Monitor what True Bed Levelling is doing?

t0bias

@fcwilt good point! I used the probing-point coordinates for the leadscrews as well by mistake..

t0bias

I adjusted that very line in my config.g to this:

M671 X-49:333 Y140:140 S0.75	; define the two Z leadscrew positions (right: -49/140, left:333/140) and max correction of 0.75

But now even weirder I get the following error:

G32
Error: Some computed corrections exceed configured limit of 0.75mm: -2.095 -0.356

And even I don't exactly know the height-difference; I can absolutely guarantee there is by no means a difference of two millimeters? And why are there two values anyway?

Veti

@t0bias said in Monitor what True Bed Levelling is doing?:

Error: Some computed corrections exceed configured limit of 0.75mm: -2.095 -0.356

you set a limit in M671. That limit has been exceeded

t0bias

@Veti thanks.

I do understand the meaning of the error; but that's what I am saying - there is no way, there is over 2mm of height difference. And before I changed the probing positions, that error never showed up..

Veti

post a picture of a g29 that should show the tilt

engikeneer

@t0bias the values reported are the corrections at the motors/leadscrews so will be a bit bigger than what you are seeing at the bed.

The reason you have two values is that it will correct both motors. Most likely your bed has some curve so the edges are lower then the middle (hence both are negative). This is the reason you should re-home z after doing auto bed levelling.

If you really think the amount of correction is too high, then the first question is on the probe readings. How consistent are they (if you repeatedly prove the same point) and are there any variations across the bed? Seeing a G29 heightmap might give us some clues.

gloomyandy

@t0bias A couple of thoughts. Firstly the -2.095/-0.356 is the correction being made at each of the stepper motors. Remember that these are outside of your bed and so the correction needed will be larger than the actual error at the edge of the bed. Secondly the correction being made to both steppers is negative (so in effect the tilt part of the correction will only be -1,739) this probably means that your bed is at a slightly different height in the centre (where you are setting z=0) than at the points you are probing so there will be a movement in both motors to try and correct for that.

One other thing, your are using G30 s-2 to set your z=0 point, I'm not sure that is correct are you sure that this should not just be G30 (with no s parameter)? That is the usual way of setting the z=0 point based on the probe offset.

Edit I see that @engikeneer got there before me!

3dML

@t0bias
Use a

G1 Xnnn Ynnn
G30

where X and Y is the tool position of the first probe point not the probe position as G30 does not use the G31 probe offset and G30 Pnnn does.

This will set the first probe point height to 0 to minimize the lead screw compensation.

After lead screw compensation rehome at the normal location to set Z offset.

Edit: Also add a G29 S2at the beginning of the file to clear Height Map before bed leveling. Reenable Height Map with a G29 S1 at the end of the file.

Update: Corrected the first G30 as stated in a later post.

fcwilt

@3dML said in Monitor what True Bed Levelling is doing?:

Use G30 Xnnn Ynnn
where X and Y is the **tool position** of the first probe point not the **probe position** as

To the best of my knowledge, at least as of 3.1.1 firmware, the X and Y parameters are ignored when no P parameter is present.

Frederick

3dML

This post is deleted!

fcwilt

@3dML said in Monitor what True Bed Levelling is doing?:

You posted about using G30 Xnnn Ynnn which suggests that the X and Y are used - but they are not.

Frederick

3dML

@fcwilt
That is true. My bad.

I use

G1 Xnn Ynn  
G30

I forget that it is two commands. I will update my post. Thanks for pointing that out.

fcwilt

@3dML said in Monitor what True Bed Levelling is doing?:

G1 Xnn Ynn
G30

 I forget that it is two commands. I will update my post. Thanks for pointing that out.

Now I happen to think that the X and Y parameters SHOULD work which would be consistent and avoid having to determine the correct X and Y values for the G1.

Frederick