Weird Bed Mesh Behavior after Z-Leveling

dc42

You should remove the M401 and M402 commands from homez.g because deploying/retracting the probe is automatic. I don't understand why you have the G4 P500 command in homez.g, because the G1 XY command won't execute until the G1 Z8 command has finished.

@jstevewhite said in Weird Bed Mesh Behavior after Z-Leveling:

It's worth noting that if I do:
power on
G28
G32
G32
G28 Z
G29
I get the -0.1 here. But if I do this:
G32
G28 Z
G29
The mesh no longer is offset -0.1.

Are you quite sure you can reproduce that?

jstevewhite

@dc42

The G4 P500 is in the wrong place there. I needed it for the servo I was using and it got inadvertently moved. I'll pull the M401 and M402, thanks. I should probably audit my configuration monthly because of all the experimentation I do

natewalk can reproduce it repeatedly. I only did it twice (the whole cycle, power on to bed 0.1mm low ) but it did do it both times. When the print that's currently running is done, I'll try it again and update.

jstevewhite

@dc42 -

power on
G28
G32
G32
G28 Z
G29

Ok, reproduction is easy. Every time I follow that procedure I get this map:

Imgur

It turned out to take more than a few G32/G28 Z/G29 cycles to return to the normal map, which is:

Imgur

But now, no matter how many times I run G28 Z, G32, or G29, I get the same ( good ) result.

Note that our leveling probe and our mesh probe use the same bed points. Our general assumption is that the 1/4" MIC6 tooling plate is flat (which is born out by actually printing).

Also, I have not tried a print with this map as I'm afraid of a nozzle crash.

veng1

In the first height map, all locations are less than 0.0. What setting would move the center up to Z = 0.0?

jstevewhite

@veng1 said in Weird Bed Mesh Behavior after Z-Leveling:

In the first height map, all locations are less than 0.0. What setting would move the center up to Z = 0.0?

That's the $64,000 question, eh? No setting changes between the first picture and the second.

dc42

@jstevewhite, please can you repeat those tests, but after each G28, G32 and G29 command (and after any individual G30 commands you do), run M122 and look at the "comp offset" value in the Move diagnostics. This is a new field that I added in firmware 2.03. Please check what value is reported there, and whether any nonzero values are related to the height offset shown in DWC.

I would like to get to the bottom of this before I do a 2.03.01 firmware release.

jstevewhite

@dc42

We'll have to get @natewalck to provide that data. Somehow, mine suddenly stopped doing this. I repeated it fifteen or twenty times when Nate told me about it, and then for the last repetition you asked for. But somehow running M122 seems to have fixed it permanently. LOL. Not sure how that's even possible.

Changes: I took out the M401/M402 and the dwell. Nothing else except M122.

Now I power cycle, G28, G32, G32, G29 and get the proper map.

natewalck

G28
G32
G32
G28 Z

Then for mesh:

G91
G1 Z5
M557 X40:290 Y42:332 S250
G29

Yielded this:

natewalck

Note: This does test did NOT use a G28 Z after the second G32 but before the G29, so it is not exactly the same as the post just before this one.

Doing the same thing, but with a M122 after EACH command yielded this:

Also here is the M122 output after each command.

M122 After G28
0_1561079030629_M122_after_G28.txt

M122 After G32 First run
0_1561079035542_M122_after_G32x1.txt

M122 After G32 Second Run
0_1561079039717_M122_after_G32x2.txt

M122 After G29 (same commands as previous post)
0_1561079095120_M122_after_G29.txt