Auto_Bed_Leveling: Enhanced G29 command

[fabsweb]

Hello Roxy,

-- 1 --
No photo, because I achieved to print well your bed leveling test object.

To do this I had to :
Find the new realy good value for 'Z probe offset' because of the new servo.
Set the XYjerk to 1 (10 before), because the PLA was not glueing in the corners off the object. (Couldn't detect this on smallers objects)

So, it's not an auto bed leveling problem, no ?
I think it's a problem with the belts, too old ? bad quality ? bad tension ?

-- 2 --
M48 in 9 places. (Don't think is usefull you spend time with this now (?))

M48 X 20 Y 20
09:17:26.687 : Mean: 9.402938
09:17:26.687 : Standard Deviation: 0.003792
09:17:26.687 : echo:endstops hit: Z:9.40

M48 X 20 Y 100
09:20:38.015 : Mean: 9.420688
09:20:38.015 : Standard Deviation: 0.010226
09:20:38.015 : echo:endstops hit: Z:9.43

M48 X 20 Y 180
09:23:49.484 : Mean: 9.334564
09:23:49.484 : Standard Deviation: 0.002696
09:23:49.484 : echo:endstops hit: Z:9.34

M48 X 100 Y 20
09:27:03.859 : Mean: 9.261376
09:27:03.859 : Standard Deviation: 0.011367
09:27:03.859 : echo:endstops hit: Z:9.29

M48 X 100 Y 100
09:30:15.359 : Mean: 9.179813
09:30:15.359 : Standard Deviation: 0.000688
09:30:15.375 : echo:endstops hit: Z:9.18

M48 X 100 Y 180
09:33:26.875 : Mean: 9.321436
09:33:26.875 : Standard Deviation: 0.002306
09:33:26.875 : echo:endstops hit: Z:9.32

M48 X 164 Y 20
09:36:41.328 : Mean: 9.171938
09:36:41.328 : Standard Deviation: 0.007783
09:36:41.328 : echo:endstops hit: Z:9.16

M48 X 164 Y 100
09:39:52.859 : Mean: 9.166501
09:39:52.875 : Standard Deviation: 0.007727
09:39:52.875 : echo:endstops hit: Z:9.16

M48 X 164 Y 180
09:43:04.281 : Mean: 9.373126
09:43:04.281 : Standard Deviation: 0.005513
09:43:04.281 : echo:endstops hit: Z:9.37

Regards.
Fabrice

[Roxy]

That is strange... You have two places where the Standard Deviation jumped up to .01mm That should still be fine, even if you are printing at .2mm layers. But it is strange it jumped up.

It is very important to get the Z-Probe Offset from the nozzle correct. I usually start with some extra which causes the melted filament to not even touch the bed. And then move it down very slowly and carefully until it gets some what smeared into the bed. Hopefully when you get through that process, it is going to print well every where.

[fabsweb]

Hello Roxy,

Thank you again and merry Chrismas.

Regards. Fabrice.

[kenkelso]

Hello and thanks for all this info. My only problem is that the 2 attached files won't download, is something broken or am I just being dumb?

KenK

[Roxy]

I don't know what is going on with .STL file downloads. But the source code is in the post. You can cut and past that into OpenScad and generate the .STL file yourself.

[peonking]

Hi

I got the auto bed leveling working mechanically with Marlin on Mendel max with Gen7. It uses a mechanical endstop.

Its able to perform G29 and produce a correctional matrix. Topology map also shows my left side of the bed is higher than the right side.

When i set the nozzle to Z=0 using G1, it will almost flush against the right side of bed. While moving the nozzle to the left side, nozzle is moving up along the z axis slightly. However, gap between the nozzle and the bed gets larger and larger.

I've tried 2 points grid up to 4 points grid and they yield the same results.
Here is the result of G29 V4
Eqn coefficients: a: -0.00134520 b: -0.00004551 d: 20.06528472
Mean of sampled points: 19.93807411
planeNormal x: 0.001345 y: 0.000046 z: 1.000000

Bed Height Topography:
+-----------+
|...Back....|
|Left..Right|
|...Front...|
+-----------+
+0.10193 -0.01074 -0.15207
+0.11593 +0.04793 -0.07074
+0.05793 +0.01659 -0.10674

Corrected Bed Height vs. Bed Topology:
+0.05356 +0.04178 +0.00134
+0.06278 +0.09567 +0.07789
+0.00000 +0.05956 +0.03711

Bed Level Correction Matrix:
+0.999999 +0.000000 -0.001345
-0.000000 +1.000000 -0.000046
+0.001345 +0.000046 +0.999999


I don't know how to fix this.

Thanks so much.

[Roxy]

Do it with a 5 x 5 grid and post that topographical map. That will tell us a lot more.

G29 n 5 V 4 T

[peonking]

Thanks for the quick reply.
Here is the prints.

>>> G29 n5 V4 T
SENDING:G29 N5 V4 T
G29 Auto Bed Leveling
Bed X: 15.000 Y: 30.000 Z: 19.825
Bed X: 52.000 Y: 30.000 Z: 19.868
Bed X: 89.000 Y: 30.000 Z: 19.881
Bed X: 126.000 Y: 30.000 Z: 19.858
Bed X: 163.000 Y: 30.000 Z: 19.806
Bed X: 163.000 Y: 82.000 Z: 19.835
Bed X: 126.000 Y: 82.000 Z: 19.934
Bed X: 89.000 Y: 82.000 Z: 19.916
Bed X: 52.000 Y: 82.000 Z: 19.918
Bed X: 15.000 Y: 82.000 Z: 19.873
Bed X: 15.000 Y: 134.000 Z: 19.900
Bed X: 52.000 Y: 134.000 Z: 19.899
Bed X: 89.000 Y: 134.000 Z: 19.914
Bed X: 126.000 Y: 134.000 Z: 19.913
Bed X: 163.000 Y: 134.000 Z: 19.857
Bed X: 163.000 Y: 186.000 Z: 19.857
Bed X: 126.000 Y: 186.000 Z: 19.940
Bed X: 89.000 Y: 186.000 Z: 19.937
Bed X: 52.000 Y: 186.000 Z: 19.947
Bed X: 15.000 Y: 186.000 Z: 19.895
Bed X: 15.000 Y: 238.000 Z: 19.911
Bed X: 52.000 Y: 238.000 Z: 19.915
Bed X: 89.000 Y: 238.000 Z: 19.905
Bed X: 126.000 Y: 238.000 Z: 19.903
Bed X: 163.000 Y: 238.000 Z: 19.851
Eqn coefficients: a: -0.00021261 b: 0.00022899 d: 19.87855720
Mean of sampled points: 19.89031982
planeNormal x: 0.000213 y: -0.000229 z: 1.000000


Bed Height Topography:
+-----------+
|...Back....|
|Left..Right|
|...Front...|
+-----------+
+0.02035 +0.02501 +0.01435 +0.01301 -0.03899
+0.00435 +0.05701 +0.04701 +0.04968 -0.03365
+0.00968 +0.00835 +0.02368 +0.02301 -0.03299
-0.01765 +0.02768 +0.02568 +0.04368 -0.05565
-0.06499 -0.02232 -0.00965 -0.03232 -0.08432


Corrected Bed Height vs. Bed Topology:
+0.03770 +0.05024 +0.04744 +0.05397 +0.00984
+0.03361 +0.09414 +0.09201 +0.10254 +0.02708
+0.05085 +0.05738 +0.08059 +0.08779 +0.03965
+0.03543 +0.08862 +0.09449 +0.12036 +0.02889
+0.00000 +0.05054 +0.07107 +0.05627 +0.01213


Bed Level Correction Matrix:
+1.000000 +0.000000 -0.000213
+0.000000 +1.000000 +0.000229
+0.000213 -0.000229 +1.000000


I've also did a repeatability test at the last spot, if it helps
>>> M48
SENDING:M48
M48 Z-Probe Repeatability test
Mean: 19.961666
Standard Deviation: 0.019730

[peonking]

Hi Roxy,

I don't mean to rush.

Do you need more data from me?

Thanks

[Roxy]

I've been driving all day. I just got a chance to hop on the internet. Your bed looks very level. But with that said, the from right corner is a little bit low and the back left is a little bit high.

And... You do seem to have a bump in the back right. Check out this line:

+0.00435 +0.05701 +0.04701 +0.04968 -0.03365

And look down the 3rd and 4th column. That bump isn't huge, but it definitely is there.