Kayak Building Materials Testing

Sam McFadden, a grad student at UC Davis offered to do some material testing. I had some panels I had made several years ago with the intention of breaking into small pieces so I sent him some of the sample which he proceded to bust. The following is a brief presentation of the results.

 

The Instron test machine set up with two supports and the center loading device. A sample in the machine prior to testing The same sample after testing

All testing was performed on an Instron Model 1321. The samples were broken on a 3-point fixture with the load frame moving at 0.65 inches/second. Six different lay-ups were tested with varying strip thicknesses and fabric coverings. Each configuration was cut into 8 pieces and tested with the strip orientation either parallel to the length of the sample of at 90° to the length. Each configuration was broken by bending both with the center load on the front face and with the load on the back face. The panels were constructed of Northern White Cedar with cove-and-bead strips. The epoxy used was from MAS. The "Front" face of each sample has one fill coat over the initial wet-out coat on the fabric. The "Back" face just has the wet-out coat. These samples were made without a seal coat on the wood prior to glassing. All fabrics, except the AeroFab, are oriented such that the fibers are running parallel and perpendicular to the length of the test sample.


A sample after testing shown at approximately full size (assuming 72 dpi display screen)

The test samples were approximately 1 1/2 in. by 5 in. The supports in the fixture were set at 4 in. separation and the data below is corrected for width. Samples 1 through 4 of each lay-up were had the wood strips running the length of the sample and samples 5 through 8 had the wood strips oriented across the width of the samples. The tests were performed with the center load applied to the front face on samples 1, 2, 5, and 6 (Front in compression) and with the center load to the back face samples 3, 4, 7, and 8 (Front in tension).

Photographs of all the test samples may be viewed on a seperate page. Clicking on the links throughout the text below will display a picture of the sample in question. You will notice that the samples are marked with a code. For example the sample above is marked "A3T", this is sample A3 and we are viewing the "Tension" side. The tension side is opposite the center load, or inside the boat if you are hitting a rock. A "C" in the code indicates the "Compression" side and means it is the side adjacent the center load. This would correspond to the side which hits the rock on the boat.

Since performing these tests, Sam has created a drop-test fixture. I still had more of the original sample material. He was able to test these samples using his new fixture. Thanks again to Sam for his work doing these tests.


Sample A

1/4" Wood - 4 oz Front, 4 oz Back

Panel Density: 8.47 oz/ft^2

Samples A1 through A4 show failure of the fiberglass on the face opposite the center load with a clean break. The wood is broken clean through almost to the face adjacent the center load.

The fiberglass adjacent to the center load shows a small amount of stress in the form of whitening around the glass fibers.

Samples A5 through A8 show failure of the fiberglass on the face adjacent the center load. In samples A5, A6 and A8 the fiberglass has lifted up off the wood and folded. On sample A7, the failure appears a small crack with just a little bit of lifting of the fiberglass. In samples A5 and A8 this failure occurred at the joint between two strips. In samples A6 and A7, the failure is in the middle of the strip.

On all four strips A5 through A8, the face opposite the load shows no sign of stress.


Sample B

3/16" Wood - 6 oz Front, AeroFab Back

Panel Density: 8.82 oz/ft^2

AeroFab is a 9 oz biaxial E-Glass with fibers oriented at 45°

Samples B1 through B4 show failure of the fiberglass on the face opposite the center load. In samples B3 and B4, the fiberglass and wood is broken clean through almost to the face adjacent the center load. With samples B1 and B2, the break in the AeroFab is uneven and broad. Not all the fibers are broken and the break does not appear to go completely through the wood.

In samples B1 and B2 the fiberglass adjacent to the center load shows a signs of stress in the form of whitening around the glass fibers in an area about 1/2" wide. In samples B3 and B4 the fiberglass adjacent the center load shows no visible signs of stress.

Samples B5 and B6 show failure of the fiberglass on the face adjacent the center load where the fiberglass has lifted up off the wood and folded. In samples B7 and B8, the AeroFab is dented into the wood and it shows a sharp line of stress where the center load pressed against the sample.

Samples B5 and B6 show no visible signs of stress on the face opposite the center load. Samples B7 and B8 show a slight amount of whitening around some of the glass fibers.

 


Sample C

1/4" Wood - 2 Layers 6 oz Front, 6 oz Back

Panel Density: 10.28 oz/ft^2

 

Samples C1, C2 and C4 show failure of the fiberglass on the face opposite the center load. In samples C1 and C2, the fiberglass and wood is broken clean through almost to the face adjacent the center load. With sample C4, the break in the 2 layers of fiberglass is uneven and broad. Not all the fibers are broken and the break does not appear to go completely through the wood.

In samples C1, C3 and C4 the fiberglass adjacent to the center load shows a signs of stress in the form of whitening around the glass fibers. In C1 this whitening is uneven, with samples C3 and C4 it is straight and clean. Sample C2 has an area of delamination about 1.6 inches wide on the face adjacent to the center load on one side of where the load was applied. Sample C3 shows little sign of stress on the face opposite the center load.

Samples C5 has delamination of the fiberglass from the wood over a 3/4" wide area on the face adjacent to the center load, the fiberglass shows a small amount of stress but is not broken. Sample C6 shows a small amount of stress but is in good condition on the face adjacent the center load. Sample C7 has a 1/4 in. wide section of fiberglass and wood which has lifted on the face adjacent the center load, the glass is partially broken. On Sample C8, the fiberglass on the face adjacent the center load is cleanly broken.

On the face opposite the center load samples C5, C6 and C8 show delamination of the fiberglass from the wood. The fiberglass is intact on all the samples (C5 through C8) with just slight signs of stress. C7 did not delaminate on the face opposite the center load.

 


Sample D

1/4" Wood - 6 oz Front, 6 oz Back

Panel Density: 8.98 oz/ft^2

 

Samples D1 through D4 show failure of the fiberglass on the face opposite the center load with a clean break. D3 shows signs of stress surrounding the break.

On the face adjacent the center load on samples D1 and D2 fiberglass has lifted slightly off the wood and there is signs of stress in an uneven line across the sample. Samples D3 and D4 show a clean line of stress on the face adjacent the center load.

In samples D5 and D6 the fiberglass has delaminated and folded upwards on the face adjacent the center load. The delamination is about 1/4" wide. In samples D7 and D8, the fiberglass has broken cleanly on the face adjacent the center load, in both cases the break occurs at the glue joint between two strips.

The face opposite the center load is unscathed in samples D5, D6 and D8. There is no visible sign of stress. In sample D7, the fiberglass has delaminated from the wood for about 3/4". The fiberglass shows slight signs of stress but is in generally good shape.

 


Sample E

3/16" Wood - 4 oz Front, 4 oz Back

Panel Density: 7.21 oz/ft^2

 

Samples E1, E3 and E4 show failure of the fiberglass on the face opposite the center load with a clean break. The wood on sample E1 has 3/4" long split running diagonally through the thickness of the strip. On samples E3 and E4, the break through the wood is clean and straight. Sample E2 is broken clean in half. The wood is broken quite sharply with little splintering.

The face adjacent to the center load on sample E1 shows a small amount of stress going diagonally about 1/2 way across the sample. On samples E3 and E4 the fiberglass adjacent to the center load shows a slight line of whitening around the glass fibers.

In samples E5 and E6 the fiberglass has delaminated and folded upwards on the face adjacent the center load. The delamination is about 1/4" wide. In samples E7 and E8, the fiberglass has broken cleanly on the face adjacent the center load.

All four samples E5 through E8 appear unscathed on the face opposite the centerload.


Sample F

1/4" Wood - 4 oz Front, Kevlar Back

Panel Density: 9.23 oz/ft^2

 

Samples F1, F2 and F3 were all tested with the centerload on the "Front" or fiberglass side. F4 was tested with the Kevlar side in compression.

Samples F1 through F4 show failure of the Kevlar or fiberglass on the face opposite the center load with a clean break. The wood is broken clean through almost to the face adjacent the center load.

The cloth adjacent to the center load shows a small amount of stress in the form of whitening around the fibers.

In samples F5 through F8 the cloth has delaminated and folded upwards on the face adjacent the center load. The delamination is about 1/4" wide.

All four samples F5 through F8 appear unscathed on the face opposite the centerload.

 

Comparative Analysis


I find the above curves a little overwhelming, and they can be hard to interpret. Some of the curves go high, showing they absorb alot of force, while others go far to the right, showing they can absorb a large deflection. It is hard to determine which is better. However, if you look at the area under the curve you can get an idea for how much total energy the samples absorb. There are several ways of absorbing energy. One is to be stiff - hard to bend, requiring a lot of force move a little bit. Another method is to be resilient - easy to bend, but able to bend a lot without breaking. The area under the curve takes both factors together and combines them into one number.

 

Total Absorbed Energy: This data compares the area under the curves across the full range of the test. This favors samples which don't break completely and continue to absorb energy after the sample has broken. For example Sample C looks pretty good here. Higher is better.

The above treats all the samples equally. It doesn't cut lighter layups any slack. You can compare their strength relative to their weight by dividing the above by the panel weight.

Total Energy divide by Density: The sample weight is proportional to it's density. This compares energy absorbtion relative to a uniform weight and is similar to strength-to-weight ratio.

The above graphs compare the samples across the whole range of testing and accounts for their ability to continue to absorb energy even after they have broken. Most of the samples started to fail at some point near where the curves reached their maximum force. The failure may start as a slight compression of the surface indated by a slight drop in the force before preciptous drop or gradual fall off. This initial drop off can be considered the first failure, although some samples my show signs of damage before the maximum force. The energy to cause this first failure is area under the curve up until the maximum force.

Absorbed Energy to Maximum Force: How much energy the sample can absorb before it begins to fail.

Again, the above is a direct comparison where below is a "pound-for-pound" comparison.

Absorbed Energy To Maximum Force, divided by Density: Roughly strength-to-weight ratio for initial failure. Or if you had two samples of the same weight the one with the higher value could take more punishment without any damage.

Conclusion


It is hard to draw any conclusion based on a few samples, but in an effort to create a composite number for each different lay up I have averaged together all the data for each layup.

Combined "Strength"-to-Weight to First Failure: Averaged data for each sample. "Strength" here is the ability to absorb a force without damage. Data is normalized for sample density.

Looking at the previous graphs it is obvious that there is a lot variation in the data due to grain orientation, and direction of bending. However, I will give a shot at drawing some conclusion. Not surprisingly,incorporating hi-tech materials such as Kevlar or the non-woven glass "Aerofab" create stronger panels pound for pound. The data suggests that adding more glass adversely effects the strength-to-weight ratio but the results are so close together that it is probably within the margin of error of the testing.


 

High Performance Wooden Sea Kayak Designs
Send snail-mail to:

Guillemot Kayaks
54 South Rd
Groton, CT 06340

Phone/Fax: (860) 659-8847

Email Guillemot Kayaks
© Nick Schade - All Rights Reserved