After experimenting with longer magazines, multiple projectiles and so on I realized that multiple layers are surely possible in a magazine cannon but will leave some non-ignited TNT in the projectile part of the magazin cause you'd use the power or two shots for one - one projectile less used but still in the magazine. So if you got bad luck, the magazine will blow with all the sticky pistons used. Furthermore, adding more than two lines of re-loadable projectiles/charges (not counting the one in the barrel) gives the same problem as a sticky piston elevator when attempting the move all pistons back to their compressed state. So I dropped it and used the model with two lines in the magazine and one in the barrel. The current one provides convenient top-level access for all control and a stair to go down the loading area. For long-rang shooting, there's also a switch for a more flat trajectory realised with some sort of different choke like used in shotguns. It's as blast-resistant as possible in case of using the wrong magazine levels (you'll need to hit it in the right order), you'll need to replace the slab as well as the trajectory piston and the magazine pistons as well. I tend intend adding a more convenient, less inconvenient method for cycling cause it would involve some memory cells that simply take too much space. You can alter the explosion time of the projectile quite fine-grain using repeaters. It plays a sound when priming the projectile, so can atleast some know how long it takes to detonate. One can do pretty deep tunnel with long detonation times - it really works lovely. Guess I'll keep as it is right now and start something else to build. Think I'm trying this programmable sequencer another time. Next is the next study day on schedule and I don't wanna go in there with no good memories about spent time, I don't to program all day long.