BLTouch Probing Accuracy

This is a followup from the previous post. I had a lot of issues with the BLTouch because of inaccurate probing. Also see the Reddit post below from another user.

https://www.reddit.com/r/3Dprinting/comments/7zl8xy/bltouch_extremely_inaccurate_with_auto_bed/

Tests before uploading Marlin

To do the tests, I first send the G28 GCode command with Pronterface to auto home the printer. Then I used the G29 command to mesh autolevel.

Test 1:
0 1 2
0 +0.075 +0.058 -0.218
1 +0.108 +0.041 -0.205
2 +0.043 +0.055 -0.183

Test 2:
0 1 2
0 +0.113 +0.105 -0.176
1 +0.103 +0.068 -0.166
2 +0.028 +0.053 -0.163

Test 3:
0 1 2
0 +0.125 +0.100 -0.158
1 +0.085 +0.068 -0.166
2 +0.023 +0.048 -0.121

Probing speed

I set the probing speed to a much lower value. This changed a little bit, but you can still see the difference between two points (for example 1,0 on test 4 and test 5).

Test 4:
0 1 2
0 -0.049 -0.083 -0.373
1 -0.083 -0.148 -0.382
2 -0.156 -0.149 -0.362

Test 5:
0 1 2
0 -0.072 -0.123 -0.350
1 -0.091 -0.117 -0.375
2 -0.162 -0.138 -0.345

This looked promising, but the time to do the measurements increased a lot, so I changed the Z_CLEARANCE_MULTI_PROBE to 1 mm instead of 5.

This changed the speed drastically, in a positive way, but not the BLTouch pin deploys the second time onto the bed, so I ended up setting the above described clearance setting to 2 mm.

The last few tests after the upload and clearance setting:

Test 6:
0 1 2
0 +0.012 -0.057 -0.288
1 -0.029 -0.050 -0.305
2 -0.094 -0.078 -0.255

Test 7:
0 1 2
0 +0.016 -0.011 -0.295
1 -0.033 -0.084 -0.300
2 -0.099 -0.075 -0.276

Test 8:
0 1 2
0 +0.009 -0.044 -0.273
1 -0.030 -0.088 -0.280
2 -0.087 -0.050 -0.276

The (1,0) and (1,1) are different. The rest is almost the same every test. I’ll stick with this configuration because it’s at least better than first.