Mini IR calibration question How are people getting good calibration?



  • The best I can get is a deviation of 0.232
    How is it people are getting such good calibration using the mini-IR? I've followed all the instructions, my delta is level (tricklaser max metal build)
    my bed.g file is from Escher3d's bed wizard
    and I used dc42's instructions to adjust the H value in my bed.g to make up for any effector tilt.

    My mini IR measurements from Z0 at each test point are consistently repeatable.

    No matter what I do, or where I start from I never get better than 0.232. Running auto calibration 6 times in a row gives me: Deviation before 0.232 after 0.232

    I can't get my bed flat at all.



  • What factor of calibration are you using? I was around your numbers till I did a few iterations of S8. My numbers also improved when I picked up a geckotek bed and stopped using a piece of white paper for the sensor to tap.

    my numbers come out at .070 range, with bed leveling I would imagine a .232 could print flat… how have your prints come out with the .232? At the end of the day numbers are just numbers



  • Run a round using S-1, and post your results. You may be running into an issue I've been trying to track down with calibration on a rostock max v2, this sounds very similar. More details here: http://forum.seemecnc.com/viewtopic.php?f=111&t=11441&sid=425da29f751bdddac4a8921388385d29

    The most likely current culprit is belt tesnion leading to incorrect steps/mm, and you'll get the up/down pattern described there.



  • @kragar I know that my z-tower belt is tighter than x and y, but I haven't been able to get them the same. How does one fix this?

    @jnofsinger: I'm using the recommended 6. I'll try the S8 - maybe I should check out the geckotek bed, too.


  • administrators

    How much effector tilt did you have to correct for? If you have a large amount of it, that will cause calibration difficulties even if you correct the trigger height. See https://duet3d.com/wiki/Calibrating_a_delta_printer.



  • So, it turns out one of the problems is that the onyx bed is way too shiny. I started from scratch - and redid the effector tilt corrections in the dark (the light of my laptop, and the paper test by feel)
    most of the differences were much closer.
    with two at -0.68 and two at -0.38 and the rest quite close.

    Running auto calibration in the dark after 3 attempts my deviation was 0.072 - that's the first time I've been under .2

    I'd still like it to be closer, though.


  • administrators

    Doing calibration in the dark isn't necessary. The IR sensor has a high degree of immunity to ambient light. The only common situation it can't cope with is direct sunlight (no clouds in front of the sun) reflecting off the bed into the sensor.

    A trigger height adjustment of 0.68mm is very high and suggests that your printer has significant geometrical errors, or that the joints don't have sufficient movement at the edges of the bed.



  • sandy - if you're probing onto the onyx, the white paint will definitely throw it off & do weird things. For a while I had black paper between the onyx & the glass. That was OK, but I've since painted the bottom of the glass with high temp spray point (for BBQ grills) and it made my results much more repeatable and consistent.

    I highly recommend it, I did two coats right onto the glass, then baked it in the oven to cure it.


Log in to reply