G29 failing height map has a substantial Z offset

Phaedrux

Please post your homeall as well.

What order are you running your files?

harlock328

I run home all
Then level with probe (it does bed leveling and g29). Tried doing level. home z then g29 same problem. 2_TrueLevel_Then_Probe.g. Also tried just homing the g29 no joy.

harlock328

I should have mention from the start; that g29 was working fine but my bed level on corners were off little. From pic when it was working

I centered my rails and re-aligned the Zs and re-ran bed level only. The bed was flat. .003 dev after that. then I did the offset procedure. That's when this mess started. Sorry not stating this from beginning. What I get for trying for perfect LOL.

engikeneer

@harlock328 I now understand the config-user.g file - it is essentially a macro which is called from within your config.g file in the line:

M98 P"config-user.g"                    ; Load custom user config

Nothing crazy there, just means you have to look in two places to see how your printer is configured.

As I understand it:

you are setting your Z height by probing the centre of the
you are auto leveling the bed based on probing the four corner points
you are then creating a height map for bed mesh compensation based on probing 4 points (one in each corner)

The Z height and the auto leveling are both correct, but I think you are using the mesh compensation (G29) wrong. You need to probe more points to get anything useful from it. The point of mesh bed compensation is to account for unevenness in your bed surface, you have already levelled it at the corners with G32 bed leveling.

If your bed bows up in the middle, that will cause a Z-offset between the bed mesh you've made (only at the corners) and the z homing point (in the middle).

You set up the spacing of the points to probe using the M557 command - you have this at the end of your config:

M557 X50:200 Y50:200 S150 S150          ; Set Default Mesh (conservative)

Personally, I'd recomend probing at least 50mm spacing (minimum), prefereably finer, by setting the S parameter (note that you only need one!). Also, probably push it closer to the edges of the bed to get better coverage. It might take longer to run, but will give much better results and probably doesn't need redoing that often if your printer is pretty rigid (which a Railcore should be )
https://duet3d.dozuki.com/Wiki/Gcode#Section_M557_Set_Z_probe_point_or_define_probing_grid
I'm thinking something more like this based on your axis limits:

M557 X20:220 Y20:220 S30           ; Probe the bed at 30mm spacing

One other thing - I would make sure that you add M561 to the start of your homing files too. That way you will get more consistent results if you run things in different orders. In the my homez and homeall files on my printer, I have M561 as the first line, then G29 S1 as the last to load the heightmap saved on disk. That way I know bed mesh compensation is on even if I re-home my printer (e.g. when starting a new print) but won't affect my homing.

harlock328

@engikeneer
That's when g29 worked.. Now I getting "Warning: the height map has a substantial Z offset. Suggest use Z-probe to establish Z=0 datum, then re-probe the mesh."

fcwilt

@harlock328 said in G29 failing height map has a substantial Z offset:

@engikeneer
That's when g29 worked.. Now I getting "Warning: the height map has a substantial Z offset. Suggest use Z-probe to establish Z=0 datum, then re-probe the mesh."

In your M557 specify 9 points using P3:3 and be sure the middle point is the same point as used setting the Z=0 datum which appears to be 150,150.

See if that doesn't get rid of the message.

Frederick

engikeneer

@harlock328 even if you get that warning, it should still have been successful and created a heightmap (it just has a Z-offset). Can you show us the heightmap with 30mm spacing?

Phaedrux

I would try deleting the heightmap.csv file entirely.

Then run G28 to home and G29 to probe the bed. Don't do G32 yet. That should produce a heightmap and there should be no warning about an offset. If that works correctly, then try G28, G32, G28 Z, and then G29 and that should also produce a heightmap with no offset.

harlock328

@Phaedrux
Did you steps; confirm no heatmap seen. ran g28 than g29; same error. Heatmap looks the same![alt text]( image url)

harlock328

@fcwilt
This is my M557 X45:282 Y46:254 P2:2 S240 S240; what should it look like for 150,150 (I just starting to lean rrf; from marlin and not expert in that either)

fcwilt

@harlock328 said in G29 failing height map has a substantial Z offset:

@fcwilt
This is my M557 X45:282 Y46:254 P2:2 S240 S240; what should it look like for 150,150 (I just starting to lean rrf; from marlin and not expert in that either)

For starters you use either Paa:bb or Saa:bb, not both.

Now you can use just Paa or Saa and it treats it as Paa:aa or Saa:aa, so I just use the aa:bb form.

As to testing with 9 points to be sure that the middle point is the same as 150,150 I think that this would work

M557 X50:250 Y50:250 S100:100

because 50 + 100 = 150 and 150 + 100 = 250.

I don't know what your probe X and Y offsets are, so remember to take them into account when moving the probe to 150,150 when using G30.

Frederick

Phaedrux

The files you are editing are in the /system folder, right?

Please send M98 P"config.g" and post the results.

If you got to the general tab in DWC, what version does it report as?

harlock328

@Phaedrux
I put changes in config.user.g config.g says Do NOT make changes here and to modify config.user

Board: Duet WiFi 1.02 or later + DueX5
Firmware: RepRapFirmware for Duet 2 WiFi/Ethernet 2.05.1 (2020-02-09b1)
Duet WiFi Server Version: 1.23

Phaedrux

I realize that, please send M98 P"config.g" in the gcode console and post the results.

For the DWC version please check under machine specific.

harlock328

@Phaedrux
Talking about Duet Web Control? If so Duet Web Control 2.0.7

Phaedrux

Yes thanks. Just trying to confirm that your version of DWC is compatible with your firmware version.

Running M98 P"config.g" will execute your config file and any other called macros and will echo any error messages to the gcode console that might otherwise get missed during the bootup process due to timing. I want to know if it's throwing any errors.

harlock328

@Phaedrux
ouputs.txt

fcwilt

@harlock328 said in G29 failing height map has a substantial Z offset:

I put changes in config.user.g config.g says Do NOT make changes here and to modify config.user

I think you will find that you are better off combining those two into just config.g.

And keep related things together.

Frederick

Phaedrux

Well this may be part of the problem.

Error: Bad command: M208 S1 Z-5 :set minimum Z travel

You can delete that command from the config files. It gets overwritten a few lines down anyway, so it doesn't have the intended effect anyway.

Still no smoking gun on why there is an offset.

The only thing that might be causing issues is the split M558 commands with some values spread all over the place, but that bug was in RRF3 and not RRF2.

At this point I would suggest backing up your config files and then creating a fresh basic config using the web configurator and testing with that. If that works correctly then we know the error is somewhere in the added complexity of your current config.

If this is a config provided by the railcore users, you may want to try checking with their discord chat to see what they think.

fcwilt

@Phaedrux said in G29 failing height map has a substantial Z offset:

Well this may be part of the problem....

In my experience the Z offset message occurs when all the points of the height map have a Z value significantly different from the Z value as determined by G30.

In this case he has probed the corners but has not included a point at or very near the point used with G30.

Which is why I suggested using a 9 points (3 by 3) such that the 5th point is at the same location as the "G30" point.

Frederick