Page 1 of 2 12 LastLast
Results 1 to 15 of 25
  1. #1
    Join Date
    Nov 2008
    Location
    somewhere
    Posts
    152

    Default Options for high positioning accuracy?

    Hi Guys and Gals,

    I'm in the thinking it through stage on a new mill, and I'm looking to design and build a small mill (about 300x300 cut area) which would be dedicated to cutting prototype printed circuit boards and panels. I already have a 6040cnc, which does a pretty good job on the boards, but I do run into some issues when dealing with small surface mount components and high densities of thin tracks. I've also had a few moments when I've had little time to work on my hobbies, and being able to run two mills at once (one cutting pcbs, one cutting other stuff) would be great.

    Ideally I'd like to have the ability to make very small movements and to have a high positioning accuracy, and I'm perfectly happy to sacrifice speed in order to get it.

    Has anybody got any suggestions for a suitable drive method (i.e ballscrew, trapezoidal lead screw, threaded rod, rack and pinion etc)?

    Does anybody else do much work cutting pcbs and if so do they have any suggestions?

  2. # ADS
    Google Adsense Advertisement
    Join Date
    Always
    Location
    Advertising world
    Posts
    Many





     
  3. #2
    Join Date
    Feb 2004
    Location
    Oxley, Brisbane
    Age
    79
    Posts
    3,041

  4. #3
    Join Date
    May 2003
    Location
    Perth WA
    Posts
    3,784

    Default

    The most accurate is ground ballscrews and buying new will set you back a lot of money. You can get them on ebay second hand but you are taking a risk as you cannot test them remotely.
    There are two critical measurements in any axis regardless of the type of linear motion. First you have backlash which is the measurement of the change in direction from forward to reverse movement. There will always be minute backlash from forward to reverse as to eliminate that totally would mean that the axis would be bound. The other variation is the true length verses actual length of the distance travelled. Both these factors are dependant on many things and accuracy will depend on how much you are willing to spend to achieve perfection.
    So there is a balance between cost and accuracy. A C7 ballscrew is rolled and if accuracy is your prime motivator then a ground screw is preferred and pay a lot more. Threaded rod, trapezoidal, and rack and pinion are your least preferred method in that order. I am not knocking these methods as they have other advantages over ballscrew.
    I boils down to how much you are willing to pay for the mechanicals as well matching the electronics to the mechanical resolution.
    Cheers,
    Rod

  5. #4
    Join Date
    Nov 2008
    Location
    somewhere
    Posts
    152

    Default

    Thanks for the info guys. Very helpful

  6. #5
    Join Date
    May 2005
    Location
    Cockatoo Vic
    Posts
    996

    Default

    Unlike Rod above, I would put rack and pinion further up the list. I am a fan. I have machined some pretty good PCBs with my R&P router.

    However for a machine the size that you need ballscrews would be much better suited. And I agree, stay away from threaded rod and acme. Acme can be done well but whatever you do, the accuracy is not as durable

    Just as important as drive method is linear bearings. I am no fan of round rod bearings, supported or otherwise. For me Linear rails from Hiwin or similar are really the only way to go for an accurate and durable machine.

  7. #6
    Join Date
    Nov 2008
    Location
    somewhere
    Posts
    152

    Default

    I'd interpreted Rod's list as threaded rod (worst) to R&P as the best of the second choices.

    Thanks again for the info guys, your helping me clarify my requirements very nicely.

    Greolt would you mind expanding on the advantages of Hiwin style profile rail systems over the round rod systems? I'm assuming they are more rigid and have a high load bearing capacity, but is there anything else?

  8. #7
    Join Date
    May 2005
    Location
    Cockatoo Vic
    Posts
    996

    Default

    Yes they are rigid and have a very much higher load bearing capacity than we load them with in a typical router.

    Round rod bearing have balls that run between the shaft and outer shell. That means that the point of contact is infinitely small. The ones that I have seen have a rather crude grubscrew adjustment. There are most likely better examples.

    Linear rails have ground tracks that the balls run against, both on the rail and inside the block. Most have four sets of rolling balls, some have two sets.

    As illustration this Hiwin youtube is the best I could find. It goes on a bit but near the start is some good illustrations.

    https://www.youtube.com/watch?v=2I44OT7c_MY

    I once ordered some Chinese round rod bearings for a project. As soon as they arrived and I saw what they were like, I chucked them in the bin and bought some Hiwin rails.

    Mounting and alignment of linear rails needs to be precise and is less forgiving than round rod bearings.

  9. #8
    Join Date
    Aug 2010
    Location
    Glenbrook NSW Australia
    Posts
    705

    Default

    if you want accuracy you do need good rails and good measurement. What we did in our cnc was to use glass scales which measure where the bit is not guessing where the the screw has move it too. we used the scale from hare and forbes https://www.machineryhouse.com.au/D722
    Can get them down to 1um
    vapourforge.com

  10. #9
    Join Date
    Feb 2004
    Location
    Oxley, Brisbane
    Age
    79
    Posts
    3,041

    Default

    Quote Originally Posted by rusel View Post
    if you want accuracy you do need good rails and good measurement.
    Can get them down to 1um
    Now THAT is accuracy.
    Bob Willson
    The term 'grammar nazi' was invented to make people, who don't know their grammar, feel OK about being uneducated.

  11. #10
    Join Date
    Nov 2008
    Location
    somewhere
    Posts
    152

    Default

    yup, 0.001mm would be very nice.

  12. #11
    Join Date
    May 2003
    Location
    Perth WA
    Posts
    3,784

    Default

    I have no doubt that a machine set up with glass linear scales has a huge advantage but it still boils down to the electronics and mechanicals to deliver that accuracy.
    Cheers,
    Rod

  13. #12
    Join Date
    Aug 2010
    Location
    Glenbrook NSW Australia
    Posts
    705

    Default

    you are quite right Rod, you need the backup to make it work
    So we used the 5um scales with no backlash nuts on 20mm screws and the servo motors are directly coupled to the screws. we could make it better by closing the loop by adding encoders to the servos.
    vapourforge.com

  14. #13
    Join Date
    Jun 2010
    Location
    Canberra
    Posts
    769

    Default

    Quote Originally Posted by foobillious View Post
    I do run into some issues when dealing with small surface mount components and high densities of thin tracks.
    How thin a track or how close are you talking? I knock out a PCB on my 6040 once in a blue moon, and it strikes me that the limitation tends to be the cutting tool. I use engraving bits that are meant to be 0.1mm, but just look pointy to me. As far as I can tell, the machine seems capable of 0.1mm resolution.

  15. #14
    Join Date
    Nov 2008
    Location
    somewhere
    Posts
    152

    Default

    yeah, the 6040 does a pretty good job at the moment - I've run into some issues trying to cut tracks around small parts (in this case TSSOP). I can do it, but I have to be pretty careful about the layout. As you've noted, the limiting factor is the cut width

    I've been using some 60 degree v carve bits which seem to have a cut width of 0.24mm when cutting deep enough to cut all the way through the copper. Unfortunately that is too fat for some lead spacings. I've got some 20 degree bits which are supposed to be 0.1mm, but I run into issues as the tip is too fine and it just snaps off (can't seem to plunge slow enough that it doesn't occur, even when aiming to cut through the copper in 2 passes). I've ordered some 40 degree bits, which I hope will do better.

    Anyway, the 6040 does a good job, limited by my current tooling. If I'm going to do a second mill though, then I'd like to know how to do it right. I can often get hold of small component sizes a lot more cheaply, even in small lots, so being able to actually use them would be great. I'd love to build something with everything that has been suggested above (though I'm not sure why the closed loop on the servo is necessary if you've already closed it with the glass scales), but the cost gets high awfully quickly. I'd need to build a lot of boards in order to justify that sort of outlay.

  16. #15
    Join Date
    Aug 2010
    Location
    Glenbrook NSW Australia
    Posts
    705

    Default

    If you want pcb I get them made in china in lots of 30 or more small pcbs cheaply even if i only need a few boards and they can be silk screened have vires and coated but do take a few weeks to get
    vapourforge.com

Page 1 of 2 12 LastLast

Similar Threads

  1. blade positioning
    By 3 toed sloth in forum BANDSAWS
    Replies: 18
    Last Post: 26th February 2015, 06:06 PM
  2. I need high accuracy on my bandsaw
    By ThinBoards in forum INCRA JIGS
    Replies: 24
    Last Post: 15th January 2006, 11:31 AM

Posting Permissions

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