There are many theories. One is the "excessive volune/bad powder? one. Lots of space, the powder ignites in a slow burn to a deflagration (primers detonate, powder deflagrates) and creats excessive pressure. Never been proven.
The one I like. Loose crimped bullet. Primer detonates. Powder begins to deflagrate and the bullet pops loose way too soon. Plugs the bore but still in the cylinder. Pressure builds with excessive resistnce caused by the forcing cone.and cylinder blows part.
Any other theories?
BTW, somebody mentioned the powder crumbling. This reminds me of the true story when Elmer Keith blew up a Single Action Army in 45 Colt. He didn't over load it in a true ense. He "ground up" black powder so the case would hold more. The handgun let go real good.
First of all, in ballistic terminology, you CANNOT have a "detonation" with small arms smokeless powder! The definition of a detonation is where a pressure front of 25,0k0 ft/sec is created. There is NOT enough chemical energy in smokeless powder in a metallic cartridge to cause a true detonation.
Now, having said that, DETRUCTIVE OVERPRESSURES CAN be created under the proper loading conditions. Depending on the burning rate of the smokeless powder, it takes 5,000 - 7,000 psi of pressure before smokeless powders will reach stable, prgoressive burning. Below that pressure, the powder can deflagrate (ignite and begin to burn), but if not obtained quickly enough, the powder can go out or almost go out. If "shot start" (the bullet begins to move) before the pressure comes up high enough, the resulting increased combustion volume can drop the pressure below the point where enough force is created to engrave the bullet into the rifling, and the bullet stops. If this happens where the bullet stops in the forcing cone of a revolver, you now have a "closed bomb" condition, and the pressure quickly goes up to where the powder starts to burn progessively. That means the higher the pressure, the faster the burning rate, whcih raises the pressure even more. When the pressure exceeds the ultimate tensile strength of the cartridge case, it will probably rupture where it has been stretched the thinnest, releasing not only a lot of energy, but also producing jets of flame that act like a cutting torch on the thin steel walls of the cylinder. Generally, the top of the cylinder then separates taking the topstrap of the gun with it to parts unkown!
Because it takes about 7,000 psi to expand a brass cartridge case (especially .45 LC, where most of the chambers are quite a bit oversized), it
is difficult, if not impossible to see premature shot start using an Oehler M43 pressure measuring system, but not impossible. It can usually be seen as a pressure spike, followed by a drop in pressure and a subsequent increase to maximum pressure. A marked spike was observed by an ammo maker using an M43 with a .40 S&W pistol. No, no overpressure, but that load was never brought to market! A South African ballistician also detected premature shot start using equipment not used in the U.S. While the oft-quoted double-charge in the .38 Special 148 gr 2.8 gr Bullsye loading as blowing up guns, such a charge could produce pressure that would blow the old M10 S&W's, that may have occurred/ But experiments run in presure guns showed that such a double charge would produce less pressure than a .357 Magnum, yet I have seen at least one Colt Python with a blown cylinder and .38 Spl. brass stuch in the cylinder. It is also true that the target shooters who used the 148 grain full wadcutters tended to use no crimp, believing it was detrimental to accuracy. We have also seen light loads using the mid-burning powders in .45 LC especially, where low bullet pull (light/n crimp plus too little neck tension) were associated wtih overpressures that blew Colt-type replicas.
Bottom line: stay above the minimum, and CRIMP, CRIMP, CRIMP!