This is just one guy's experience and his conclusions... any resemblence to the truth... well you know.
Back in the late 60's when I started shooting a 40mm, I was a starving college student and would pick up a 50# box of "B" blasting from Margraf Explosives in Rancho Cordova, CA when visiting the mother-in-law (after all the trip had to have some productive purpose). We were paying about $65 for 100# (or maybe that was for the 50# box). "A" blasting was rather thoroughly researched at that time as it was $85, made with KNO3 instead of NaNO3 for an oxidizer, had the same chemical composition as sporting powder (%'s of S, C, and KNO3) and with the granulation charts, one could figure out equivalencies between "A" and Sporting. Densities were in different units but after running the conversion the densities were also fairly close 'tween "A" blasting and the "g" series Sporting grades. I never could get "B" to swell a skirt of a mini-ball slug even though I tried different masses, alloys, skirt thicknesses and powder charges. My mold had three different cavity plugs to vary skirt thicknesses and two rings that could be put around the plug to control the depth of insertion into the cavity and thus vary the mass and OA slug length (pics below). "B" is (was) noticeably slower in the pressure rise than "A" or Sporting. "A" is also called fireworks powder by GOEX and "B" was discontinued about the time DuPont quit the black business.
I still do most shooting with "A" blasting 4Fa which is very equivalent in all ways, including performance, to Fg sporting. And it is usually 2-4 dollars a pound cheaper than sporting from the GOEX distributor. The only bad news about "A" is there appears to be a wider spread in performance between powder lots... beings we can only get/buy a 50# box at a time, by the time we shoot that up, the distributor usually has a different lot of powder and it is necessary to re-evaluate it's perforance in relationship to the previous lot and vary the powder charges (or aim points) accordingly. Again, the military has the same problem with their propellants and thus invented the velocity collars on all modern ordnance to rapidly ID the variance and automatically feed it into the fire control computer so he can comp their aiming points.
The good news about the Ordnance Manual info is it tends to verify what I found in experience... with KNO3 powder and a lead slug, the pressure must get pretty extreme at the bottom of the bore where the walls are thickest because I was able to get very reliable "upset" performance (the swelling of the skirt into the rifling near the bottom of the bore so it gets a good spin and stability before exiting the muzzle).
DD:
I also believe it is black powder pressure that retains it pressure when it meets resistance until the resistance is over come either by the movement of the resistance or the failure of the resistance... (Some body correct me if I am wrong or correct the part I have wrong.)
Or in other words the projo moves or the piece explodes...?? That agrees with my experience.