Hi Gerry,
I aree with your point on a well built tortion box not developing warping, however, you and I have both working in wood for years, not everyone builds to the same standards. I'm not saying anyone is intentionally sloppy. But building in wood is a skill that is developed over years, and a metal worker who has never worked in wood (or rarely does) may not have developed that skill set. My wood skill are good and I'll be the first to admit my metal work skill suck, not because I'm sloppy, but because I haven't developed the feel and intuition for it. I look forward to seeing your complete design.
On point 2 that is good to hear that the pipe may not be that much of a structural member, and I did read the point about the laminated ply making up the main structure of the roller.
As for point 3, and I invite you to try this yourself. Epoxy has very good strength in bonding metals for translational stresses. But it has fairly poor peel strength. take two pieces of aluminum or any other metal (cleaned bonding surfaces of course) and glue then together. Try to pull then apart in a way that one has to slip across the others surface. Tough to do. Now, take a knife and slide it under one edge. Torque the knife to start the peeling process. They will peel apart pretty easily. This is my point about changing out the metal strips. The epoxy will anchor the metal strip very well when the whole mechanism is clamped together by the rollers and you will get no longitutinal slipping of the metal strip. But when the piece needs to be changed out you can peel it off fairly easily by applying a stress that it would never see in the normal course of operation of the machine.
If you cut it to small you can always nail another piece on the end, but if you cut it to big... then what the hell you gonna do?
Steven