A little reverse engineering would likely expose whatever they're up to. Watching current in to the laser would tell if they're actually controlling power going to it in with anything other than "on/off" or not. Limiting current in to the laser would be relatively easy, so a control that results in "less than 100%" output to the laser is relatively simple.
There's a few unknowns (but again, with their architecture they can't really *hide* anything since it ultimately goes to the laser and galvos) as to how it actually operates-- I would suspect that it would all come down to "energy per unit area". In theory you could:
* have a fixed laser output power and a fixed drawing speed and draw the same pattern multiple times for longer curing
* have a variable laser output and fixed drawing speed and draw the pattern the same number of times but with different power settings for different thicknesses
* have a fixed laser power and draw the same pattern, but do it with a variable drawing speed/slew rate (faster or slower beam movement) resulting in more or less energy going in
* any other combination or permutation of the above
A hour with an oscilloscope would answer those questions. I'm probably about out of warranty before too many more weeks here.
I suppose another approach would be to stick a USB analyzer on the bus and download the smallest possible job at each of the different material/thickness settings and see what differs and if there's some sort of 'cure profile' data that goes by in the clear.