Page 1 of 3 123 LastLast
Results 1 to 15 of 35
  1. #1
    Join Date
    Mar 2017
    Location
    Brisbane
    Posts
    245

    Default Angular contact bearing source for ball screw supports?

    Anyone got a good angular contact bearing supplier for the fixed end on a ball screw?

    I've been searching around and it's doing my head in. As soon as I find something decent looking (good contact angle above 15 degrees, matched set of 2, DF, ABEC 7 or so) the price tag ends up being 1/3 the cost of my mill.
    Most of the stuff I find on ali-express is good but lacks one desirable feature.

    e.g. https://au.rs-online.com/web/p/produ...E&gclsrc=aw.ds

    I'd like a matched pair as apparently if you get that you don't have to fool around with making a spacer between them.

  2. #2
    Join Date
    Oct 2007
    Location
    Sydney
    Posts
    3,112

    Default

    I'd suggest you're over-speccing your part

  3. #3
    Join Date
    Mar 2017
    Location
    Brisbane
    Posts
    245

    Default

    Not really, there are close ABEC-7 angular contact bearings on Aliexpress but they dont have all the recommended specs, e.g. contact angle. I find ithard to believe a contact angle from 15 to 25 degrees should result in a 900% price increase. I'm comparing aliexpress with local suppliers of NSK tho.

  4. #4
    Join Date
    Jun 2007
    Location
    Vic
    Age
    48
    Posts
    544

    Default

    I replaced the quill bearings for the bridgeport clone and got mine from Bearing Wholesalers

    20160805_200212 (1).jpg

    20160805_205952.jpg
    Attached Images Attached Images

  5. #5
    Join Date
    Oct 2007
    Location
    Sydney
    Posts
    3,112

    Default

    I think you are OVER-SPECCING this.

    If quite reasonable home lathes and milling machines can get often impressive results with nothing more than trailer bearings, then I cannot for the life of me see why a home rolled CNC machine would need ABEC7 angular contact bearings of such specific contact angles on its ballscrews!

    The internet if a wonderful thing, we have so much data available to us only a mouse click away. The downside is there is no master editor to oversee what is being published, and many complete BS stories are perpetuated and become internet "fact". For some reason, bearings, and in particular ABEC7 seem to attract the nutjobs like moths to a flame. When inline skating was all the rage I recall a whole generation swearing the World would no longer turn unless they had their ABEC7 bearings in their inline skates! Data is not information, information is not knowledge, and knowledge is not experience. Beware of the brigade who sweat the small data stuff, make an N=1 example and then try to convince the world that's the only way to do it!

    You are speccing leadscrew bearings that are almost certainly many times higher quality than the rest of your machine. Probably by a factor of 100! If you're happy to do that, good luck to you, nothing wrong with that other than the cost, so pay your 700 bucks per ball screw and move on. If you're not happy to do that, and by "doing your head in" it seems you're not, then go back to basics and think about what you're trying to achieve and the machine it is going on. There is a lot of information (as opposed to data!) on spindle design, which with just a few minutes read will give you enough basics to understand how to achieve rigidity in a spindle assembly. For example, yes, to achieve a specific rigidity you can pay a truckload and have higher contact angles for the spindle/bearing assembly .... or you can just have larger bearings! To put this in perspective, this is what Tormach use on their machines! https://www.tormach.com/store/index....30&portrelay=1 https://www.tormach.com/store/index....32&portrelay=1 The 1100 uses ABEC 7202AC /P5 bearings, between 5 and 20 bucks off the usual sources.

    Personally I don't know why you're even going down this road in the first place. Ball screw support bearings are almost ubiquitous from China these days, and a quick look on Aliexpress suggests they will set you back somewhere between 20 bucks and 100, for a complete assembly. However if you did want to copy a known design, then Tormach will have a parts manual and there will be enough information off that as to how to fabricate the assembly. They will likely fabricate their own and throw in some 5 buck bearings so as to save cost. They have all manner of high quality machining facilities at their disposal, so why not. However for the rest of us doing this as a onesie, then it's often far better to just buy a ball screw support assembly like this already fabricated from Aliexpress etc and get the machine finished. https://www.aliexpress.com/wholesale...upport+bearing I expect there are almost as many unfinished CNC machines out there as there unfinished yachts!

  6. #6
    Join Date
    Mar 2017
    Location
    Brisbane
    Posts
    245

    Default

    Basically I've been reading cnccookbook and the instructions on there are that you want decent ballscrew supports to reduce backlash.
    Backlash is the no. 2 accuracy enemy after spindle runout. Bearing wise, you want large axial resistance for ballscrew support bearings and large radial resistance for spindle bearings.
    Backlash on the mounting ballscrew connection is eliminated via double ball nuts with preload (no problem. just specify it when ordering)
    It's a fair bit of reading but basically the elite bearings for support are "TAC" bearings, for your larger cnc machines and can cost oodles.
    ABEC 5 or 7s are apparently the next best step for home gamers. Something like 7204 B DU (B is angle, DU is duplex) then P4 for the ABEC rating. Bigger angle on the angular conact bearing is better (15 is low, 60 is high, B is 40).
    After much searching I found these for $50 - https://tinyurl.com/y8ts765k - very little difference in ABEC rating for cost and they're also available with a 40 degree contact angle.
    HOWEVER, after your post I checked out Tormach and their bearings seem to be $20 a pop for replacement and AC (20 degree angle I think ?).

    After eliminating this sort of backlash you need to remove friction/stiction. I'm intending to scrape the surfaces and install one shot oilers, with Hiwin rails on the z.

    Big project. Half of it is reading though !

  7. #7
    Join Date
    Feb 2016
    Location
    NSW
    Posts
    117

    Default

    You've overspecced it out the wazoo. For all practical purposes you need nowhere near the level of bearing you've specced just to eliminate end play on the screw. You've taken one little kernel of truth (reducing end play on your leadscrew improves accuracy), but not taken into account its overall impact on the picture, which is very little. Focusing on the contact angle to that degree is folly.

    Unless you're building a full-size VMC with a very heavy table and you want to work on very heavy workpieces, there's no way your screw is going to see enough axial thrust to justify anthing more than a 72xx AC bearing pair.

    A plain jane 7200 AC bearing has a dynamic load rating of 700kgs. I'm willing to bet your mill will never see a quarter of that load. Pick a nice pair of AC bearings that are cheap and fit your screw, install them with the correct preload, and your axis will have a long happy life, and you won't be tearing your hair out if in 5 years time you need to replace them.

    For example, this $17 bearing at RS (who are always horrendously overpriced) will take a literal TON of axial thrust: https://au.rs-online.com/web/p/ball-bearings/4088458/

    For reference, with the highest axis acceleration I've EVER configured into my mill (about 0.5 m s-2), that bearing alone could support a screw pushing a table and workpiece weighing 20 tons.

  8. #8
    Join Date
    Mar 2017
    Location
    Brisbane
    Posts
    245

    Default

    Thanks fellas, I'll let you know how I go. I've done enough ballscrew research for a while and I'm going to give it a break (almost dreaming about ballscrews).
    In my defense the info came from here - https://www.cnccookbook.com/anti-bac...backlash-nuts/

  9. #9
    Join Date
    Feb 2016
    Location
    NSW
    Posts
    117

    Default

    Yeah, but you missed the central thrust of what the article is saying - yeah, you can spend a fortune on a duplex pair of ABEC 7 AC bearings, but you don't have to, because the difference between them and cheap angular contacts is going to be one poofteenth of nothing in terms of runout, so as long as you're willing to spend the (tiny modicum) of effort setting the preload on a non-duplex pair yourself, you can get away with $20 bearings and still have more than acceptable performance.

    More money is not always more better when it comes to speccing bearings. There's an awful lot of very accurate grinding spindles around running in $15 angular contact Magneto bearings.

  10. #10
    Join Date
    Aug 2008
    Location
    near Rockhampton
    Posts
    6,216

    Default

    You can guarantee the bearings on aliexpress are not ABEC rated. You would not have a clue what you are buying.

    You need to look at your design and just use standard grade off the shelf bearings.
    Gold, the colour of choice for the discerning person.

  11. #11
    Join Date
    Oct 2007
    Location
    Sydney
    Posts
    3,112

    Default

    Quote Originally Posted by j3dprints View Post
    Basically I've been reading cnccookbook and the instructions on there are that you want decent ballscrew supports to reduce backlash.
    Backlash is the no. 2 accuracy enemy after spindle runout. Bearing wise, you want large axial resistance for ballscrew support bearings and large radial resistance for spindle bearings.
    Backlash on the mounting ballscrew connection is eliminated via double ball nuts with preload (no problem. just specify it when ordering)
    It's a fair bit of reading but basically the elite bearings for support are "TAC" bearings, for your larger cnc machines and can cost oodles.
    ABEC 5 or 7s are apparently the next best step for home gamers. Something like 7204 B DU (B is angle, DU is duplex) then P4 for the ABEC rating. Bigger angle on the angular conact bearing is better (15 is low, 60 is high, B is 40).
    After much searching I found these for $50 - https://tinyurl.com/y8ts765k - very little difference in ABEC rating for cost and they're also available with a 40 degree contact angle.
    HOWEVER, after your post I checked out Tormach and their bearings seem to be $20 a pop for replacement and AC (20 degree angle I think ?).

    After eliminating this sort of backlash you need to remove friction/stiction. I'm intending to scrape the surfaces and install one shot oilers, with Hiwin rails on the z.

    Big project. Half of it is reading though !
    Yes I'm well aware of WHY it's done, but as PM agreed, you are taking a valid principle and using it waaaay outside any practical value, just because it might say so on an internet site, doesn't making it a fact, and that's half the point of my post. If you were designing a highly accurate 5 tonne CNC controlled measuring machine you will need to sweat this stuff, the reality is however you're not, and from what I've read so far of what you're using, if you could finish up with a result 90% of something like a Tormach 1100, you could hold your head up very high and be proud of that achievement. The 1100 or even the smaller 770 are quite fine for 99.99% of people who would want that size/price/quality machine and I've seen them (well actually earlier versions) run. The finish and accuracy would satisfy the majority of users, all at a speed that's acceptable for a home shop. You want to accelerate like a Brother Speedio or have the accuracy of a DMG machining centre, that's fine, but be prepared to build a several tonne machine and tip tens of thousands of dollars worth of parts into it. In the meanwhile, Tormach use 5 buck bearings, and they seem to do just fine!

    I've bought all manner of bearings through Aliexpress etc and every one has been fit for purpose. For sure if it was a critical application, difficult/expensive to get to etc I would go for a decent brand name (I prefer SKF or Tmken, but that's probably just because I'm old/grumpy/don't like change), but for something like this pretty much any bearing of that design will be fine. The cheaper bearings may not last as long, but I've found them fine. If you look around you will see countless machines with the purpose designed balls screw bearing blocks, and if it were me that's how I'd try to design the machine. Once it's running, if you find they aren't up to scratch you can always use the machine itself to machine better alternatives and throw in some 20 buck bearings.

    Edit: Just as an aside, I had a quick look through that article you linked to, and it seemed pretty clear to me that the author was saying most people won't need the level of precision theoretically available. I'd suggest looking at the actual tolerance figures he's providing, and making decisions based on that hard data. I'm not sure about the figures he supplies, as they don't seem to correlate with my own, but either way, the numbers are VERY small. Even with the worst rated bearing available, you're talking about RADIAL bearing runout in the single digit micron range. Due to the pre-load and angles I'm guessing the practical runout would be less. Even if the paired bearing's tolerances stacked, I'd suggest something in the single digit axial and radial runout range is so ridiculously low, that unless you're operating these in a temperature controlled room with an impressively ground lead screw, you have bigger fish to fry!
    Last edited by Pete F; 13th Dec 2017 at 05:45 PM. Reason: typo

  12. #12
    Join Date
    Oct 2007
    Location
    Sydney
    Posts
    3,112

    Default

    Quote Originally Posted by j3dprints View Post
    Basically I've been reading cnccookbook and the instructions on there are that you want decent ballscrew supports to reduce backlash.
    Backlash is the no. 2 accuracy enemy after spindle runout. Bearing wise, you want large axial resistance for ballscrew support bearings and large radial resistance for spindle bearings.
    Backlash on the mounting ballscrew connection is eliminated via double ball nuts with preload (no problem. just specify it when ordering)
    It's a fair bit of reading but basically the elite bearings for support are "TAC" bearings, for your larger cnc machines and can cost oodles.
    ABEC 5 or 7s are apparently the next best step for home gamers. Something like 7204 B DU (B is angle, DU is duplex) then P4 for the ABEC rating. Bigger angle on the angular conact bearing is better (15 is low, 60 is high, B is 40).
    After much searching I found these for $50 - https://tinyurl.com/y8ts765k - very little difference in ABEC rating for cost and they're also available with a 40 degree contact angle.
    HOWEVER, after your post I checked out Tormach and their bearings seem to be $20 a pop for replacement and AC (20 degree angle I think ?).

    After eliminating this sort of backlash you need to remove friction/stiction. I'm intending to scrape the surfaces and install one shot oilers, with Hiwin rails on the z.

    Big project. Half of it is reading though !
    I'm sitting at my desk, supposedly studying (but will use any excuse not to do it!) and was thinking about this, and think I can see the problem here. From your comments, it seems you are confusing backlash and bearing runout. They are NOT the same error.

    Backlash is the end play in the bearing arrangement that will allow the leadscrew (ball screw in this case) to move from side to side. In a conventional machine it comes from two sources (mainly) and they are the leadscrew nut and the leadscrew bearing arrangement. Leaving aside the ball screw and nut, as that's a whole different kettle of fish, that leaves the bearing arrangement. This is the reason for using double bearings that are pre-loaded. In a properly designed bearing arrangement you will effectively have zero backlash to the level of your preload for that bearing pair and the arrangement. So in other words, you can exert a force on the leadscrew of a certain amount, and the bearing pair will be rigid enough to resist that force. Simply increasing the precision of the bearings will not in itself increase the stiffness of that same arrangement. Having higher precision bearings may however allow you to run a higher preload, and that will increase the stiffness.

    Having said that, the stiffness of the bearing arrangement is already very high for a properly designed pair like this, and with even a remote modicum of care, simply won't be an issue for a machine like this. With very high acceleration and heavy table loads you may need to consider this, but the steppers/servos you could use on a machine like that would be extremely unlikely in my opinion to develop sufficient forces to be capable of overwhelming the bearing pair preload. Thinking about that further, even if they could theoretically overwhelm the preload, it would be a transient acceleration and the rest of the machine would not be capable of actually machining at those acceleration rates. Once the acceleration reduced, the bearing preload would return and the accuracy return to normal (all that just an aside, not something you would need to consider).

    The other area of error is the runout of the bearing and that has nothing to do with backlash. The runout will be determined by the bearing class and present itself as a runout figure in the specs. The result of this will be a radial movement of the ball screw (which can be discounted) and an axial movement. The numbers here are teensy, by any metric, and will be totally overwhelmed by the tolerances of the ball screw itself, and the fact it expends and contracts with changes in temperature. The precision of the bearings would be such a small error, I would defy anyone to present meaningful errors in the finished machined work that could be attributed to ball screw axial runout from the bearing pairs. The leadscrew may shift a few micron back and forth through runout, but that is way below the threshold of noise for a machine like this.

    Hopefully that makes sense and it's now a little clearer as to what you're aiming for and why.

  13. #13
    Join Date
    Jun 2016
    Location
    Sydney
    Posts
    283

    Default

    The trouble with Alibaba and AliExpress is that whatever is advertised is not what you end up with.
    I was looking for a Crossed Roller Bearing, and Alibaba had several suppliers in the range of $1 to $20. OK, fine, hard to believe, but I asked for quotes. The quotes ranged from $95 - $230 per unit. There was some hand-waving about their advertised prices ('ah so sorry'), but by the end I did not believe ANY of them.
    I didn't buy from them either.

    Cheers
    Roger

  14. #14
    Join Date
    Aug 2010
    Location
    Near Bendigo, Victoria, AUS
    Age
    72
    Posts
    3,102

    Default

    could you use double row angular contact bearings? They are cheap as they are used in modern car non-driven wheels. They are available in all sorts of sizes straight from the catalogue, you don;t have to worry about adjustment and preload and that's built in to match the bearing and you need a lot less space. Usually the internal tracks are in two halves and once torqued to specs they have no backlash.....
    Cheers, Joe
    retired - less energy, more time to contemplate projects and more shed time....

  15. #15
    Join Date
    Jun 2016
    Location
    Sydney
    Posts
    283

    Default

    Double row ACBs instead of CRBs? In principle, but I want an almost zero height profile.

    I have a different way of solving the problem which so far has cost me $5 for SS balls (eBay) and a small DGB from the stock bin. The secret is that while a commercial bearing may be rated for 5,000 RPM and 200 kg load, I need something for a <20 kg load and <4 RPM.
    Prototype works fine.

    I could almost use a lazy susan table, except that they can lift and rock. Not sure about their 'rattle' either. Not good enough.

    Cheers
    Roger

Page 1 of 3 123 LastLast

Similar Threads

  1. Now ball Screw Re-fit
    By Sterob in forum CNC Forum
    Replies: 15
    Last Post: 14th Jul 2016, 06:11 AM
  2. Ball Screw problem
    By Sterob in forum CNC Forum
    Replies: 12
    Last Post: 22nd Jun 2016, 09:51 PM
  3. Replies: 3
    Last Post: 13th May 2015, 07:19 PM
  4. Ball bearing oilers
    By Jim Ferrous in forum METALWORK GENERAL
    Replies: 14
    Last Post: 17th Sep 2013, 06:56 AM
  5. Used large ball bearing/race anyone? I want one...
    By jhovel in forum METALWORK GENERAL
    Replies: 46
    Last Post: 26th Feb 2012, 06:01 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
  •