Good question… I admit I never cared about what the maximum possible speed is since my focus is very much on mechanical stability - it is well known that FDM parts get more brittle if you print them very fast. And it has literally been decades since I had to do such a calculation for the last time.
What I can give you however is the following: after an intensive run-in period where I literally had to regrease the linear slides every two or three prints (!) since they started running less smoothly, my printer runs perfectly fine so far with the original black linear slides using the following values:
Firmware settings (Caution! These are taken from Reprapfirmware - it is up to you to translate them back to Marlin settings!):
M201 X1000 U1000 Y1000 (max. acceleration in mm/s²)
M204 P900 T900 (max. combined acceleration in mm/s²)
M566 X342 U342 Y342 (max. jerk in mm/min)
Prusaslicer settings:
max accelerations for X and Y 1000mm/s²
max acceleration for extruding / retracting / travel : 900 mm/s²
max jerk for X and Y 5.7 mm/s
max speed for X and Y 350mm/s - but: the max speed I actually print with is 200mm/s.
Constraints of the above:
- First, you need to acknowledge that (unless the Marlin version Snapmaker uses has a bug there) a single acceleration vector in a two-axis system (X and Y) can physically never be larger than the combined total acceleration vector defined in M204. This means that the 10000 mm/s² acceleration values Snapmaker define in their M201 setting and state on their website is nonsense anyway since the combined acceleration defined in their firmware is just 1000mm/s²… or was at least when I threw out the Marlin controller of the J1 to get a decent web interface and full native support for open slicers.
Thus, any acceleration setting in the slicer above 1000 should be limited to 1000mm/s² already today. - Second, any possible max acceleration values highly depend on your print head - its weight and centre of mass, the lever arm the printing nozzle has, the forces it creates when rumbling over a blob of a printed material (which again depends on the material, the print speed and the shape of the nozzle you use) and such.
- Third, I have no clue if the mentioned run-in has any effect by cold-hardening the linear rails or the roller balls…
For comparison: my J1 print heads still have the heavy twin 5015 part cooling fans since I did not find the time to replace those with the better variants that exist by now. On the other hand, I have the self-made hotend upgrade i3sven, MrBean and me had cooked out, which reduces the lever arm of the nozzle and therefore lateral forces when printing. And I have set Prusaslicer to avoid crossing curled perimeters which minimises rumbling into material blobs.