Marlin Auto Bed Leveling Reverse Behavior

Hi guys. I'm using Marlin RC7 and I enabled Auto Bed Leveling Grid option. I'm using a proximity sensor LJ12A3-4-ZBX and an MK3 Heated Bed. But the problem is that when the nozzle should go up it goes down and when the nozzle should go down it goes up. I have a drawing where I have show the movement of my coreXY axis. Does the assembly of my Z axis motors like on the old Mendel printers can affect this process? My printer works fine and execute all the movement as I order.

Is there a reason you are not running bugfix_v1.1.0 ? For sure that is better than RC-8 and that is better than RC-7

I tried ABL liniar, biliniar, 3 points, UBL, MBL, nothing is working. Where the nozzle should go down it goes up and reverse. I used marlin bugfix v1.1.0. What's wrong? What's reversed? What more complex tests should i do in order to see where the problem is?

You need to install PronterFace. And using its target in the upper left hand corner... Can you move the X and the Y axis in the correct directions? I presume this will work OK.

After you verify that... Can you move the Z-Axis with Pronterface? When you tell it to go up... Does it go up? Same question for down. When you tell it to go down, does it go down?

If both up and down move but they move in the wrong direction, you need to invert the Z-direction...

My axis move in the right direction so this is not the problem. I think........i might change the place of the probe mounting. Now it's about 50mm to the right from the nozzle tip and 2 mm in the front of it. Maybe I should get it closer to the nozzle. And I'm not sure but my probe seems to be not exactly perpendicular to the bed. It's slightly off..........Can this be a cause of my problem?

And the marlin bugfix v1.1.0 is moving so slow. A lag is present in the LCD menu and for example if I use the Stop print function the LCD get blank. I must hit the encoder button one more time in order to see something on LCD.

Bed Level Correction Matrix:
+0.999999 +0.000000 -0.001229
-0.000006 +0.999988 -0.004905
+0.001229 +0.004905 +0.999987
X:125.00 Y:99.99 Z:3.61 E:0.00 Count A: 22560 B:2328 Z:1189

This is what I get. How do I interpret this?

Those are pretty good numbers in your Correction Matrix. With numbers that good, you almost don't need Auto Bed Leveling.

The way you interpret those numbers is the starting [X,Y,Z] coordinate is multiplied by the Correction Matrix to generate a new coordinate.

Look up Linear Algebra and coordinate transformation in Wikipedia.

Hey DOYT
I'm having the exact same problem...
Did you find a solution?

Hello DOYT, hello Lyngbo,

have you found a solution? I (maybe) have the same issue.

My printer works good, only ABL doesn't work. The left side of my bed is currently nearer to the nozzle then the right side. That is what the ABL mesh told me and what is correct:

Send: G29 V4
Recv: G29 Auto Bed Leveling
Recv: Probing mesh point 1/12.
Recv: Bed X: 10.000 Y: 77.000 Z: -0.296
Recv: Probing mesh point 2/12.
Recv: Bed X: 102.333 Y: 77.000 Z: -0.149
Recv: Probing mesh point 3/12.
Recv: Bed X: 194.667 Y: 77.000 Z: 0.076
Recv: Probing mesh point 4/12.
Recv: Bed X: 287.000 Y: 77.000 Z: 0.337
Recv: Probing mesh point 5/12.
Recv: Bed X: 287.000 Y: 182.000 Z: 0.356
Recv: Probing mesh point 6/12.
Recv: Bed X: 194.667 Y: 182.000 Z: 0.112
Recv: Probing mesh point 7/12.
Recv: Bed X: 102.333 Y: 182.000 Z: -0.134
Recv: Probing mesh point 8/12.
Recv: Bed X: 10.000 Y: 182.000 Z: -0.324
Recv: Probing mesh point 9/12.
Recv: Bed X: 10.000 Y: 287.000 Z: -0.325
Recv: Probing mesh point 10/12.
Recv: Bed X: 102.333 Y: 287.000 Z: -0.179
Recv: Probing mesh point 11/12.
Recv: Bed X: 194.667 Y: 287.000 Z: 0.069
Recv: Probing mesh point 12/12.
Recv: Bed X: 287.000 Y: 287.000 Z: 0.315

Recv: Bilinear Leveling Grid:
Recv:       0      1      2      3
Recv:  0 -0.296 -0.149 +0.076 +0.337
Recv:  1 -0.324 -0.134 +0.112 +0.356
Recv:  2 -0.325 -0.179 +0.069 +0.315
Recv: 
Recv: echo:busy: processing
Recv: X:287.00 Y:210.00 Z:5.00 E:0.00 Count A:39760 B:6160 Z:4281
Recv: ok

While printing the nozzle drives in the wrong direction, driving in the bed on the left side and have a gap between the nozzle and bed on the right side. I tested everything from inverting the driver (and the cable) to swapping alpha and beta (corexy -> coreyx). Nothing worked for me.

Yesterday I inverted the mesh grid manual with M421 for every value

M421 I0 J0 Z0.296 
M421 I0 J1 Z0.324 
M421 I0 J2 Z0.325
...

Send: M420 V
Recv: Bilinear Leveling Grid:
Recv:       0      1      2      3
Recv:  0 +0.296 +0.149 -0.076 -0.337
Recv:  1 +0.324 +0.134 -0.112 -0.356
Recv:  2 +0.325 +0.179 -0.069 -0.315

What should I say, the result was great, a perfect first layer.

Best,
Wilm

Does anyone have an idea how to fix this?

I seem to have to same problem with the "inverted" correction through ABL.
Compensation seems to be reversed (compensate upwards where it should compensate downwards and vice versa).

It looks like there are many ppl experiencing this problem. Part of it seems to be right handed / left handed coordinate system. I am printing on an Ender 5, with 0/0/0 in the upper left front. I also use TMC2209 on a MKS SGEN L with sensorless homing enabled. Perhaps thats a possible error source, with lots of inverting going on with these drivers?

Thanks in advance
LaKi

Edited 2 time(s). Last edit at 03/28/2020 12:51PM by LaKi.

Update: I got this issue fixed for me.

As I am using sensorless homing with my board the diag pins for x and y are hardwired to x min and y min. For homing etc. this wasnt an issue as marlin seems to be able to recognize this.
I solved this by changing the pins in the pinout file of my board (changed x/y+ to x/y-)..

Now ABL gives repeatable results over pretty much the whole bed.

Hi i have the same issue and i’ve too the mks gen l board but i do not find the file that you have modified. Where i need to change the pin output?i have to change something else in config h?

Edited 1 time(s). Last edit at 04/01/2020 07:26AM by rainito.

Pin files are found under Marlin -> src -> pins -> your chip -> your board.

I am not aware of any other changes I made to solve this. For some reason it wont let me attach my config files...

Ok thanks. So in the config.h you haven’t made any change? The direction of the stepper will not change with pin exchange?

Quote
LaKi
Pin files are found under Marlin -> src -> pins -> your chip -> your board.

I am not aware of any other changes I made to solve this. For some reason it wont let me attach my config files...

Maybe you need to zip your config file to upload it.
So you put the x and y endstop pins to the max endstop pin in the mks board?

Ok i change the pin (x/y max to x/y min)in the pins file of my board and i need to put the endstop switch to the max endstop plug but the problem is always there. Maybe my problem is the probe sensor? I have an lj18a3-8-z/bx i don’t remember if this is a 5v or a 6-36v. The wrong voltage can be the problem?

I dont think the voltage should be a problem, as long as in general the probe is working.
See attached files for reference

Edited 1 time(s). Last edit at 04/02/2020 04:37AM by LaKi.