1. Got a question or need help troubleshooting? Post to the troubleshooting forum or Search the forums!

Unresolved Autolevel issues

Discussion in 'Troubleshooting' started by cosber, Aug 14, 2015.

  1. cosber

    cosber Active Member

    Joined:
    Jun 20, 2013
    Messages:
    300
    Likes Received:
    87
    Since I got my machine back from Robo for the rod upgrade, my autolevel is acting strange. When it does the 9 point calibration, on point #3, the front right point, the print head doesn't come down all the way. The other 8 points do fine, (although it does seem like it pushes down the plate alot on the back end). Because of this, I'm getting good layer in the front of the print, but the back is too low. If I let it run, as the head rises, the back end fills in after about 5-6 layers. I'm assuming it has to do with that front #3 calibration point not coming down all the way to the bed. Any suggestions?
     
    #1 cosber, Aug 14, 2015
  2. mark tomlinson

    mark tomlinson ༼ つ ◕_ ◕ ༽つ
    Staff Member

    Joined:
    Feb 21, 2013
    Messages:
    23,915
    Likes Received:
    7,338
    It should come down until the nuts are thrown (the switches on the Z nuts triggered).

    Perhaps those are triggering incorrectly? Did you review the autolevel sticky thread?
     
    #2 mark tomlinson, Aug 14, 2015
  3. Mike Kelly

    Mike Kelly Volunteer

    Joined:
    Mar 11, 2013
    Messages:
    6,967
    Likes Received:
    2,277
    Do a paper leveling
     
    #3 Mike Kelly, Aug 14, 2015
    Frankn likes this.
  4. Frankn

    Frankn Member

    Joined:
    Mar 1, 2015
    Messages:
    261
    Likes Received:
    22
    First, it can only level so much. What you are really doing is compensating for lack of parallelism between the X carriage shafts and the work bead.
    Measure the height of the rods on each side from the work bed. adjust them manually by lifting the carriage and turning the nuts until you get it as close as you can. Then follow Mikes suggestion and configure them with the paper test. Frank
     
    #4 Frankn, Aug 14, 2015
  5. cosber

    cosber Active Member

    Joined:
    Jun 20, 2013
    Messages:
    300
    Likes Received:
    87
    As far as the rods, I would think that all 3 right side calibration points would be off, not just the one. I haven't had a chance to do the paper leveling again. I'll give that a shot first when I am able. It's funny though how the machine was "perfectly calibrated" when Robo installed the new rods.
     
    #5 cosber, Aug 17, 2015
  6. Frankn

    Frankn Member

    Joined:
    Mar 1, 2015
    Messages:
    261
    Likes Received:
    22
    You learn that things "SEAT" after a while. No adjustment on a 3D printer is permanent. The amount of glue or hair spray can make a difference. I just ran 2 days straight with perfect prints. After a days rest, prints were popping loose, you figure! I had to wash the board, respray, paper test, and reset Z offset to -.06. Back to perfect prints. Yes, I paper test after I spray and it dries. Don't paper test until I have problems again. Frank
     
    #6 Frankn, Aug 17, 2015
  7. WheresWaldo

    WheresWaldo Volunteer ( ͠° ͟ʖ ͡°)
    Staff Member

    Joined:
    Feb 18, 2015
    Messages:
    5,905
    Likes Received:
    3,593
    @Frankn I am going to suggest something too you because I know you will understand what I am saying, and perhaps this can help @cosber also. Stop playing with zprobe_offset (M565)! You come from a background where you worked on machines that required precision. So how about this, think of zprobe_offset as a way to assure that a primary layer height is exactly the layer height you asked for, not the layer height you want. Let's say you want a 0.2 mm layer. Most people will ask for a 0.23 - 0.3 mm layer and then use zprobe_offset to squeeze it down until it measures 0.2 mm. While this approach works, your printer's brain (and the slicer that you used) still thinks the layer is 0.23 - 0.3 mm in height. So immediately you have introduced an error in your model size. When you finish printing you scratch your head wondering why your 20 mm cube in only 19.7 - 19.8 mm tall. then your plastic shrinks (all plastics shrink as they cool)* some over 1% in all directions and now your 20 mm cube is 19.6 mm high or shorter. At least give your print a chance to be correct. if you want 0.2 mm layers then ask for 0.2 mm layers and adjust the flow rate for the first layer to push more plastic than in subsequent layers. Almost all slicers allow this, you end up with a 0.2 mm layer height and a squished first layer that adheres perfectly well to the bed.

    The common practice that most people use would be like you servicing a copier paper feed mechanism knowing you need 0.1 mm clearance and setting it to 0.2 mm on purpose because it is going to be used during the summer and you want to account for thermal expansion. You get my drift.

    * Only one compound in existence expands when it cools and this is significant. Water as it becomes ice (4° - 0° C) actually increases in volume and becomes less dense, that is why ice floats.
     
    #7 WheresWaldo, Aug 18, 2015
  8. Frankn

    Frankn Member

    Joined:
    Mar 1, 2015
    Messages:
    261
    Likes Received:
    22
    _wire snake.jpg Waldo, Wow, that was a long explanation.
    I use Z offset to get a flat smooth bottom layer that sticks well to the bed. I am presently making one of those 'snake' wire looms for the X carriage wiring. It requires tight dimensions and it is working out great. Most of my works like raptor claws are large, but rough in nature.
    Yes, I am use to setting things up with a micrometer and VOM, not a computer code. I live and learn as I go along. My main goal now is making Robo more user friendly. I just made a set of 1" high feet with a layer of silicone and cork to adsorb the vibration and sound transmission. The clear acrylic sides are in the works also. More light inside and quick access for either side. Just having fun. lol Frank
     
    #8 Frankn, Aug 18, 2015
    Last edited: Aug 21, 2015
  9. cosber

    cosber Active Member

    Joined:
    Jun 20, 2013
    Messages:
    300
    Likes Received:
    87
    When you do the 9 step autolevel, how far does the head push down on the plate before it goes to the next point? Shouldn't it just touch the plate to get its calibration? Mine pushes down a good (I'm guessing) 10 mm on most of the points. I recall it has always done this, though not as much, even when I was getting good prints. I did the paper level again and at least all the points are being measured now. I haven't gotten the calibration right though and I haven't messed with it because I get so frustrated when I take it in for repair or upgrade, and when it comes back, despite being told it is perfectly calibrated, it's not. I'd rather be flying my drone.
     
    #9 cosber, Aug 28, 2015
  10. WheresWaldo

    WheresWaldo Volunteer ( ͠° ͟ʖ ͡°)
    Staff Member

    Joined:
    Feb 18, 2015
    Messages:
    5,905
    Likes Received:
    3,593
    10 mm is really a lot, more likely the flex in the build plate is closer to .5 - 1 mm. It does a double tap kind of dance, The first time it hits one of the nine points it just looks for the trigger, the second time which is much slower, it takes the measurement. If you have a terminal window open, it will show the results of the auto-tramming. The tramming is surprisingly robust and can account for a significant amount of offset from one point to another, but it only measures the points it touches. If your build plate has a wave in it, it can miss something in the middle. This is why it is a good practice to get the build plate as parallel as possible to the X/Y axis.
     
    #10 WheresWaldo, Aug 28, 2015
  11. Frankn

    Frankn Member

    Joined:
    Mar 1, 2015
    Messages:
    261
    Likes Received:
    22
    The nozzle just touches, or sits on the plate until the carriage lifts off the switch. It really doesn't push on the plate.
    This switch action tells it to stop driving the carriage down then as Waldo implies, the code triggers the second dip to get the actual height. Frank
    hand print-2_edited-3.jpg
     
    #11 Frankn, Aug 29, 2015

Share This Page