Composites 2004 Convention and Trade Show or supported, and whether tabs are bonded to the

American Composites Manufacturers Association specimen ends.

October 6-8, 2004
Tampa, FL USA The “shear” compression failure mode has long been
considered ideal for basic laminate properties usage,
while a buckling failure is generally considered to be
undesirable. But in reality, all compression failures have
Compression Testing - Comparison of an element of buckling, although this can be confined to
Various Test Methods a very localized area. The challenge is to reduce the
tendency of the test specimen to buckle on a global scale
Arthur R. Wolfe, Goodrich Corporation (column buckling) during the test.
Michael Weiner, Goodrich Corporation
The Test Methods
Abstract
The different test procedures vary in two main respects –
Several test laminates were fabricated for specimen shape and method of loading. In most cases,
evaluation of mechanical properties under in-plane the test specimen is held or restrained in some sort of
compression loading. The panels were constructed using “jig” or fixture during loading. It is interesting to note
a conventional E-glass/vinyl ester knit, a carbon/epoxy that the methods that work best have been developed
prepreg, and an E-glass/epoxy prepreg. The laminates relatively recently (specifically for the composites
were cured and post-cured according to typical FRP industry), while the older and less desirable methods
manufacturing practices. were adapted for composites from thermoplastic sheet
testing procedures.
Different types of test specimens were cut from
each panel to compare test methods. For example, the The common methods in the early days of composites
carbon fiber laminate was tested for compressive testing were all “end-loaded”, in which a coupon was
strength using ASTM and SACMA procedures. All of held erect in the test fixture, with an end protruding
the methods evaluated are commonly used to determine slightly from the top. The specimen was held fairly
the compressive strength of an FRP laminate. Each of the loosely in a supporting fixture to prevent buckling. The
specimens were cut in the same orientation, with the test fixture was loose enough so that it would not bind up in
axis being the 0° reinforcement axis. the fixture during loading, but tight enough to prevent all
but the smallest out-of-plane movements. In these tests
It was shown that the results of compressive the end that was loaded would tend to “mushroom”, or
strength testing of a given laminate will vary according “broom”, when relatively high modulus materials were
to the method chosen. Some procedures are not tested. In some cases, the test specimen would be of the
appropriate for medium-to-high modulus composite dog-bone or dumbbell shape, similar to thermoplastic
laminates at all, due to the type of failure mode. Recent tensile test specimens. This was done to try to force a
methods designed specifically for high modulus failure within the gauge length area. ASTM D695 is an
composite materials yield higher, more consistent values example of this type of test (see Figures 1 & 2).
for compressive strength. Older, historical methods may
have value when testing lower modulus or thicker The Suppliers of Advanced Composite Materials
laminates. Association (SACMA) developed another end-loaded
compression test, SRM 1R-94. This method uses the
Introduction same holding fixture as ASTM D695, however it utilizes
a tabbed, straight-sided specimen (see Figure 3). The
Like many composite testing procedures, the tabbed ends reduce the probability of mushroom-type
determination of compressive properties has long been failures, while the holding fixture prevents global
considered a “black art”. The various methods can buckling as in the D695 method. While the short gauge
produce wildly different results, when the same material length of the SACMA procedure encourages the
is tested. The choice of test method will influence the preferred failure mode, its very small area precludes the
results – some procedures yield higher strengths, and use of a strain gage for modulus determination. A
others produce lower values. separate, untabbed specimen must be loaded to measure
elastic properties. SACMA is no longer in existence, so
Studies dating back to the 1980’s have looked at the this method is not reviewed or updated periodically.
influence of test method on results1. Some of the
differences between methods include how load is applied Shear-loaded test methods use a tabbed specimen, with
(shear vs end loading), whether the gauge length is free the tabbed ends held in wedges or grips, similar to tensile
grips. Shear loading eliminates the possibility of
mushroom failures. The gage length of the specimen by the method. The rate of loading was that specified in
(distance between tabs) can be opened up to allow the the procedure. All tests were conducted under ambient
use of a strain gage, if desired. Increasing the gauge lab conditions, those being 75°F (+/- 3°F) and 50%
length increases the possibility of a buckling failure. relative humidity (+/- 10% RH).
Shear-loaded methods include ASTM D3410 and ISO
14126 (see Figures 4 & 5). A minimum of 5 specimens was tested for each
sample/procedure combination. The results as shown in
A relatively new method uses a “combined loading” the tables below represent the average value for the set of
scheme, with the untabbed specimen held securely test specimens.
between wedge faces, while a cap prevents a mushroom
failure. This method (ASTM D6641) has been found Test Results
very effective for balanced (0/90) laminates. A short
specimen gauge length reduces the buckling tendency. In almost all cases the SACMA method yielded the
highest compressive strengths. For the E-glass/vinyl ester
The compressive strength of FRP skins on a sandwich infusion sample, the 4-point bending test produced at
panel can be determined using a 4-point bending test. In result that was ~20 MPa higher than the SACMA test.
this instance, the composite skin is restrained only by its The SACMA test has the shortest specimen gage length,
adhesion to the core material; no other restraining and thus the least tendency to buckle.
fixturing is used. The test is set up so as to result in a
compressive failure of the upper skin, between the Generally, the ASTM D695 test produced results 10%-
loading noses. This method is represented by ASTM 20% lower than the rectangular, tabbed type of test, even
D5467 (see Figures 6 &7) when “mushroom” failures were eliminated from the
data set.
A summary of specimen type and method of loading is
provided in Table 1. The test results are summarized in Table 2.

Experimental Procedure Conclusions & Recommendations

Typical materials used for composites parts were chosen Determination of compressive properties using the 4-
for test. Several test panels were fabricated using point bending test is more time consuming and expensive
conventional techniques. Separate panels were made than any of the other methods. There would not usually
using these raw materials: be any advantage to the use of this method for typical
composite materials testing.
- carbon/epoxy prepreg (single-skin)
- E-glass/epoxy prepreg (single-skin) The legacy test (ASTM D695) will produce lower
- E-glass/vinyl ester (single-skin & sandwich) material strength than the other methods, but may still be
of use for thick composites and low modulus materials.
The plies of fabric were all laid up in the same
orientation with respect to each other; i.e., the 0° No significant difference was seen between the SACMA
reinforcement axis laid down in the same direction for test and the relatively new combined loading test (ASTM
each ply. The prepreg layups were then vacuum bagged, D6641).
and cured according to manufacturer’s instructions in an
autoclave. The E-glass/vinyl ester panel was fabricated The choice of test method is dependent on cost, time,
using a wet resin infusion technique. A sandwich panel ease of sample and specimen preparation, and the
was also infused using the same skin laminate schedule potential end use of the data. Composites engineers
as the E-glass/vinyl ester panel. should beware of attempting to get the highest design
values possible, as a few spectacular compression
After cure, test specimens were cut from each panel. All failures of composite parts will slow the introduction of
tests were performed along the 0° reinforcement axis. composite materials into new markets.
Great care was taken to maintain this alignment when
cutting the specimens. The specimens were machined
according to the instructions, schematic diagrams, and
tolerances as given in the test procedure. Where tabs
were required, they were cut from sheet stock of a
commercially available fiberglass material (G-10).

The compression testing was performed on an Instron

universal test machine, using holding fixtures as required
Figure 1 – ASTM D695 Test Specimens Figure 4 – ASTM D3410 Test Apparatus

Figure 2 – ASTM D695 Support Jig (also used

for SACMA SRM 1R)

Figure 5 – ASTM D3410 Test Apparatus

Figure 3 – SACMA SRM 1R Test Specimens

 Carbon/Epoxy EGlass/Epoxy EGlass/Vinyl
Prepreg Prepreg Ester
Infusion
ASTM 644.5 345.6 331.9
D695
ASTM 703.8 337.1 -
D3410
ASTM - - 443.1
D5467
ASTM - 377.1 -
D6641
SACMA 811.8 378.2 423.4
SRM 1R
Table 2. Compression Strength Test Results
Figure 6 – 4-Point Bend Test Apparatus (MPa)
Note: Values represent the average of 5-8 test specimens

References

1. “Compression Test Results- A Tough Nut to

Crack”, Advanced Composites, July/August
1989, pp. 57-63.

2. “The Influence of Test Method on The

Compressive Strength of Several Fiber-
Reinforced Plastics”, Journal of Advanced
Materials, Volume 25, No.1, October 1993, pp.
35-45.

Acknowledgements
Figure 7 – 4-Point Bend Test Apparatus
The authors would like to thank the lab personnel at
Cincinnati Testing Laboratories and OCM Test Labs,
both of whom provided no-cost testing services for this
Test Method Specimen Type Loading presentation.
Condition
ASTM D695 Dog-bone End, side support Authors
ASTM D3410 Rectangular, Shear, no side Art unwittingly joined the composites industry after
tabbed support graduating from college in 1983. Unable to escape, he
ASTM D5467 Sandwich panel 4-point bending has since worked in a variety of testing laboratories, and
currently coordinates the composites lab at the
ASTM D6641 Rectangular, may Combined end & Engineered Polymer Products division of Goodrich
be tabbed shear Corporation, Jacksonville, FL. Mike is a Senior Engineer
SACMA SRM 1R Rectangular, End, side support at Goodrich-EPP, and has been involved in the design,
tabbed fabrication, and testing of many large-scale composite
ISO 604 Rectangular End, no side structures.
prism support
ISO 14126 Rectangular, may Shear or end
be tabbed
Table 1. Test Methods for Composite Materials