zoffset changing every time

Tinchus

Hello. Im having this issue on a new printer we assemble at school: we have duet3 on sbc mode, with 3.4.1
We are trying to calibrate the Z offset but something is wrong (and we are using the exact same configuration than in the other 2 printers, all hardware is the same). We stablished a z offset, we print a first layer test and it is OK.
Z offset is saved on config-override and printer is restarted. G31 tell me the value is active
But a new print shows like the Z offset is wrong, and I have to adjust the baby steps again to get a good first layer. I adjusted later the new value to config override... and then a new print forces me to baby step again!

Is there a ny report about this? Could be a corrupted firmware? As I said, config is the same, hardware is the same than in other 2 printers, and this never hapenned on them.

engikeneer

@tinchus post your homing files and start code from a print. Likely something there is causing your issue

Tinchus

@engikeneer homez.g homey.g homex.g homeall.g

Start Gcode is:

G21
M83
M561 ; clear any bed transform
G28
T0
G10 S0
M400
M141 S80
G10 S<MAIN1 - 50>
M116 P0 S5
G1 X175 Y175 F2000
G30; Home z again with probe
G29
G10 S<TEMP>
M116 P0

jay_s_uk

@tinchus what probe are you using? and what printer?

Tinchus

@jay_s_uk it is a custome made corexy, and I use an unductive probe, wich I have check and it is working ok

jay_s_uk

@tinchus which inductive probe?

Tinchus

@jay_s_uk dont have the brand right now. It is german, high quality and hight temperaure (can stand 150 degrees C). It is very precise: 0.03mm of error and a sensing distance of 4 mm

Phaedrux

Your homing files probe at one location, and your start gcode probes at another. They should be the same.

How did you set the G31 trigger height?

Please post your config.g and config-override.g

What happens if you run a macro like this to measure the trigger height repeatedly? Does the value stay consistent, or does it drift?

G1 Z5
M400
G30 S-1
M400
G1 Z5
M400
G30 S-1
M400
G1 Z5
M400
G30 S-1
M400
G1 Z5
M400
G30 S-1
M400
G1 Z5
M400
G30 S-1
M400
G1 Z5
M400
G30 S-1
M400
G1 Z5
M400
G30 S-1
M400
G1 Z5
M400
G30 S-1
M400
G1 Z5
M400
G30 S-1
M400
G1 Z5
M400

Tinchus

@phaedrux @phaedrux The values do not drift al all, the sensor for sure is not the problem. The average error of the measurements in this test is 0.003 mm .

I noticed the "error " on the homing files, but some days ago I asked about this and the answer made notbothering in change it, because: this is the 3rd machine assembled at the scholl, the oter 3 are identical in every sense, they have the same config but not the problem. The first thing I thought was "sensor failure". So I switched and used a sensor from the other printer, and the problem was still there, so the sensor was removed as the suspect ja

Tinchus

@phaedrux said in zoffset changing every time:

How did you set the G31 trigger height?

to set the trigger height:

.- Home all
.- Cancel any existing bed compensation by sending M561
.- Send M564 S0 to allow movement lower than where the firmware thinks Z=0 is.
.- G1 Z1
.- Command the print head down in small steps until the nozzle just touches the bed
.- Send G92 Z0 to define that height as Z=0
.- Command the nozzle up 5mm (G1 Z5)
.- move zprobe to nozzle location:
---G91
---G1 X-7.5 Y-27
---G90
.- Send G30 S-1 to probe the bed without resetting the Z=0 position (repeat 3 or 4 times to average)
Read off the Z height from the web interface. That is the trigger height.

Norder

@tinchus said in zoffset changing every time:

Command the print head down in small steps until the nozzle just touches the bed

Was the nozzle tip clean?
A tiny bit of filament can falsify the result.

Tinchus

@norder the nozzle is clean.

Phaedrux

@tinchus said in zoffset changing every time:

to set the trigger height:

With the nozzle and bed at print temp?

@tinchus said in zoffset changing every time:

G29

Can you post a photo of your heightmap being displayed?

Norder

The print bed and the nozzle should be heated up during this measurement.
Best with the temperatures that are used for the most part.

It is best to use a feeler gauge or a metal setting plate for this measurement.
If this setting plate has a thickness of 0.5 mm, for example, then later subtract this value from the release height.

I do it like this...
Heat up the print bed and nozzle.
Homen (G28).
Run the nozzle directly over a level screw of the print bed with a Z height of 0.05mm.
Place a 0.05mm spring steel sheet between the nozzle and the print bed and adjust the level screw of the print bed so that the adjustment sheet is slightly clamped.
Make sure that there is no filament at the nozzle tip!

Now I move the sensor exactly over the spot where the nozzle was before.
I measure 10x the trigger height and then determine the subtotal.
This is the exact Z offset of the probe!

Google Translate
-- Original Text --

Das Druckbett und die Düse sollten bei dieser Messung aufgeheizt sein.
Am besten mit den Temperaturen die zum größten Teil genutzt werden.

Nehme für diese Messung am besten eine Fühllehre oder ein Einstellblech aus Metall.
Hat dieses Einstellblech zB eine Stärke von 0,5mm, dann ziehe diesen Wert später von der Auslösehöhe wieder ab.

Ich mache es so...
Druckbett und Düse aufheizen.
Homen (G28).
Düse direkt über eine Levelschraube des Druckbettes fahren mit einer Z Höhe von 0,05mm.
0,05mm Federstahlblech zwischen Düse und Druckbett legen und die Levelschraube des Druckbettes so einstellen dass das Einstellblech leicht geklemmt wird.
Dabei darauf achten das kein Filament an der Düsenspitze ist !

Jetzt bewege ich den Sensor genau über die Stelle wo vorher die Düse war.
Ich messe 10x die Auslösehöhe und ermittele danach die Zwischensumme.
Das ist der genau Z-Offset der Sonde !

Tinchus

@norder Lets focus. We are talking about a heated chamber printer, is not a prusa clone. Have you tried to put a gage under he nozzle while the nozzle is at 400 degrees C and the chamber is at 120 degrees celcius? The way to do it is on cold, THEN you add and offset rlated to temperature.
As I said: this is not the first time I do it and we have another 2 printers, exactly the same in hardware and configuration, and the procedure on them was the same, and this behaviour is not happening on them

Tinchus

@phaedrux No, I do it on cold, knowing that later the real value will need to be adjusted with babystepping, and I use the babystepping extra to correct the z trigger gotten on cold. It is not possible to do this with the chamber hot, it is at 120 degrees celsius, you would burn your face and fingers...
But going back to original question: im not having a problem with the mesh, once I adjust with baby stepping, the first layer is excelent.
The problem is later when I turn off the printer and start a new one: I have to baby step again, EVEN after I already corrected the offset on the config, and I already checked using G31 that the new offset is acive

Norder

@tinchus
Then the next step would be to measure the geometry and check whether components are deformed or have loosened connections that could cause this behavior.

Phaedrux

@tinchus said in zoffset changing every time:

I have to baby step again, EVEN after I already corrected the offset on the config, and I already checked using G31 that the new offset is active

@phaedrux said in zoffset changing every time:

Please post your config.g and config-override.g

@tinchus said in zoffset changing every time:

It is not possible to do this with the chamber hot, it is at 120 degrees celsius, you would burn your face and fingers...

Understood. However, if you're using the babystep method to modify the Z offset, are you waiting the same amount of time for heat soak to occur? I could see variance in temperature and time resulting in different Z positions. You meanion other printers, are they all identical?

Tinchus

@phaedrux Yes, the probing on the ced start ONLY after 10 minutes of waiting to be sure all temp is stable (actually the variation inside the cmaber is only 3 degrees at the front and back, very stable). I do the baby stepping when the print has already started and printing a brim wich I use to fine tune the first layer by eye (there is a window to see). The other 2 printer are exactly the same, same hardware, same software, same configuration. There is no even 1 single screw different. I have checked the hardware to see if there is something loose, it is all ok.
And the offset variation is very stable: I i print and baby step 0.2 mm, next time the printer is ON and ofsset already corrected and active on config, I will have to baby steps another 0.2mm again... same value as before.

Very weird.

How can I update the firmware to a specific version? Im think on reflashing the firmware? coud it be corrupt or something like that?

Norder

@tinchus
Have you tried copying the most important files such as config.g, homeall.g etc. from one of the two printers 1:1 without any problems and then overwriting the files on the problem printer?
A tiny mistake quickly creeps in that is overlooked.
If the three printers are identical, this shouldn't be a problem.

Google Translate
-- Original Text --

Hast Du schon versucht die wichtigsten Dateien wie config.g, homeall.g etc. von einem der beiden Druckern ohne Probleme, 1:1 zu kopieren und dann die Dateien auf dem Problem-Drucker zu überschreiben ?
Es schleicht sich schnell mal ein winziger Fehler mit ein den man übersieht.
Wenn die drei Drucker identisch sind, sollte das kein Problem darstellen.