Magnets + Z-Probe

= FAIL

The auto bed leveling readings just didn't seem right. So I did a test. I ran g29. Then added a magnet in the far right corner. Ran G29. Removed the magnet. Ran G29. Here are my results:

(pay attention to column 4, row 3)

1st Run: no magnet:

SENDING:G29
Bilinear Leveling Grid:
      0      1      2      3      4
 0 +0.795 +0.362 +0.227 +0.553 +1.102
 1 +0.438 +0.102 +0.012 +0.250 +0.765
 2 +0.025 -0.370 -0.442 -0.205 +0.305
 3 -0.573 -0.952 -1.102 -0.843 -0.365

2nd Run: with magnet:

SENDING:G29
Bilinear Leveling Grid:
      0      1      2      3      4
 0 +0.827 +0.370 +0.227 +0.555 +1.125
 1 +0.462 +0.105 -0.005 +0.235 +0.775
 2 +0.032 -0.382 -0.462 -0.220 +0.312
 3 -0.578 -0.963 -1.132 -0.860 -0.645

3rd Run: no magnet:

SENDING:G29
Bilinear Leveling Grid:
      0      1      2      3      4
 0 +0.838 +0.365 +0.220 +0.553 +1.135
 1 +0.467 +0.107 -0.005 +0.232 +0.780
 2 +0.035 -0.382 -0.462 -0.223 +0.307
 3 -0.575 -0.963 -1.132 -0.857 -0.360

So magnetic flex plate is out.

andrew leroy

Sep 5Edited: Sep 5

I removed all the magnets, taped the flexplate to my bed, and ran another 3 consecutive G29 commands:

>>> g29
SENDING:G29
Bilinear Leveling Grid:
      0      1      2      3      4
 0 +0.818 +0.427 +0.365 +0.588 +1.135
 1 +0.452 +0.090 -0.025 +0.247 +0.748
 2 -0.055 -0.482 -0.555 -0.325 +0.212
 3 -0.610 -1.032 -1.152 -0.945 -0.448
>>> g28
SENDING:G28
>>> g29
SENDING:G29
Bilinear Leveling Grid:
      0      1      2      3      4
 0 +0.805 +0.413 +0.357 +0.595 +1.117
 1 +0.442 +0.080 -0.035 +0.237 +0.738
 2 -0.048 -0.473 -0.555 -0.322 +0.195
 3 -0.623 -1.035 -1.165 -0.950 -0.455
>>> g28
SENDING:G28
>>> g29
SENDING:G29
Bilinear Leveling Grid:
      0      1      2      3      4
 0 +0.805 +0.413 +0.355 +0.590 +1.112
 1 +0.438 +0.080 -0.037 +0.237 +0.738
 2 -0.045 -0.467 -0.553 -0.320 +0.202
 3 -0.623 -1.030 -1.160 -0.950 -0.450

[edit]

With these numbers saved, I think I should be alright with the magnets back in place. Just need to make sure there isn't one near the home location. ...and hope I don't need to collect new readings any time soon.

Michael G

Sep 5

So are you saying that the magnets interfere with the probe? I'm wondering if it's the magnets pulling the flex plate to a not true plate. Flexing could be all over the place. I need to take pictures but mine changes by .75mm over the course of a cm or 2. It's ridiculous.

Oh, I also got a 310x310x3mm borosilicate glass plate to build on. The inductive probe is not able to make the difference! UGH, it's supposed to be 4mm for the trigger point.

andrew leroy

Sep 5Edited: Sep 5

Exactly. You can see in my OP column 4, row 3 there's a 0.3 (mm?) difference with the magnet added directly under that probe point. And in my second set of readings you can tell where the magnets were as some points are almost identical to those in my 1st set, while others off by a deal more.

[edit]Some of that may be due to how the magnets hold the flexplate compared to the tape. But I clearly showed that magnets do affect the probe readings. With a bit more testing I could probably tell the extents of the affect.

My bed mapping is better with the magnets off; but, unfortunately, the flex plate has a little warp to it. The tape just can't hold it flat enough for a better reading. I made my own heated bed out of an 1/4 in aluminum sheet and some nichrome wire. I might try printing directly on the aluminum.

Michael G

Sep 5

Ok. Gotcha. I'm thinking of replacing the inductive probe. Haven't decided to make a BLTouch mount or get the 8mm one. Goal being to use a glass bed.

andrew leroy

Sep 5

BLTouch looks promising. You seem to do your research so I assume marlin has some way to integrate the extra wire for the control signal and enable/disable it appropriately.

I'm not all that impressed with the inductive sensors. There's too many variables that can affect its accuracy.

Michael G

Sep 5

I'm not running the MKS Gen L v2.0 board in my printer. Mine died. I have a BTT SKR 1.4 Turbo. It's very similar to the MKS Gen L.

You can still hook up the BLTouch. There's a servo port and you can use the z-axis limit switch port to trigger it. I think, though, after looking at the board that the "AUX 1/2" would also work. The z-axis limit switch is the easier way to go but I want to say I remember that there is an advantage to using another pin.

TeachingTech did a video on this board in the Ender 3.

Also, here's the docs on the board: https://github.com/makerbase-mks/Datasheet/blob/master/English%20datasheet/MKS%20Gen_L%20DataSheet.pdf

If you need firmware compiled with it setup just let me know. I don't think this is too hard to do.

andrew leroy

Sep 6

Thanks Michael.

If I can print Gareth's Y-axis stabilizer laying flat, I'm good with what I got. ...for now. I'm considering going glass at some point. Been reading a lot of hype for it. I'll probably get a probe upgrade at the same time. I'm sure marlin can handle it. Just a matter of doing my homework.