. .
Page 1 of 2 12 LastLast
  1. #1
    ian's Avatar
    Lives in London, United Kingdom. Last Activity: 16-07-2022 Has been a member for 2-3 years. Has a total post count of 25.
    Hi All

    I'm new to the forum and the CNC'ing and am currently building my 2nd iteration machine - it's a steep learning curve!

    I need to achieve 0.01mm accuracy and precision on my machine and are building based on Hiwin 15mm rails, two rails per axes, duel carriages per rail and 1605 ballscrews, C7 grade and aluminium extrusions. I've build a first version and am getting the following accuracy and backlash (measured using a dial indicator):

    X: 0.15mm (800mm rails/ballscrew)
    Y: 0.04mm (800mm rails/ballscrew)
    Z: 0.04mm (125mm rails/ballscrew)

    I'm investigating why the X is worse than the other two axes, checking the fittings, plates, etc. But I do want to improve the overall accuracy and want to know whether fitting a second ballnut will improve accuracy and reduce backlash? If so, does the second ballnut need to be tensioned relative to the first ballnut?

    Also to add, I know that accuracy depends on a lot more that just the ballscrew and ballnut, but I want to get clarity on this.

    Shared knowledge, experience and constructive critique are welcomed!

  2. #2
    Hi Ian,

    You're missing a key bit of information - how are the screws retained axially?

    Typically people use a fixed block at one end (often the driven end) and a floating block at the other. These are often of the FK/FF or BK/BF variety.

    You might be able to feel the screw moving backwards and forwards as you pull on it with 0.15mm of backlash, if so that would imply you need to look at the screw rather than the ball nut.

  3. #3
    Quote Originally Posted by ian View Post
    Hi All

    I'm new to the forum and the CNC'ing and am currently building my 2nd iteration machine - it's a steep learning curve!

    I need to achieve 0.01mm accuracy and precision on my machine and are building based on Hiwin 15mm rails, two rails per axes, duel carriages per rail and 1605 ballscrews, C7 grade and aluminium extrusions. I've build a first version and am getting the following accuracy and backlash (measured using a dial indicator):

    X: 0.15mm (800mm rails/ballscrew)
    Y: 0.04mm (800mm rails/ballscrew)
    Z: 0.04mm (125mm rails/ballscrew)

    I'm investigating why the X is worse than the other two axes, checking the fittings, plates, etc. But I do want to improve the overall accuracy and want to know whether fitting a second ballnut will improve accuracy and reduce backlash? If so, does the second ballnut need to be tensioned relative to the first ballnut?

    Also to add, I know that accuracy depends on a lot more that just the ballscrew and ballnut, but I want to get clarity on this.

    Shared knowledge, experience and constructive critique are welcomed!

    It is doable to get it close.
    I've got my machine down to 0.025mm. It uses gibs/ways so is fairly tight unlike linear rails and my longest is X at 800mm.
    I did 2 things.
    1: Used decent angular contact bearings size 12x32x10mm in custom made mountings for my 1605 DFU screws.
    2. I packed (shimmed) the double nuts with thin (0.01mm) aluminium foil (tin foil) between the back nut and center spacer until they were slightly bitey.

    These 2 ways seemed to be key. You have to be carefull where you buy your screws as some don't bother with the spacer bit so it's useless.
    My Y screw needed packing by 0.04mm and my X needed 0.07mm. Bit of a difference there. I reckon your X will take a fair bit.
    It's well worth taking the time to do these.

    You can still have issues with spindle run out too especially with ER collets. Generally I can get it to 0.01mm in the ER collets or 0.004mm direct collet.
    I generally add 0.01mm oversize to my drawings and it comes out pretty bang on.

    Otherwise you may have to get some proper expensive grounded C3-C5 type screws. And angular contacts to fit them.
    Last edited by dazp1976; 22-05-2020 at 01:03 AM.

  4. #4
    ian's Avatar
    Lives in London, United Kingdom. Last Activity: 16-07-2022 Has been a member for 2-3 years. Has a total post count of 25.
    Quote Originally Posted by AndyUK View Post
    Hi Ian,

    You're missing a key bit of information - how are the screws retained axially?

    Typically people use a fixed block at one end (often the driven end) and a floating block at the other. These are often of the FK/FF or BK/BF variety.

    You might be able to feel the screw moving backwards and forwards as you pull on it with 0.15mm of backlash, if so that would imply you need to look at the screw rather than the ball nut.
    Hi Andy UK

    It has a fixed FK at one end and a floating block at the other end which I cut myself. There is no play on the ballscrew, in fact on inspection it seems that there is a bit of compression on the ballscrew and I will spend some time on the weekend to troubleshoot it.

  5. #5
    ian's Avatar
    Lives in London, United Kingdom. Last Activity: 16-07-2022 Has been a member for 2-3 years. Has a total post count of 25.
    Quote Originally Posted by dazp1976 View Post
    It is doable to get it close.
    I've got my machine down to 0.025mm. It uses gibs/ways so is fairly tight unlike linear rails and my longest is X at 800mm.
    I did 2 things.
    1: Used decent angular contact bearings size 12x32x10mm in custom made mountings for my 1605 DFU screws.
    2. I packed (shimmed) the double nuts with thin (0.01mm) aluminium foil (tin foil) between the back nut and center spacer until they were slightly bitey.

    These 2 ways seemed to be key. You have to be carefull where you buy your screws as some don't bother with the spacer bit so it's useless.
    My Y screw needed packing by 0.04mm and my X needed 0.07mm. Bit of a difference there. I reckon your X will take a fair bit.
    It's well worth taking the time to do these.

    You can still have issues with spindle run out too especially with ER collets. Generally I can get it to 0.01mm in the ER collets or 0.004mm direct collet.
    I generally add 0.01mm oversize to my drawings and it comes out pretty bang on.

    Otherwise you may have to get some proper expensive grounded C3-C5 type screws. And angular contacts to fit them.
    Thanks - it is good to know it is achievable and realistic - I wasn't sure whether I had the wrong expectations of what can be realistically achieved. I'm settling in for some time consuming fiddling to get it dialed...

    Did you tension your ballscrew axially?
    Do you ha e a picture or design for the contact bearing housing and mount?

    Thanks in advance

    Ian

  6. #6
    Quote Originally Posted by dazp1976 View Post
    Otherwise you may have to get some proper expensive grounded C3-C5 type screws. And angular contacts to fit them.
    There's no point buying expensive ballscrews unless the whole machine is built to the same very high standard.

    To be honest it makes me laugh when I see figures quoted in the 0.001 range or higher from DIY builders because they are just fooling them selfs. Just leaning on the bench while measuring will affect most DIY machines as it deforms to match the bench movement at this level or precision. Measurments taken in the morning will be different than in the after noon and as the machine heats up, most don't even own measuring devices that are calibrated to that level.!

    I've got precision ground granite table and calibrated granite squares etc with calibrated measureing devices and wouldn't claim or expect those kind of precision from a bolted together aluminium profile or a welded steel framed machine that hasn't been heat treated and had machined surfaces etc.

    Precision of 0.01mm is about as good as you should expect from a DIY machine with C7 screws and bearings. Even then only expect this over a short distance because the thread pitch will vary over the length. Anything above this is a bonus and will most certainly vary over a length not much more than 300mm.!

    Keep it real folks, then you won't be dissapointed.
    -use common sense, if you lack it, there is no software to help that.

    Email: [email protected]

    Web site: www.jazzcnc.co.uk

  7. #7
    Quote Originally Posted by JAZZCNC View Post
    There's no point buying expensive ballscrews unless the whole machine is built to the same very high standard.

    To be honest it makes me laugh when I see figures quoted in the 0.001 range or higher from DIY builders because they are just fooling them selfs. Just leaning on the bench while measuring will affect most DIY machines as it deforms to match the bench movement at this level or precision. Measurments taken in the morning will be different than in the after noon and as the machine heats up, most don't even own measuring devices that are calibrated to that level.!

    Precision of 0.01mm is about as good as you should expect from a DIY machine with C7 screws and bearings. Even then only expect this over a short distance because the thread pitch will vary over the length. Anything above this is a bonus and will most certainly vary over a length not much more than 300mm.!

    Keep it real folks, then you won't be dissapointed.
    Yeah. I'm happy enough with the 0.025mm I got it to. It's took me a long time of taking it apart, tweaking and rebuilding. The amount of thin alu foil shimming everywhere is ridiculous!!! lol.
    The machine is over travelling a bit on proper calibration.
    I've been cutting a few 300x150mm squares and they've actually been coming out at 299.98x148.98mm. I'll leave it at that
    I won't need to go over 400x180mm in size so I add extra onto the edges of drawings. I'll leave it there as being within +/- 0.05mm for anything is good enough for me.

    I do still need to work on the head. I need to make a spacer for it that's 0.14mm difference from end to end to bring it more square. I've got the column tilted atm. It's level with the table but not as it lifts and lowers. There is also still an issue with X where the screw mounts aren't perfectly straight with it. This is causing a tilt in the table as it travels full length so I'm limited to 300 of the 500mm available.

    Yes I've had accidents with crashing dial indicators and there's stil loads of tweaking to do.
    Last edited by dazp1976; 22-05-2020 at 06:26 PM.

  8. #8
    I've been cutting a few 300x150mm squares and they've actually been coming out at 298x148mm. I'll leave it at that
    Did you mean that? as that is 2mm out so the steps per will need to be adjusted.
    ..Clive
    The more you know, The better you know, How little you know

  9. #9
    Quote Originally Posted by dazp1976 View Post
    The machine is over travelling a bit on proper calibration.
    I've been cutting a few 300x150mm squares and they've actually been coming out at 298x148mm. I'll leave it at that
    I won't need to go over 400x180mm in size so I add extra onto the edges of drawings. I'll leave it there as being within +/- 0.05mm for anything is good enough for me.
    WTF 2mm error is massive I wouldn't expect that amount of error if a blind man built it, (no offense if your blind by the way). Something is seriously wrong if you have that much error.!
    -use common sense, if you lack it, there is no software to help that.

    Email: [email protected]

    Web site: www.jazzcnc.co.uk

  10. #10
    Quote Originally Posted by JAZZCNC View Post
    WTF 2mm error is massive I wouldn't expect that amount of error if a blind man built it, (no offense if your blind by the way). Something is seriously wrong if you have that much error.!

    Sorry I meant it was 0.02mm out 299.98 x 149.98mm!! Just measured them again after realising I need some sleep.... lol

Page 1 of 2 12 LastLast

Thread Information

Users Browsing this Thread

There are currently 1 users browsing this thread. (0 members and 1 guests)

Similar Threads

  1. Cutter accuracy
    By Neale in forum Tool & Tooling Technology
    Replies: 21
    Last Post: 02-09-2017, 03:23 PM
  2. Cut depths and spindal speeds to improve finsh time
    By Fivetide in forum Tool & Tooling Technology
    Replies: 23
    Last Post: 21-05-2015, 08:39 AM
  3. Improve Kress TIR or get an upgrade?
    By Hauk in forum Kress Milling Motors
    Replies: 3
    Last Post: 23-02-2014, 06:18 PM
  4. trying to improve my board game
    By cubikoman in forum General Discussion
    Replies: 28
    Last Post: 12-05-2012, 06:57 AM
  5. NEW MEMBER: cubikoman needs to improve quality of his boardgame
    By cubikoman in forum New Member Introductions
    Replies: 2
    Last Post: 16-02-2011, 12:17 PM

Tags for this Thread

Bookmarks

Posting Permissions

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