Adjust stop criterion for QGL and probe measurments

General Discussion
1

Hi there,

The convergence criterion (remaining Error) for probe/QGL is ridiculusly low! It makes no sens to enter values like 0,005 (= 5 micron! You cannot measure this an that environment! Don’t believe digits, showing such numbers). I propose to set the value to stop the QGL at 0,05-0,025. Calculate the heat enlongation of ABS : 0,1mm/m * K
e.G. Voron Afterburner X-Rail to nozzle length: 50mm. Heat up during measurement Room Temp-40 deg C (maybe more?) = 20K. Elongation = 0,1 * 50 / 1000 * 20 = 0,1mm (!) Find your 5 Microns here!

Or do it another Way round: let’s say, the increase in heat during one QGL run is 2 C. The error of this is: 0,1*50/1000 * 2 = 0,01mm, and this is only the error of temperature.

Alternative: Heat up your system properly (30 mins?) but i am too nervous to do so…
So with this value, you can measure as long as you like. Finding a position for this means: you did not measure enough :wink:

2

This value is adjustable in your config. Set it to whatever you like. I use .01mm with no issues.

1 Like
3

I know and I did. But I thing one could change the defaults.

4

There is no default retry_tolerance value. Retries are disabled by default, and it’s up to the user to define the number of retries and the tolerance.

5

OK, my fault: I downloaded a configuration (i forgot from wich source) and I didn’t compared it with the klipper config. The klipper config says 0,1mm for the probe. So you are right, the proposed value is 0,1mm and nothing else…

If got the value samples_tolerance: 0.05 in the [probe] section and the value retry_tolerance: 0.05 in the [quad_gantry_level] section. If I change that last value, the printer changes its behaviour during leveling, but i will doublecheck this, maybe I am wrong. See my configuration block in printer.cfg (and the old tolerance value I commented out):

[quad_gantry_level]
gantry_corners:
   -60,-10
   410,420
#  Probe points
points:
   50,25
   50,275
   300,275
   300,25

speed: 100
horizontal_move_z: 20
retries: 5
#retry_tolerance: 0.005
retry_tolerance: 0.05
max_adjust: 10

Maybe this came from the klicky probe? I forgot…

I just quickly checked, if values are dicovered, that are not related to a section, and it does. So my comclusion is: there is a tolerance value in the QGL section, but it is not documented (I cant find it in the docs). Maybe it is inherited from somewhere, but this is not easy to find out.

Option 'some_bullshit' is not valid in section 'quad_gantry_level'
6

I’m not sure what you are reporting. See Configuration reference - Klipper documentation

7

Yes my fault! I just took the config from “somewhere” and did not compare it to the reference. Will do better in the future :wink:

1 Like