Duet3 not changing Z height on second layer

[eduard]

@phaedrux Thanks for your reply.
The printer works good except this weird second layer problem. To avoid this problem I using raft. At printing the raft two layers print on each other without moving the Z axis, but the third and beyond is good.
I can move the Z axis by command, or by baby steps and it is very accurate. While I print with raft, the calibartion cube in Z axis is under +-0.03mm deviation.
I checked everything what I can imagine. I do not have a Z probe, leveling manually.

Here is the config file:config.g

Bests,
Eduard

[Phaedrux]

Can you share your homing files and a sliced gcode file?

[eduard]

@phaedrux Sure, here they are:
https://www.dropbox.com/s/4foc8pba25khhqj/Archive.zip?dl=0

[Phaedrux]

G1 H1 Z-180 F100 ; move Z down until the endstop is triggered
;G1 H2 Z10 F300
G90                      ; absolute positioning
G92 Z0                   ; set Z position to axis minimum (you may want to adjust this)

At the end of your homeall/homez it's forcing the z position to be Z0 meaning the nozzle would be touching the bed. Is that actually the case? You should find the actual offset and set that.

[eduard]

@phaedrux The Z is set to 0.2mm from bed.

[Phaedrux]

Not according to your homing files

[eduard]

@eduard I have to set G92 Z0.2 ?

[Phaedrux]

Yes, set it to whatever the actual distance is from nozzle to bed when the endstop is triggered.

[eduard]

@phaedrux OK, I set it now to Z0.2 and I give it a try.

[eduard]

@eduard unfortunately this was not the problem. At the second layer, the bed still does not move. The third is fine and so on.

[Phaedrux]

the file you shared has a raft?

[eduard]

@phaedrux Unfortunately, I no longer remember exactly what was and what wasn’t, but now I’m sending a new one that doesn’t have raft. During yesterday's test, I printed this and the second layer went to the first one here, too, without raft.

xyzCalibration_cube (1).gcode

[rjenkinsgb]

You file has the correct Z moves, from what I can see.

Try doing a basic test:
Reference (home) the printer, move up in Z a few mm then down until the nozzle is just noticeably gripping a sheet of paper.

Then move away just 0.1mm in a single move and see if the paper is loose?

If not, the problem is mechanical; either something is sticking or there is backlash in the Z mechanics.

The moves to get a light grip on the paper must be one way only for the test to be valid, no back & forth; if you overdo it, move eg. 5mm clear and start again, as the whole point is to test a single reversal of Z direction.

Likewise when moving away from it the 0.1mm; no further, or start over.

[eduard]

@rjenkinsgb Thank you for your reply.
The nozzle distance from bed was set with a precision 0.2mm metal gauge. The movement of the bed was calibrated with a 0,001mm precison indicator (+-0.015mm)
Mechanical failure is completely ruled out.
All already layers except the second are executed and perfect. When I use raft, the print is perfect and in all directions.

[eduard]

@eduard
This is what the table leveling process I follow looks like:
The duet is set above 3 points on the bed:
M558 P0; No Z probe
G30 P0 X90 Y0 Z-99999; probe near an adjusting screw
G30 P1 X150 Y185 Z-99999; probe near an adjusting screw
G30 P2 X30 Y185 Z-99999; probe near an adjusting screw and report adjustments needed

At each point, I adjust the duet to Z0 and adjust the head bed distance to 0.2mm with the bed adjustment screw. Then OK and restart the process to check back.
The thickness of the first layer in the slider is set to 0.25mm.

[rjenkinsgb]

@eduard said in Duet3 not changing Z height on second layer:

All already layers except the second are executed and perfect.

That does not rule out backlash - all layers have Z moving in the same direction except the first, which follows a move in to the 0.2mm clearance; ie. that first layer change is the reversal point, where motion would be lost.

Do a test with the dial gauge, as you have one:

Step 0.1mm several times towards the head & note the needle movement - then do a single 0.1mm step the opposite direction and see if you get the same movement??

If that is correct, then it clears the mechanics.

[Nothing personal, but having worked on machine tools for decades, it's a familiar problem].

[eduard]

@rjenkinsgb Sure, I don't take personal
If there was a backlash, the motor shaft would move, but the bed would not. However, the motor shaft does not move at the second layer.
I measured the backlash and about 0.01mm (the bed moves on a ball screw)

[rjenkinsgb]

@eduard
That could still be mechanical stiffness, giving the same effect.

The same test would prove or disprove that.

[rjenkinsgb]

Have you posted your config.g file yet?

[eduard]

@rjenkinsgb When I built it, I measured the backlash using the method you wrote. I have measured it several times since then and get the same value. This is unlikely to be where the problem comes from. Mainly because the bed goes higher (endstop) to the first layer and then comes lower, 0.2mm.