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!)

I still question if this is 'a bug', or perhaps a bad pinda or something else. I have been printing at many different temperatures for some time now, and I have not noticed any kind of 'bad first layer' at all. I just print, and print, and print- and no worries. Usually if Prusa sees there truly is a bug, at least they will respond to your pull request. I wonder if the team thinks there is even a bug?

I wonder too!

But it is not a hardware issue I am really quite sure:

They replaced my pinda. Did not help.
The replaced kindly my whole MK3. Did not help.
So I started to look into the software.

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

I think i just came up with the real problem!

The MK2S uses a hardware endstop for homing the z axis. So it is always the same height no matter what. Then the original code to do a - instead of + in mesh bed leveling could be correct!

Maybe that is why it is more complicated to incorporate the fix.

I do not know. Currently I am just glad we have a seemingly working fix and I am sure the devs at Prusa Research will come up with the optimal solution

Edit: damn. The MK2S does not have hardware endstop on z.

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

this bug is DRIVING ME NUTS ALSO... i hope the staff sees this and there is a new firmware ASAP!!! i upgraded the the newest today and the live Z bug is maddening.

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

So I just started an ABS print with PINDA temp of 50C, worked perfectly - not needed to dial in live adjust Z.

😛 it just works

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

great investigation so far, stahlfabrik. You have much more patience than I. 🙂

I'm just glad I do not appear to suffer this problem. I do fiddle with my live-Z (sometimes) when changing from PLA to PETG, for example. But mostly I am able to just print. I noticed you mentioned a grey PINDA. I would describe mine as black. Are there two versions out there?

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

Are there two versions out there?

Yes. Early Mk3s have a black cable and black tip. Current ones have a grey cable and grey tip. Prusa changed suppliers (source: web support when I was sent a replacement for my broken black one and asked if it was the correct item)

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

just want to give some observations on my mk3 using FIRMWARE 3.1.3 (MARCH 8, 2018)

i have the grey PINDA. when the machine is cold and i start it up for the first print the LIVE Z works fine.

If right after the first print i start a new one the LIVE Z becomes closer to the print bed and the first layer is messed up.

I didn't try to use any 3rd party firmware.

i just completed the swap to new X axis parts. version R2/B6 and installed them.

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

just want to give some observations on my mk3 using FIRMWARE 3.1.3 (MARCH 8, 2018)

i have the grey PINDA. when the machine is cold and i start it up for the first print the LIVE Z works fine.

If right after the first print i start a new one the LIVE Z becomes closer to the print bed and the first layer is messed up.

I didn't try to use any 3rd party firmware.

i just completed the swap to new X axis parts. version R2/B6 and installed them.

That is the exact problem I am battling! And if you would try a third time and the PINDA got even warmer the live z would be to close again to the bed...

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

great investigation so far, stahlfabrik. You have much more patience than I. 🙂

I'm just glad I do not appear to suffer this problem. I do fiddle with my live-Z (sometimes) when changing from PLA to PETG, for example. But mostly I am able to just print. I noticed you mentioned a grey PINDA. I would describe mine as black. Are there two versions out there?

Thanks a lot! Is your printer in an actively heated chamber?

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

additional Notes:

i found out not only does the machine have to be cold before starting up a new print, but you have to cycle the power.

there is a change on the main menu that might be bug related. on a power up the access to the Live Z is different then if i finish the print and leave the machine on. the machine goes all nutzo on the live Z when left powered on from the first print and then let it cool to start a new one. cycle the power and it became more normal again..

my feeling is this bug was masked during Alpha/beta testing because of the textured PEI plates, on a rough surface it would be REALLY hard to see this problem, it is only because of the smooth PEI plates that you can see the problem right away.

i also have my collision detection set to OFF because of the layer shift bugs, it was suppose to be fixed the last 2 Firmware changes but i had it happen again and i'm not risking a 20 hour+ print with that function causing a layer shift problem. that was another bug driving me insane.

i wish i bought another MK2S and waited about a year for a MK3 🙁

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

stahlfabrik
Could you provide the source please. I would like to try adding hyperleveling to your code
Cheers

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

I do not understand where you are entering and saving your new uStep values of 5, 11, 11,13, &14?

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

stahlfabrik
Could you provide the source please. I would like to try adding hyperleveling to your code
Cheers

Isn't the code available on his github? https://github.com/stahlfabrik/Prusa-Firmware/tree/MK3_GCODE_TOOL_FOR_TEMP_COMP_ACCESS

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

stahlfabrik
Could you provide the source please. I would like to try adding hyperleveling to your code
Cheers

I linked the pull requests in the original post. In my GitHub in the branch „STAHLFABRIK“ you can find all three code changes at once.

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

I do not understand where you are entering and saving your new uStep values of 5, 11, 11,13, &14?

Hi Douglas
You use the M667 gcode:
M667 ? Query the current table
M667 ! Reset the table to default
M667 Z set all values to zero. This is very very handy for manual temperature calibration because it disables temperature compensation.

Finally
M666 S5 I0 sets table Index 0 (aka 40C) to 5 usteps

Is it working for you?

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

Hi stahlfabrikand thanks for your effort.
I've just tried your code and seems work very well!
I've done just a quick test and not the complete procedure you suggest (sure more reliable).
I have not used the 666 GCODE but instead simply put the Z in home position and on the center of the bed for a while once reached the correct bed temperature (should be useful a quick switch on the 666 code to just read the actual PINDA temperature...).
Then I adjusted e noted down the Z offset for the various bed temperatures and dialed the corresponding value (very roughly ma similiar to yours) in the eeprom with the 667 command. The values I entered are
>>>M667 ?
SENDING:M667 ?
PINDA cal status: 1
index, temp, ustep, um
n/a, 35, 0, 0.00
0, 40, 5, 12.50
1, 45, 11, 27.50
2, 50, 12, 30.00
3, 55, 14, 35.00
4, 60, 15, 37.50
This is the result, the white is generic ABS 255 ext/100 bed and the grey is the PLA Prusa silver 215 ext/60 bed.

I completely agree that this calibration should be included in the standard MK3 menu.

Regard, Vincenzo, Rome

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

Awesome!

But I would advise you to use the M666 gcode to wait until the pinda has reached the real temperature

For a bed temp of 60C my pinda would not get warmer than 45C.

So with this small optimization in your calibration you will get the total perfection:-)

I printed petg today. Did not have to adjust live z again. IT JUST WORKS:-)

So happy you can confirm my fix. Great!

Edit: thanks for posting your values. Again very very much lower than default values.

Edit2: the M666 gcode is especially very helpful for the calibration. The approach without the gcode is ok later. So just home. Warm bed. Wait. Pinda temp will be higher than 35C then likely. And since you calibrated perfectly before you do not have to care😀

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

Yes, as said before I've done a quick test just to verify that the fix works. The next step will be a more precise calibration using the 666 code to be sure to reach the correct PINDA temperature.
Also as you can see from the pictures I should lower a little more the nozzle (the value used for the test was -485 and a I think should be better -500).
In any case I hope the fix and the GCODES or an equivalent menu procedure will be included in the firmware soon.
Thanks again!
Regards, Vincenzo, Rome

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

SIMPLY AMAZING a 99% improvement using stahlfabrik firmware my XYZ Calibration went from slightly skewed to PERFECTLY PERPENDICULAR (I think the Mesh Bed Leveling now works) printed a Z calibration file and it was almost perfect. 😀 😀 😀 😀

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

Yes, as said before I've done a quick test just to verify that the fix works. The next step will be a more precise calibration using the 666 code to be sure to reach the correct PINDA temperature.
Also as you can see from the pictures I should lower a little more the nozzle (the value used for the test was -485 and a I think should be better -500).
In any case I hope the fix and the GCODES or an equivalent menu procedure will be included in the firmware soon.
Thanks again!
Regards, Vincenzo, Rome

For the calibration I think it is not 100% necessary to get 100% the perfect layer for each temperature but to have the same quality layer each time: I dialed live z up to see the split lines and then slowly down until the splits beteen the lines were just not happening anymore. That MIGHT be the perfect layer height. But important is that this moment is very easy to recognize, to compare and to repeat.