Y axis confused: will move beyond 200... even to 237mm

Discussion in 'BigBox General Chat' started by R Design, Mar 6, 2016.

  1. R Design

    R Design Well-Known Member

    Joined:
    Oct 7, 2015
    Messages:
    918
    Likes Received:
    183
    Motion system was working perfectly.

    Then at some point head suddenly crashes into back wall. Cut power and then noticed after an auto-home nozzle was forward of the centre.

    If - using Rumba knob - I move the print head forward manually it goes perfectly to Y=0; but if I move it backwards, it will move beyond Y=200: near the ooze bucket I've been able to go as high as y=237

    The level bed command works OK and goes to a max Y=157mm

    After an auto-home X=150 but Y=68, every time.

    What is going on?

    And for a while there were "Err: too far!" errors displaying after running commands.
     
  2. R Design

    R Design Well-Known Member

    Joined:
    Oct 7, 2015
    Messages:
    918
    Likes Received:
    183
    Thinking that perhaps the Y axis parameter had become corrupted in memory also tried "Failsafe Restore" which apparently refers to the original config.h file.

    No improvement.
     
  3. Greg Holloway

    Greg Holloway Administrator
    Staff Member

    Joined:
    Sep 4, 2015
    Messages:
    565
    Likes Received:
    397
    hi,

    i'm a little confused as to what the problem is.

    the max travel for Y is set to 240 to give the head enough space to dock over the ooze bucket.

    the head will be at 150 by 68 as that is the position the nozzle will be after auto home. remember there is an offset between the nozzle and the probe.

    do you mean that you can not move further back than 157?
     
  4. R Design

    R Design Well-Known Member

    Joined:
    Oct 7, 2015
    Messages:
    918
    Likes Received:
    183
    Thanks Greg.

    The head goes all the way back and have now - in desperation - successfully launched a print and been watching it. Second benchy and, now the nozzle properly adjusted, X belts re-tensioned and X grub screw tightened, it's going down PERFECTLY and nearly finished!

    OK so what happened above?

    Is it possible that the head crashed into the back because - perhaps - I did a "bed level" having moved the head manually (and not done another auto home first)?

    That that then spooked me and I thought there was something wrong with the Y axis? I was convinced the print now underway would not go down in the middle of the bed and yet there it is.

    Having a bit of a job understanding what normal behaviours are for the Rumba / Marlin. Couldn't find an online guide / forums?

    ps would love to level the glass bed (no instructions on that, nor on tightening it up) and to understand optimal fan height (I remember you said 0.5mm but that seems risky because of the height of the bed clips).
     
  5. Greg Holloway

    Greg Holloway Administrator
    Staff Member

    Joined:
    Sep 4, 2015
    Messages:
    565
    Likes Received:
    397
    the important thing to remember is that it is really quite difficult to actually break it physically, short of taking a hammer to it of course!

    is something goes wrong kill the power, check things over, make sure the print area is clean, turn it back on, run auto home and then everthing should be back to normal again.

    knowing what to do comes with experience, and you can only get experience by doing!

    with regards levelling the bed move the nozzle to a corner, raise the bed through the prepare > move axis > 0.1mm > Z, place a bit of paper under the nozzle and adjust the height so you can move the paper freely and still feel a little resistance from the nozzle touching it. move the head to another corner, you might need to disable the motors (prepare menu) to move in X&Y btw. this time loosen the corner and adjust the height of the bed by moving the pillar up and down. adjust it with the paper to get the right height. do this for each corner, usually i do each one twice. after the initial adjustment using the lcd do them all manually.

    it will take a bit of practice but it's pretty straightforward. Once your done run autohome, check the z offset, do an auto level if you like then go print!
     
    R Design likes this.
  6. R Design

    R Design Well-Known Member

    Joined:
    Oct 7, 2015
    Messages:
    918
    Likes Received:
    183
    Great!

    I think I can do that. ;)

    Is it me or is there is something erratic about Marlin.

    Having successfully completed bench 2 and examined the ship's "portholes" decided to tighten Y motor grub and redo.

    Run autohome: works as expected.

    Run level bed: it moves to the front right corner, only does not go as far right as expected... (With hindsight I should have stopped it there)

    Then when it gets to the back left corner it goes a bit too far, such that the IR sensor is off the back of the bed.

    That has for predicted consequence a nozzle crash into the bed before I can get to the kill switch.

    Power on, autohome, bed level - all OK.

    Why can't I trust Marlin?
     
  7. R Design

    R Design Well-Known Member

    Joined:
    Oct 7, 2015
    Messages:
    918
    Likes Received:
    183
    Yes, that is repeatable (I just repeated it since writing the above - now doing a 1cm cube):

    - print finishes and printhead is in ooze bucket;

    - auto-home; (works as expected)

    - level bed; (moves to front right, but does not go far enough to the right....)

    Is this just known general buggyness in Marlin?

    (I have no idea how reliable our build is.)
     
  8. Mike Kelly

    Mike Kelly Volunteer

    Joined:
    Mar 7, 2014
    Messages:
    444
    Likes Received:
    89
    Is the cube dimensions correct? This may have been asked already but did you verify the pulleys for the X and Y motor are correct?
     
  9. R Design

    R Design Well-Known Member

    Joined:
    Oct 7, 2015
    Messages:
    918
    Likes Received:
    183
    Thanks Mike.

    The thread was really about odd behaviour in Marlin, the fact that it does some odd things, in a way that can be predictable.

    I was sort of wondering how buggy it actually is?

    Yesterday I discovered another odd behaviour: if the carriage touches the back wall - and presumably the belt slips slightly - then when it comes forward to the front of the box it will IGNORE THE ENDSTOP. In other words, it seems that it's not always "listening" to the input from the microswitch, that sometimes it just charges ahead.

    As for the 10mm cube, it was 10.2mm in X and Y directions. In the scheme of things I could certainly live with that for prototyping, but I think people are doing much better? What should I be aiming for?
     
  10. Mike Kelly

    Mike Kelly Volunteer

    Joined:
    Mar 7, 2014
    Messages:
    444
    Likes Received:
    89
    I'm not sure I get what you mean by buggy. I have noticed a bug with the manual control menu where it would sometimes jump around and if you keep moving it will cause marlin to freeze.

    But ignoring the endstops is expected. The machine assumes no loss of steps so it only uses the endstops during homing. It's also only an 8 bit processor so it's hard to know if it could handle the added check in the process loop.

    10.2 is pretty good and could be chocked up to measurement error. It's also pretty small in general so errors amplify
     
  11. R Design

    R Design Well-Known Member

    Joined:
    Oct 7, 2015
    Messages:
    918
    Likes Received:
    183
    I've been really impressed by the precision of the BB.

    However it seems that the interior of holes enjoys twice the error of the external dimensions of a part.

    So if you have a 10.2mm cube (10mm); then your holes will be out by 0.4mm (2 * 0.2mm) on the internal diameter.

    Do you see the same?
     

Share This Page