Live Z huge variance over the bed, bad Pinda or ?
I have one of the first delivered MK3s, and from the beginning, my Live Z has been very inconsistent from print to print. Even using the same gcode file, same print settings, and same ambient temperature, and no physical or settings changes to the printer, I had to change the Live Z for almost every print to get a good 1st layer, ranging from -.620 to -.780 . But at least I was able to print...
Now, I can only get small items to print correctly in the center of the bed, and every print takes 5-6 tries to get a good 1st layer because of huge variances across the bed.
I should add this is my 3rd printer (first Prusa) and I never had this much trouble getting a good 1st layer with the other 2 printers. (One manual level, one BL Touch which is amazingly consistent.)
Here's what I've tried:
1. Firmware is b150 RC5
2. Reset to factory defaults (All data)
3. Selftest runs ok - All correct
4. Calibration runs ok - Your axes are perpendicular, congratulations!
5. Under Support, Skew is 0.12 and belt status is X: 240, Y: 270
6. Same results both before and after running Temperature Cal (PINDA temp calibration)
7. Same results if I clean the bed with acetone (only), 90% IPA (only), acetone followed by IPA, or IPA followed by acetone
8. Same results both before and after using fine steel wool to lightly scrub the PEI
When running the first layer calibration, I can get it ok on the left and bad on the right or vice versa, but no value for Live Z gets it to be correct over the entire print. (This picture was all done with the same Live Z value). Hopefully you can see the square printed at the end (lower left) is good but for the rest of the traces, some parts are way too high and some parts are way too low.
When printing anything larger than 1-2 cm, everything towards the back left corner is way too high, and everything towards the front right corner is too squished.
Did anyone have this problem and manage to solve it? Should I ask for a replacement PINDA?
Is there any way to get a readout of the mesh bed leveling results?
Thanks.
Re: Live Z huge variance over the bed, bad Pinda or ?
Try this...
http://help.prusa3d.com/l/en/article/3tSXh6nWJd-pinda-probe-testing
Re: Live Z huge variance over the bed, bad Pinda or ?
You could attempt a Temp. Calibration and see if it helps. What it does is calibrate the Pinda against temperature fluctuations as inductive sensors drift when the get hot.
The feature is hidden in the very buttom of the calibration menu. The test itself only takes about 15 minutes.
Re: Live Z huge variance over the bed, bad Pinda or ?
Thanks RotaryDane, but as I mentioned in the original post, I did try the PINDA temperature calibration (several times in fact).
Thanks george.b2, that's what I needed. I'm not sure what to make of this, but here are some different runs of G80/G81.
Taking for example the last point on the first line, it goes 0.203, 0.085, 0.168, 0.215. That's a huge variance, right? Especially from the 1st to the 2nd run, that's a change of 0.12mm, more than half of a layer height.
I'll ask Prusa support if this looks like a bad PINDA or what else could cause it.
Measured points:
0.14583 0.15565 0.16537 0.17500 0.18454 0.19398 0.20333
0.12574 0.13285 0.13930 0.14509 0.15023 0.15470 0.15852
0.11407 0.11526 0.11597 0.11620 0.11597 0.11526 0.11407
0.11083 0.10287 0.09537 0.08833 0.08176 0.07565 0.07000
0.11602 0.09569 0.07751 0.06148 0.04760 0.03587 0.02630
0.12963 0.09371 0.06239 0.03565 0.01350 -0.00406 -0.01704
0.15167 0.09694 0.05000 0.01083 -0.02056 -0.04417 -0.06000
Measured points:
0.06750 0.09370 0.11065 0.11833 0.11676 0.10593 0.08583
0.06620 0.08630 0.09855 0.10296 0.09954 0.08827 0.06917
0.07037 0.08290 0.08867 0.08769 0.07994 0.06543 0.04417
0.08000 0.08352 0.08102 0.07250 0.05796 0.03741 0.01083
0.09509 0.08815 0.07559 0.05741 0.03361 0.00420 -0.03083
0.11565 0.09679 0.07238 0.04241 0.00688 -0.03420 -0.08083
0.14167 0.10944 0.07139 0.02750 -0.02222 -0.07778 -0.13917
Measured points:
0.13167 0.14981 0.16315 0.17167 0.17537 0.17426 0.16833
0.12528 0.14043 0.15000 0.15398 0.15238 0.14519 0.13241
0.12278 0.13225 0.13599 0.13398 0.12623 0.11275 0.09352
0.12417 0.12528 0.12111 0.11167 0.09694 0.07694 0.05167
0.12944 0.11951 0.10537 0.08704 0.06451 0.03778 0.00685
0.13861 0.11494 0.08877 0.06009 0.02892 -0.00475 -0.04093
0.15167 0.11157 0.07130 0.03083 -0.00981 -0.05065 -0.09167
Measured points:
0.11500 0.12333 0.13500 0.15000 0.16833 0.19000 0.21500
0.09083 0.09340 0.09941 0.10889 0.12182 0.13821 0.15806
0.07833 0.07281 0.07133 0.07389 0.08049 0.09114 0.10583
0.07750 0.06157 0.05074 0.04500 0.04435 0.04880 0.05833
0.08833 0.05969 0.03765 0.02222 0.01340 0.01117 0.01556
0.11083 0.06716 0.03207 0.00556 -0.01238 -0.02173 -0.02250
0.14500 0.08398 0.03398 -0.00500 -0.03296 -0.04991 -0.05583
Re: Live Z huge variance over the bed, bad Pinda or ?
took me quite awhile with G80 as well. There's another post about about modifying the code even further to use all 9 leveling points, but I kinda expected PINDA to be more accurate. The time it took to do all of that I could have just strapped on a dial caliper and turned a few screws :/
Re: Live Z huge variance over the bed, bad Pinda or ?
Same issue here!
I received a replacement pinda which I had not the time to install yet. I hope this will solve the issue.
My g80, g81 values are jumpy too.
Re: Live Z huge variance over the bed, bad Pinda or ?
I have the EXACT same issue.
Replaced both the Pinda and einsy itself.
Issues persist and stI'll exist.
Re: Live Z huge variance over the bed, bad Pinda or ?
Was the bed cold during all of those tests or cooling from first to last? That does seem to be a huge static variance if the bed temp was the same. Was this with the steel sheet on or without? Just wondering for comparative purposes. Any details would be appreciated. It may be interesting to try it with both temp compensation on and off as well.
Re: Live Z huge variance over the bed, bad Pinda or ?
Hej jettoblack!
So you did resolve this issue with a fixed firmware someone proposed to prusa as a pull request?
Could you elaborate on your experiences with that fixed firmware?
Do you still have to adjust live z for every print?
Re: Live Z huge variance over the bed, bad Pinda or ?
We're referring to these threads on the firmware github:
https://github.com/prusa3d/Prusa-Firmware/issues/421
https://github.com/prusa3d/Prusa-Firmware/pull/439
After applying TZB's firmware with this pull request merged in, I'm still seeing some Live Z inconsistency from one print to the next, but at least the mesh bed leveling seems to be more consistent over the surface of the bed now, so I actually got some successful prints.
So this problem does appear to be related to PINDA temperature calibration in firmware.
What puzzles me is that I have another 3d printer with an inductive sensor without temperature calibration and it doesn't have any inconsistency issues at all. I don't understand why the temperature calibration, which was supposed to make the sensor more consistent, is making it much less consistent for some users.
Re: Live Z huge variance over the bed, bad Pinda or ?
....
Measured points:
0.11500 0.12333 0.13500 0.15000 0.16833 0.19000 0.21500
0.09083 0.09340 0.09941 0.10889 0.12182 0.13821 0.15806
0.07833 0.07281 0.07133 0.07389 0.08049 0.09114 0.10583
0.07750 0.06157 0.05074 0.04500 0.04435 0.04880 0.05833
0.08833 0.05969 0.03765 0.02222 0.01340 0.01117 0.01556
0.11083 0.06716 0.03207 0.00556 -0.01238 -0.02173 -0.02250
0.14500 0.08398 0.03398 -0.00500 -0.03296 -0.04991 -0.05583
doesn't look so bad - my printer works fine and has similar reports. Although this isn't directly related to the trouble you've reported you could consider trying to level up the right side of your machine. The left side looks pretty flat (.115 to .145) but the right side varies from .215 to negative .0558
I think this can be addressed by raising the back end of the right-hand Y-rod just a little bit. You could do a test leaving a small scrap of notebook paper underneath one corner of the spring steel sheet during G80 mesh level and observe the difference. If you like the result you could loosen the screws and lift the block a little bit and re-tighten. This might be possible or it might not, depending where you placed it the first time.
I am working on *tweaks* and considering the use of a shim system or simply printing a different holder with a slight offset to move the rod up.
Again - I don't think this is directly related to what you are reporting but if you are trying to get things perfect you'd want the left and right Y rods to be parallel to each other.
For getting the material to stick - are you cleaning with alcohol between each print? I always do this and also have noticed sometimes a very small adjustment to live-Z can make a difference between bad and good.
Re: Live Z huge variance over the bed, bad Pinda or ?
On my MK2s on the latest firmware for that printer i found that adjusting Z during a print dosen't stick between prints BUT... the number on the LCD SAVES but every print i adjust by the same amount to get to where i previously adjusted it. Once i got down to 0.0mm then it still started printing too low, same place as before like there is another value calculated and stored but not updated.... only doing the offset calibration from menu reset it properly (not even Z calibration, just the zig zag pattern). Also if i calibrate XYZ and ignore the "your z offset hasn't been set" (since it was set sorta) it does the opposite. once i was apparently -1mm after adjusting it for every print and it still printed at the same layer height, that should be jammed nearly into the bed... but it was too high... lol
firmware is broke some place....
Re: Live Z huge variance over the bed, bad Pinda or ?
For the "Measured points" grid, is the first line referring to the front of the printer (Y = 0), or the back ( Y = max)?
On recent Marlin FW, the bed mesh is reported with the first line (top line) of the report being at Y = 0, and the last line (bottom line) of the report is at Y = max. This is backwards compared to viewing the printer from top-down where Y = 0 is at the bottom, and Y = max is at the top.
For Prusa FW, is it the same, or reversed?
I suppose I will find out next time I try to adjust it. 🙂
Re: Live Z huge variance over the bed, bad Pinda or ?
I updated to 3.1.1 RC5 B150 the other day to hopefully fix issues with Octoprint (which so far it seems to), but now I'm having a similar issue. Every print I do it seems like the Live Z is completely off. The screen still shows the value it used to be, but every print it seems wrong and I have to adjust it more. This issue wasn't happening before I updated to this firmware. I'm at work currently, but I'll attempt some of the suggestions here when I can.
Re: Live Z huge variance over the bed, bad Pinda or ?
For the "Measured points" grid, is the first line referring to the front of the printer (Y = 0), or the back ( Y = max)?
On recent Marlin FW, the bed mesh is reported with the first line (top line) of the report being at Y = 0, and the last line (bottom line) of the report is at Y = max. This is backwards compared to viewing the printer from top-down where Y = 0 is at the bottom, and Y = max is at the top.
For Prusa FW, is it the same, or reversed?
I suppose I will find out next time I try to adjust it. 🙂
I really think it is like this, first numbers to print out are the back
here are my numbers right now on a print bed that was warmed but is now down to 32C
back
0.14167 0.18806 0.21944 0.23583 0.23722 0.22361 0.19500
0.21157 0.26984 0.30644 0.32139 0.31468 0.28632 0.23630
0.25157 0.31644 0.35564 0.36917 0.35703 0.31922 0.25574
0.26167 0.32787 0.36704 0.37917 0.36426 0.32231 0.25333
0.24185 0.30413 0.34064 0.35139 0.33638 0.29561 0.22907
0.19213 0.24521 0.27644 0.28583 0.27338 0.23909 0.18296
0.11250 0.15111 0.17444 0.18250 0.17528 0.15278 0.11500, front right
front, left
AND with a small 1cm square piece of paper stuck under the right front corner (near the encoder knob)
0.14417 0.18528 0.21333 0.22833 0.23028 0.21917 0.19500
0.21306 0.27259 0.30917 0.32278 0.31343 0.28111 0.22583
0.25528 0.32407 0.36472 0.37722 0.36157 0.31778 0.24583
0.27083 0.33972 0.38000 0.39167 0.37472 0.32917 0.25500
0.25972 0.31954 0.35500 0.36611 0.35287 0.31528 0.25333
0.22194 0.26352 0.28972 0.30056 0.29602 0.27611 0.24083
0.15750 0.17167 0.18417 0.19500 0.20417 0.21167 0.21750
of course when you think about raising or lowering the rod you have to think backward because the bearing is at the back of the machine when you are at Y0 : )
Re: Live Z huge variance over the bed, bad Pinda or ?
Yes themzlab you are right about that.
I was able to play with the Y axis rods to get them a little more parallel than before. This shows an improvement in the bed level mesh points being closer to each other. However I still have the issue with Live Z needing to be tweaked on every print. I still think this is a PINDA temperature calibration issue.
Re: Live Z huge variance over the bed, bad Pinda or ?
Try turning the temperature adjustment off and running a few tests back to back. You can in the alternative calibrate the Pinda via the menu and not use the built-in temperature table. There are time I desire to go back to my (borrowed) simple method of raising the probe to 60 or 80 during the heating cycle via script, and then making simple adjustments based on bed temp IE -~.005mm for 100C bed (ABS) or no change for PLA basing it off of the PLA calculation.
Re: Live Z huge variance over the bed, bad Pinda or ?
I now had the time to swap my PINDA probe for the replacement probe that Prusa support kindly provided.
The sad thing is that it did not improve anything about my situation.
Still live z is wrong with every print. Still there is a huge temperature dependence - the warmer the printer is, the higher I have to set live z. I cannot get any decent print from this machine because the important first layer is always wrong. No chance of getting it dialed in.
I tried both with and without temperature calibration.
By the way: I compared the output from G80, G81 with different ambient temperatures. Are they supposed to be somehow the same when lets say comparing between 16C ambient and 20C ambient? Because my values are stable, when I do 5 measurements in quick succession - but quite different when the ambient temperature changes. I guess that is why I am facing this problem?
cold, 16C hotend and bed:
Measured points:
0.52250 0.42306 0.34556 0.29000 0.25639 0.24472 0.25500
0.52667 0.43127 0.35059 0.28463 0.23340 0.19688 0.17509
0.50389 0.40840 0.32355 0.24935 0.18580 0.13290 0.09065
0.45417 0.35444 0.26444 0.18417 0.11361 0.05278 0.00167
0.37750 0.26941 0.17327 0.08907 0.01682 -0.04349 -0.09185
0.27389 0.15330 0.05003 -0.03593 -0.10457 -0.15590 -0.18991
0.14333 0.00611 -0.10528 -0.19083 -0.25056 -0.28444 -0.29250
warmer, 20C hotend and bed
Measured points:
0.40750 0.35380 0.30296 0.25500 0.20991 0.16769 0.12833
0.43278 0.37001 0.30828 0.24759 0.18794 0.12933 0.07176
0.43028 0.35794 0.28612 0.21481 0.14402 0.07374 0.00398
0.40000 0.31759 0.23648 0.15667 0.07815 0.00093 -0.07500
0.34194 0.24896 0.15936 0.07315 -0.00968 -0.08913 -0.16519
0.25611 0.15205 0.05476 -0.03574 -0.11946 -0.19641 -0.26657
0.14250 0.02685 -0.07731 -0.17000 -0.25120 -0.32093 -0.37917
Has anyone any ideas?
Re: Live Z huge variance over the bed, bad Pinda or ?
I have the exact same problem. Send an email with a lot of pictures to Prusa yesterday, hope they will react.
Today I tried to print the arms of the T-Rex that was posted yesterday. Have a look at the pictures, even on 5 cm the skirt on the left is completely different from the skirt on the right. Print keeps failing cause the nozzle catches on the print and it separates from the skirt and always on the same height. Dunno if that is a completely different problem or if it is related to the first layer.