Klipper Losing Position

[concentric_killa]

After building my 2.4 in the 350 config, I have successfully printed various things without issue. 

My issue now, is that occasionally the X-Y position shifts with a different layer. It doesn't consistently happen either. I can print various calibration tests from Cura, and the issue doesn't occur. Sometimes when the printer goes to a new layer, the X-Y position is completely wrong.

Does anyone know of anything to test/ tighten/ calibrate? I used to be able to leave the printer un-attended, but now I have to babysit this thing since it might start printing in 'space' . 

 

 

[Simon2.4]

Does it shift a few mm or a few cm?

do you hear a sound when it happens? ->collision with the part, belt skipping,

are your belts properly tensioned and equal on both sides.

Just throwing ideas here, I had that happen once where the printer just shifted by a few cm and kept printing. It was on a platter full of small parts so it all bridged nicely and kept going mostly. I'm fairly sure I heard a collision noise when it happened and had a loose part on the edge of the platter.

[concentric_killa]

It was at least 1 or 2 cm. How do you properly tension the belts? On the voron documentation, they mention some sort of frequency monitor. Are we talking about a guitar tuner, or can a simple caliper measurement be taken with the gantry in the center of the bed? 

I initially re-tensioned them the other day and that is when these problems started. I thought they weren't too tight, but then when I tried to QGL, it sounded like the motors were skipping. I immediately hit the e-stop on the web interface. I then loosened the tensioners. I then did the homing routine and eventually got it to successfully QGL without the motors 'skipping'. I then ran a 1st layer calibration from thingiverse, and that printed successfully. 

I am in the middle of a 3hr print and it has randomly lost position 3 times. It is strange. Some layers, it transitions just fine. Other layers, it just takes off in a random direction and starts printing. 

 

[concentric_killa]

I have, even combed through my current print Gcode to see if cura added something. I have been adjusting setting for printing PLA (retraction, etc.)  From the looks of it, nothing seems strange. 

[Simon2.4]

It's better to adjust belts to be closer to 'too loose' than the opposite. It's a bit of an issue on the voron because of the way its assembled its really hard to vibrate the belts without them rubbing somewhere. So getting a tone is difficult for a spectrum analyser. I've been thinking about designing a belt tension meter that's printable for a while now; I think It's time I get to it.. This is going to be tough but I have good ideas how to make/calibrate one.

[ajscheid]

Check your wiring if tensioning of belts doesn’t help; sounds like a cable break to me

[concentric_killa]

4 minutes ago, ajscheid said:

Check your wiring if tensioning of belts doesn’t help; sounds like a cable break to me

That does sound plausible. I have already replaced my probe wiring with silicone. The probe would trigger at certain parts of the bed and then other parts it wouldn't trigger.

Ill keep an eye on that. I might even try doing some test prints at a slow rate. All my prints have been at 0.2. The printer books. Sometimes I think it is a little too fast.  

[claudermilk]

7 hours ago, Simon2.4 said:

It's better to adjust belts to be closer to 'too loose' than the opposite. It's a bit of an issue on the voron because of the way its assembled its really hard to vibrate the belts without them rubbing somewhere. So getting a tone is difficult for a spectrum analyser. I've been thinking about designing a belt tension meter that's printable for a while now; I think It's time I get to it.. This is going to be tough but I have good ideas how to make/calibrate one.

You mean something like this? https://www.prusaprinters.org/prints/46639-tension-meter-for-the-gt2-belts-of-i3-mk3s-or-prus

Edited February 17, 2022 by claudermilk
typo

[compman2]

I have had an issue with layer shift and it ended up being that my run current on x and y were too low.  I used a default configuration file that were for different stepper motors.  This could be an issue.  Follow the formula:

(max amps stepper motor * .707) * .8 = max running current

You probably want to start one or two tenths below the formula.  If there are still skipped steps raise a tenth.  Feel the steppers right after a print completes to see if you stepper motors are hot.  If you can't leave you finger on the stepper for a couple seconds they are too hot and run current should be lowered.

Also the TMC2209 drivers have a max current of 1.2amps so I wouldn't run them over 1amp.  The higher the current the more heat the stepper drivers generate.  My motors were 2amp max current.  I am running them at .9amps could go lower probably but they are still only warm to the touch.  I was having problems with layer shift when set to .6 amps and .7 amps.

 

 

Edited February 17, 2022 by compman2

[concentric_killa]

22 minutes ago, compman2 said:

I have had an issue with layer shift and it ended up being that my run current on x and y were too low.  I used a default configuration file that were for different stepper motors.  This could be an issue.  Follow the formula:

(max amps stepper motor * .707) * .8 = max running current

You probably want to start one or two tenths below the formula.  If there are still skipped steps raise a tenth.  Feel the steppers right after a print completes to see if you stepper motors are hot.  If you can't leave you finger on the stepper for a couple seconds they are too hot and run current should be lowered.

Also the TMC2209 drivers have a max current of 1.2amps so I wouldn't run them over 1amp.  The higher the current the more heat the stepper drivers generate.  My motors were 2amp max current.  I am running them at .9amps could go lower probably but they are still only warm to the touch.  I was having problems with layer shift when set to .6 amps and .7 amps.

Thanks for the info. I have 2A steppers. In my config file, my motors are set to 0.8A. I don't remember them being too warm after a 3 hr print. I guess i could run them at .9 and see how they are. 

I will see if I can run a print position test to see if I can get a snapshot in a klipper log on when this happens. I was thinking that maybe there is a buffer error in klipper and the Octopus board. Just grabbing at straws here.  Up to this point, I haven't had any issues. I am leaning towards an idler or loose belt. But the fact that it isn't consistent makes me think it is software. 

[compman2]

It may not be a current issue but it was magic for me.  I only had it occasionally and stepping up the current was the trick.  Good luck running this down.

[Simon2.4]

3 hours ago, claudermilk said:

You mean something like this? https://www.prusaprinters.org/prints/46639-tension-meter-for-the-gt2-belts-of-i3-mk3s-or-prus

yes exactly.. thanks!

[concentric_killa]

Ok, 

So I ended up loosening my XY belts and then I downloaded 3 different frequency apps on my phone and tightened them up (110hz ~2lb). I then re-did my z-end stop calibration with the bed at 100C and Head at 240C. 

I have managed to print two jobs successfully without the head going off into a random direction. 

 

 

[Simon2.4]

33 minutes ago, concentric_killa said:

3 different frequency apps

Can you point to the one that worked for you. I'm glad to hear you've (probably) resolved that issued; print shifts are a deal breaker. So were the belts too loose or too tight?