G29 Vs G32

Veti

G31 P500 X-30 Y50 Z2.90 ; Set Z probe trigger value, offset and trigger height

is the probe really to the left and behind the first nozzle?

G10 P0 X-9 Y0 Z0.0 ; Set tool 0 axis offsets

why do you have a tool 0 offset? on a Cartesian printer there should not be an offset.

M305 P0 T100000 B4138 R2200 ; Set thermistor + ADC parameters for heater 0
M305 P1 T100000 B4138 R2200 ; Set thermistor + ADC parameters for heater 1
M305 P2 T100000 B4138 R4700 ; set thermistor + ADC parameters for heater 2

This is wrong. 2200 is for maestro not 4700. 4138 is the default and very likely wrong for your thermistors.

M557 X15:250 Y6:320 P3 ; Define mesh grid

P3 is not enough for such a large bed.

M671 X15:250:250:15 Y6:6:240:240 P0.5
M350 X16 Y16 Z16 E16:16 I0 ; set drive interpolation

why are you not using interpolation?

from the configuration i am deducting that you are not using this feature.
https://duet3d.dozuki.com/Wiki/Bed_levelling_using_multiple_independent_Z_motors

in that case ignore G32 and just use G29.

Veti

also post the repeatability of your probe
https://forum.duet3d.com/topic/6962/m48-measure-z-probe-repeatability-and-print-to-serial-output

Phaedrux

@lyndon said in G29 Vs G32:

G1 X15 Y6 F6000

Ideally you should probe the center of the bed in your homeall.

lyndon

@Veti Thanks for your replies.
is the probe really to the left and behind the first nozzle?
A: It is yes.
why do you have a tool 0 offset? on a Cartesian printer there should not be an offset.?
A: I have a duel switching extruder and each nozzle when engaged are 18mm X appart. So I set center of head as 0. Hence nozzles ate -9mm and +9mm from center.
This is wrong. 2200 is for maestro not 4700. 4138 is the default and very likely wrong for your thermistors?
A: my bad I have changed to 2200.
P3 is not enough for such a large bed.
A: total agree this grid is temporary for the purpose of trying to figure out my mesh issues as a direct comparison to the G32 results/locations.
why are you not using interpolation?
A: I was but had issues with extruder. My error was:

M350 X16 Y16 Z16 E16:16 I0 ; set drive interpolation

Should have been:

M350 E16:16 I0                                  ; configure microstepping without interpolation
M350 X16 Y16 Z16 I1                             ; configure microstepping with interpolation

Corrected now.
from the configuration i am deducting that you are not using this feature?
A: Yes this is true, I only used the G32 to manually level the bed. As a result I was able to level the bed well. Following this I ran a G28 then G29 then print and the resulting print was terrible, as mentioned earlier the nozzle was above he bed by 0.3mm rear left but digging into the bed front right. Exactly as G29 mesh bed height maps says it should. HOWEVER the bed was almost perfect according to the results of G32.

This is my issue?? When the bed is as level as results of G32 does G29 seem to apply a multiple of 10 ?

in that case ignore G32 and just use G29.
A: totally agree, and I have never used G32 in the past, but I can print reasonably well NOT using G29. HOWEVER my bed is not flat enough so I wan't to uses G29 again.

lyndon

@Phaedrux said in G29 Vs G32:

Ideally you should probe the center of the bed in your homeall.

Hi, I do normally, I have just changed that so the X Y of home Z is on the same point as the G32 and G29 in the hope it would help me understand what's going on with the G29 mesh bed levelling.

Veti

my suggestion is to start of with the basics.

forget about the dual extruder for now and just focus on the first extruder.

your offset from the nozzle is quite large.
if after a g29 your print is higher on one side than on the other that suggests that the z offset changes with the position of the printhead.
this could be due to tilting or other imperfections.
you still have not posted anything about the repeatability of your probe.

so post a picture of a detailed g29.
and indicate across the picture where the print is in the air and where it is ok and where it is squashed.

lyndon

Results of repeatability test:

M98 P"0:/macros/BlTouch/Repeatability test"
G32 bed probe heights: -0.190 -0.188 -0.185 -0.197 -0.195 -0.197 -0.195 -0.195 -0.195 -0.197, mean -0.194, deviation from mean 0.004

G29 25 point probed:

G29
Warning: the height map has a substantial Z offset. Suggest use Z-probe to establish Z=0 datum, then re-probe the mesh.
25 points probed, min error -0.303, max error -0.100, mean -0.201, deviation 0.053

First layer of test print includes G29 as above:

First layer of test print WITHOUT G29:

lyndon

Further to the above this is the completed print with mesh bed compensation:

Compared to this, without mesh bed compensation:

Clearly for some reason the mesh bed compensation is working against me, any help appreciated.

Veti

@lyndon said in G29 Vs G32:

G29
Warning: the height map has a substantial Z offset. Suggest use Z-probe to establish Z=0 datum, then re-probe the mesh.

you have a big difference from when you are homing to when you are probing.
can you remove T-1 from your homeall. and try again.
also please post your config-overwrite.g

lyndon

Hi, I have remove T-1 from homeall.g and here is the Latest G29:

25 points probed, min error -0.072, max error 0.152, mean 0.034, deviation 0.067

Contents of config-overide.g:

; config-override.g file generated in response to M500 at 2019-12-27 09:42
; This is a system-generated file - do not edit
; Heater model parameters
M307 H0 A36.2 C173.9 D2.8 S1.00 V11.6 B0
M307 H1 A468.5 C177.5 D9.8 S1.00 V12.1 B0
M307 H2 A434.1 C167.6 D6.7 S1.00 V12.1 B0

Going to re run that test print now..

Veti

@lyndon said in G29 Vs G32:

Going to re run that test print now..

before you do recheck your offset.
please measure exactly and confirm that the offset for the first nozzle ist
X-21 Y50
use a ruler below the nozzle and take a picture both in the y and x direction.

then set the probing distance between points to something very small. i.e. 15mm and then try printing again.

lyndon

Sorry mate print already started before I read your reply.
I have re measured the probe offset and it's actually X-45 and Y55. I have corrected these in config.g. As you can see from the tool head photos, it's not the easiest measurements to take. So I marked paper on the bed the position of the nozzle and probe, hope it makes sense.
Image showing probe left of nozzle:

Image showing probe behind nozzle:

Measurements of probe offset:

New height map 42 points probed:

Console output of G29:

42 points probed, min error -0.123, max error 0.110, mean 0.012, deviation 0.054
Height map saved to file heightmap.csv

And here is the finished print for the above changes:

Veti

your bed seems very uneven.
from your config i can see that your bed is about 250mm.
but i can see only 6 points. thats is 1 point every 40mm given the irregularities of your bed, that is not enough.

when you updated your offset, did you remember to account for your offset that you configured in the tool config?

lyndon

Hi the bed is 300 x 300 but due to the large head the printable area is about 250 yes.

But the bed is NOT as bad as the G29 mesh is saying. Hence the print I did without using mesh bed compensation. However it is pot luck what area of the bed I print on, ie. different model sizes. So I would prefer to use mesh bed comp.

Oh yes I didn't take the tool offset into account, should I subtract that from the probe offest or add it ?

Also I have just updated firmware to 2.05 after reading this:

Bed compensation did not take account of the XY offset of the printing nozzle from the head reference point

on github. But not had time to play since, I hope to have time later today.

T3P3Tony

@lyndon said in G29 Vs G32:

Oh yes I didn't take the tool offset into account, should I subtract that from the probe offest or add it ?

See this documentation:

https://duet3d.dozuki.com/Wiki/Test_and_calibrate_the_Z_probe#Section_Measuring_Probe_X_Y_Offset

Veti

that documentation assumes that the tool point is a nozzle.
he defined his x0 y0 to be between his 2 nozzle and is making his life a lot harder that way.

Danal

@veti, I strongly agree with your last post. The "Control Point" from G-Code's perspective needs to be 0,0,0 aligned with ONE of the nozzles. Everything else is an offset from that base.

It is almost essential that he do this, so that all the people trying to help him, and OP himself, are "on the same page".

@lyndon Does this make sense?

lyndon

Hi all. what and idiot I am, I hadn't changed the tool offset since replacing my chimera hot end for this duel switching head. DUH.
with the new head the offset is more like 2mm. So for now I have set tool offset to 0 as suggested, and since the head change I agree.

I have now re levelled the bed and am running a detailed G29 and will post results later.

lyndon

This post is deleted!

lyndon

Hi All. here is the latest height map data:

G29
210 points probed, min error -0.175, max error 0.220, mean -0.003, deviation 0.092

Height map:

Result from G32:

G32
Manual corrections required: 0.00 turn up (0.00mm) 0.01 turn up (-0.01mm) 0.12 turn up (-0.06mm) 0.07 turn down (0.03mm)

Print:

Must admit this is the best print using mesh bed compensation I've got so far but no where near as good as the print achieved without it ??

If you study the height map in relation to the print you can see the lower points on the height map are in deed where the nozzle is touching the bed. Which suggests mesh bed compensation is working. But the bed is not as low as G29 thinks it is, at those points.

Again I ask why does a Z of a G32 probe point on the bed NOT equal a Z of a G29 probe of the same point ??