CanadianPyro

I use a method that's very similar to this one. Works well.

Hydrogen & oxygen doesn't detonate (a supersonic reaction propagated by a shock wave) , it deflagrates (a subsonic reaction propagated by heat).

The deflagration can transition to a detonation given sufficient run-up distance. A stoichiometric Oxygen/Hydrogen mixture has a very high flame front velocity, so the distance is quite small. I seem to recall a test in which DDT was achieved in an open-ended pipe just 10" long and a few inches in diameter.

I believe Ammonium Perchlorate is only classified as a HE when the particle size is below a certain point, somewhere around 15µm. Above that, it's classified an an oxidizer only.

Not touching on the HE properties question, since that type of discussion is banned on these forums. Don't worry about using Google search; information is still free at the time of this writing.

When I roll tubes, I usually wrap a wooden dowel with a material that PVA glue won't adhere to such as thin polyethylene sheet, then roll the tubes around it and allow them to solidify on the dowel. This prevents deformation, and once dry, the tubes can be slid off the dowel.

my thought was that if the pressure wave is moving slower then there more chance of a uniform break on the case rather than it breaching on a weak point.

It's actually the opposite. If the pressure builds more slowly, the casing will tend to fail at its weakest point, and release all of the gas through that small opening. That's why relatively slow burning compositions like black powder can fire a shell into the air without blowing up the mortar.

With fast flash powder, the gas is generated so quickly that the inertia of the end plugs is sufficient to keep them in place while the internal salute pressure continues to rise well above the failure pressure of the plugs. Fast flash powder can be used in thick-tubed salutes with paper thin end plugs, and still produce a uniform burst of the tube, even if its failure pressure is many times higher than that of the end plugs.