zprobing followed by printing crashes bed

[PeteIHughes]

I am having a lot of difficulty with zprobing followed by printing.

bottom line - when printing, the head crashes into bed by a mm or so (enough to gouge aluminum bed and rip 3m tape but still tries to print)

where am I going wrong ?
Is this probing saved somehow ?
Does Reboot lose probing info.

Pics available if needed

Many Thanks

Pete

Info

I have a homemade delta with a duet wifi.
400mm H x 220mm x 220mm.
Very solid construction (all aluminum 40x20 - no play in carriages / hotend)
DuetEthernetFirmware-1.18.1stable
DuetWifiFirmware-1.18.1stable
JohnSL type probe connected to probe connector

I calibrate the zprobe per Calibrate the Z probe trigger height.
I zero out Z in G31 in the config.g
jog the head to touch a piece of paper (this is at 0.1mm z height)
send a G92 Z0
Head up 10mm
send G30 S-1 multiple times
This results in a g31 value -0.451 consistently
I then put this value into config.g like so
G31 X0 Y0 Z-0.451 P250 ; set the zprobe height and threshold

Then I run G29
145 points probed

then I send a M500

Then I try printing and the head crashes the bed
(i guess by about a mm)
primary layer height 0.2500mm
first layer height 59%

here are what i think are the relevant bits of my files

config.g
;*** Machine Specifics
M665 R220.1 ; delta radius
M665 L454.2 ; diagonal rod length
M665 B205 ; printable radius
M665 H402 ; homed height
M666 X0 Y0 Z0 ; endstop adjustments

;*** Z Probe and Compensation
M557 R135 S20 ; define mesh grid
M558 I1 P5 X0 Y0 Z0 H6 ; a switch not used for homing (johnsl three-pin to zprobe connector)
G31 X0 Y0 Z-0.451 P250 ; set the zprobe height and threshold

M208 S1 Z-4.0 ; minimum z
M501

from config.g.bak
;*** Machine Specifics
M665 R220.1 ; delta radius
M665 L454.2 ; diagonal rod length
M665 B205 ; printable radius
M665 H402 ; homed height

;*** Z Probe and Compensation
M557 R135 S20 ; define mesh grid
M558 I1 P5 X0 Y0 Z0 H6 ; a switch not used for homing (johnsl three-pin to zprobe connector)
G31 X0 Y0 Z0 P250 ; set the zprobe height and threshold

M208 S1 Z-4.0 ; minimum z
M501

from bed.g

; Auto Calibration For Delta Printer

M561 ; clear bed transform
G31 X0 Y0 ; don't want probe offset
G28 ; home printer
M104 S210 T0
M109 S210 T0
G4 S5

; 13 points, 6 factors, probing radius: 177, probe offset (0, 0)
G28
G30 P0 X0.00 Y177.00 Z-99999 H0
G30 P1 X113.77 Y135.59 Z-99999 H0
G30 P2 X174.31 Y30.74 Z-99999 H0
G30 P3 X153.29 Y-88.50 Z-99999 H0
G30 P4 X60.54 Y-166.33 Z-99999 H0
G30 P5 X-60.54 Y-166.33 Z-99999 H0
G30 P6 X-153.29 Y-88.50 Z-99999 H0
G30 P7 X-174.31 Y30.74 Z-99999 H0
G30 P8 X-113.77 Y135.59 Z-99999 H0
G30 P9 X0.00 Y88.50 Z-99999 H0
G30 P10 X76.64 Y-44.25 Z-99999 H0
G30 P11 X-76.64 Y-44.25 Z-99999 H0
G30 P12 X0 Y0 Z-99999 S6

G1 X0 Y0 Z100 F15000 ; move head up to 100

from the start of my Simplify3d sliced calibration xyz cube
; G-Code generated by Simplify3D(R) Version 4.0.1
; Aug 10, 2018 at 5:49:26 PM

G90
M82
M106 S0
M140 S60
M190 S60
M104 S190 T0
M109 S190 T0
G28 ; home all axes
G92 E0
G1 E-1.0000 F1800
G1 Z0.147 F1000
; process Process1
; layer 1, Z = 0.148
T0
; tool H0.250 W0.400
; skirt
G1 X24.600 Y-21.776 F4800
G1 E0.0000 F540
G92 E0
G1 X24.600 Y21.776 E2.8971 F1980
G1 X24.509 Y21.997 E2.9130
G1 X21.997 Y24.509 E3.1493
G1 X21.776 Y24.600 E3.1652
G1 X-21.776 Y24.600 E6.0623

[PeteIHughes]

Just did a M115
Turns out:
M115
FIRMWARE_NAME: RepRapFirmware for Duet WiFi FIRMWARE_VERSION: 1.19 ELECTRONICS: Duet WiFi 1.0 FIRMWARE_DATE: 2017-08-14

No homez.g file
No homeall.g

But homedelta.g

; Home Delta Printer

G91 ; use relative positioning
G1 S1 X1800 Y1800 Z1800 F2600 ; move all carriages up to the endstops
G1 S2 X-4 Y-4 Z-4 ; move carriages down 4mm
G1 S1 X4 Y4 Z4 F200 ; move carriages slowly up 4mm to the endstops
G1 Z-4 F2000 ; down a few mm so that we can centre the head
G90 ; back to absolute positioning
G1 X0 Y0 F2200 ; centre the head and set a reasonable feed rate

[PeteIHughes]

Something just struck me as odd.

Why does the head touch a piece of paper firmly at 0.1mm z height.
And yet the G30 S-1 report -0.451 consistently.
Shouldn't they be roughly the same ?

The JohnSL probe kit included 6 rubber round pads (2 for each sensor).
These seem to make pressing on the bed feel a little dampened or "squishy".
Is this the problem ?

Thx

Pete

[PeteIHughes]

Some pics
printer top
head crash
probes
probes under bed

[T3P3Tony]

@peteihughes Thanks for the comprehensive information.

you should calibrate the delta printer before you probe the bed for mesh bed levelling. It is not clear if you have done this yet. Basically calibrating the delta sorts out all the variables for M666 and M665 that define the geometry of the printer. You have a Bed.g with 6 factor calibration setup already so have a look here:

https://duet3d.dozuki.com/Wiki/Calibrating_a_delta_printer#Section_Running_auto_calibration_for_the_first_time

Once the printer is calibrated nicely, trt rerunning G29 (the mesh bed levelling) and post a picture of the heightmap here.

[PeteIHughes]

@t3p3tony
Thx for the reply.

This printer is a few months old.
I have printed many fairly accurate benchys and calibration cubes with it.
But I was never thrilled with the first layer (usually a "fraction" low in one corner),
so I decided to go with the JohnSL for zprobing.
Never z probed before.

If I set G31 Z param to 0.1, even now, I get good prints.

Prior to zprobing,
my config.g file did not have
G31 set to Z-0.451
It was set to zero or not even there.

And there was never a crash.

Now it does, because I followed
https://duet3d.dozuki.com/Wiki/Calibrating_a_delta_printer

I zero out Z in G31 in the config.g
jog the head to touch a piece of paper (this is at 0.1mm z height)
send a G92 Z0
Head up 10mm
send G30 S-1 multiple times
This results in a value -0.451 consistently
I then put this value into config.g like so
G31 X0 Y0 Z-0.451 P250 ; set the zprobe height and threshold.

And every print crashes.

So, I believe my issue is with the G30 S-1 returning -0.451,
and me setting G31 Z to -0.451.

I know this is not a Duet problem per-se.
But I don't know where else to look.

As I mentioned, if I set G31 Z param to 0.1, even now, without z-probing,
I get good prints.

Is this a JohnSL problem, G31 P250 threshold or what.

I though zprobing was going to be an end to my issues, but alas.

[Phaedrux]

If you tell the printer to move to Z0, where does the nozzle end up?

A negative value for G31 Z means that the probe is triggering below the surface of the bed, which makes sense given that your probe has some squishy pads and depends on the bed pressing down.

I'm not familiar with the JohnSL probe, but you may be able to increase it's sensitivity by increasing the probe speed slightly in your M558 F command or decreasing the trigger value in G31 P command.

Right now your M558 command doesn't have a speed parameter. I'm not sure what your default Z axis feed rate is, but it may be too high or too low.

Right now your G31 P trigger value is 250. Try 100 or 50 and repeat the G30 S-1 test and see if the value returned gets close to 0.

Personally, I wouldn't use a piece of paper to measure the Z0 position if you can visually move the nozzle to just touch the bed for a true representation of Z0.

[PeteIHughes]

@phaedrux and @T3P3Tony
Great info guys.

Currently in the middle of a print, so will try all suggestions in the morning.

Thanks Again

Pete