Unreliable Z height calibration on V-Core 3 w/BLtouch & Mini 5+
-
@juliann move the the postion you want to test, then send G30 S-1
https://duet3d.dozuki.com/Wiki/G30
Send it a few time to see the result. -
12/10/2021, 14:48:02 G30 S-1 Stopped at height 3.174 mm
12/10/2021, 14:47:51 G30 S-1 Stopped at height 3.171 mm
12/10/2021, 14:47:37 G30 S-1 Stopped at height 3.172 mm
12/10/2021, 14:47:24 G30 S-1 Stopped at height 3.175 mm
12/10/2021, 14:47:09 G30 S-1 Stopped at height 3.171 mm
12/10/2021, 14:46:51 G30 S-1 Stopped at height 3.172 mm -
@juliann so pretty repeatable probing, it does not look like its probe repeatability causing the issue (although maybe it stops becoming repeatable when it shot or something). Do you always probe with everything at the same temperature? How much variation between prints are you seeing?
-
@juliann said in Unreliable Z height calibration on V-Core 3 w/BLtouch & Mini 5+:
12/10/2021, 14:48:02 G30 S-1 Stopped at height 3.174 mm
12/10/2021, 14:47:51 G30 S-1 Stopped at height 3.171 mm
12/10/2021, 14:47:37 G30 S-1 Stopped at height 3.172 mm
12/10/2021, 14:47:24 G30 S-1 Stopped at height 3.175 mm
12/10/2021, 14:47:09 G30 S-1 Stopped at height 3.171 mm
12/10/2021, 14:46:51 G30 S-1 Stopped at height 3.172 mmYou can do a variation that reports not the height but the variation and deviation:
Assuming the point of interest is X=180,Y=180.
G30 P0 X180 Y180 Z-99999
G30 P1 X180 Y180 Z-99999
G30 P2 X180 Y180 Z-99999
G30 P3 X180 Y180 Z-99999
G30 P4 X180 Y180 Z-99999
G30 P5 X180 Y180 Z-99999
G30 P6 X180 Y180 Z-99999
G30 P7 X180 Y180 Z-99999
G30 P8 X180 Y180 Z-99999
G30 P9 X180 Y180 Z-99999 S-1Frederick
Printers: A FT-5 with the 713 upgrade bits, a custom MarkForged style, a small Utilmaker style, a small CoreXY from kit and a E3D MS/TC setup. Various hotends. Using Duets (2 and 3) running 3.4.6
-
Yes, I did the calibration at the same bed temperature I always use but the variation is not always the same, sometimes is a bit challenging to me getting the right calibration spot again.
What would be the best method/ procedure to minimise this issue?
Thanks again!
-
Thank you @fcwilt
This is the result:
G32 bed probe heights: -0.079 -0.077 -0.077 -0.078 -0.078 -0.076 -0.078 -0.078 -0.078 -0.077, mean -0.078, deviation from mean 0.001
-
@juliann the probe looks to have minimal variation, well under what should need babystepping to compensate for.
what is your start gcode doing? maybe you are homing the printer after probing and the endstops are less repeatable
-
-
@juliann so no G28 is inserted by the slicer?
Also can you post an image of the height map you are using?
Is there a variation between different runs of the same print? -
You are correct, there is no G28 inserted in the slicer.
Yes, the variations are present between different runs of the same print.
-
@juliann said in Unreliable Z height calibration on V-Core 3 w/BLtouch & Mini 5+:
Thank you @fcwilt
This is the result:
G32 bed probe heights: -0.079 -0.077 -0.077 -0.078 -0.078 -0.076 -0.078 -0.078 -0.078 -0.077, mean -0.078, deviation from mean 0.001
I'm a bit late to the party so if I repeat what you have already been told I apologize.
When you use a single G30 at a fixed XY location you are setting the Z=0 Datum.
For the fixed XY location I suggest using the center of the bed.
It is important to do this when creating your height map and when loading your height map.
Now setting the Z=0 Datum can be done as part of homing Z or homing ALL as long as it occurs before creating the height map or before loading the height map.
If you change Z probe settings in either M558 or G31 it would be best to create a new height map.
If you change the XY point you use when setting the Z=0 Datum you should also create a new height map.
Frederick
-
@juliann ok so quite a lot of variation across the bed is shown by the height map.
Try adding a single G30 at the bed centre into your start gcode before you load the height map and see if it if makes it more consistent (it might not be exactly right and still need baby stepping. its more about seeing if the amount of baby stepping needed changes.
-
-
-
After following your instructions I think I got an improvement because now there is not a baby stepping variation anymore between different runs of the same print.
But... always there is a but... if I turn off and on the printer and then try to print anything, all the calibration is gone, so I need to recalibrate the baby steps again to get those nice first layers that I used to have.Thanks again!
Julián. -
@juliann said in Unreliable Z height calibration on V-Core 3 w/BLtouch & Mini 5+:
After following your instructions I think I got an improvement because now there is not a baby stepping variation anymore between different runs of the same print.
But... always there is a but... if I turn off and on the printer and then try to print anything, all the calibration is gone, so I need to recalibrate the baby steps again to get those nice first layers that I used to have.Thanks again!
Julián.If you find you need to adjust baby-stepping once at power on and then it works fine you should be able change the Z Trigger Height parameter in G31 by the same amount.
If you are baby-stepping closer (a negative value) add the absolute of that value to the Z Trigger height parameter.
If you are baby-stepping away (a positive value) subtract that value from the Z Trigger height parameter.
Recall that increasing the Z Trigger Height parameter will result in the nozzle ending up closer to the bed.
In any case when you change G31 you need to be sure it is executed to have the changes take effect.
Frederick
-
Sorry for the delay, I did a lot of tests! After applying all the tweaks you gave me the Z height improved a lot, I have a new printer!
There is just one specific and persistent issue which is only present after powering on the printer.
Its behaviour is the following:
- The printer is already calibrated and all previous prints were perfectly fine.
- Power off and power on the printer
- I try to print anything and the first layer is much more closer to the bed that usual.
- Cancel that print (Emergency stop on DWC) and clean the bed
- so I print again the same file without touching any setting and now the first layer is perfect again.... (and of course I can continue printing without issues)
What I am missing?
Thanks a lot for all the help!
Julián. -
Hi,
Well clearly something is being done under those circumstances that is not being done on power up and thus some setting is not getting set correctly.
Do you have any of the following files in your Sys folder:
- cancel.g
- pause.g
- stop.g
If you do please post them here using the </> tag.
Frederick
-
Thank you very much Frederick!
No, I don't have any of these files in my system folder.
Julián.
-
@juliann said in Unreliable Z height calibration on V-Core 3 w/BLtouch & Mini 5+:
Thank you very much Frederick!
No, I don't have any of these files in my system folder.
Julián.
Hmm...
I don't know all the "side effects" of hitting "Emergency Stop".
It's possible that, in the code executed at the start of a print, the order of execution is incorrect.
Here is a very simple, artificial example of what I am getting at using just two commands:
M703 T0The command M703 will fail if a tool is not selected but the tool is not selected by the T0 command until after the M703 command has already failed.
If I execute that same pair of commands a second time the M703 will work since the tool is now selected.
So if you could post the first 20 lines or so of a file that doesn't print after power on but prints the second time perhaps we can see something.
Frederick
