Bed mesh zero reference point?

sirkris

-Printer: voron 2.1 with 4 indipendent z motors
-NPN inductive probe for z offset

Hi guys, I don't understand the bed mesh compensation behavior:

I make first a multi-pass level gantry,cleaning all hightmaps before
(good leveling result with a error around 0.003mm)
Than I make a bed mesh scan: It is preceded from a single z probe centered at x=0 and y=0 to set the z offset relative at the center (for real z offset for inductive probe is set to 0,because I use the mechanical switch for final z offset adjustments)

All z probes....the first single and the bed mesh scan are performed with the same speed, the same current for stepper...etc...etc

Ok so...why I have a resulting map that has a deviation on the center point other than zero?
I would expect that the center point was for me the "reference" point, and therefore all the map was shifted down to have a zero at the center.

Gerrard

@sirkris, Firstly, this isn't a problem, it shouldn't effect your print. Someone else will be able to confirm, but I suspect the reference point is taken as the average (or some other normalization) of all z heights - which may mean no single point on your bed is Z 0.000mm.

Download your heightmap.csv file (from the /sys folder), and take a look at your Z values - you may be able to figure out how it's rationalised.

Phaedrux

@sirkris said in Bed mesh zero reference point?:

(for real z offset for inductive probe is set to 0,because I use the mechanical switch for final z offset adjustments)

Can you better explain what you mean by this? Showing gcode sequence would be best. Otherwise we just have to guess what is actually happening.

sirkris

@Gerrard oh yes...I think so, but it is for understand better the behavior! That you say could be a good explication: "average" of the all z heights.
Yes, I want to take a look at csv.

sirkris

@Phaedrux Yes, my configuration is a little bit complicated, I have several macro that call other macro, I can post all the duet dump.

Anyway: my printer uses two heights probes, one is a mechanical switch and the second one is a inductive probe. The inductive probe is used for all measure on the plate, for example bed leveling, bed mesh, an height measure at a precise point on the plate. The mechanical switch is used only for find the real last height and to set the z offset. It is usefull because hard switch is more stable at temperature variations than a inductive (or bltouch) probe.
I have to post all the macro, because i can't upload a single rar package...my real configuration use several folders...I am not so masochist zhop_up.g zhop_down.g use_mslow.g use_mfast.g use_islow.g use_ifast.g probe_zm_wipe.g probe_zm.g probe_zlevel.g probe_zi.g probe_y.g probe_x.g park_xy.g mswitch_xy.g iswitch_safexy.g homezm_wipe.g homezm.g homezi.g homez.g homey.g homex.g homeall.g heightmap.csv config-override.g config.g coarse_level.g center_xyz.g brush.g bed_mesh_scan.g bed.g

Phaedrux

I'll take a look at your configs once I have a bit of time to concentrate. The voron configs are a bit spread out.

Veti

can you test your repeatability?

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

dc42

When you run mesh bed probing, the print head is shifted by the G30 X and Y probe offsets so that the probe itself is at the probe points. So your single probe with the nozzle at bed centre was probably not at any of the mesh probe points.

sirkris

@Veti Hi, good point!...Done several times and I can say SD on 10 probes is about 0.006mm.

sirkris

@dc42 Hi, yes It is true. In the mean time I noted this by myself and I have adjusted the single probe by G30 with the right value of x and y shift to have the same point to the 0,0 point probed from the mesh probing. (in my case 0,-29)
I have also adjusted better in the same way the 4 bed leveling points to be similar at the 4 corners of the bed mesh.
I can say now that works as expected.
It was very very very tricky:
when I was ignoring the x,y probe offset (applied from default on the mesh, but not on the single first G30 probe) the point of the first probing fell between the 0,0 point and the 0,70 point on mesh that have a strong deviation with different sign too: +0.042 and -0.046.
(wtf of luck)
...and the 4 corners at about 0.002mm!!!!(so in the real far away from the real 0,0 point about -0.003mm)
(wtf of luck)
So you can understand, I saw "something" same a mesh "grounded" on the 4 corners!
When I had put the first g30 at the front left corner, similar at the first point probe from the mesh (always ignoring the shift) I saw a mesh shifted down, but also in this time not with the first point at zero as expected!
...than the repeatability error of the probing does the rest.
So such a mess in my brain.