Microstepping for smart effector



  • I can't seem to get my sensitivity below 40 without false trigger and I can't seem to get my deviation under .06 or so (S6). In the marketing material for this lovely device its mentioned that sensitivity works lower on duet hardware with its sweet sweet microstepping. Which brings me to the question of what settings would benefit my probing results the most? 1.8 steppers, duet wifi, 12v power supply, metal corners, very exactly measured and stable build, but still get get that deviation or sensitivity setting down…


  • administrators

    The first thing to check is whether the false triggering when you increase the sensitivity is caused by noise from the hot end fan. So try probing cold, with the hot end fan not running, and see if that allows greater sensitivity.

    Have you fitted the spirit level to your effector, and if so, does it show a tilt that varies with XY position? Effector tilt is a more likely cause of higher than expected deviation than lack of sensitivity.



  • Yes, I don't have noticable effector tilt, haydn's rods are very accurate and my extrusions are cut very precisely. Your effector and carriages are obv even.



  • Your suggest is to do the probing with a hot bed but cold hotend (so no fan) then heat the hotend and G92 my way to a first layer height to accommodate for expansion from heating? I'm sure this would get me much better results and allow a lower sensitivity, but it'd be a pain, especially if expansion of the hotend varies between PLA and PETG temps, my z offset wouldn't be accurate. That said, I have to set my z= height manually with G92 after each calibration as it is now because my z offset isn't consistent between probes, so I guess I'll try it!


  • administrators

    No. I am suggesting you test probing cold with the fan off, at higher sensitivity, to establish whether it is interference from the hot end fan that is limiting the sensitivity you can use.



  • Indeed with probing with the hotend and fan off, I can get through a 16 point calibration with sensitivity set to 11. Odd, I'm using an original e3d fan but I guess for some reason it generates extra noise. Interesting though that at sensitivity at 11 and visually verifying no early or missed probe hits, my deviation is still at .129 (S6)


  • administrators

    I think there is some other reason for the high deviation then. If you have fitted the spirit level and you can see that there is no observable change in effector tilt with XY position, here are some possibilities:

    • Bed is not flat (check it against a straight edge in several directions, for example a steel rule)
    • Towers not parallel (are you using metal vertices?)
    • Backlash caused by belts or friction. If this is significant, then if you run mesh levelling the height map may show distinct ridges.


  • It's worth pointing out that I can actually print a very nice first layer at .1 layer height across the entire bed, so whatever that deviation represents, it's compensating well for it. I'm using metal corners and 2040, Everything is cut very accurately, my bed is MIG6 aluminium 12mm thick, silicone heater underneath, glass on top, with or without PEI (swap the glass occasionally). It's possible my bed mount holding up the aluminum isn't straight, I checked it with a square to the towers when I set it up, but its been moved since then and it's possible thats out of straight now. I'll measure it up when I get home and see. As always, appreciate the support. I'd like to get that fancy deviation down to what everyone else seems to be experiencing. Should I be using a different S5/6/7/9 etc?


  • administrators

    S6 generally gives good results, but you could try S8 to see if it improves the deviation.

    Are you quite certain that the spirit level doesn't show any tilt at any of the probe points?


Locked
 

Looks like your connection to Duet3D was lost, please wait while we try to reconnect.