Aramid tube

Even at the keystone there is no change in direction for the stress. That is simply a mis-conception. A free body diagram shows it clearly. The keystone is in compression and it is a static system.

 

If you used metal tubes as the form/mold as opposed to using a mandrel, you would end up with a finished outside surface. Using a mandrel probably means individually finishing the outside of each tube. I'm envisioning something like an innertube in the metal tube/mold to apply pressure outward. The big problem would be getting cloth inside the tube in a reasonably orderly way so you would get consistent results. I believe all GRP layups shrink to varying degrees when curing, so that would help with removal from inside a tube as opposed to the outside of a mandrel, where laminate shrinkage tends to make removal harder.

Maybe use a two piece mold.

Possibly the cloth could be rolled on a bias and inserted into the mold and then expanded, in the same way you get out of a Chinese finger trap.

 

finished tubes

Thanks SamSam...interesting suggestion...the sleeve...thanks. I'm pretty convinced the air mandrel is a workable solution. This is my plan so far.
The weave of the aramid cloth naturally resists "racking" (here we go again) because it does not flex. To me this means any "flat stretch" over a spinning mandrel (reel) "square" to the bolt of cloth (also a reel) can be put under light tension (using PWM controlled, opposing DC motors to maintain tension) and rolled on a spinning mandrel (the tubing reel) without much confusion (of fibres/layers). As to 45 degree warps, I'm thinking it would be relatively easy to set up an additional reel (the 45 would have to be hand rolled because right now as I cannot think of an alternative) to feed the 45 degree warp in between the 90 degree wraps. The tension of the primary bolt (reel) should be enough to retain the position of the 45 degree warp. Of course the problem would be precisely cutting the length of the 45 degree warp. Further the process would rely on making the transition from reel tension to vacuum holding of the warps steady in place. In this case, because the surface of the aramid is relatively slick, releasing the tension or pressure to atmosphere could cause untoward shifting of the layers...so this may make a case for an unfusion approach. Certainly it would help control the resin:fibre ratio and support more quantifiable testing. This would be relatively easy to build, compact in size, and an optimum cycle time could be established and tended while doing other things in the shop...I''m there anyway, mostly piddling.
I also had some ideas of how to split the sleeve/mandrel that might serve as a good starting point for squaring the pull of the starting wrap.
Thanks again for your suggestions.

 

It might be you're re-inventing the wheel. I googled 'aramid fiber tubes' and there were a lot of places that sell it, a lot of Alibaba sites. This is the only place that had prices... https://www.rockwestcomposites.com/materials-tools/vacuum-bagging-materials/resin-systems and they are horribly expensive. It could be everything they sell is expensive, you can compare their resin prices to the real world and compare. There are also places that sell braided aramid sleeves of various sizes and unlimited lengths. Like the finger trap.

I imagine there are a lot of uses for that type of thing, golf club shafts, fishing rods, who knows what else, so I imagine industry has got it down pretty good. It would be handy for small shops to have a way to convert scraps into useful products though.

 

tubing manufacturers

Thanks for the link...it is HORRIBLY expensive. I might have expected filament wound to be around that altitude given the machine time. I have never used anything but the brother's resin. Their customer support and research is famous. They recently (meaning I just noticed) advertise custom resin blends for specific flexibility requirements. THis is interesting as well.
I am sure they have it down. I don't fancy myself competition at all, just my usual po'boy approach to getting where I want to be. In this case the tubing (or whatever profile) is just the first step in developing structural fab methods.
Interesting you should mention scraps...a have a bunch of scraps as well...just unsure how to utilize them correctly (highest and best use kind of thing). I never throw anything away.
Thanks again.