EXPERIMENTS WITH FOAM SANDWICH CONTRUCTION MATERIALS

[Robert Hughes]

A while ago I uploaded a discussion of the engineered carbon fibre crossbeams I've fitted to my Imagine Multihulls glassfibre Tiki 26, Zest.

http://wharrambuilders.ning.com/forum/topics/engineered-carbon-fibr...

At the time, I mentioned I was experimenting with materials for a new cockpit. I wanted to build a replacement that was significantly lighter than the existing one, while incorporating lockers under the seats, eliminating places where water could splash up in rough weather, and modifying the engine well. I had decided to use a foam sandwich construction, with structural PVC foam cores, but was undecided whether to use glass or carbon fibre, and polyester or epoxy resin. I did consider kevlar, but experiments showed that cutting, finishing etc would be more difficult, and it's not as stiff as carbon. I had received conflicting advice from various sources as to which materials would be best, and concluded that the choice really depended on how I would actually use them in practice. As an example, I would be using a wet layup method without vacuum bagging, and no oven. I thought the best way to decide would be to prepare some samples and test them.

The first picture shows some of the samples being prepared. The fabric is laid up on the foam cores by hand, just as you would normally lay up fibreglass. The fine white fabric is peel ply, which I used to help consolidate the matrix and exclude air. When it is peeled off it leaves an ideal surface for further bonding processes.

After the samples were laminated they were all trimmed to a uniform size, using a small bandsaw with a fine toothed metal cutting blade.

After trimming I examined the cut edges. This picture shows two samples of carbon fibre, the only difference being that the upper one is epoxy resin, whereas the lower one is polyester. It looks like the epoxy holds the carbon fibres more strongly so they cut cleanly, but the polyester gives a more ragged edge as short bits of fibre pull out.

This shows the test rig I used to measure the stiffness of the samples. Ignore the mug of tea, it's nothing to do with the tests, it's just there because I'm British and can't work without tea.

Each sample was supported on a pair of wooden blocks on a work bench, then progressively loaded with lead weights in the middle. The deflection was measured from below, through a hole in the work bench.

Each lead block weighed around a kilogram, 2.2 pounds. It was surprising how much weight some of the samples took, and it got a bit scary crouching down below to take the measurements!

The first graph is really just a test of the method, and illustrates the theory of sandwich structures or I-beams. Glass was laminated with polyester resin over three different thicknesses of core. You can see the dramatic increase in stiffness as the core thickness goes up, due to the load bearing faces being further apart. Thicker sandwiches are much stiffer for their weight, particularly as the core material is light. The fact that the data points fall on fairly smooth lines shows the measuring method works well enough for what I'm trying to do.

The next graph shows the comparison of polyester vs epoxy resin, for glass and for carbon fibre. In both cases the epoxy structure is slightly stiffer. You will also notice the lines for polyester stop just above 10kg applied load. The reason is that I stopped adding weights at that point, because very sharp high pitched snapping noises (“tink!!”) started to occur and I felt it was unsafe to continue. I believe these were individual fibres snapping as microscopic distortions took place in the matrix and the load was no longer shared properly across all fibres. The fibres that became more heavily loaded then failed. This never occurred in the tests of the epoxy samples, even with a further three weights added. This suggested that epoxy would certainly give superior results, in the application I envisaged.

The next graph compares the results of the carbon and glass fibre samples, both using epoxy resin. You can see the carbon is very much stiffer, the glass bending about 50% more under the same load. It should be noted that the glass cloth is much heavier.

To allow for the differences in cloth the results for carbon vs glass were recalculated by adjusting for weight, the results being shown below. The glass now bends about 2.5 times as far as the carbon for a laminate of the same weight. Please note that this is a fairly crude comparison, because there are many other potential variables such as different grades of glass, weaves of cloth etc available. It was sufficient for my purpose though, as it demonstrated a carbon structure would be much stiffer for the same weight, not just a marginal improvement.

I wanted to test a sample to destruction, but in view of how much weight was required in the bending tests I decided to use a much smaller sample, for safety reasons. This rig was set up on my main workbench, with a simple piece of wood laid across the test piece so that the weights did not fall off sideways. To give an idea of the stiffness of these materials note that the sample uses 10mm foam and is only about half an inch wide, weighing very little. For the same weight of material, mahogany could be 2.4mm thick, or aluminium 0.7mm. The wood would have snapped by now, and the aluminium would have bent down to the bench. To achieve good stiffness with those materials other approaches are used, such as ribs and bulkheads in wood, or clever extruded sections or tubes in aluminium. I can now understand why the America's Cup boats are built in carbon...

This picture shows a carbon/epoxy sample after loading to failure. Note the break is in the face under compression, not the one under tension. This suggested I should make the upper surface of the cockpit floor a bit stronger than the lower surface. The break is very clean and sharp; I also tested a carbon/polyester sample and the break in that was more ragged.

In conclusion, these tests gave me the information I needed, to decide to use mainly carbon fibre and epoxy resin for my new cockpit. While working with the materials I also saw that the carbon can be quite brittle in some circumstances, so I used a hybrid laminate for the floor, to avoid damage from anything like a dropped anchor or outboard motor.  The tests were not comprehensive, nor were they rigorous enough to publish in a scientific or engineering journal, but they were sufficient for my purposes and I hope you find them interesting.

Construction of the cockpit is now well under way, and I'll write that up in due course. I will say two things at this point though, it's feeling impressively light, but it's taking a lot of work...

[Roger]

Hi Robert,

Thanks for sharing your experiment - such valuable info. Don't your results also show that epoxy/glass would also be plenty strong enough to do the job, and be less vulnerable to dropped anchors et?

Roger

[Robert Hughes]

Thanks Roger.  The answer to your question is "probably"...  What I've actually used on the floor is a layer of glass with one of carbon / kevlar hybrid cloth on top...  Most likely over-engineered but hey ho!  More about that when the cockpit is finished.

Roger said:

Hi Robert,

Thanks for sharing your experiment - such valuable info. Don't your results also show that epoxy/glass would also be plenty strong enough to do the job, and be less vulnerable to dropped anchors et?

Roger

[Roger]

Hi Robert,

Did you also test the glass/epoxy combo to destruction? Even though it deflects more, I wonder if the glass will fail at a higher load than the carbon?

Cheers,

Roger

[Robert]

Hi Robert,

Thanks for publishing this. I am however not a huge fan of carbon fiber unless it is really needed. Sanding and cutting carbon is significantly more hazardous to human health than glass fiber as the fibers cannot be degraded by your body while glass fiber is broken down in the body.
I found these videos that discuss when to use glass and carbon fiber interesting:
https://m.youtube.com/watch?v=j7bQYEt_qtU
https://m.youtube.com/watch?v=KHXVf0SaJpA

R

[Valery Gaulin]

High tech material on a Wharram!!! Why not...

[Frank Higgott]

Nice work Robert, I found your experiments really interesting, it certainly got me thinking as to how I can apply some of this in the future, so thank you.

One thing I did find quite spectacular was your ability to balance all of those lead weights so well on top of that half inch wide piece of test material! Well done!

[Ian Bamsey]

Great work Robert, and really interesting. I look forward to your reports on progress.

I think it is to be expected that the failure mode is in the upper, compression, part of the beam. The carbon is particularly strong in tension. I think the compression surface fails due to delamination at a micro scale and there may be advantages in a thicker construction on the top layer. 

Well done.

[honey badger]

Good man Robert,your process of elimination is awesome.

[Jeremy Walker]

For practical info on sandwich panel catamaran construction, refer Derek Kelsall and his KSS system. He started with this type of composite construction almost 50 years ago and has kept at it with updates ever since.