Z endstop comparison

Discussion in 'Calibration, Help, and Troubleshooting' started by Alex9779, Sep 15, 2016.

  1. Alex9779

    Alex9779 Moderator
    Staff Member

    Joined:
    Sep 4, 2015
    Messages:
    2,405
    Likes Received:
    731
    Ok so I did some further research about the two Z endstops I have available at the moment.
    I was a little annoyed by the fact that the inductive sensor is temperature dependant so I have to home with the built plate always at the same temperature at which I levelled the bed.
    So if the bed was "cold" I had problems to quickly restart a print. That's why my profiles now have a wait for the bed to reach 50° C also on cool down because that's pretty much the lowest temperature despite of NinjaFlex which I print at 40° C.
    I chose 50° because it only take about 10 minutes for the bed to cool down to this from 85° C, to room temperature it is much much longer. (I had an issue recently where my end of print script didn'T wait to the set 30° C even it should and the print finished at 50°, I am still searching for a cause of this...)

    So now to what I tested with the endstops.
    I switched over to the IR, leveled the bed and after homing I used a feeler gauge to test the distance between nozzle and bed.
    It was pretty nice that the "normal" distance was about 1mm, the feeler gauge had a little resistance between the nozzle and the bed.
    Then I homed and homed and homed. And I can tell the distance varied. Sometimes the distance increased ( i saw an air gap between the gauge and the nozzle), sometimes it decreased (the gauge had a higher resistance). "Sometimes", well I had home sequences where it was greater 10 in a row, then 10 where it was ok, 5 lower, 1 ok, 1 lower, 1 higher, pretty much random...

    So at first I tried to put black tape on the surface. At first it looked as it helped but consecutive homing showed that the results are again random...

    Then I thought it could be the firmware so I was going to exclude as much as mechanical and electrical influences as possible.
    I added a second driver for the Z axis and splitter the Z motors, compiled a firmware for this. Fortunately this is just one line in the config to active a second Z motor driver.
    But also no result, still random height after homing.

    The variance is roughly about +-0.1mm...

    Next try was to lower the approaching speed so I set the divider to 8 with is damn slow to approach to the sensor but also no effect, still random heights after homing...

    Then I tried to exclude the external lighting. Dark room, LEDs switch off, was pretty dark, bed was always cold.
    But also no effect, still that random height after homing...

    So I was out of ideas for the IR and tried the inductive probe again.
    I used the same cable, my setup allows this, then I just change the "safe homing spot".
    And I homed and homed again and measured the height with the feeler gauge.

    And the result was a constant height after each homing sequence. I was not able to "feel" a difference when sliding the gauge in the gap between bed and nozzle...

    So my conclusion is that it is not the firmware or the RUMBA or the cable. It must be the sensor.

    I wanted to order a spare PEI plate anyways soon so I think I am going for the black one and then I will redo the test with the IR but so far my impression is that the IR has some problems.

    So back to the inductive probe for now with the hassle of keeping the bed on the same temp for leveling and homing...
     
    Ephemeris likes this.
  2. Mark Curry

    Mark Curry Well-Known Member

    Joined:
    Jan 17, 2016
    Messages:
    81
    Likes Received:
    6
    Just for fun I tried this on my glass bed. I auto-home and cant get a .05mm feeler gauge under the nozzle (0.05 is the smallest my set has). So then I step +0.1 on the Z and can then get a 0.13 to slide under the nozzle, but the 0.15 drags slightly on the gauge (I am guessing the nozzle height is actually around 0.135-0.140). I repeated 10 times and found the same every time. Exactly the same. Some times I leave Z at 0.10 sometimes I set it to 0.0, but irrespective after auto-home the height is exactly the same. Wish I had a larger set of feelers to see that the height is exactly at auto-home.

    Perhaps the IR reflection is better off a polished surface rather than coated?.
     
  3. Stian Indal Haugseth

    Stian Indal Haugseth Well-Known Member

    Joined:
    Sep 11, 2015
    Messages:
    573
    Likes Received:
    95
    I think you both are on to something here. Thinking of spreading of light there will be more "spillage" of light with a diffuse surface. Sharper light reflecting the better.

    But this effectively ruins using any kind of treatment on the glass.

    Someone should try to make a laser sensor angled like the IR but with dual slits at a distance to stop any light beams astray.
     
  4. Alex9779

    Alex9779 Moderator
    Staff Member

    Joined:
    Sep 4, 2015
    Messages:
    2,405
    Likes Received:
    731
    Ya I think my problem is the silver PEI alu plate. I don't doubt that you get pretty steady homing with the glass, I had no problems so much with it too.
    But for me it shows that the conditions for the IR are really limited...
     

Share This Page