Advice to fix BLtouch drift?

supernovae · 14 Dec 2019, 19:47

Running this test:

; Probe testing single point.
G30 P0 X150 Y150 Z-99999 F120
G30 P1 X150 Y150 Z-99999 F120
G30 P2 X150 Y150 Z-99999 F120
G30 P3 X150 Y150 Z-99999 F120
G30 P4 X150 Y150 Z-99999 F120
G30 P5 X150 Y150 Z-99999 F120
G30 P6 X150 Y150 Z-99999 F120
G30 P7 X150 Y150 Z-99999 F120
G30 P8 X150 Y150 Z-99999 F120
G30 P9 X150 Y150 Z-99999 F120
G30 P10 X150 Y150 Z-99999 F120
G30 P11 X150 Y150 Z-99999 F120
G30 P12 X150 Y150 Z-99999 F120
G30 P13 X150 Y150 Z-99999 F120
G30 P14 X150 Y150 Z-99999 F120
G30 P15 X150 Y150 Z-99999 F120
G30 P16 X150 Y150 Z-99999 F120
G30 P17 X150 Y150 Z-99999 F120
G30 P18 X150 Y150 Z-99999 F120
G30 P19 X150 Y150 Z-99999 F120
G30 P20 X150 Y150 Z-99999 F120
G30 P21 X150 Y150 Z-99999 F120
G30 P22 X150 Y150 Z-99999 F120
G30 P23 X150 Y150 Z-99999 F120
G30 P24 X150 Y150 Z-99999 F120
G30 P25 X150 Y150 Z-99999 F120
G30 P26 X150 Y150 Z-99999 F120
G30 P27 X150 Y150 Z-99999 F120
G30 P28 X150 Y150 Z-99999 F120
G30 P29 X150 Y150 Z-99999 F120
G30 P30 X150 Y150 Z-99999 F120
G30 P31 X150 Y150 Z-99999 F120 S-1

And getting this:

PM M98 P"0:/macros/test-z-probe"
G32 bed probe heights: -0.032 -0.045 -0.045 -0.060 -0.067 -0.065 -0.075 -0.080 -0.090 -0.095 -0.092 -0.110 -0.117 -0.110 -0.120 -0.135 -0.137 -0.150 -0.145 -0.157 -0.168 -0.165 -0.175 -0.185 -0.185 -0.185 -0.205 -0.195 -0.202 -0.215 -0.230 -0.240, mean -0.

seems to just increase

BLtouch v2.2 (antclabs)

config.g

; Z-Probe
M574 Z1 S2 ; Set endstops controlled by probe
M307 H3 A-1 C-1 D-1 ; Disable heater on PWM channel for BLTouch
M558 P9 H5 F120 T6000 R0.2 ; Set Z probe type to bltouch and the dive height + speeds
G31 P25 X-24 Y-19 Z4.70 ; Set Z probe trigger value, offset and trigger height
M557 X20:250 Y20:250 S75:75 ; Define mesh grid

HyperCube COREXY 300x300

supernovae · 14 Dec 2019, 21:26

This is on firmware 2.05 btw

Phaedrux · 14 Dec 2019, 22:33

Try adding this to your M558 command A10 R0.75 S0.003 B1

That will have a slightly longer pause between travel moves and probing dives. It will force a tighter tolerance between consecutive probe results. And it will turn off the heaters during probe dives. You may notice that it will probe each point a minimum of 2 times and a maximum of 10.

supernovae · 14 Dec 2019, 22:44

I’ll give that a shot. Someone mentioned there may have been a bug in 2.03 and to downgrade e to 2.02 and if that fixes it then 2.05 may still have that bug. I’ll confirm and experiment more tonight when I get back home to printer

dc42 · 14 Dec 2019, 23:01

As the trigger height keeps falling, that may mean that the Z axis is overshooting when the probe triggers because it can't stop fast enough. What is your Z steps/mm? Does a lower probing speed fix it?

supernovae · 15 Dec 2019, 04:32

Here is my gcode for stepping, steps per mm and speeds. I can try a lower probing speed - off to experiment

M350 X16 Y16 Z16 E16 I1 ; Configure microstepping with interpolation
M92 X160 Y160 Z400 E837 ; Set steps per mm - set to 837
M566 X600.00 Y600.00 Z250 E2600 ; Set maximum instantaneous speed changes (mm/min) (jerk)
M203 X12000.00 Y12000.00 Z250.00 E3600 ; Set maximum speeds (mm/min)
M201 X600 Y600 Z80 E3000.00 ; Set accelerations (mm/s^2)
M906 X1200.00 Y1200.00 Z900.00 E900 I30 ; Set motor currents (mA) and motor idle factor in per cent
M84 S30 ; Set idle timeout

.9 degree steppers on X/Y
1.8 on Z (dual z)

Phaedrux · 15 Dec 2019, 05:01

Your settings look very reasonable for the Z axis and probe.

supernovae · 15 Dec 2019, 15:45

So I enabled the multiple checks per point if the variation is greater than .03 and one time it never finished checking the level. I haven’t had too much time to test more but I made sure the bltouch was secure, straight and moved it up a tad higher so there was more room in case it was too close to bed but still seeing the same symptoms.

Slowed z way down too to make sure bed wasn’t moving quicker than the bltouch could reset and also set a delay between each probe

I’ll run another 30 test and report back

Phaedrux · 15 Dec 2019, 16:05

@supernovae perhaps time to clean the pin.

jens55 · 15 Dec 2019, 16:08

Do you actually mean 0.03 or did you mean to say 0.003?
I had to set my BLTouch to 0.005 to get the probing to finish (up from 0.003) but then it worked fine. 0.03 (rather than 0.003) is pretty bad ....

supernovae · 15 Dec 2019, 17:16

yeah, .003 - sorry, I was typing on my phone (i dunno how people do that all the time.. give me a keyboard!! lol)

i'll try cleaning the pin. I just switched to PEI build on my glass and stopped using glu-stick - maybe there is some residue on pin.

Phaedrux · 15 Dec 2019, 17:49

@supernovae said in Advice to fix BLtouch drift?:

maybe there is some residue on pin.

I find wisps of filament start to gather on the pin sometimes.

supernovae · 15 Dec 2019, 19:15

This is what i get now that I try and have it set to A5

G28 Z
Warning: Z probe readings not consistent

supernovae · 16 Dec 2019, 01:12

@Phaedrux said in Advice to fix BLtouch drift?:

@supernovae perhaps time to clean the pin.

is there a way to safely remove & clean pin or just have it deployed and gently clean it deployed?

percar · 16 Dec 2019, 02:26

First lets look at the realistic results that the printer is able to produce
Your z axis has a resolution of 250 steps per 1mm that equates to a resolution of .0025mm now also remember that this is in micro steps which is not really all that accurate chances are the best resolution that you could achieve would be .003mm. BL touch also has a best case repeatability of .01 and actually if you read their documentation it mentions .05 mm. If you are getting better than this it is a bonus. If you are getting resolution better than this think of it as a bonus.

Phaedrux · 16 Dec 2019, 02:34

@supernovae said in Advice to fix BLtouch drift?:

@Phaedrux said in Advice to fix BLtouch drift?:

@supernovae perhaps time to clean the pin.

is there a way to safely remove & clean pin or just have it deployed and gently clean it deployed?

There is a set screw at the top of the probe. You remove that and the pin will come out the top. The set screw is also how the trigger distance for the electromagnet is set.

supernovae · 16 Dec 2019, 15:10

@percar said in Advice to fix BLtouch drift?:

First lets look at the realistic results that the printer is able to produce
Your z axis has a resolution of 250 steps per 1mm that equates to a resolution of .0025mm now also remember that this is in micro steps which is not really all that accurate chances are the best resolution that you could achieve would be .003mm. BL touch also has a best case repeatability of .01 and actually if you read their documentation it mentions .05 mm. If you are getting better than this it is a bonus. If you are getting resolution better than this think of it as a bonus.

I'd be OK getting a reading that was +/- within some margin of error range, but the fact my variance keeps drifting further and further suggests something else. I'd expect my G32 bed probe heights to report more random distributions within the margin of error rather than going from 0.032 to .240 probing the same spot

supernovae · 16 Dec 2019, 20:40

The really odd thing is, if i turn off, and not change anything and power back on, the first read is the same - every time... working on cleaning the bltouch tip to make sure its not friction causing height change

supernovae · 23 Dec 2019, 20:33

Found out my problem. I noticed one side of my printer was always sagging towards end of print, ran motor unlock and i could rotate left z and feel steps, but right z just spun - i guess my right z motor had failed.

Replaced my right z motor, and now bltouch is happy and i don't have to run g29 between each print... kind of surprised i got the quality i did with that motor having been bad or faulty for so long...

but, i'm back in business... no more crazy drift, nice level bed and great prints again. took the time to move to reprap3 firmware and enjoy it too!