RRF3.2Beta4.1 Heightmap vs. Spindle/Driver Mapping

3dML

@Frederik
It looks like the second pass is better then the first.
If your setting are not prefect it may take several more rounds depending on the tolerance you want.

This is my output and it starts out much closer then your first attempt. I only have two Z motors instead of three hence only two probe points. As @Phaedrux said I also use the default S value and it works fine.

Leadscrew adjustments made: 0.005 0.005, points used 2, (mean, deviation) before (0.005, 0.000) after (-0.000, 0.000)
Leadscrew adjustments made: -0.007 0.016, points used 2, (mean, deviation) before (0.004, 0.004) after (0.000, 0.000)
Leadscrew adjustments made: -0.012 0.026, points used 2, (mean, deviation) before (0.006, 0.006) after (-0.000, 0.000)
Leadscrew adjustments made: -0.014 0.047, points used 2, (mean, deviation) before (0.015, 0.010) after (0.000, 0.000)
Leadscrew adjustments made: -0.080 0.095, points used 2, (mean, deviation) before (0.004, 0.029) after (-0.000, 0.000)
Leadscrew adjustments made: -0.098 0.245, points used 2, (mean, deviation) before (0.066, 0.056) after (0.000, 0.000)

Frederik

@Phaedrux I was reading on the antclabs web page that the v3. 1 is measuring within 0.005 mm and the results im getting are different every time I generate a Heightmap

Here are the pictures of my head and in green the position of my spindles and white the position of mgn15 rails to eliminate any z wobble

Frederik

3dML

@Frederik The different Heightmaps is because your Leadscrew deviation is not 0 before you generate the Heightmap.

Frederik

@Phaedrux one moment. I probing single spots and checking the values before and after the z probe code change

Frederik

@3dML I delete the Heightmap and compensation bevor generating a new mesh.

Is that wrong?

And after startup only a g32 - 3 point level +loading of Heightmap.csv (bed.g)

Frederik

@3dML the showed Heightmap was generated at room temperature. To have a baseline when starting test at different standard bed temperatures. Like 60, 90, 100 etc

How Lo g would you give the bed before starting the measurements, so that the temp is as evenly as possible

3dML

If you run the 3point calibration multiple time it should return a smaller deviation each time hence the reason I use a while loop and an if statement to check the deviation value. It will continue to run until it passes the tolerance set by the if statement.

You could put this code in a macro and run it multiple time to see how small you can get the deviation.

G30 P0 X40 Y80 Z-99999          ; Probe near the front left lead-screw
G30 P1 X215 Y395 Z-99999        ; Probe near the rear lead screw
G30 P2 X415 Y80 Z-99999 S3      ; Probe near the front right lead-screw

Frederik

@Phaedrux

Changed bltouch config to H3 F60 T6000 A10 B1 R0.5

3dML

@Frederik said in RRF3.2Beta4.1 Heightmap vs. Spindle/Driver Mapping:

And after startup only a g32 - 3 point level +loading of Heightmap.csv (bed.g)

generate a Heightmap

If you can do a 3point calibration the a deviation of 0 or as close as you can get it you should not have to generate a new Heightmap every time you do a 3point calibration.

The time to temp all depends on your setup.

I have a 750w heater that covers the entire 300mm bed with a temp sensor to control the heat in a slot cutout of the magnet for the flex sheet so very close to the top of the bed. I also have a temp sensor on the heater pad itself so once they both stabilize its good to go. If you are only measuring temp at the heater pad it will take a while for the top of the bed to reach an even temp unless the bed is very thin.

@Frederik said in RRF3.2Beta4.1 Heightmap vs. Spindle/Driver Mapping:

@Phaedrux

Changed bltouch config to H3 F60 T6000 A10 B1 R0.5

What was the max Z speed for this test? The last 4 readings look good. Too high of a Z speed can cause the shift between the second and third reading.

Frederik

@3dML Z Probing speed F60 normal in the RRF config tool was F120

3dML

@Frederik said in RRF3.2Beta4.1 Heightmap vs. Spindle/Driver Mapping:

@3dML Z Probing speed F60 normal in the RRF config tool was F120

Yes that is used for Z- speed when probing but M203 sets the max speed which is used for Z+. Too high of max speed will cause higher deviations when probing but not enough to notice when printing.

I have found that slowing down the maximum Z speed can help with tolerance. Try adding a M203 Z100 before any bed probing moves and M203 Z1000 after the bed probing is finished to restore the speed back to your config settings. You may be able to push the speed a little higher Z100.

Frederik

@3dML its probing with 1mm/s. is that slow enough ?

The Last G32 after changing the settings

Leadscrew adjustments made: 0.380 0.168 0.089, points used 3, (mean, deviation) before (0.244, 0.085) after (0.000, 0.000)
Leadscrew adjustments made: 0.064 -0.009 -0.213, points used 3, (mean, deviation) before (-0.044, 0.083) after (0.000, 0.000)

Leadscrew adjustments made: 0.215 0.207 -0.272, points used 3, (mean, deviation) before (0.044, 0.169) after (0.000, 0.000)
Leadscrew adjustments made: 0.011 0.098 -0.100, points used 3, (mean, deviation) before (-0.013, 0.066) after (0.000, 0.000)

Leadscrew adjustments made: -0.087 0.158 0.672, points used 3, (mean, deviation) before (0.216, 0.222) after (-0.000, 0.000)
Leadscrew adjustments made: -0.018 -0.124 0.255, points used 3, (mean, deviation) before (0.060, 0.126) after (-0.000, 0.000)

Leadscrew adjustments made: 0.305 0.199 0.353, points used 3, (mean, deviation) before (0.304, 0.054) after (-0.000, 0.000)
Leadscrew adjustments made: 0.017 -0.064 0.012, points used 3, (mean, deviation) before (0.002, 0.031) after (0.000, 0.000)

Leadscrew adjustments made: 0.278 0.250 0.293, points used 3, (mean, deviation) before (0.279, 0.015) after (0.000, 0.000)
Leadscrew adjustments made: 0.012 -0.013 -0.004, points used 3, (mean, deviation) before (0.003, 0.008) after (0.000, 0.000)

Leadscrew adjustments made: 0.264 0.270 0.289, points used 3, (mean, deviation) before (0.274, 0.008) after (0.000, 0.000)
Leadscrew adjustments made: 0.011 -0.011 -0.009, points used 3, (mean, deviation) before (0.000, 0.007) after (0.000, 0.000)

G29
81 points probed, min error -0.806, max error 0.757, mean -0.124, deviation 0.338
Height map saved to file heightmap.csv

xmin,xmax,ymin,ymax,radius,xspacing,yspacing,xnum,ynum
50.00,400.00,50.00,390.00,-1.00,38.90,37.80,9,9
 -0.028, -0.333, -0.536, -0.719, -0.794, -0.806, -0.694, -0.498, -0.213
  0.086, -0.210, -0.423, -0.591, -0.686, -0.655, -0.535, -0.386, -0.109
  0.123, -0.133, -0.348, -0.498, -0.593, -0.561, -0.411, -0.277, -0.096
  0.277, -0.020, -0.213, -0.400, -0.465, -0.453, -0.370, -0.169,  0.052
  0.300,  0.019, -0.156, -0.299, -0.374, -0.309, -0.263, -0.114,  0.103
  0.395,  0.137, -0.045, -0.197, -0.278, -0.281, -0.146, -0.010,  0.215
  0.523,  0.232,  0.018, -0.110, -0.236, -0.192, -0.128,  0.062,  0.233
  0.651,  0.365,  0.130, -0.022, -0.146, -0.135, -0.048,  0.122,  0.346
  0.757,  0.547,  0.280,  0.062, -0.042, -0.017,  0.075,  0.206,  0.420

RepRapFirmware height map file v2 generated at 2020-12-23 01:13
xmin,xmax,ymin,ymax,radius,xspacing,yspacing,xnum,ynum
45.00,410.00,40.00,390.00,-1.00,365.00,350.00,2,2
  0.423,  0.344
  1.333,  1.363

th last is a heightmap on the 4 corners as far out as possible

3dML

@Frederik
Yes if multiple G30 S-1 in the same spot return a very small variation these results mean the config is not correct.

If this is your only G31 this may be your problem. I just notice you have no X or Y offset for your BLTouch.

G31 P1000 X0 Y0 Z0.81

Frederik

@3dML is the offset measured from the nozzle or always from x0 y0 ?

3dML

@Frederik
G31 is the probe offset from the nozzle. If you position the nozzle at x0 y0 the G31 offset would be the position of the BLTouch.

Frederik

@3dML when my printhead is homed at x0 y0 (left front corner) the bltouch is around x-3 y-2
the nozzle x35 y8

this is are the values for another test with 4:4 Probing Points

xmin,xmax,ymin,ymax,radius,xspacing,yspacing,xnum,ynum
45.00,410.00,40.00,390.00,-1.00,121.67,116.67,4,4
  0.020, -0.735, -0.706, -0.097
  0.252, -0.427, -0.378,  0.354
  0.536, -0.148, -0.128,  0.676
  0.903,  0.127,  0.082,  0.945

3dML

@Frederik said in RRF3.2Beta4.1 Heightmap vs. Spindle/Driver Mapping:

@3dML when my printhead is homed at x0 y0 (left front corner) the bltouch is around x-3 y-2
the nozzle x35 y8

Post the files you use for homing. The DWC tool position should report the exact position of the nozzle.

3dML

@Phaedrux said in RRF3.2Beta4.1 Heightmap vs. Spindle/Driver Mapping:

M671 defines the position of the lead screws
The order must match the order of the motors defined in M584
In bed.g you probe as close to the lead screws as you can in the same order as the lead screws are defined.

Have you verified these settings are correct and the motors are positioned in the same order in the M671 as in the M584?

Frederik

@3dML

I added your if loop to a makro. Works almost perfekt.
Would it be possible that the measured values are displayed as message or in the console after each move? I stopped the makro after around 10min without knowing if the values are decreasing or what's happening

I checked the limits bevor and it's running.

I also tested several mesh sizes, the resulting height maps are above.

I think, these values are reasonable to some point, I sliced a X from center of the bed 350mm to the 4 corners with a 50x50 rectangular pattern 150mm from the center.

I increased the z trigger height to 1.81mm for safety and started a really slow test pattern. Today after Work I will verify it with a good old magnetic base + digital indicator on different spots along the gantry + x and y along he bed corners. Perhaps I can shim it, modifie the mounts or something else