entire bed too high or too low

robm

I have seen this condition with the entire bed too high or too low before, but I cannot understand what the configuration or error is that generates it. Normally I just keep fiddling and the problem goes away, but I would like to know where to look to address it.

It seems like it should not be possible? This is on a delta with a mini-IR sensor, glass bed covered with blue painter's tape. After running G32 the result is:

3:15:05 PM G32
Calibrated 6 factors using 10 points, deviation before 2.712 after 0.056

Going to bed centre Z=0 puts the nozzle on the bed with reasonable drag on a sheet of paper. The mesh calculation is run without re-homing the carriage, but just tried with re-homing and the result is the same.

So what can cause the mesh compensation result to be entirely shifted up or down?

Check your probe trigger height. Its likely off by 0.8mm as indicated by the mean error?

Z parameter here https://duet3d.dozuki.com/Wiki/Gcode#Section_G31_Set_or_Report_Current_Probe_status

robm

As in the original post, going to z=0 puts reasonable drag on a sheet of paper following g32 or just re-homing. I set this with g31:

G31 P500 X-9 Y20 Z1.600    ; Z probe trigger value, probe offset from nozzle (lower val raises nozzle) and trigger height - glass

Also, this just sets the physical distance between the nozzle tip and the Z-probe trigger point - so (without physically crashing the nozzle and causing skipped steps and bed damage) the mesh software doesn't know any more than what the probe tells it?

dc42

How are you homing Z? One cause of getting a height map like would be that you are homing the bed using a homing switch; but the Z=0 position defined by the homing switch is not the same height as the Z=0 position defoned by the Z probe.

robm

Thanks David, indeed that fixed it - although my 'extra careful' measurement was only 2mm off from what I originally had for the M665 H parameter.

dc42

@robm said in entire bed too high or too low:

Thanks David, indeed that fixed it - although my 'extra careful' measurement was only 2mm off from what I originally had for the M665 H parameter.

If you have a delta, then before running mesh probing, you should run auto calibration and save the parameters, either by sending M500 if you are using config-override.g, or by editing them into the M665 and M666 commands in config.g if you are not.

robm

Ah! Thanks again, I (clearly) missed that trick to get measured values.