I returned my MINI2 and got a TAZ 6 from craigslist… it was all sorts of busted but fixing it up was an awesome project all on its own and no regrets, it prints great now. Much larger bed allows one to not print adjacent to the corners and it actually doesn’t have the same fundamental problem as the belt-driven mini2 because the Z axis movement is implemented using the tried-and-true leadscrew way. Leadscrews give more z-wobble, but there’s solutions to this, google “Piercet anti-wobble” if you’d like to know more.
p.s. it’s still not perfect, but I just got a BLTouch system which will probe the bed itself and not those stupid washers (which combined with an impossible-to-fully clean nozzle fail at least half the time), should get consistent and accurate leveling.
Add “M425 F1 Z” after “M420 S1” to resolve this problem. No more left side squishing. Without this command it will measure, but not apply the backlash on Z that was measured with G29. I confirmed this by using OctoPrint to issue a G29 and then a M425 only to have it say backlash was inactive.
The left side weighs more and needs a little extra movement in the stepper to match the right, which is why it’s squished on the left. When commanded to move up, say 0.2 mm, the left side only moves about 3/4 that, and the backlash will add the extra steps needed to make it move a full 0.2 mm. I mean, I guess that’s what it is doing.
I tried just about everything to fix this, even loosening/tightening the z axis belts to see if that could be it. And it’s NOT easy tensioning any belts on the Mini 2. Shimming between the washer and the PEI glass plate sorta, maybe did a little to alleviate it but there’s only so much you can do since there’s no real give in that area. Which makes no sense why it would solve it, since I literally played musical chairs with all the bed mounting components, randomizing their placement several times and each time the left side was still being squished. I measured each washer - each was the same thickness, and since it’s precisely the washer’s thickness that is the defacto “z offset” getting the squish on the left can only be caused by something in the SOFTWARE.
I ran through an entire 1 kg spool running 100x100 test squares trying to fix this problem.
What I have found after adding BLTouch to my Mini2, was that the rightmost 20mm or so exhibited readings consistent with X axis twist. I believe this is due to the cable chain not being perfectly aligned to the carriage, so as the bend gets sharper, it exerts greater push on the top of the carriage, making the carriage tip slightly forward, but enough that the readings were off with a BLTouch mounted about 30mm ahead of the nozzle. When measuring with the nozzle on the washers, this was compensated for, but with BLtouch, the errors were amplified.
It’s always the first probe point that seems to probe a little farther down than it should, I think it’s the momentum of the gantry being at it’s heaviest point right there. I’m going to get the firmware and change the probing speed to see if that prevents the over-probing at that spot.
// X and Y axis travel speed (mm/min) between probes #define XY_PROBE_FEEDRATE 6000 // ← changed
// Feedrate (mm/min) for the first approach when double-probing (MULTIPLE_PROBING == 2) #define Z_PROBE_FEEDRATE_FAST 480 // ← changed
// Feedrate (mm/min) for the “accurate” probe of each point #define Z_PROBE_FEEDRATE_SLOW 60 // ← changed
And it’s no longer overshooting the first washer. It slowly goes down and touches without any deflection of any kind. First layer is printing now and it’s so far even on all sides on the 100x100x1 object I have been printing.
Used the master branch at GitLab as that is the 2.0.9.0.17 version already on my machine prior.
Yeah that’s unfortunate. To go through that effort and only to have it make it worse. I wonder if there is a lighter stepper motor that can be used for the x axis instead of the honking Moons motor it’s using now.
That motor is sized based on the standard toolhead, which also has a pretty large Moons motor on it. If you went with a much smaller orbiter, sherpa, or other toolhead, you could get away with a smaller x-axis motor.
