1st layer problems - in depth look at software PINDA problems (and solutions!)  

Re: 1st layer problems - in depth look at software PINDA problems (and solutions!)

any news update?

Not from me!

Re: 1st layer problems - in depth look at software PINDA problems (and solutions!)

I still am not sure that I understand why the sign is correct
Like I said, the purpose of the offset variable in out FW is to keep track of how higher above the bed the probe triggers with respect to some reference temperature, 35C in the official FW (and I may change it back to 35C too, it is probably insignificant). When you do your calibration, you keep track of how much you must increase the live adjust z value in order to get it to the same height it was before. These two numbers are equal in magnitude, but have opposite signs. If "our" offset is e.g. positive (PINDA triggers higher), you must compensate by moving it lower - so "your" offset would be negative.
If I put it in another way - in your very first post where you describe your calibration there is a step where you subtract the two live z offset values. How did you decide which of the two numbers is the one that gets subtracted? Do it the other way and you should be happy with the sign that we have there:-) This is the sense in which our sign is correct - it corresponds to what is saved in offset variable. Unless there is something I keep missing, it should be like that.

No, no, I disabled the line were the offset is set from 0 to the return value of the temperature compensation function.
Allright, don't mind the objection then. You wrote "I commented out the buggy line of code", which gave different impression.

So I used Lukas' firmware and used the new temperature calibration method three times.
Thank you very much for going through it, I really appreciate it and it was helpful. I have thoroughly examined the results and compared them to the values you report to actually work. We are doing the same thing on several printers that we have here. What I can say rightaway is that PINDA parameters can be very different piece to piece, which may be part of the problem. However, I will not speculate about all the possibilities and ideas now, not before we are more sure that we understand what's happening. I will PM you with some additional info so I don't clutter the discussion here.

hopefully calibration will be automatically run with the new version and the current z-offset will be set to 0
Yes, if we really change the meaning of values in the EEPROM, we will do what we can in order to prevent any misinterpretations after fw is upgraded or downgraded. However, we have not finally decided that the values in EEPROM will really change, maybe the problem I describe will not be present in the new official fw after all. That would be the cleaniest solution.

'youtube' version of the firmware
Trust me I am very sorry for the issues that you (and others here) have. But I can assure you there is no "youtube" version of the fw. Although in this thread it looks like none of printers works, there are many people who don't experience the issue, because the calibration works fine for them. We now thoroughly test several randomly chosen printers, some of them you don't have to touch at all. I'm in no way saying you should be happy about the fact that someone else has no issues - your printer must too work fine. But if you suggest we provide better printers to the ones reviewing them, we are not. Hopefully we'll make to solve the issue soon.

also while lukas is here, there are still issues with the extruder with ringing and moire patterns
Sorry, I'm not actually a fw developer and therefore cannot tell you about current status of fw. All I know is the issue is known and has been worked on.

Re: 1st layer problems - in depth look at software PINDA problems (and solutions!)

Lukas, many thanks for the update and the efforts underpinning it!

What I can say rightaway is that PINDA parameters can be very different piece to piece, which may be part of the problem. However, I will not speculate about all the possibilities and ideas now, not before we are more sure that we understand what's happening. I will PM you with some additional info so I don't clutter the discussion here.

It's helpful and encouraging to know that you have a particular focus. Especially that it is (currently at least) the PINDA and potential performance variation piece to piece. I've been dreaming about some kind of standard calibration technique that each user can deploy to get base values for a correction where the variability between ambient conditions can be eliminated or at least mitigated to some degree. So I wish you'd "clutter the discussion" with your additional thoughts and experiences as your testing and evaluation opportunities are clearly wider than ours.

My kit is to be delivered today so I hope soon to get some real-world experience of this issue...

Re: 1st layer problems - in depth look at software PINDA problems (and solutions!)

Since my attempt to physically level the bed did not entirely take care of the left hand side bed squish all I can say is that the bed leveling does not function correctly. The bed was now better flat leveled but not parallel with the X-axis. The z-axis movement had noticeably increased to compensate for the non-parallel surface to the X-axis. The first layer was always still thinner on the right side of the bed and could not be tuned out.

Something in the system somehow tells the mess leveling to always be CLOSER on the right side. The mesh fine tuning does not help and the adjustment directions are jacked besides. Hope Prusa gets this stuff fixed soon.

Re: 1st layer problems - in depth look at software PINDA problems (and solutions!)

For me the PINDA gives even more issues when printing ABS with bed set to 100C.
While PLA@60C is mostly spot-on perfect layer, ABS needs different live-z, AND I get unleveled bed with worse adhesion at the back.

What I think is happening is that while performing mesh calibration from front to back of the bed the PINDA's temp goes up because of 100C bed and because of incorrect temp response the values start to shift at the end of mesh calibration, resulting in higher Z at the back.

Grey PINDA here, no derbis between the heatbed and steel sheet

Re: 1st layer problems - in depth look at software PINDA problems (and solutions!)

I was browsing the github for the firmware yesterday and I saw they have re-vamped the PINDA temp, and method for initial calibration. They may have done more.

I tried a unoffical build of it, but it was VERY unhappy. (massive stepper motor whine) and I would rather not cause anything to break on it!

But the 12 minute initial calibration, rather than lines back and forth, it's little circle motions? (I didn't print anything with it, so I don't know)

Re: 1st layer problems - in depth look at software PINDA problems (and solutions!)

Hi Lukas

Thanks again for the follow up posting and the PM.

I just want to write publicly that if the sign is correct in the firmware then - like you state - the offsets for my machine need to be defined negative.

The firmware in many places treats the offset as an uint16_t. Examples are the dcode 8 or of course the setup().
So that is the bug, the firmware must accept negative offsets as well.

Still I am a huge fan of my manual, print based temperature calibration. To have that build into the firmware comparable to the first layer calibration (which should be run with 35C pinda temp and thus is not good currently - by the way;-)).

The temperature calibration should disable the temp compensation and then guide the user through the pinda Temperatur Levels, each time printing a nice big First Layer that the user then manually live adjusts. The firmware then could calculate the correct offsets. Of course there should be a menu to manually set the offsets.

And then, I would propose to increase the range from 25 to 60 degrees. Because the pinda is colder than 35C for the first print after a longer pause.

Re: 1st layer problems - in depth look at software PINDA problems (and solutions!)

Since my attempt to physically level the bed did not entirely take care of the left hand side bed squish all I can say is that the bed leveling does not function correctly. The bed was now better flat leveled but not parallel with the X-axis. The z-axis movement had noticeably increased to compensate for the non-parallel surface to the X-axis. The first layer was always still thinner on the right side of the bed and could not be tuned out.

Something in the system somehow tells the mess leveling to always be CLOSER on the right side. The mesh fine tuning does not help and the adjustment directions are jacked besides. Hope Prusa gets this stuff fixed soon.

I have noticed the exact same thing!

My right side is also a lot closer to the nozzle than the left. To top it off, I have a low spot on my bed (a warp of sorts) just to the right of the middle mounting bolt on the left side (verified using a straight edge). I also have a similar low spot at the middle mounting bolt at the front side. I have tried all manner of Live Z tuning (taking care to preheat my PINDA each time), I have tried Bed Level Correction, and I have even gone so far as to try shimming the bed. Nothing seems to be able to help.

I'm most disappointed with Bed Level Correction. Not only does it provide a very tiny adjustment window, when applying a bed level offset, you have to apply it along an entire side even if that entire side isn't out of spec. Not only that, but it's counter intuitive if you have to apply bed level correction on more than one side at a time. It works against you in some cases. I would very much prefer a more granular approach to bed level correction.

So I'm stuck in a similar situation. I've gotten it down to a semi-OK first layer, but the right side is still rather close while the left side of my bed sometimes has gaps between infill lines.

Re: 1st layer problems - in depth look at software PINDA problems (and solutions!)

I experienced exactly the same problem with the first layer calibration as described by Stahlfabrik.

The DIY MK3 Kit is my first 3D printer, so I discover the world of 3D printing, when I first used the printer after assembly with no failure, and first layer calibration done, I printed only little objects in PLA to test and acquire experience, It works good even if I had sometimes to restart the print to achieve a good first layer, I thought it was due to my lack of experience until I decide to change for PETG and to print larger objects.

With PETG I experienced more first layer failures, can’t achieve uniform first layer for larger objects, and detect inconsistent layer calibration, calibrating hundred times Z axis or XYZ axis changed nothing.

I decided to print a PETG test rectangle of 150 x 30 x 1,5mm to try fine calibrate the first layer. when I start printing I always preheat first for 15mn but the extruder far of the printbed because I raise it after a print to clean the nozzle.
During the first attempt I tuned the gap to -0.590 to achieve an almost good result at the end of the test rectangle, then I started another print to validate my tunings, but at half the print the nozzle came to close to the printbed and the extruder begins to produce noise and no more material was extruded, then I had to return to -0.570/-0.560 to achieve the print.

There is clearly a weird relation between the temperature and the PINDA calibration, the gap between the nozzle and the bed decreases as the temperature raise.

I have done my test with PRUSA PETG, default slic3r PE configuration for 0.20mm layer and the black cable PINDA.

I hope PRUSA staff will solve this problem soon, for the momment I am not ready to try custom firmware.

Re: 1st layer problems - in depth look at software PINDA problems (and solutions!)

To top it off, I have a low spot on my bed (a warp of sorts) just to the right of the middle mounting bolt on the left side (verified using a straight edge). I also have a similar low spot at the middle mounting bolt at the front side. I have tried all manner of Live Z tuning (taking care to preheat my PINDA each time), I have tried Bed Level Correction, and I have even gone so far as to try shimming the bed. Nothing seems to be able to help.

Talk to support, with pictures of a straight edge across your bed showing the low spots. They'll send you a replacement. Mine was painfully dipped on the left side between the screw posts and I could never get a good first layer. The replacement bed wasn't perfect, but was way better and now I get proper first layer adhesion.

Re: 1st layer problems - in depth look at software PINDA problems (and solutions!)

To top it off, I have a low spot on my bed (a warp of sorts) just to the right of the middle mounting bolt on the left side (verified using a straight edge). I also have a similar low spot at the middle mounting bolt at the front side. I have tried all manner of Live Z tuning (taking care to preheat my PINDA each time), I have tried Bed Level Correction, and I have even gone so far as to try shimming the bed. Nothing seems to be able to help.

Talk to support, with pictures of a straight edge across your bed showing the low spots. They'll send you a replacement. Mine was painfully dipped on the left side between the screw posts and I could never get a good first layer. The replacement bed wasn't perfect, but was way better and now I get proper first layer adhesion.

I'll certainly give that a shot. Thanks for the advice.

I think I will have to compile a bit more evidence to be able to plead my case, however. The low points are an extreme annoyance to me (I am a bit OCD with things like that), but because I can, reluctantly, fudge some of my print settings to make infill in those areas look better and reduce the splitting between infill lines (increase extrusion multiplier, increase extrusion width), my worry is support may not consider it problem enough to warrant a replacement even though despite everything I do, the right side of my first layer is quite a bit more squished and the left is barely squished enough (especially in the low spots). According to what I've got for the output of the G81 command, my bed is barely outside of the bounds of acceptance if the >200 micron difference is the guideline by which support goes by. IIRC, mine is in the area of 240 to 280 microns difference between high and low which I thought was rather good in comparison to some others that have posted their results in the forums here. I suppose all I can do is try and hope for a favorable outcome.

Otherwise, I'll be exploring adding back shims to maybe make things a bit more tolerable (though I didn't notice any drastic improvement when I shimmed the bed before). Worst case scenario is I dredge up the $90 USD (I'm Canadian, so it's like a good $120 for me) plus shipping, to purchase a new heat bed and hope for luck with getting a good one.

Re: 1st layer problems - in depth look at software PINDA problems (and solutions!)

I decided to print a PETG test rectangle of 150 x 30 x 1,5mm to try fine calibrate the first layer. when I start printing I always preheat first for 15mn but the extruder far of the printbed because I raise it after a print to clean the nozzle.

Have you tried connecting your printer to your computer with the USB cable and using Pronterface (included in the drivers download from Prusa) to talk with the machine? You can see the temperatures of the Extruder, Bed, and Pinda probe displayed. Your waiting the fifteen minutes lets the Pinda heat up, but as your print is printing it heats up allot more; so that when your starting your next print the Pinda hasn't cooled down to the same temp it was when your started the first print. This is throwing the measurements of the bed mesh off. If you are running Pronterface you can see the temp and wait till it falls to the same temp as when you started the first print (usually you need to aim for 35C Pinda temp). You should then be good to go to start the next print at that point without having to readjust the live Z.

Hopefully the temp calibration gets sorted out with the next official firmware update; or if your brave (I'm not when it comes to programming) you can use stahlfabrik's firmware and tune it to work without waiting for it to cool down.

Re: 1st layer problems - in depth look at software PINDA problems (and solutions!)

Hello- I got my printer less than a week ago now, and have had nothing but issues since I got it...from the common problem people seem to have with the motor overheating and causing the printer to jam, all the way to the PINDA problem detailed in this thread-

I've tried everything, and nothing works, with almost every print i have issues.

So am I to understand that all of this is technically faulty firmware/parts? is this something they will fix in the near future?

I've read over many forums describing all my issues, but every fix leads to another break..I'd appreciate any advise any of you would have.

Re: 1st layer problems - in depth look at software PINDA problems (and solutions!)

I do not know nothing about jams caused by overheating motors. This is not the thread to talk about this issue.

The pinda temperature thing is partly a software issue and a issue about getting the optimal temperature correction offsets into your machine. These seem to be different for every machine. So that part is manageable and I expect a good official solution in the future.

Re: 1st layer problems - in depth look at software PINDA problems (and solutions!)

@stahlfabrik
I am curious what you set your scripted GCode PINDA wait temperature to before printing if you are just printing in PLA? I know the settings in the firmware are zeroed out for 35c but do you initialize your PINDA at maybe 40c for a normal print just so it will be a little warmer?

I finally got access to the update my settings in your firmware and things are working better for me. I tried some PETG prints last night and I think Live-Z needs to be adjusted a little higher for that filament type. I will test some ABS tonight. The whole live-z thing has me a little curious though because I am just guessing on what it should look like when it's dialed in perfectly.

My start was -.660 and I followed that up with new settings I stored to the EEPROM of 6, 12, 20, 24, 24

Thanks

Re: 1st layer problems - in depth look at software PINDA problems (and solutions!)

Hey,

So i have a different yet maybe related problem: My Live Z jumps.
What i mean is that the liveZ i have to set for a good first layer jumped by 0,15 two weeks ago. And a couple of days ago it jumped back.
I had it at -0,61 (for pla at least) it then jumped and i had to set it to -0,47. Had it running for at least a week with daily prints without any problems or changes to liveZ
All of a sudden with a new print it was not sticking anymore and after adjusting to -0,61 it was all back to normal and running quite well.

Its a bit anoying not knowing whether or not its gonna work but more importantly if it would jump again i might drag my nozzle on the pei and damage something bevor i can react...

Anyone else experience something similar?

cheers
Oliver

Re: 1st layer problems - in depth look at software PINDA problems (and solutions!)

@stahlfabrik
I am curious what you set your scripted GCode PINDA wait temperature to before printing if you are just printing in PLA? I know the settings in the firmware are zeroed out for 35c but do you initialize your PINDA at maybe 40c for a normal print just so it will be a little warmer?

I finally got access to the update my settings in your firmware and things are working better for me. I tried some PETG prints last night and I think Live-Z needs to be adjusted a little higher for that filament type. I will test some ABS tonight. The whole live-z thing has me a little curious though because I am just guessing on what it should look like when it's dialed in perfectly.

My start was -.660 and I followed that up with new settings I stored to the EEPROM of 6, 12, 20, 24, 24

Thanks

Hi Douglas

To be honest I just use the wait gcode for the manual print based temperature calibration where it is very useful.
Day to day I use preheat and home xyz a few minutes before a print and check the support menu that the pinda is above 35C.

So if I would use the gcode day to day then yes, I would set it to 35C definitely!

Best regards

Re: 1st layer problems - in depth look at software PINDA problems (and solutions!)

@Jakub: Could you please move this thread to the "firmware" folder? TIA!

Re: 1st layer problems - in depth look at software PINDA problems (and solutions!)

@lukasmatena: how is this effort coming along? Any idea when we could expect the next release?

Re: 1st layer problems - in depth look at software PINDA problems (and solutions!)

Thanks for all the hard work, looks like you are on to something for sure.

This may explain why those of us in cooler rooms have a hard time getting a good 1st layer. My room is around 22C / 68F, so if I don't intentionally warm up the PINDA before printing, the first mesh point will be read with the probe temp still around 22C. Then it starts to warm up as it hovers close to the warm bed during probing subsequent points. Calibration doesn't kick in until 35C so those subsequent points have different temperature variances but aren't getting any calibration adjustments. Either the temperature table needs to be extended to cover colder temps, or as you said, the PINDA always needs to be warmed up to at least 35C before probing.

^^ this is how i interpreted things myself - and I have been doing a manual warm up of the system before printing to account for it.

Thanks for spelling it out so clearly!