> purpose was to drive at the max the 6L6GC
6L6 are not hard to drive. Almost any reasonable driver can over-drive them.
And the U-1075 would need more drive than the 1268 because of higher 6L6 voltages.
It's a cross-word puzzle.
Put in a tube. Not enough gain. Put in another tube. Too much gain. Throw in a tone-stack. Or a mixer. Less gain. You move things around and change them until you have a reasonable gain-structure from input to output. You also want a reasonably minimum cost, which suggests dual-triodes, but you have to get the right halves in the best places.
The U-1075 has a lot of inputs, so has a lossy mixer, and has to start with high-gain preamps and individual gain controls. After the mixer the level is too weak to go to a tonestack, but too much gain added may overload. So a low-gain stage. Then tonestack. The output stage is lifted direct from Fenders, needs pretty strong input, so need a little gain after the tonestack. While two low-gain stages seems complex, they are just one dual-tube. This also allows a 2.2K+0.3uFd cathode network, twice, to provide a lift and shelf in the midrange, a classic guitar EQ.
The 1263 looks like two hi-gain stages before the tone/volume network but the NFB around these stages means a very clean input with moderately high gain. All the tone/volume loss is in the middle, and then it needs fairly high gain to the output. The 1263 is unusual for a guitar amp in that all the gain happens in just two NFB sections. It might be very low coloration before over-load.
Also I am sure the guy who designed the Japan-made UniVoxes had the same problem as everybody here. He just wanted to try something different each time.