Different first layer results every new print  

Re: Different first layer results every new print

Did you change that in the printer or the slicer?

At first i did the change on the fly using the printers tune menu and when it prints the first 2 layers i had the extrusion set to 100 from 95 some times a bit more but i have been bouncing between 5 different spools of PLA and 3 ABS filaments of IPrint-3d Filament purchased in Canada.
I am slowly making profiles in slic3r for each of my filaments.

I have made it a habit to watch the first 2 layers very carefully before i walk away from it.

Re: Different first layer results every new print

I had been having this problem for quite a while and actually started a similar topic a couple days ago because I couldn't get a consistent first layer to save my life. A response from that topic pointed me in the direction of temperature differences at the start of prints which caused the inconsistencies.

The fix that seems to work for me is placing a pause in the starting script between heating and the mesh bed leveling, allowing the temperatures to steady before the calibration is performed.

I home all the axes, heat my bed to 60ºC and extruder to 150ºC, use "G4 S60" to wait for 60 seconds and let things stabilize, perform the mesh bed leveling, and then heat to the desired temps for printing.

It's still early days since I've implemented the change but I'm seeing much improved results. Hope this helps!

Re: Different first layer results every new print

Same problem here.
It appears to me that (at least in my case) the print start temps are important here. Eg. if I start a print with the head and bed cooled to room temps (in my case 20-25*C) and set te Z height then after that every print started from cooled down printer comes with more or less the same first layer. But when I start a new print if the printer is not colled down then the first layers Z height is too high. Seems like the the PINDAs temp offsets are way off.

Re: Different first layer results every new print

That is exactly the problem. The problem with your workaround is that even very small variations in temperature seems to have huge differences. When I repeat first layer calibrations quickly, the machine gets warmer and warmer, so each retry I have to adjust live Z (raise it). So I like the idea of creating a startup g-code workaround but that will never lead to a consistent perfect first layer. Maybe it is good enough. But still it is diametral the opposite of what the PINDAv2 in the MK3 is supposed to achieve. It is like "maybe it would be a good idea to patch out temperature compensation code out of the MK3 firmware to basically act as if there was a MK2 PINDAv1 only available". Honestly, maybe that is also a good workaround as the temperature sensing capability of the PINDA seems to be the problem for us having this issue.

Do you guys know how to read Arduino code then I suggest to join me in investigating the code. Because IMHO, it is not a hardware bug (since I replaced PINDA, then the whole MK3). Well maybe it is dependent on the way the machine is assembled. Maybe a micron in the way the steppers are installed is all the difference between issue and no issue. But I guess that even if that is so the firmware should compensate for that!

And I DO hope that Prusa engineers look into this issue finally as well! It is really a 😳 blocker for me.

Talking about the code: I do not know yet if the code of the bed leveling is the issue or maybe just the code which calculates the temperature compensation. In Marlin_main.cpp there is the function "float temp_compensation_pinda_thermistor_offset(float temperature_pinda)" maybe the bug lies in here would be nice to hunt it down. Help needed!

Re: Different first layer results every new print

That is exactly the problem. The problem with your workaround is that even very small variations in temperature seems to have huge differences. When I repeat first layer calibrations quickly, the machine gets warmer and warmer, so each retry I have to adjust live Z (raise it). So I like the idea of creating a startup g-code workaround but that will never lead to a consistent perfect first layer. Maybe it is good enough. But still it is diametral the opposite of what the PINDAv2 in the MK3 is supposed to achieve. It is like "maybe it would be a good idea to patch out temperature compensation code out of the MK3 firmware to basically act as if there was a MK2 PINDAv1 only available". Honestly, maybe that is also a good workaround as the temperature sensing capability of the PINDA seems to be the problem for us having this issue.

Do you guys know how to read Arduino code then I suggest to join me in investigating the code. Because IMHO, it is not a hardware bug (since I replaced PINDA, then the whole MK3). Well maybe it is dependent on the way the machine is assembled. Maybe a micron in the way the steppers are installed is all the difference between issue and no issue. But I guess that even if that is so the firmware should compensate for that!

And I DO hope that Prusa engineers look into this issue finally as well! It is really a 😳 blocker for me.

Talking about the code: I do not know yet if the code of the bed leveling is the issue or maybe just the code which calculates the temperature compensation. In Marlin_main.cpp there is the function "float temp_compensation_pinda_thermistor_offset(float temperature_pinda)" maybe the bug lies in here would be nice to hunt it down. Help needed!

Congratulations on finally accepting the fact it's not an issue with the hardware/pinda. Lol. Only took 6 weeks to get you to accept fairly logical fact.

Re: Different first layer results every new print

I home all the axes, heat my bed to 60ºC and extruder to 150ºC, use "G4 S60" to wait for 60 seconds and let things stabilize, perform the mesh bed leveling, and then heat to the desired temps for printing.

Hi JDre357, could you please share the GCODE that you have added to your slicer's startup script? Would be interesting to see! Thank you!

Re: Different first layer results every new print

I think I found a workaround in the firmware - but this needs more verification. So far I get far more consistent first layers than with the stock firmware. I disabled the PINDA's temperature compensation in the firmware. This is of course only a workaround, but if you could be so nice to confirm that it solves or improves your issues as well that would be great!

Then I could do a pull request or create an issue on GitHub for the developers to investigate.

Yesterday I failed to even start a print because of the first layer was so inconsistent. Now with the workaround I actually could fine tune the first layer quite well and now the print is finally running! (There is some imbalance from front to back of my heatbed now but maybe it is another issue:-) - like I should maybe repeat XYZ calibration) Second layer looking great!

Please check out my firmware build, give me feedback!

Disclaimer - please do at your own risk - like the prompt of the firmware advises you to do!

Re: Different first layer results every new print

So, i had this issue when i tried to change the startup gCode in an attempt to prevent it from leaving dots at each of the height test points. I started the preheat, and after calibration waited for temp, rather than wait before.

Immediately after making this change i had this issue.
The filament on the nozzle jacks the leveling up. The cold hard filament don't give like hot filament.

I fixed it by pre-heating BEFORE calibration.
Solved my filament oozing to bed two ways. I lowered my pinda probe. (About .4mm or .5mm from print instead of almost 1mm as using ziptie)
AND running the cooling fan @ max during calibration. (Then immediately turn it off after)

Edit: Setting nozzle to 160c during calibration caused this issue. Filament was still too hard.
Nozzle MUST be at full temp (maybe minimum for material?) during calibration. (From my tests)

Edit2: Do you guys have any issues with the live z-calibration test? (I think that's it's name, at work)

Re: Different first layer results every new print

I home all the axes, heat my bed to 60ºC and extruder to 150ºC, use "G4 S60" to wait for 60 seconds and let things stabilize, perform the mesh bed leveling, and then heat to the desired temps for printing.

Hi JDre357, could you please share the GCODE that you have added to your slicer's startup script? Would be interesting to see! Thank you!

Here's my starting sequence for S3D. It's not a perfect fix but it has made a huge difference in my first layer consistency.

M115 U3.1.2-RC1 ; tell printer latest fw version

G28 W ; home all without mesh bed level

M104 S150 T0 ; this will take the layer 1 temperature for extruder 0
M140 S[bed0_temperature] ; this will take the layer 1 temperature for bed 0

M109 S150 T0 ; wait for extruder temp
M190 S[bed0_temperature] ; wait for bed temp

G4 S60 ; wait 60 sec. for temp to stabilize

G80 ; mesh bed leveling

M104 S[extruder0_temperature] T0 ; this will take the layer 1 temperature for extruder 0
M109 S[extruder0_temperature] T0 ; wait for extruder temp

G1 Y-3.0 F1000.0 ; go outside printing area
G92 E0.0
G1 X60.0 E9.0 F1000.0 ; intro line
G1 X100.0 E12.5 F1000.0 ; intro line
G92 E0.0

Re: Different first layer results every new print

Thank you!

I compiled another version of my custom firmware that tries not a workaround anymore but and actual bugfix.

I am not yet 100% sure if I actually found THE bug.

But I really did not sleep much thinking about the problem. In the end I changed only one "-" to a "+" in the firmware:-)

So please checkout my second version to fix that issue. PLEASE try it and tell me if it works for you as well. It seems to work for me (much better than the stock firmware!) - but I have not done a lot of testing. Actually I am still fine tuning the first layer of my first big print with this firmware...

Please lets refer to this firmware as the "second attempt"

EDIT: Fingers crossed!

Re: Different first layer results every new print

As my fix works perfectly for me I created a pull request on GitHub. Please test my firmware and if you have any feedback please let me know. If you like my custom firmware that includes only that small fix (compared to the stock firmware) please support my pull request!

Re: Different first layer results every new print

As my fix works perfectly for me I created a pull request on GitHub. Please test my firmware and if you have any feedback please let me know. If you like my custom firmware that includes only that small fix (compared to the stock firmware) please support my pull request!

Can you post here the function that you changed so i can read it?
I'm a programmer so i can tell if that + or - makes sense there

If other people are ok with default firmware they may have problems with yours

EDIT: I found your pull request under github, it seens that - or + depends on what effect temperature have in pinda probe.
I will test your firmware and post back

Re: Different first layer results every new print

There is definitely a difference if the printer is "cold" or preheated before printing.
The two top layer tests in my picture were made directly from the "cold" machine.
The lower box was made when the printer was preheated for 10 minutes before printing. Then the print was absolutely perfect.

The prints are made with exactly the same conditions except that the printer in the third test was preheated.
I have not yet tested the modified firmware, I'm waiting and see if Tiago has any more comments.

Re: Different first layer results every new print

I invite you to try my bug fix custom firmware.
I am still totally happy with this.

And if it works for you we can all chant loudly so that this finds it way into the official build.

Thanks!

Re: Different first layer results every new print

As my fix works perfectly for me I created a pull request on GitHub. Please test my firmware and if you have any feedback please let me know. If you like my custom firmware that includes only that small fix (compared to the stock firmware) please support my pull request!

Can you post here the function that you changed so i can read it?
I'm a programmer so i can tell if that + or - makes sense there

If other people are ok with default firmware they may have problems with yours

EDIT: I found your pull request under github, it seens that - or + depends on what effect temperature have in pinda probe.
I will test your firmware and post back

I am sooo curious about what you will report back. Happy you found the pull request.

Re: Different first layer results every new print

There is definitely a difference if the printer is "cold" or preheated before printing.
The two top layer tests in my picture were made directly from the "cold" machine.
The lower box was made when the printer was preheated for 10 minutes before printing. Then the print was absolutely perfect.

The prints are made with exactly the same conditions except that the printer in the third test was preheated.
I have not yet tested the modified firmware, I'm waiting and see if Tiago has any more comments.

That is exactly the behavior which made me mad that you observe as well.

If you iterate quickly to find the perfect first layer the pinda gets warmer and warmer and you dial and you dial and you dial live z. And when it cools down again, everything is way off
Have not witnessed this behavior anymore since I installed the custom bug fixed firmware 😀

It is just like the printer works as advertised 😂

Re: Different first layer results every new print

So I did my first ABS print with the bug fix.

With the old firmware, more heat would have meant more squish so I would have to dial the first layer height less negative. And a lot less negative.

Here with the bug fix, the first abs layer had a little to less squish which means that my PINDA was stronger than the pre set-temperature calibration (compensation) table.

So what SHOULD help here for me is doing a temperature calibration. I will try that tomorrow.

EDIT: I aborted again and do temperature calibration now

EDIT2: The temperature calibration leads to MUCH weaker compensation values. I think that is why the bug fixed firmware rightfully overcompensates for higher temperatures. So the behavior that I observe for ABS is not a sign that my fix does work poorly but that for my PINDA, the temperature calibration stock table is to strong.

So my next steps: Repeat temperature calibration multiple times, take the mean value and put it to EEPROM. Then I expect that the complete temp range works PERFECTLY 😎

Does anybody know how one does use the "D8" gcode? My MK3 will not know it... Would be soooo helpful for fuzzing with the temperature calibration table.

Before Temp-calibration:

temp_pinda=35C temp_shift=0um temp_shift=0usteps
temp_pinda=40C temp_shift=32765um temp_shift=-1usteps
temp_pinda=45C temp_shift=17um temp_shift=7usteps
temp_pinda=50C temp_shift=27um temp_shift=11usteps
temp_pinda=55C temp_shift=55um temp_shift=22usteps
temp_pinda=60C temp_shift=97um temp_shift=39usteps

Re: Different first layer results every new print

I have tried the firmware but prints not attach to bed, so i have to recalibrate live z, but now i don't have power where printer is 🙁
Tomorrow i hope to fully test this, so far same calibration don't work which makes sense, in my case i will need to lower the live z a bit more

Also consider to remove the warning message about custom firmware, it's annoying, block printer and octoprint from working (Need to push the LCD button). If printer on a enclosure is even worst

Re: Different first layer results every new print

For legal reasons I will not fuss with this message:-) I like it that it is there. But I get your pain:-)

You could compile your own custom version of my custom version 😆

Yes you must now calibrate live z. Then it should be stable.

EDIT: I did not put the message there. It is the default behavior for Prusa firmware on GitHub. There must be an easy way to get rid of this warning. But I do not want to mess with peoples printers (releasing builds that act like an official build). Please understand

Re: Different first layer results every new print

For legal reasons I will not fuss with this message:-) I like it that it is there. But I get your pain:-)

You could compile your own custom version of my custom version 😆

Yes you must now calibrate live z. Then it should be stable.

EDIT: I did not put the message there. It is the default behavior for Prusa firmware on GitHub. There must be an easy way to get rid of this warning. But I do not want to mess with peoples printers (releasing builds that act like an official build). Please understand

Yes, i know there's a flag, by default is undefined or CUSTOM, the correct flag have GOLD defined, which is Prusa use on official firmware