Welcome to Our Community

Some features disabled for guests. Register Today.

accuracy issue ooznest ox (what am i doing wrong)

Discussion in 'CNC Mills/Routers' started by tore micaelsen, Mar 7, 2016.

  1. tore micaelsen

    Builder

    Joined:
    Feb 25, 2016
    Messages:
    27
    Likes Received:
    1
    Hi all!

    Just buildt my ooznest ox and it looks great.
    Going to be using this to engrave stuff, but im getting a issue with text not appearing as it should.

    You see the issue good on the 0's and the = sign. It doesnt stop where it starts.
    Have steps/mm at 26.67 with 1/8 microstepping.
    have configured max feedrate at as slow as 500 now, and accel at 50.
    Have tried loose belts and very tight belts, no difference.

    Anyone have any idea what could be causing this?

    -Tore
     
  2. Bad Sequel

    Bad Sequel Well-Known
    Builder

    Joined:
    Aug 19, 2015
    Messages:
    83
    Likes Received:
    32
    Might be your belt pulleys slipping on the motor shafts. Double check the nuts holding the pulley to the shaft.
     
  3. tore micaelsen

    Builder

    Joined:
    Feb 25, 2016
    Messages:
    27
    Likes Received:
    1
    They are firm and tight, with the set screw directly on the flat part of the shaft.
     
  4. David the swarfer

    David the swarfer OpenBuilds Team
    Staff Member Moderator Builder Resident Builder

    Joined:
    Aug 6, 2013
    Messages:
    3,239
    Likes Received:
    1,815
    that is definitely backlash, 'something' is loose and shifting when the carriage reverses direction.
     
  5. tore micaelsen

    Builder

    Joined:
    Feb 25, 2016
    Messages:
    27
    Likes Received:
    1
    Looks to me like it has to be in X, you agree?
     
    GrayUK likes this.
  6. David the swarfer

    David the swarfer OpenBuilds Team
    Staff Member Moderator Builder Resident Builder

    Joined:
    Aug 6, 2013
    Messages:
    3,239
    Likes Received:
    1,815
    yes, looks like X.
     
  7. Jimmybuckets

    Jimmybuckets Well-Known
    Builder

    Joined:
    Aug 15, 2014
    Messages:
    234
    Likes Received:
    36
    Looks to be identical in all the attempts? Are you having issues with other cut files? Backlash usually doesnt repeat identically over and over again. It would be more random. Try rotating it and cutting it in the Y and see what happens.
     
    #7 Jimmybuckets, Mar 8, 2016
    Last edited: Mar 8, 2016
  8. tore micaelsen

    Builder

    Joined:
    Feb 25, 2016
    Messages:
    27
    Likes Received:
    1
    identical... replaced the set screws with normal m3 hex, and double checked that its firmly in place, and its stuck, tried loosening the belt, or tighten the belt, same trouble.
    This is only run with the same cut file, will try another now.
     
  9. Jimmybuckets

    Jimmybuckets Well-Known
    Builder

    Joined:
    Aug 15, 2014
    Messages:
    234
    Likes Received:
    36
    Try the same file but rotate so it cuts along the Y.
     
  10. tore micaelsen

    Builder

    Joined:
    Feb 25, 2016
    Messages:
    27
    Likes Received:
    1
    Ran a vertical test looks like the "=" sign has the same issue that direction too :)

    [​IMG]
    20160308_144159.jpg
     
  11. Jimmybuckets

    Jimmybuckets Well-Known
    Builder

    Joined:
    Aug 15, 2014
    Messages:
    234
    Likes Received:
    36
    Might be your font my brotha
     
  12. tore micaelsen

    Builder

    Joined:
    Feb 25, 2016
    Messages:
    27
    Likes Received:
    1
    Have used that font before on a mpcnc mill and it turned out great. Difference there i gress is the steps / mm. Gt3 on ox and i had Gt2 on the mpcnc.
    Does it matter that the steps/mm get decimals? The ox har 26.67 on X and Y
     
  13. Jimmybuckets

    Jimmybuckets Well-Known
    Builder

    Joined:
    Aug 15, 2014
    Messages:
    234
    Likes Received:
    36
    Try just drawing a few circles of known size. Do some squares as well. Some a something like 2", 3" ,4",5" and see how they come out.
     
  14. Jimmybuckets

    Jimmybuckets Well-Known
    Builder

    Joined:
    Aug 15, 2014
    Messages:
    234
    Likes Received:
    36
    If you steps per mm were off they would be off for everything. Your **** would just come out the wrong size. Start at your bit and trace backward. Grab the bit collet and move it around (make sure its unplugged). Make sure it is snug and that something in your z isn't loose.
     
  15. tore micaelsen

    Builder

    Joined:
    Feb 25, 2016
    Messages:
    27
    Likes Received:
    1
    1457446419578893970368.jpg

    small square is 10x10mm big is 20x20
     
  16. Jimmybuckets

    Jimmybuckets Well-Known
    Builder

    Joined:
    Aug 15, 2014
    Messages:
    234
    Likes Received:
    36
    If your x and y are all in order that only leaves z left. Give it a yank and see whats loose sir.
     
  17. tore micaelsen

    Builder

    Joined:
    Feb 25, 2016
    Messages:
    27
    Likes Received:
    1
    cant find anything loose :( and if you look at the last picture the inner circle has issues in Y on the small square and in the big has issues in X.. im lost.
    The squares looks ok tough, it looks like it doesnt show any kind of backlash or trouble there
     
  18. Jimmybuckets

    Jimmybuckets Well-Known
    Builder

    Joined:
    Aug 15, 2014
    Messages:
    234
    Likes Received:
    36
    I once had an issue like that and the Z would kind of jam into place and then break free under certain cutting loads. It was frustrating to track down. Cutting forces are pretty high in a CNC...don't be afraid to yank on things to see what is loose. Measure you squares and circles and see how accurate they are.
     
  19. tore micaelsen

    Builder

    Joined:
    Feb 25, 2016
    Messages:
    27
    Likes Received:
    1
    the circles are ofcourse not accurate since they dont even end where they start, but the squares are looking good.
    Have been yanking and pulling but cant see any wobble of any kind in the router itself or the fastening of it.
    How much friction should acctually be on the wheel's? should you be able to turn a wheel by hand with a little force (just moving the wheel so it slides on the rail) or should it be stuck? or should it turn with one finger ..? not really sure anyone got any tips?
    Mine i have to use two fingers and some force to make it slip against the rail.
     
  20. Jimmybuckets

    Jimmybuckets Well-Known
    Builder

    Joined:
    Aug 15, 2014
    Messages:
    234
    Likes Received:
    36
    Sounds like your wheels are OK. Are you climb cutting or conventional cutting or using a mix? What feed speeds are you using? What depth of cut are you using? Try slowing way down and then taking very shallow cuts. Turn off mixed tool pathing so you are either running climb cutting or conventional cutting. Do a run of both styles of cutting with the new lower speed and reduced depth of cut. See what happens.
     
  21. tore micaelsen

    Builder

    Joined:
    Feb 25, 2016
    Messages:
    27
    Likes Received:
    1
    only engraving, at 0.5mm depth and 4 mm high letters, have run so slow that each letter takes 30-40 seconds, still the same problem.
    Tried running the same squares and circles with a marker now, no resistance. same issue.
    20160308_205022.jpg
     
  22. Jimmybuckets

    Jimmybuckets Well-Known
    Builder

    Joined:
    Aug 15, 2014
    Messages:
    234
    Likes Received:
    36
    What are you using to produce G-Code?
     
  23. tore micaelsen

    Builder

    Joined:
    Feb 25, 2016
    Messages:
    27
    Likes Received:
    1
    tried both vectric and estlcam
     
  24. Jimmybuckets

    Jimmybuckets Well-Known
    Builder

    Joined:
    Aug 15, 2014
    Messages:
    234
    Likes Received:
    36
    If they are both producing similar results then it's most likely got to be mechanical. If your drivers were under performing you wouldost likely be missing steps but it doesn't look like you are missing steps. And your code looks good in the simulator? Is this a new problem or is the machine brand new to you?
     
  25. tore micaelsen

    Builder

    Joined:
    Feb 25, 2016
    Messages:
    27
    Likes Received:
    1
    Brand new problem on a brand new ox. BUT the bob and drivers is the same as i used on my mpcnc that had no issues like this.
    Code looks good in camotics and mach3 preview
     
  26. Jimmybuckets

    Jimmybuckets Well-Known
    Builder

    Joined:
    Aug 15, 2014
    Messages:
    234
    Likes Received:
    36
    Have you tried actually pulling and pushing a cutting bit that is chucked into the router? These are me reaching at this point, but could be run out on your spindle/router. Maybe the current on the drivers are not adjusted and/or you have a bad ground and gettings some weird stuff going on.
     
  27. tore micaelsen

    Builder

    Joined:
    Feb 25, 2016
    Messages:
    27
    Likes Received:
    1
    Runout is excluded since a sharpie ziptied next to the spindle produces the same result.
    Have tried lowering and raising the amps.
    Will try with a arduino and cnc shield tomorrow and see if that makes any difference.
    Suspecting the drivers (china tb6600 hy-div-256-5A ) not the most praised drivers.
     
  28. Jimmybuckets

    Jimmybuckets Well-Known
    Builder

    Joined:
    Aug 15, 2014
    Messages:
    234
    Likes Received:
    36
    yeah let us know
     
  29. GrayUK

    GrayUK Openbuilds Team Elder
    Staff Member Moderator Builder

    Joined:
    May 5, 2014
    Messages:
    1,668
    Likes Received:
    627
    tore micaelsen.
    Give this a try. Do what you have already done, but, do it at the most extreme opposite part of the work area. Make demands on your machine that rarely get used. This will use different mechanical parts of the machine. If you usually use bottom left for your Zero, then use top right, or thereabouts.
    It should define the area at fault. :thumbsup:
    Gray
     
  30. tore micaelsen

    Builder

    Joined:
    Feb 25, 2016
    Messages:
    27
    Likes Received:
    1
    Tried with cnc shield and the same drivers, No difference. Tried different places on the plate, No difference.

    Set up a dial gauge and ran alot of different length of moves back and forth, only deviated 1-2 hundreds of a mm. BUT consistently between those, didnt add up. I.e. it was 1 up, 1 down etc.
    Ran this test on both X and Y same result.
     

    Attached Files:

Share This Page

  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.
    Dismiss Notice