›

autoleveled delta is always .5mm high on -x

HenryLoenwindHenryLoenwind
in Repetier-Firmware
So I have a nice and standard delta (Anycubic Kossel) with a bltouch. I run the autoleveling (G32 S2), no distortion correction, and as result the print is about .5 over the bed on the -x side and scraped into the bed on +x. I tried raising the -x side, but the result is the same (so the autoleveling is working, it's just not giving me the result I want...

I think I'm doing something wrong here, I just have no idea what. :(

Version: the dev from the configurator
Endstop offsets from G132 

The last G32 run:
4:41:09.311: Z-probe:7.505 X:-60.00 Y:-16.00
4:41:15.604: Z-probe:7.573 X:-60.00 Y:-0.40
4:41:21.921: Z-probe:6.268 X:-60.00 Y:15.20
4:41:28.256: Z-probe:7.868 X:-60.00 Y:30.81
4:41:34.599: Z-probe:7.965 X:-60.00 Y:46.41
4:41:40.971: Z-probe:8.057 X:-60.00 Y:62.00
4:41:47.682: Z-probe:7.135 X:-36.00 Y:-16.00
4:41:53.935: Z-probe:7.253 X:-36.00 Y:-0.40
4:42:00.214: Z-probe:7.413 X:-36.00 Y:15.20
4:42:06.501: Z-probe:7.520 X:-36.00 Y:30.81
4:42:12.807: Z-probe:7.675 X:-36.00 Y:46.41
4:42:19.128: Z-probe:7.795 X:-36.00 Y:62.00
4:42:25.805: Z-probe:6.808 X:-12.00 Y:-16.00
4:42:32.021: Z-probe:6.960 X:-12.00 Y:-0.40
4:42:38.257: Z-probe:7.115 X:-12.00 Y:15.20
4:42:44.519: Z-probe:7.313 X:-12.00 Y:30.81
4:42:50.798: Z-probe:7.470 X:-12.00 Y:46.41
4:42:57.095: Z-probe:7.625 X:-12.00 Y:62.00
4:43:03.743: Z-probe:6.543 X:12.00 Y:-16.00
4:43:09.915: Z-probe:6.655 X:12.00 Y:-0.40
4:43:16.113: Z-probe:2.592 X:12.00 Y:15.20
4:43:22.331: Z-probe:7.003 X:12.00 Y:30.81
4:43:28.571: Z-probe:7.192 X:12.00 Y:46.41
4:43:34.835: Z-probe:7.385 X:12.00 Y:62.00
4:43:41.417: Z-probe:6.035 X:36.00 Y:-16.00
4:43:47.543: Z-probe:6.268 X:36.00 Y:-0.40
4:43:53.694: Z-probe:6.480 X:36.00 Y:15.20
4:43:59.871: Z-probe:6.678 X:36.00 Y:30.81
4:44:05.191: Z-probe:5.420 X:36.00 Y:46.41
4:44:11.418: Z-probe:7.095 X:36.00 Y:62.00
4:44:17.933: Z-probe:5.542 X:60.00 Y:-16.00
4:44:23.991: Z-probe:5.788 X:60.00 Y:-0.40
4:44:30.094: Z-probe:6.060 X:60.00 Y:15.20
4:44:36.219: Z-probe:6.280 X:60.00 Y:30.81
4:44:42.374: Z-probe:6.540 X:60.00 Y:46.41
4:44:47.685: Z-probe:5.320 X:60.00 Y:62.00
4:44:47.687: plane: a = -0.0149 b = 0.0093 c = 6.5683
4:44:47.698: Transformation matrix: 0.999890 0.000000 0.014858 0.000139 0.999956 -0.009334 -0.014857 0.009335 0.999846
4:44:47.698: Printer height:296.80
4:44:47.698: CurrentZ:6.57 atZ:7.10
4:44:47.713: Info:Autoleveling enabled

This was done with the severely raised bed as described above. And this was the run before that, only slightly raised bed:

4:29:53.289: X:0.00 Y:0.00 Z:297.400 E:224.9375
4:30:05.298: Z-probe:7.583 X:-60.00 Y:-16.00
4:30:11.601: Z-probe:7.653 X:-60.00 Y:-0.40
4:30:17.932: Z-probe:7.820 X:-60.00 Y:15.20
4:30:24.282: Z-probe:7.978 X:-60.00 Y:30.81
4:30:30.639: Z-probe:8.077 X:-60.00 Y:46.41
4:30:37.013: Z-probe:8.200 X:-60.00 Y:62.00
4:30:43.750: Z-probe:7.253 X:-36.00 Y:-16.00
4:30:50.023: Z-probe:7.398 X:-36.00 Y:-0.40
4:30:56.318: Z-probe:7.538 X:-36.00 Y:15.20
4:31:02.641: Z-probe:7.785 X:-36.00 Y:30.81
4:31:08.981: Z-probe:7.938 X:-36.00 Y:46.41
4:31:15.338: Z-probe:8.097 X:-36.00 Y:62.00
4:31:22.049: Z-probe:7.060 X:-12.00 Y:-16.00
4:31:28.303: Z-probe:7.225 X:-12.00 Y:-0.40
4:31:34.581: Z-probe:7.415 X:-12.00 Y:15.20
4:31:40.872: Z-probe:6.108 X:-12.00 Y:30.81
4:31:47.201: Z-probe:7.813 X:-12.00 Y:46.41
4:31:53.543: Z-probe:7.963 X:-12.00 Y:62.00
4:32:00.222: Z-probe:6.798 X:12.00 Y:-16.00
4:32:06.443: Z-probe:6.978 X:12.00 Y:-0.40
4:32:12.696: Z-probe:7.190 X:12.00 Y:15.20
4:32:18.971: Z-probe:7.445 X:12.00 Y:30.81
4:32:25.266: Z-probe:7.630 X:12.00 Y:46.41
4:32:31.596: Z-probe:7.825 X:12.00 Y:62.00
4:32:38.225: Z-probe:6.438 X:36.00 Y:-16.00
4:32:44.399: Z-probe:6.653 X:36.00 Y:-0.40
4:32:50.614: Z-probe:6.925 X:36.00 Y:15.20
4:32:56.861: Z-probe:7.205 X:36.00 Y:30.81
4:33:03.135: Z-probe:7.420 X:36.00 Y:46.41
4:33:09.438: Z-probe:7.653 X:36.00 Y:62.00
4:33:16.008: Z-probe:5.940 X:60.00 Y:-16.00
4:33:22.135: Z-probe:6.285 X:60.00 Y:-0.40
4:33:28.316: Z-probe:6.650 X:60.00 Y:15.20
4:33:34.524: Z-probe:6.913 X:60.00 Y:30.81
4:33:40.755: Z-probe:7.112 X:60.00 Y:46.41
4:33:47.025: Z-probe:7.403 X:60.00 Y:62.00
4:33:47.026: plane: a = -0.0091 b = 0.0127 c = 7.0232
4:33:47.037: Transformation matrix: 0.999958 0.000000 0.009107 0.000116 0.999919 -0.012710 -0.009106 0.012710 0.999878
4:33:47.037: Printer height:297.33
4:33:47.038: CurrentZ:7.02 atZ:7.10
4:33:47.053: Info:Autoleveling enabled

(I'm starting to hate autoleveling. There I have my P802, which I manually leveled 2 months(!) ago, and then there are the 2 printers that aren't that stable and need autoleveling...)

Comments

  • RepetierRepetier
    First you should use Repetier-Host bed height map feature and make a map after leveling. Hard to see from the result patterns but I think you have reagion dependend height differences, not a linear error from left to right. This makes autoleveling complicated and not working 100% as expected.

    Reason is deltas are nonlinear beasts and one wrong distance can cause waves and bumps over bed position. So everyone is cursing when it comes to first time calibrating a delta. You must calibrate geometry first and then autolevel and if that is not enough enable distortion correction in firmware to solve last bumps. 

    For calibration you might google "escher delta calibration" whcih is a online tool that converts measuring results into geometry updates. Use 6 point correction.
  • HenryLoenwindHenryLoenwind
    edited April 2017
    Oh, height map. I read about that before, never seen it in the program......ok, found it. Ran it. Got the nozzle rammed into the bed. (At least it was cold...)

    Update: Strike this, I had the print radius still at 88 instead of 90.


  • HenryLoenwindHenryLoenwind
    Um, how do I convert the G30 reading into "Nozzle height error"? I'm quite sure my A tower is not 20° off...

    Here's the height map: http://i.imgur.com/yYZjQnM.png Doesn't tell me anything.
  • RepetierRepetier
    It says it measured 4mm difference. Use clipboard to copy the values as a grid into excel or texteditor then you see the values for the colors. Guess one measurement was wrong.

    One important thing is to never stop printer during during measurement. If you do restart host a sit gets the result count wrong then.

    ymin -16 is not much but I guess it is from z probe offset and required.
  • HenryLoenwindHenryLoenwind
    Again, how do I convert the G30 reading into "Nozzle height error" for the escher calculator? Or can I use that calculator at all with this probe (47mm offset)?
  • HenryLoenwindHenryLoenwind
    ok, I think I got it. I replaced the bltouch with a micro switch on the nozzle for calibration. The value for the escher calculator is the G30 value minus 32. No idea why, but that works.

    Now I have this:

    image

    X: -60.00 -46.67 -33.33 -20.00 -6.67 6.67 20.00 33.33 46.67 60.00  
    y:60.00 48.95 48.88 48.86 48.87 48.86 48.89 48.90 48.91 48.92 48.95      
    y:46.67 48.93 48.91 48.88 48.86 48.88 48.89 48.90 48.91 48.93 48.93      
    y:33.33 48.91 48.88 48.85 48.85 48.86 48.89 48.91 48.93 48.93 48.92      
    y:20.00 48.90 48.86 48.85 48.85 48.86 48.89 48.91 48.95 48.94 48.95      
    y:6.67 48.90 48.84 48.84 48.86 48.86 48.90 48.93 48.95 48.93 48.94      
    y:-6.67 48.88 48.86 48.86 48.86 48.88 48.91 48.94 48.94 48.91 48.91      
    y:-20.00 48.86 48.85 48.85 48.88 48.90 48.93 48.95 48.93 48.91 48.88    
    y:-33.33 48.86 48.84 48.89 48.89 48.90 48.94 48.94 48.92 48.90 48.83    
    y:-46.67 48.85 48.83 48.88 48.91 48.93 48.96 48.95 48.94 48.89 48.86    
    y:-60.00 31.96 48.88 48.89 48.90 48.96 48.96 48.96 48.94 48.91 48.80    


    I think the first measurement is garbage because Repetier-Host runs it with G1 P1 and runs the other ones with P0.

    Next I'll switch back to the bltouch and see how the height map looks there.

    PS: Why does Repetier-Host try to measure from z-max? That's a long way down at 0/0 and a needs much stronger motors to crash through the endstops at any other xy...
  • RepetierRepetier
    G30 P1 is just P1 = enable z probe. Last point has P2 = disable z probe. Should not change result.

    For height map you need to go to Z10 or 20 before starting it. It does not change start height.
  • HenryLoenwindHenryLoenwind
    10 or 20? I see a difference of 17 between the first measurement (which is taken from whatever height G30 chooses) and the rest (which are taken from the height Repetier-Host send in the G1 command, 50 in this case). And there's is a difference between P1 and P2/P0. I see it in the raw output, too.

    Little update: I reattached the bltouch and ran a height map. it measured a 2mm difference between points. Which, after the nice height map I had with a centered probe, seemed very wrong. So I ignored it (disabled autolevel) and only finetuned z-length. The bed was spot-on level and has only drifted to about .2mm since. Which still is too much for a nice first layer without raft, but much better than what a leveling with the bltouch would have given me.

    I think I'll go back to the manually attached micro switch for leveling. That 47mm offset from the nozzle seems to be just too much for a usable result.

    But for now I stopped worrying about the bed, as the motors started losing steps again. The skipped distance got smaller and smaller as I increased the current again and again over the last weeks, but now I'm approaching max current :-(
  • RepetierRepetier
    Ok, I know the problem now. You have a variably saying what max. probe distance to bed should be. First probe enforces it and host goes back to higher z for the rest. So you need to start with Z < that value. It is a variable like z probemax  z bed distance.
Sign In or Register to comment.