Calibration Problem  

Re: Calibration Problem

Hi Patrick,

can you please make a video of the printer while doing calibration? Smarphone is enough and use https://sendvid.com

Is the probe nicely aligned with the points? Try to get a detailed view of this in the video.

Thank you!

Re: Calibration Problem

Hi Josef,

thank you for that fast answer 🙂

Here is the video: https://sendvid.com/1auu5zqn

I glued a little piece of plastic on the extruder part where the endstop hits to aligne the points on the bed with the probe (the probe was too far left), you can see it shortly in the 2nd video below at 00:13. I also noticed that the calibration sometimes is done two times in a row. https://sendvid.com/s2han67d

Also after the print is finished the printer will sometimes get in calibration routine and would hit (if I'm not fast enough) the printed part.

Re: Calibration Problem

Calibration is done 2 times in a row for me as well. Is it rechecking something?

Re: Calibration Problem

If you have the calibration 2 times in a row, just download the 1.7.1 drivers and reinstall. The bundle also contains updated test prints so you can replace them on your SD card.

patrick.h do I understand correctly, that you got the printer to finish the calibration?

Re: Calibration Problem

No, the printer always finishes the calibration but the nozzleheight after that calibration is just wrong. Like I mentioned earlier, on the left hand side perfect and on the right hand side too high (or left crashing in bed and right perfect).
The only way to get good prints at the moment is to start the print, wait for the auto calibation and after that turning the right z-axis servo by hand a bit so that the right hand side is a bit lower.

Re: Calibration Problem

No, the printer always finishes the calibration but the nozzleheight after that calibration is just wrong. Like I mentioned earlier, on the left hand side perfect and on the right hand side too high (or left crashing in bed and right perfect).
The only way to get good prints at the moment is to start the print, wait for the auto calibation and after that turning the right z-axis servo by hand a bit so that the right hand side is a bit lower.
This is awesome I have spent almost a month trying to figure this out and you nailed it . I couldn't figure out why when I adjusted the lead screw it made no difference and it was because I was doing it before calibration. Once I made my adjustment after mesh leveling it started to print level . This is crazy man I am trying to get support to replace my bed I even offered to buy one. All I want is my printer to work ans do what it was advertised to do. The one reason a bought this mk2 was because of the bed , ohh sweet irony .

Re: Calibration Problem

I'm not quite sure if this is proper workaround. Moving only one side of Z axis by hand means you get X carriage off level so all prints are then skewed in Z axis. The X carriage rods should be precisely perpendicular to Z rods.
I was fighting the same problem initially after building new MK2 and found out that one of the stepper (by coincidence also right side) had very hard times while moving X-carriage up. The problem was in very tight hole for trapezoid nut on right side. The trapezoid nut was squeezed and didn't allow smooth moving. I solved that out by turning trapezoid nuts upside down (they are now not inserted anywhere, just secured by screws).
Does your printer demonstrate the same behaviour while set into power mode instead of silent mode? Are your Z rods precisely parallel? If you remove trapezoid nuts and try to move X carriage up and down by hand, does it move smoothly with almost no resistance all way up and down?

Re: Calibration Problem

David

Great thought but unfortunately, both quiet and high power modes use the same power setting for the Z motors...

There was an issue with firmware versions prior to the current latest 3.0.5 RC whereby the Z motors could momentarily turn off during the auto calibration stage and cause this problem.

Best thing James could do it to upgrade to 3.0.5 and run the calibration routines on that then try again.

Peter

Re: Calibration Problem

james, you are not alone and I am having an *identical* problem.

I too, have emailed support several times over a week ago and have not heard back. I've also posted in various threads here on the forum but so far have gotten dead ends.

If I watch the Z motors while the unit is printing, they are slowly moving when the X-carriage is moving from left to right so I can see that *a* correction is being done, but unfortunately it's not quite correct. I am beginning to guess that PINDA's induction calibration points are not so precisely placed, and thus so many people are having variations.

Several people on this forum have reported this issue so far.

We need a setting in the firmware to test for and define individual calibration point offsets, to supplement auto-calibration. Something where you do the paper-drag-test for each point, and then those settings are saved in EEPROM for the life of the bed.

Otherwise, this problem is going to continue.

Re: Calibration Problem

Babak

I totally agree with you.

However I would add that this problem is limited to a very few people. The vast majority of Mk2 users are printing without these issues.

Prusa Research are well aware of this issue and will be able to fix the problem once they have been able to replicate it - which is not necessarily an easy thing to do, given the limited extent of the problem and which could still be a build issue where the only "fault" is in the manuals and/or the user (and I hope you don't take that the wrong way, but it is a possibility, however unlikely).

Peter

Re: Calibration Problem

Thanks for the responds here I have made a video of my issue. I feel that the bed is the issue. I have tried so many things to get the printer to actually work ans this trick of moving the lead screw after mesh leveling worked.

Re: Calibration Problem

Babak

I totally agree with you.

However I would add that this problem is limited to a very few people. The vast majority of Mk2 users are printing without these issues.

Prusa Research are well aware of this issue and will be able to fix the problem once they have been able to replicate it - which is not necessarily an easy thing to do, given the limited extent of the problem and which could still be a build issue where the only "fault" is in the manuals and/or the user (and I hope you don't take that the wrong way, but it is a possibility, however unlikely).

Peter

Please watch my video to see that the printer is doino what it's suppose to ans that the calibration points in the bed are not correct

You can look at the filament lay out on the bed and see that the top section is square as can be. As far as the y axis being slid into the slots far enough. I have tried moving the frame in and out of the slots but when I do the mesh leveling accounts for it and still prints high on the left side .

Re: Calibration Problem

Hi James

Yes, I saw your video and agree that there is something wrong. Did you see my post regarding large Z Adjust values? I don't know if it helps or no, but I think it may be worthwhile trying my suggestions: http://shop.prusa3d.com/forum/original-prusa-i3-mk2-f23/tried-everything-can-t-get-bed-level--t1249-s10.html#p9878

If you can try it out please let me know if it helps.

Thanks

Peter

Re: Calibration Problem

Hey there guys,

I had the exact same problem as what you described, although on opposite sides (the left being too high and the right being too low). so here is what I did...

Sensor was already calibrated.

Some background: Typical hair pulling and head scratching on why one side was always about 200 microns off. I reached out to Prusa through the chat and it was suggested to either print in smaller chunks and glue together, or try loosening the screws a bit to give more. Well the first was not an option, so I tried loosening the screws. The issue remained.

Goofing with the left side of the board to 'lift' the bed:
I tried putting a washer under the screws to the bed, the issue remained
I had a razor blade as a shim, so I tried again, but the issue remained.

Sensor: it senses magnetic fields...

Well, why not:
Grab a magnet, stuck it to the razor blades on the rear/front screws of the board.
Bed leveling was kicked out saying that it was sensing the board too high.
So I moved the magnets further from the sensor: The calibration was way off and the left was even higher!!!!

AWESOME!

I moved the razor blades over to the right side and stuck the magnets on there.
Bed height calibration and...
...Good news it did even out the corners (well the right was really high), but the center front z axis was off by 100microns.
I figured the magnets were a bit much so I went with bucky ball magnets on the razor blades
This evened out the board some, but it was not perfect. I needed an even smaller influence on the sensor

As we all learned in 3rd grade science, if you move a magnet along a piece of metal you can make that piece a magnet.
So I took my big magnets back out and stroked the razor blades.
I doubled checked and the blades would hold the washer to them.
I ran the calibration and low and behold the bed was plus or minus 40microns.

Sure the centers to the edges are a bit closer than I would like, but I can now use the whole of the bed to print with.

Let me know if you have similar luck with Magnetic Shimming. <--- We should call it that.

If the peeps from PRUSA read this, I would strongly suggest giving us the ability to manually correct for magnetic calibration on the sensor to the bed. It appears that there is enough of a tolerance between the sensor and board manufacture to have a handful of us wishing for a bit more control over the board.

Here are my final Z calibration numbers using a 130 micron feeler gauge (.005 inch)

R .732-802-780
C .782-.775-.795
F .825-.777-.832

Rows are left to right
Columns rear, center, front.

Not perfect but it beats being 200 microns off and not sticking to the board.

Re: Calibration Problem

Where did you put the blades??? Between the nut and the bed or between the nut and frame??? I am having the same issue but the left is lower and right is higher.

Re: Calibration Problem

I had similar problems, and ended up disassembling the Y-axis assembly and reassembling. The power supply mount was very tight before, but after redoing everything, checking to make sure all four corners were straight and level after tightening and that the rails were not being pinched (roll them with your fingers - they should spin in the bracket fairly easily, but not have a gap) - the bracket that attaches to the power supply lined up and things are behaving a bit better.

Not to say the bed provided by Prusa is perfect (it definitely is not), and auto-levelling is perfect (which is definitely is not).

The last thing I would do would be to manually move a linear rod to correct a problem that is probably a bad y-axis assembly.