MK3S 3.7.0 RC1 firmware 7x7 bed leveling fails 50% of the time
I uploaded the 3.7.0 RC1 firmware to try out the new 7x7 mesh bed leveling. It's super fast. But on the last row generally towards the middle 3 points it seems to fail. Not every time. Maybe half the time it stops on the back row and gives me a "Some problem encountered. Z-leveling enforced..." Sometimes it works just fine.
I tried 3 taps? Senses? Whatever that is (the number of times it checks each spot). I tried 3 and I tried 5. And I get 50% of the time it works, 50% of the time I get that error. Then if I let it continue it wants to re-calibrate the Z axis.
I redid the calibration wizard 2x through the full process and it comes up ok.
If I set it to the 3x3 mesh bed leveling, it always works.
I know it's an RC, but any thoughts? Anyone know what I should check? Or set? Or see the same thing?
Re: MK3S 3.7.0 RC1 firmware 7x7 bed leveling fails 50% of the time
IIRC that warning appears if the meshbed leveling finds z-point differs too much from reference values position stored in EEPROM.
You may need to redo XYZ calibration to update the stored EEPROM bed Z reference values to better match current conditions.
Re: MK3S 3.7.0 RC1 firmware 7x7 bed leveling fails 50% of the time
I re-ran the entire calibration wizard, twice. Each time I did the XYZ calibration, right? I didn't run the XYZ calibration separate. But I assume going through the entire wizard it would be the same thing.
And after it errors it redoes the z calibration.
Made no difference.
Re: MK3S 3.7.0 RC1 firmware 7x7 bed leveling fails 50% of the time
OK. The max variance allowed for z is 1 mm.
Resync Z by hitting top stop was done during your XYZ recalibration. So, I don't that is the problem
Perhaps PINDA is a tad too high above nozzle.
I would set it between 0.7 to 1.0 mm above nozzle, redo XYZ calibration.
Does setting mesh to turn on magnet compensation help?
Re: MK3S 3.7.0 RC1 firmware 7x7 bed leveling fails 50% of the time
You can try this solution on the Czech forum ...
MK3 + Fusion 360 + SlicerPE
Re: MK3S 3.7.0 RC1 firmware 7x7 bed leveling fails 50% of the time
So they are playing around with not heating the heatbed during mesh.
If you're going to do that, I'd go the opposite direction and makes sure the PINDA and heatbed are at temperature during mesh.
Here is how I do it in custom start g-code
M83 ; extruder relative mode
G28 W ; MUST FIRST home all without mesh bed level or else risk nozzle crash!
G0 X70 Y70 Z0.2 F2000; PINDA warms faster near bed
M140 S[first_layer_bed_temperature] ; set bed temp
M104 S160 ; preheat extruder to no ooze temp to help warm PINDA
M190 S[first_layer_bed_temperature] ; wait for bed temp
M109 S160 ; wait for extruder temp
M860 S35 ; wait for PINDA temp. For most stable PINDA temp during mesh, set temp here 1-2 deg below your actual PINDA temp at end of mesh cal.
M104 S100 ; drop extruder to lower temp to minimize PINDA temp rise during mesh
G80 N7 R1 ; mesh bed leveling with extra points but this requires 3.6.0 firmware
G1 Y-3.0 F1000.0 ; go outside print area
G92 E0.0
G1 Z3 ; raise head safely clear of bed before heating extruder
M104 S[first_layer_temperature] ; we heat extruder AFTER mesh to minimize ooze
M109 S[first_layer_temperature] ; wait for extruder temp
G1 Z0.7 ; move slightly above bed for easy clean intro line
G1 X60.0 E9.0 F1000.0 ; intro line
G1 X100.0 E12.5 F1000.0 ; intro line
G1 E-0.8 F2100; anti-stringing retract and wipe
G1 X99.0 E0 F1000.0 ; -1mm intro line @ 0.00
G1 X110.0 E0 F1000.0 ; +10mm intro line @ 0.00
G1 Z0 ;lower head back to normal Z after intro line
G92 E0.0
M221 S{if layer_height==0.05}100{else}95{endif}
Re: MK3S 3.7.0 RC1 firmware 7x7 bed leveling fails 50% of the time
3.7.0-RC1 was failing more than 50% of the time for me, sometimes at the front of the bed, sometimes at the back, I even tried switching to 9 point calibration with no better results, Went back to 3.6.0 and no problem with bed leveling. I think it may be something to do with the speed the printer triies to measure each point, even on 9 point the three probes are very rapid, back on 3.6.0 with only 9 point available the probing is a slower, more delibrate operation. Perhaps the probe just can't keep up with the probing speed and gives three wildly different results sometimes which messes up the calibration?
Re: MK3S 3.7.0 RC1 firmware 7x7 bed leveling fails 50% of the time
Does setting mesh to turn on magnet compensation help?
That was with magnet compensation on. I didn't try it off. Maybe I should try turning it off?
Re: MK3S 3.7.0 RC1 firmware 7x7 bed leveling fails 50% of the time
You can try this solution on the Czech forum ...
It appears that the recommendation revolves around the temp of the bed during the mesh bed leveling. That's what guy.k2 posted about as well.
Luckily mine works fine in 3x3 leveling. I don't want to muck about with custom g-code to get 7x7. I'll try re-adjusting the pinda. And recalibrating. Or wait for 3.7 official. 🙂
Re: MK3S 3.7.0 RC1 firmware 7x7 bed leveling fails 50% of the time
I don't want to muck about with custom g-code to get 7x7.
It's pretty trival. You just paste in my startup g-code into the custom G-code printer settings once. Thereafter, it gets applied to all sliced objects.
Heatbed temperature does make a difference to the shape of the print bed. It is amazing that the standard startup manages to work while completely ignoring heatbed temp. The custom code, gets the bed to printing temperature before mesh. That makes the bed closer to the shape it is during printing.
Re: MK3S 3.7.0 RC1 firmware 7x7 bed leveling fails 50% of the time
I don't want to muck about with custom g-code to get 7x7.
It's pretty trival. You just paste in my startup g-code into the custom G-code printer settings once. Thereafter, it gets applied to all sliced objects.
Heatbed temperature does make a difference to the shape of the print bed. It is amazing that the standard startup manages to work while completely ignoring heatbed temp. The custom code, gets the bed to printing temperature before mesh. That makes the bed closer to the shape it is during printing.
Yes, you are absolutely right. It's not that big of a deal to do custom g-code. But this is a printer I use for work, I would have to configure and setup 4 or 5 different people's slicers and maintain those and make sure they all do the same as I'm doing.
Also it's an RC1 firmware, and Prusa making the end user muck about with custom g-codes for a new feature doesn't sound like them. If it's a firmware bug or a problem with the basic process, I'd be shocked if they release 7x7 officially and it requires the end user to put in custom g-code. So I was thinking I would rather wait than muck about with custom g-code for an RC1.
Unless others are having luck with the 7x7 and it's something wrong with my printer, or something physically wrong.
Re: MK3S 3.7.0 RC1 firmware 7x7 bed leveling fails 50% of the time
The print bed changes shape with temperature. 3 x 3 leveling gets away with not catching the change because it samples less area. When you go with 7 x 7 meshed, one pretty much HAS to do it with the bed at print temperature. Otherwise, there are going to be variances.
The only way to get around it in firmware would be to make G80 always preheat the bed before taking any measurements. I kind of doubt they would do that in firmware as a default behavior. Probably going to come down to whether one wants the entire print bed most usable (with 7 x 7) vs more limited area needs met with 3 x 3. Without preheating, it's not going to match print temp shape, but that's probably not a user priority if they elect to remain with just 3 x 3 sampling
I'm thinking they leave 3 x 3 as default, no heating.
Could conceivably add a preheat conditional with selecting higher sampling density, but it's really a lot easier to redefine the printer definitions and startup g-code.
Re: MK3S 3.7.0 RC1 firmware 7x7 bed leveling fails 50% of the time
I've got the same issue. I've done a few prints on the new firmware and it was fine, now I can't print anything because it goes into an infinite loop. It tries to 7x7 mesh level, fails on the back middle point, gives the "Some problem, enforcing Z calibration" message, does Z calibration, tries to mesh level again, fails with the same error - infinite loop.
I've gone back and done XYZ calibration from scratch, that all worked fine, but printing still fails at the same point, going into the infinite loop of 7x7 failing, starting calibration, etc.