Hi Nick,

If you analyse a beam (composite or just a metal or wood) you will find that the slope you see in these curves doesn't change with the direction you push in even for an unsymmetric lay-up (2 layers on one side - 1 on the other). What happens is that the neutral axis (the stress is zero in this plane) is shifted towards one side. It is not so easiy to predict the failure point of these laminates. You have to establish the stress levels for failure in compression and tension as well as for interlaminar shears and peel stresses (may become more important when you have narrow specimens). Then you can predict properly what is happening.

Hand lay-ups are notorius for showing large spreads in failure points. A vacuum bagged sample offers more consistancy - or even better an autoclave sample. You get better repeatable conditions. With hand lay-up your epoxy to resin ratio is inconsistant. Further you find that the glass layer floats and is further away from the wood, which affects the stresses which the fiberglass sees. As I mentioned before, distortion of the glass can have a significant influence. You will always see that unidirectional fabric is stronger than bidirectional fabric (assuming the same number of fibers in one direction) when loaded in only in the direcion of the unidirectional fibers. This is due to weaving pattern you have in a fabric. A straight fiber is stronger.

But lets get back to what matters. You mentioned that the cedar is strong enough by itself for a kayak. I agree. The problems are impacts and abraision. I think that it would be very interesting to compare the damage created by different lay-ups under impact. Minimizing delamination between the wood and glass is the key.

Great thoughts Nick.

Andreas

: Sam McFadden did the work. I just supplied some samples and posted the
: results.

: Obviously as you increase the percentage of the layup that is glass, the
: amount of strength contributed by the glass will increase. But
: interestingly the fact that one side of the sample given had 1 layer of 6
: oz and the otherside had 2 layers did not appear to effect the slope of
: the curve when the samples were bent in the other direction. In other
: words the location of the reinforcement different layer relative to the
: load did not appear to effect the stiffness. The point of final failure
: was all over the place so with the number of samples Sam had to test it is
: hard to see the relative importance of the reinforcement location on
: strength.