> What I ridiculed the was lack of inclusion of West's disclaimer. West said
> the data had too much scatter to be of use. I looked at West's data and
> computed the tensile load in the coth (simple composite beam formula) of
> each sample at failure. There were 4 clusters in the data, 100% (scaled
> to), 83%, 75%, and 60%, there should have been 1 cluster. It looks like 4
> batches of epxoy were mixed (perhaps 4 people did the work or some such).

I've got to remember that this bbs will not show old fashion emoticons. Should have used the :-) to ensure there was no barb or sting, for none was included. Just wanted to get that clear.

Now, about the data. I don't recall you saying anything about lack of disclaimer, but I could be wrong. I also do not recall West saying the data was of no use. I think they said instead that further testing to obtain a larger sample pool was needed before drawing any strong conclusions.

I do not quite follow what you are saying about four clusters. For there were four sizes of wood tested, with four runs of different glass amounts upon each. What clusters are you speaking off? For in fact, there should be no repeats to gather clusters of single point data because each point was unique with no repeats. The results were perfectly reasonable in that more oz of glass equates higher strength, and same the plywood. I do not see how you have concluded that four different epoxy samples or applications were used.

> For my purposes I think it is accurate enough. In any case it may be
> better than nothing. I am not trying to measure strength so that we can do
> engineering calculations. That requires lots of samples and accurate
> equipment.

And that was all the West test was about George. Basic observation of glass amounts and wood. Nothing detailed, and not intended to cause a revolution in the engineering of boat building or the world of composites. Just a little info for people to look at, observe and learn from. Geared right towards the type of person who will build their own kayak or canoe.

> When I drop hammers, I am only trying to measure large relative
> differences so I can compare several layups. Refering to the West results
> above, I figure anything less than 40% is going to be in the noise of day
> to day construction variation.

On a single day, yes I would roughly agree with that. However, it is one more point in the large data pool if you keep gathering data and compiling it. If you are not, you are working against yourself and your interests of a detailed study. More data makes trend and noise recognition far more accurate and easier.

> I am not going to try to determine which is stroonger 6oz cloth or 2
> layers of 3.2oz cloth (they lay up to the same total weight). But I can
> determine if they fail differently.

While I agree with this point singularly, that was not what the West test was about. For you had no data of that type to compare. What they did allow you to observe was that generally one layer of 6 oz cloth is not as strong as two layers of 6 oz cloth. They also allowed you to observe that it was really the oz of cloth, and not the number of layers that was giving the strength. They also allowed the reader to observe that yes in deed, thicker plywood is in fact stronger plywood, as a general rule. And left the weight and flexibility issues alone.