[3.5.0-rc1/rc3] Issue with Probe offsets

[MaxGyver]

I am not sure if this is user error or a bug...

EDIT: It was user error. Turns out my Z-axis lost steps during initial Z-homing. It was almost unnoticeable and I only found out by reconfiguring my whole setup from the ground up.

I noticed that the Z-Probe offset is not applied when the Z0 Datum has not been established. On my Printer without the Z0 datum probed the nozzle touches the bed at Z=0.35mm, so there should be no big difference between Z0 right after booting/homing and the probed Z0 datum.

I use the euclid probe so the probe switch extends quite far below the nozzle when deployed.
I have to use a very large dive height to establish the Z0 datum after booting/homing, in order to prevent the printer crashing the probe into the bed at full speed. With a dive height of 15mm the probing move still starts with the probe at about 3mm over the bed . After the first point is probed, the probe offsets are applied/respected. I would expect the probe offsets to be always applied when a probing move is done.

I think I noticed this problem shortly after upgrading to RRF 3.5 from RRF 3.4.6 initially I did not think much of it and fixed it by setting the dive hight to 15mm in config.g and adjusting the dive height to a lower value after probing the first point. But while streamlining my probing routines I stumbled on this again...

Duet3 6HC , RRF 3.5-rc3

This is my probe configuration:

M558 K0 P8 C"^124.io1.in" H15 F120 T42000 A5 S0.03
G31 X0 Y26 Z10.55

This is my bed.g

;deselect tools
;echo "bedfourscrews"
M671 X-41:495.5:495.5:-41 Y353:353:23:23 S10 ;lead screw positions and maximum allowed correction S in mm
M561 ;clear any bed transform
M290 R0 S0 ;clear babystepping
G29 S2 ;disable mesh compensation
 
if sensors.probes[0].value[0]<1000 ;check if probe is deployed
	M401 ;deployprobe
 
; probe at each corner of the bed
M558 K0 H15
G30 K0 P0 X{(move.axes[0].min+45)} Y{(move.axes[1].max-10)} Z-99999 ; probe near a leadscrew
G30 K0 P1 X{(move.axes[0].max-20)} Y{(move.axes[1].max-10)} Z-99999 ; probe near a leadscrew
G30 K0 P2 X{(move.axes[0].max-20)} Y{(move.axes[1].min+20)} Z-99999 ; probe near a leadscrew
G30 K0 P3 X{(move.axes[0].min+45)} Y{(move.axes[1].min+20)} Z-99999 S4 ; probe near a leadscrew and calibrate 4 motor
 
 
while move.calibration.initial.deviation >= 0.03 ; perform additional tramming if previous deviation was over 0.01mm
	if iterations = 5
		abort "Probing cycle cancelled! Too many calibration attempts or inconsistent probe readings"
	else
		M558 H0.5:0.2 ; adjust probe dive height
		G30 K0 P0 X{(move.axes[0].min+40)} Y{(move.axes[1].max-10)} Z-99999 ; probe near a leadscrew
		G30 K0 P1 X{(move.axes[0].max-15)} Y{(move.axes[1].max-10)} Z-99999 ; probe near a leadscrew
		G30 K0 P2 X{(move.axes[0].max-15)} Y{(move.axes[1].min+20)} Z-99999 ; probe near a leadscrew
		G30 K0 P3 X{(move.axes[0].min+40)} Y{(move.axes[1].min+20)} Z-99999 S4 ; probe near a leadscrew and calibrate 4 motor
		M558 H2:0.5 ; reset probe dive height
echo "Plattform calibration successfull. Deveation between points: " ^ move.calibration.initial.deviation ^ " mm"
 

[oliof]

how would the machine know where the head is so it can apply probe offsets before homing? Is there something I am missing?

[MaxGyver]

@oliof said in [3.5.0-rc1/rc3] Issue with Probe offsets:

how would the machine know where the head is so it can apply probe offsets before homing? Is there something I am missing?

The machine is homed to Zmax after booting. Then the probing routine is started.