Page 1 of 3 123 LastLast
  1. #1

    I have a warco mini mill which has recently been converted to CNC using cncfusion kit 4.
    The machine works great with backlash <= 0.03mm and smooth table travel. But, I have a major problem.

    When the machine was being used manually I never checked the squareness of the X and Y axis.
    But now with the resulting resolution of the CNC I have noticed a major flaw:

    1) When I machined a large circle (50mm plus) the resulting circle was not round. This is a classic backlash problem but proved to be not the case.
    2) I machined an 'L' shape edge (200 by 70mm) into 3mm plastic and the resulting square edge was out by about 0.3mm on its longest side.
    3) Because of the above I decided to measure the squareness of the X and Y axis using the following method:
    a) Squared up a 400mm length of straight edge (25mm by 10mm profile +- 0.01mm over 300mm) along the X held in a vice.
    My best effort achieved +- 0.03mm repeatability over 300mm which meant the X table travel was ok.
    b) Made sure the vice was solid and replaced the bar with an engineers 150mm square held in the vice.
    c) I now measured the Y axis movement along the right angle formed by the square. The results were alarming.
    Over only 75mm in the Y direction the error was 0.32mm which means the Y axis was running at an angle to the X axis. This is hard to believe especially when the X and Y are running so smoothly under CNC control.

    I repeated the above measurements by setting up to the Y axis first and then checking the squareness of the X . Same error found.

    Any ideas how to resolve this problem.

  2. This one?

    The question is how the table attaches to the Y-axis. On my bigger one the table is machined with integral dovetails that run on ways machined in the Y-saddle, similar to this one of an X2 If so its hard to see how they could get misaligned. All the pictures I can find of the mini-mill (X2 clone) seem to suggest its much the same unless the saddle is in two parts? Are the gibs tightened up? A loose gib could allow the table to twist on the saddle, or the saddle to twist on the y-ways.

    Click image for larger version. 

Name:	gibadjust.jpg 
Views:	1557 
Size:	22.7 KB 
ID:	1678
    Last edited by irving2008; 31-01-2010 at 09:57 PM.

  3. #3
    Thats' the one.
    The X/Y saddle has dovetails machined presumably!! at 90 degrees same as the X2.
    I am trying very hard to see how the X and Y can be out of square, it's also very difficult to do an actual
    measurement of the complementary dovetails on the saddle to check their squareness.

    The gibs were nice and tight and threadlocked and the steppers had an easy life moving the table (3Nm) . Table also had
    no more than 0.08mm play with the X extended to far right/left.

    I should of noticed this problem about 10 months ago.

  4. when you did the test did you lock the gibs on the stationary axis?

    The other thing i would do is put the base of the engineers square in the vice but proud of the top and indicate along that in the X to make sure its still aligned that way - it should be. with the Y-gib locked. Then move to the other leg of the square, lock X and release y and then indicate Y. that way you haven't got any disassembly between tests... or do the same things with a 123 block on the 3+2 sides.

    If the ways are not machined square I suppose it would be possible to have them recut and fit a wider gib-strip (if need be). You could remove the table then indicate the Y axis movement relative to a square clamped to the x-way... that would give a clue as to which way is mis-aligned

  5. #5
    I visited the garage about 45 mins ago to recheck everything again and
    it's true the Y is not square to the X and again I got >0.3mm over about 60mm along the engineers square. I even checked the engineers square and its fine.

    Tomorrow I will try you idea of indicating the Y movement relative to the X.

    Let you know the result.

  6. #6
    I did another test tonight with the table removed and checked with a square
    how true the Y movement was. The results show a 0.4mm error over about 65mm.

    I repeated yesterdays test and set the dial to +-0.03 over 140mm on the X axis. A 6inch square held in the vice as before and I get 0.62mm error over a Y movement of 120mm. From this I can work out the angle error between the saddle X and Y dovetails.

    I now have two choices, either get someone to machine the X/Y channels true
    or possibly get software to compensate. Can mach3 auto compensate for non true XY axis!!!

  7. I would strip the table and saddle off to check there's not swarf or other crap trapped in the dovetails. I would not expect these little machines to have tapered gibs, so any foreign body on the fixed side could cause the problem with giving a tell-tail loose gib on the other.


  8. #8
    I just removed the saddle and examined the LHS dovetails and indeed the machining was not good. At the top of the dovetail was a small (0.1mm) of unmachined
    ramp shaped metal 2/3 thirds the length of the channel. I scraped most of it away with a thin file and finished with off with 600 grit.
    If I keep up this scraping process I will of course do two things - ruin the dovetail or get no near perfection!!!

    I will reassemble the lot tomorrow and retry the previous laborious calibration setup again.

  9. #9
    I have reduced the XY error to 0.4mm max over a Y travel of 98mm but the LHS Y dovetail has now in my opinion been scraped too much. So anyone know of a reasonably cheap!
    engineering firm in Bedford or surrounding area capable of machining this mini mill saddle.

    I might consider travelling up to 50 miles to get it done,
    Thanks for all your help.

  10. #10
    I had the saddle machined by a local company. They first measured the dovetails for parallelism and the results was not good. The fixed end was out by 0.43mm and the gib side was out by 1.05mm.
    Both end were machined with a 30 degree dovetail and the results are truely wonderful.

    On the machine I now get 0.03mm worst case after lapping the gib on the y axis and making slots for the set screws. In addition I replaced all the 0.125mm balls in the ballscrews with 0.1268mm and this has achieved 0.08 backlash.

    I thank you all for the helpfull replies.

Page 1 of 3 123 LastLast

Similar Threads

  1. Warco WM16 Mill - Conversion to AC motor with VFD
    By birchy in forum Milling Machines, Builds & Conversions
    Replies: 14
    Last Post: 27-06-2014, 03:48 PM
  2. CNC Fusion Kit for Warco mini mill / Amadeal XJ12-300
    By rnr107 in forum Gantry/Router Machines & Building
    Replies: 2
    Last Post: 03-07-2012, 12:18 PM
  3. Warco WM16 Conversion
    By birchy in forum Milling Machines, Builds & Conversions
    Replies: 21
    Last Post: 25-03-2011, 08:32 PM
  4. CONVERSION: Warco VMC Turret Mill - CNC Conversion
    By antsals in forum Conversion Build Logs
    Replies: 33
    Last Post: 06-03-2011, 11:25 PM
  5. CONVERSION: Buget Warco Major 5 axis conversion
    By Tweaky in forum Conversion Build Logs
    Replies: 45
    Last Post: 07-10-2010, 03:26 PM


Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts