreasing as the
core gap increased. Core gaps of 1/8 inch and greater reduced the edgewise
compressive strength more than 40 percent (table 3). The effect of a
Rept. No. 1809
-9-
1/16-inch core gap was questionable, since 4 percent reduction was obtained
in the sandwich of glass-cloth facings on cellulose acetate core, but 16
percent reduction was recorded for the sandwich with the glass-cloth facings
on a glass-cloth honeycomb core.
Fabric Lap in One Facing
None of the three fabric-lap arrangements in one facing reduced the
edgewise compressive strength of the sandwich panels, and there appeared to
be a slight increase in the strength due to the additional facing material
in the laps.
Local Areas of Delamination Within One Facing
The results on panels having local areas of delamination in one
facing were somewhat similar to those obtained for the unbonded areas of
similar size. Unbonded areas within a facing in general reduced the edge,!,
wise compressive strength. On the sandwich with glass-cloth facings on,a
glass-cloth honeycomb core, the unbonded area of 1/2-inch diameter had no
effect, but on the sandwich with glass-cloth facings on a cellulose
acetate core, en average reduction of about 24 percent was obtained. This
loss of compressive strength was mainly due to one specimen that failed at
a low load, and was probably caused by a poor bond in the facing that .did
not contain the defect. Four of the five specimens from this group failed
in the facing that did not contain the defect indicating little or no
effect of the 1/2-inch unbonded defect; however, in most of the specimens
the unbonded area could not be determined by tapping. During the curing,
resin from the glass cloth may have run between the two disks of cellophane
to produce a partial boncL It is also probable that the larger unbonded areas
were reduced in size during the curing process.
Low Resin Content in One Facing
The strength of a sandwich panel with glass-cloth facings is affected
by the resin content of the facings. The resin content of the facings was
determined before the panels were fabricated, and was based on the total
weight of cloth and resin. This value, normally about 45 percent, is higher
than the resin content of the facing after the panel is cured because some
resin is squeezed out of the facing into the core. The effect on the compressive strength of reducing the initial resin content in the facings to
35 percent is questionable, since one combination showed no reduction and
the other an 18 percent reduction. The greater reduction was recorded for
the sandwich with glass cloth on a glass-cloth honeycomb core. A reduction
in initial resin content to 25 percent caused a marked reduction in compressive strength in both combinations tested.
Rept. No. 1809
-10-
Fold in'One-Facing
The effect of a fold in the facing was not pronounced, amounting to
about a 15 percent reduction in compressive strength. Since failures on
all test specimens occurred outside of the defect, it appears that the
slight reductions were not caused directly by the fold in the facing. The
results are based on panels fabricated in a press, and the same results
may not be obtained from bag molding. During the pressing operation .any
high point, such as a fold, takes the entire load of the press until other
areas bear. This tends to squeeze the high spots down and produce a
flatter panel than would be obtained by bag molding.
Variation of Number of Glass-cloth Plies in Facings
No appreciable weakening effect was found when the number of glasscloth plies in both facings were increased from four to eight, either
gradually or abruptly. Test specimens made with the change in one facing
only, however, lost as much as 13 percent of the control strength (table 3).
The compressive results presented in table 3 are based on a four-ply
cross-sectional area.
Unglued Butt Joint in Core
No loss of . compressive strength resulted from an attempt to produce
unglued butt joints in the core (table 3). Although the individual core
sections were assembled without glue, some of the resin from the facings
'was' forced-between the core sections during the curing cycle, producing q
partially glued joint. This was true only if the sections of core material
remained in contact with adjacent sections.
Rept. No. 1809
-11-
APPENDIX I
Description of Resins and Adhesives
Adhesive A. A high-temperature-setting mixture of thermosetting
resin and synthetic rubber.
Adhesive B. A high-temperature-setting, acid-catalyzed phenolic
resin.
Adhesive C. A high-temperature-setting, two-component resin with
a thermosetting liquid and a thermoplastic powder.
Resin D. A high-temperature-setting, high-viscosity, contactpressure laminating resin of the polyester type.
Rept. No. 1809
-12-
7,
L.
9-nil
n 0
.;;11
Z
r
4l
M
P... d
0
M
HI
.0
-,3
ri
4.
0 Fi
a p o
Ea
14 5-4 4,Pr0.1
O
q
w
H
r8
O
.0
4o
0
O
C
0
.
.4-1
i
O 2'
A
rri
ii
I.
,;
7
, ..
Y
-p• 4
3 0
01
..0 \
VD
n
,
as
t---
`.0
HI
VD
FC \
If \
?
a.
'
a-,
?
a-.
a,?
44-N
a4 1
HI IFF
-A
T ?
a-.
co
u-N
ch
•
.-4
'
O
? 'V
(.,
i.
i
0
A '0 0
'rgl g
-ro
. .. t'3' ''..
8
0 A H .
..)O
o
2
i:',
S.
ii.',' ., ,?,
2
5.,
E'‘
A
CI;
CV
to
.. .16 ..
2
g
4 .-q p
''
l.
7.,,
"g
;:, ' V - -
3
.'-'
▪
2 T4
'
a) 13
0
OF
C-)
.,
gA
g
E-E
at
P•
..
0
+3
o.
.0
2
to
0
A
ISE
0
i
.--%
ti 1
aa
I-1
.-9
.--9
N't
CV
V)
a,
/,‘
2
..-n
N
44- \
4-1
''
HI
°
ON
..
..
..
..
g
un
°
ai
CV
H
4.
..
0. .0 .. 0.. .. .0 .. .• 0
In
...
..
tn
CV
1
rd
0
u,
CV
N
1 .7
A
0
( \ 1
2
N
Fr,
a,
n:1,
.1 .-1
0 N
0
24
80
,-;
,§
0
°
N.
i
-I
0,
...
1
H0
0
.-1
'N
2
.2
,o
A
,0
• • •• •• •• •• • • • • •• •• •• •• •• •• •• n • •• •• •• • • .•
-0
1-4
.1
-
0
2xxx
l0
H
▪m
..
-I
N
... \
HI
,-.7
2
CV
H
-4
OD
0
N
CV
H.\
.
HI
CIS
HI
H
H
m
w
...1.
d0
o• HI
N
OD
al
E.--
CO
41'
?-1' It
..
g ..
7..
.• • • -• •• •• •• •• •• •• •• •• •• •• •• •• •• • •
._,
01
in
if \
CV
A.'
••-I
41.
in
CV N
CV
:
4+
MI
0
4)
'4
ca
i
al
• • n • .• • • •• •• • • •• • • •• • • • • •• •• • • ••
••
2
73
2 ...,
.4,
3)
P
0
+0'da
f,3
0 g
20
4 7
m o
o •
I
0
H Ft
a
a o
.1-1
•--L
g
fa.
9"
0 N
H HI
0 9
.. . . .. '''
'; ..2.
.d
CO
a)
OF
..0
2
0
7.,
W.
R.
11
..--..
•• •• •• •• •• •• • • • • •• •• •• • • • • •• • . • • •• • • • • • • P.
1' i'''';Fi
a
.0
F:61
IX
L.'
E
co
0 \
0
.--I
If \
N
d A
C•1 0
d1 N
a,
..r.
s.
. .
2-- 14-\
Al 41 171
1.,
0.
tQ
.• •• •. •• •• .• s• •• ••
9 .
ON x,\
cm
-,
•. •• •• •• •• •• .. .. •• •• •• •• •• •• •• •• •• ••
.--,
.--1
at
o
A G.
9
?
co
V3
,
n1
&,,,
u
1 Y
•• s• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• ,
I
CF1
Cr \
H
O
2 ,,, Z' g
1
q .,
!': ..
O o. 4-f%
• 2 •0 °
Fd
VD
I, \
N-
2.7, q-,
2 N
0 1-c3.
0 g
CI
H
,....0
4'
.0
.0
OD
1'V
.--I 1-1
I ?
0.1
0,
0 7.; .251
H
Y
01 t'
f \
2 'S
-4,
1,"
...
t') 'A :1")
2.. •
0
49
09
.0
N
II •• • • •• •• • • •• • • "•• •• •• • • 0. •• •• •• •• •. • • •• •• •. .. I.
'
t Z 0 0 0
O
al
py
u
Rd
4
• • • •• H • • • • •
a 0 I.
,
3
4:2
P.
r--I
N-
.-1
HN
H
0ii
tl--1
k
a)
a
Ics
'44
A
G' lw ,., 0
Icl
1.1
g
5°
0°
,
1
re\
..-i
01
VD
OD
V
N HI
°:
e
. s
CO
r-I
4.n
CV
„ - ,
0
a,
..4
Ll \ .
t'?„' 13 3 4
'1
1,1
Kl
2
V
It \
•
0
0 9-4
VI
reN
-7
3 ro °
g 2 k' .3
2 l':,
.. .. .. ..
i.: '
2 0 k.
1..
91°4
..4
1J .. .. ..-' '- .. .
S
'44'j.
au
1
2 2
I 4-'5 -I
4-.
n
fi;
O R 1.1
.-r p.
0 CO
.9
CO
I, \
-9
3
a•
a
7
441
CO
c44 q calm
1... 1.6 w4 M.
NI VD ''
.1,
0
?
Ms
f,
.41
U
0'.
-1-
HI •
m 0
=,...-
CSI
MI CV
-
+ g
HI "9
I
CV
0 HI
0) 0
9 4
0 0
xr-I 0,1
-4
9 C3
°
'''
.. .. ..
0 I
N
0 H x
.2 0
•
',1 c; ..
A
A
0 HI
0 2
0 0 I,
0 c'
t' 'j
-I. 0
Co .o4,o•
H 0 •
a, o •-.....
o
0
.00
4, 0
.0
•
-
.9 ifs ci
....,
4.7
.4.
Pcn
HH
0 I
N
0 H
.. .. 8 3 .. .. .. 8 3 .. ..
.0 xi
'0' "
.-I
....t
fj
Q C.J I
ia 0
.7'
201 20 g•.
4and
4:
s--I
ED
.0 ---.
0
al ca
w
4+
N .-..
0
4-4
>a
.4
15 0
0
o C.4
la, 4-,
Pt 0
..
0
R.
0 >..
Hi? NT'
;•$
a)
Ii
RI
8 TS
00
•ci bi)
0
•r-1
0
4-1
0 a)
;-n
N
4-) g
a) 0
E 0
•n•-1 0)
rci
0
0
•r--1
CV
RP
..-1
0
cd
a)
127
g
Co
o
0
1
1
rd
1
• • .. • •
4-,H
0
0 CO
c)-I -I-) 0
i•-n
0) 0
0
U DI
Pa
•• •• ••
0
i
I
1
••
$.4
a)
••
CV
t-
\ 0
\
tfl
8
0
8
0
▪\
N
8
0
trn
cOn0▪
oOn
teN
r-Em.
Fr 1.
ant
C\
re\
trn
cu
••
•,I-1
4-1 0 0
O
a)
tal
1 ;A
0 0
i Q)
Pi
1 rd a) 0
0
:71i
0 P-•
Z-P al
4-I
.
4-)
H
0 co
0
0
1-1 CO
04-1 -1-) 0
ti , 0 8
PA
0 u)
..
0 CO CD
0
•n-1 0
z q,
PI
i-;
a)
.0 q, 7.1 2
0 o
gi
a,
cc
••-1
(/)
•I-1 •P-1
U
_0 al 4.)
4-1
0
0
O
\ KIN
ai
0-1
0
H
0
H
tr\
t-
to
ON
CV
CO
O
• •
•
'H
0
teN
0
OP
0
O
c‘i
\
u)
H
cH bD
r•-%
re)
.off
0
0
$.•
0 ----
q$
\
CV ge)
H 1•1l,
C)
rcI
0
rnTh
N-
0)
0
8
rn
;-1
\O
0
t--
CT
N
ON
COn
as
Na) co
.0 to
u) a)
g
\
A
0
U
• • • • • •
C.)
4-1
4-1
•
IN
a)
H
,c)
E-1
••
••
•••••••
•••••n
.s=1
,rI
0
9 •'-'
°.
(I-4
••
0 CV
O
H 0
‹4 ........
•• •• ••
1
tr1
Ky
te%
0
4i
• n
0
1.• 0
••
0
••
••
••
••
9 "1,
0 0
M
• 8
H 0
.:::4 ......-
••
ern
••
••••••••
••••••
'8
ti
0 0
,-, ..-1
0 1
CV
co .-I
co 0
A0
1:1
••
••-•n
ti
0 0
H .4-i
0 I
CV
co H
co 0
0 •
H 0
0 ......
0 0
H •.-I
0 I
CV
co ,-I
co 0
El
5•1 0
0
H
H
cH
01)
-P
0) H
(0
el-I 0
O "
—0
.4 0.
0 ..-...
0 .......
...---..
...--..
..--.
..--..
o
0
4-1
I
O
cd -ICV
N
..d 0.
P1 •----
.0
,c) 0
0 .01
0 I
0 P-4 0
0 0
+' 0 ‘.0
-I.) 0 •
0 .0 0
0
*----
...---..
4
0
0
0
0(1) 4-1
0 -P I
•-I 0 0
0 4-, Kr\
H 0 -1H 0 •
a) cd 0
0
----
,o f.)
0 El
H 0 H
0 0 1
1).1 0
0 0 0
CO 0 ....1Cd 0
•
H P 0
Ch
---
.0
0
0
"-I
I
0
0 CO
co _..10
P:1 ---
ef•d4
cd q:1
4-1 ---4-I •
0 ,c)
$),
HI
ON
0
00
0
0
0
0
0
WH
-Proom
mpm
4) S.
0 a>
OO
O
ci2p
O uj U m
N%
u-
• .
3w1
NH
HH
4
•
H
H
c)0.0
•
•
NNH
00,CT
ri
Cd)
. •
Hs.0
CO30
•
M
vOiN
• .
\O C\
CNO
H
0\
•
H
H
H
C
• • • •
CV
'3
s.0
HO
M
4.0
00
HN
•
•
•
tyl \ID OD
H
Cs0. 0
• • e•
H
C0
v.> CV
N
• •
0
0
CV
•
000
O \03.
C3H
ft • ••
00
CV ON
‘13
0
0.
N
00
‘.0 N
0 CV
0
N
\O
0
V-
0
0
\D Cri H
(0
ArN
H
0
CV
C.-)11)-0,
•• ea
0\ O
HH
3
H
Oe 00 00 so ow
00 Om 00
ee 00
00
3-1rN 01003
MH Er\
x11w04-I w1)11 •
0 NO N
O
0
v") V") CO 3
1.1W\ WI
W04-01-1 WOr1 Is\
3H C110.0C
'U
0
A
• 49
\C),0)1\
•
C\3 O\
Wv\M
P-.
bp E
O0
•0 .0
H 4-,
O 0
ri
U
Cu m
H
b0 Cd
3),C,
•
MM
M\O
H H H
CV H H H H
00 00
00 00 00 0. 00
Oa
6"
e•
H
0 4.
0W0
S. [E
(1)),),-1
poHo
O.
0 P.,
.ce
4-i cC
I0
i
0
P
a)
wl
C)
0m
O 0,
0
O•
trr
4a)
•
00
0
0
CO
bp 4.,
0
O a)
o cd
Cd
4. a)
040
4000mm
A
▪ H
O 0
HH
0H
.40
P.Hop
a.)Xce...3
454-10
0)
s.00OW 00 C\ ONH
M
.
.
•
.
.
•
w\CO
1.0\0 H
N4-m o.).4\ NN 00 0
r-I H H
WC-VD
C.)
44
..•
• WNO
.
wICYNC,- HW trN 1.0 -1.
M MO M
0'
•• DO De
o0 oo 00
00
0. 00 • 00 ee
00 00 00 so 0.
o• 60
so
o0 00
00 00
W 000 00 03 00 0
H
MNH C.-3 Owl ..0N c,..
H
.00.0 Ne...- CV3- 33 M
•
• a
• •
ea
• • •
.; •
CD
IT
33 HH 0
N wIC\CCV
HH
.. .. ..
000 00 0 00 0 o 0
w\ MW 30 3 OH --1-
.1-%
Irl
Nov-% 3C.- H WO, wt• N
s.0
•
• • •
ft •
•
• •
••
M
L11\04,-) Ow\ C- C.-0 4,-
ON
NN N HHH NH H HN H H C4
.. .. .
..
O
mo
O
cc! S.
O
z
O
O O
trl tc\ m y\ kr,v-NH
0 0
zz
H
a)
0 O
trw-N 0 0
0 0 0
z zz
a)
0
H
OH
0
••-I
01 .0
H0
0
•
0
If\IC\ 11-\11•\
oD
v\
ta0
0
..-1
H
c....
m
O
O
O
,4
O
a)
CE
w
40
0
Rio,'
H S. Pt
0,44,
4.3m
0.0.0
000
000
.0.C-1 ,C-1
a1 I
P •
.......-.0
n
a
H
WTI
w04-\
• •
1
.r1cs,
m
‘41w0rN \1114-\
m
0.
H a)
o3
to H
.00
OPE 4
00 0
,-1,) C
IF-. ...
c\14)
I
."-... CI. 0)
H O \
W H
4.1N
-51
4),
0
0
.0
•-• 0 ,-I
..-1
4.4.
4Z141W 00
4-1 0(1)
00
.0.0,0
00
S. S.
0 0 .0 a) a)
.-4.-10
0,0,
110
N3-,4 v\ v\
\\I MN
HMH
II
I I I
W h0
41 10 Fit 00
-
CO 0
Co .4
a> H
..-la
000
MO
H
0
co
w
•4
H
04
4. W
o
0
0)
o
..-1
H
10•
W
4,
0
a)
e
4-,
4-,
.-I
W
41
o
0
r-I
'0
0
•-•
a)
P
o
0
4-,3 0
4-.
Y 4.
a)
a)
0
0
...4..
-t
0
a) . •
P a> P a> a>
0
.4.0-4.0-1
00
00
...,,c1 4
4
000 OPOPP 4:> 3 $
a)
6.9
0 0 0 oS oS
0 0 0 0 0
O
4-, 4-,
03 aS aS 4.4.
0
W nd P E
•ri
>a >a >a
0 0 0
I
4. 4-o 4-n
0 a) 0 0
o
qi
C
C
1
1
Pp.()
4.4. ..-1 PP. P
P
O
..-10,-I=0
mai
4-1
4.
0
000 00 0
r-i
./-I
R7 ,0 rd
.-4 -.-4
4.0 40 104-,.-I - .I
a3 0.4-D
.4
0
a> 0 a) 0, 0,
4-, 4.
a)
0 0 0 P
4-1 Cd
a)
C
E se
mia,mfa,o. -I ,' .-1 M zoo CC
O
to
..-1
e
4-, 03 bD
I-1
l 3,
3, 3.
bD ai bD oi 03 00aaop. 000 ma) a)
40 0
0
•,..-, 000
as
IC WW OWCI,WW ....-I,I 4.4. 0 WC Cd
. .•..
m 4-0.4. .0 .0
4.4.414..P.P.).-, cc
0 0.H .0
.0c4 OH
Wm,
0.00.0.0
,40 OJ 0W ...p
H .0 WWM 00
H
r-i
03
0 0 .4-
r-IHN
H 0
0 o 2
HO mwm 00000 CC 00 COO 00 C 01 C
42) HHH ,40.100 ....4.0,40 •-1 .-I,i .-I W4-1 4-1,I 4-).-I .-40
m
P
00
4,
ao ao
,.. MN .1)isli"V ..2..ri g55 .-IH •d
00S Oaf 4.40
-o
0 .-1.,4.-[ HooHA-N aa..-laH HHH MW H 04.
P44 P,.., 4.W
a) 0 w a) a> 0 0 0 .0
O
$.3.3. \\\\\ WW W
0
.0
r-I H C•1,--1 H •-.1 .1 ,1
CO
O
..•'. :E:
Q0 Cl (4 )a r... CD 03 <4
-11
4
m.
Figure 1.--Cutting diagram illustrating method of making
up panels with various types of defective specimens and
controls.
(M 81473 F)
Figure 2.--Methods used to build up facing thickness of test
specimens. A, thickness varied gradually from four to
eight plies; B, thickness varied abruptly.
(m 8 14 74 F)