Making very good bearing fits to 2/3 microns. Lapping ? - Page 5
Close
Login to Your Account
Page 5 of 6 FirstFirst ... 3456 LastLast
Results 81 to 100 of 103
  1. #81
    Join Date
    Aug 2005
    Location
    Toronto
    Posts
    4,986
    Post Thanks / Like
    Likes (Given)
    2939
    Likes (Received)
    3271

    Default

    Quote Originally Posted by hanermo View Post
    If an electronic or manual indicator reliably and repetably shows a 1 um out of roundness, and lapping the surface makes it go away, I consider the problem solved.

    If someone thinks this is a bad idea, please let me know why.

    .
    - Please post a pic of your lap for roundness

    - how do propose to lap these 100mm lengths and have the ends square to a micron?

    The reasons have all been posted. Top of my list is that there is no way you are going to be able to measure to a micron using go/no go homemade pin gauges. next is it is very difficult to control material removal to that level of accuracy, microns. You asked about lapping, I think that could work but not in a blind hole.

    Are you accustom to machining to a tenth, ie couple of microns? Experienced at this? The collective wisdom here from a group of experience guys is that it s very bloody difficult without the right equipment (Sunnen hone and quality bore gauge for a start). If you don't except what experienced guys have said, prove 'em wrong and please post pics

    why do you think you need ABEC 9's. Is this your design? what's the application and performance required driving the decision on ABEC 9's. Surely not skateboarding

  2. #82
    Join Date
    Jun 2008
    Location
    Bellingham, WA
    Posts
    3,736
    Post Thanks / Like
    Likes (Given)
    895
    Likes (Received)
    1885

    Default

    I am going to start a new thread to see if the title will attract input from members with real-life experience with extreme high precision bearings. Because this thread has gone on for quite a while with a lot of very good information posted (like Mcgyver's above) but has focused a lot on the theoretical and probably undoable except in a very highly specialized setting. I would guess that those folks with experience are just not bothering to read this thread. I think it is an interesting question as to what applications they are used and how they are actually installed. Maybe that thread will not yield that input based on actual experience. But let's see.

    Denis

  3. #83
    Join Date
    Sep 2009
    Location
    barcelona, spain
    Posts
    2,292
    Post Thanks / Like
    Likes (Given)
    464
    Likes (Received)
    1277

    Default

    I respectfully disagree on A.
    Both SKF and Timken (and others) are quite clear on this. The tolerance is about 3-4 microns, depending a bit on which table you pick.

    I respectfully disagree on B.
    Bearings are fit every day with great sucess. Likewise SKF, Timken and other bearing manufacturers make no mention of clean rooms being needed.
    A clean workspace, gloves and materials is quite sufficient.

    I disagree on C.

    No temperature control is needed at all.
    The delta in C between the 2 pieces must be small.
    4 degrees is about 1 micron.
    The pieces will be at the same temperature to much better than 1 degree, or 0.25 microns, as such it is a non-issue. IMO.

    The bearing, shaft and bore all expand or contract at the same rate, as all are made of steel.
    We are only interested in the relative size, not the absolute size.

    Quote Originally Posted by thermite View Post
    A) "Conventional wisdom" has been TRYING to point out that the tolerance required for the BORE, not bearing manufacture, is not what you think it is.
    B) You are stuck on mechanical means of measurement a good 60 years after industry in general realized they had hit a wall and gone over to other means.
    C) You haven't said a word about the grade of temperature-controlled clean-room in which you are operating. And you NEED one for these tolerances.
    Bill

  4. #84
    Join Date
    Sep 2009
    Location
    barcelona, spain
    Posts
    2,292
    Post Thanks / Like
    Likes (Given)
    464
    Likes (Received)
    1277

    Default

    The industrial lapping manufacturers all state that the lap becomes round by the action of the lapping compound.
    American Lapping is one.
    Honing and lapping are similar statistical processes, and both seem to work well.

    I am not using "homemade pin gages".
    I suggested using industrial pin gages. These are lapped and guaranteed to be accurate to 1 micron.

    I am accustomed to machining to some microns.
    Thats why I am asking for advice, as this is getting into a hard area.

    I never said I want to make it to 1, note *1*, micron.

    I am trying to make it have less than 4 microns TIR and to a size within spec, under 4 microns.

  5. #85
    Join Date
    Sep 2009
    Location
    barcelona, spain
    Posts
    2,292
    Post Thanks / Like
    Likes (Given)
    464
    Likes (Received)
    1277

    Default

    I have no idea if the ends will be square, much less to a micron.
    Thats not a requirement, I never said that.

    Quote Originally Posted by Mcgyver View Post
    -
    - how do propose to lap these 100mm lengths and have the ends square to a micron?

  6. #86
    Join Date
    Sep 2009
    Location
    barcelona, spain
    Posts
    2,292
    Post Thanks / Like
    Likes (Given)
    464
    Likes (Received)
    1277

    Default Abec9 bearings - applications

    Application is T&C cutter grinder spindles.

    Approx value will be 70.000€ for one T&C grinder.
    This type of bearings are routinely used on better machine tool spindles as well.

    I have 14 bearings, standard, P4 (ABEC7) and Abec 9 (P2).
    Most in 25 mm inner diameter 7205 size.
    4 bearings in 7210.

    I will make the 7 spindles, all with the same techniques.
    I expect recording the spindle bearing sound, and measuring TIR from a running spindle, will tell me how well I succeeded.
    Always comparing one set of bearings to another, all made from the same engineering materials.
    The spindles are 52100 bearing steel, not a backyard auto-body part recycled.

    I have no idea how much better the P2 bearings are than the P4 in truth.
    I expect I will find out shortly.

    The grinder is my own design, using large ballscrews and AC brushless servos.
    It is not a typical hobby budget type of thing.
    I will be using industrial contactless labyrinth seals, and Kluber NBU15 spindle grease.
    I have the servos, grease, etc already.

    I have the thing about 80% done, and am willing to buy any tools needed.
    I have 11 years experience in the arena, and sold 65 machines last year in Spain, while visiting about 200 shops myself.
    So I am neither totally inexperienced, nor totally clueless.

    Many (or most) of us on this board are much better machinists than me.
    I am very grateful for all the advice and help received, and learn a lot from many of these posts.

    I expect to have the first versions running about 1 week after I get the lapping compounds and or grinding spindle.
    I will probably go and see the local NSK people, about the high precision grinder.

  7. #87
    Join Date
    Dec 2004
    Location
    uk
    Posts
    13,188
    Post Thanks / Like
    Likes (Given)
    3857
    Likes (Received)
    6046

    Default

    Actually pretty easy if you do it the way jones and shipman do there 540 spindle housing, a straight through accurate honed hole for the od fit of the bearing, then a large precision od ground and ends ground at the same time spacer tube.

    The housing is a simple id hone job, the spacer a simple grinding job.

    The spacer is actually split and has a bunch of springs, one end fixed by some bolts, other provides the pre load + copes with any thermal growth.

    If you can not get the bores true, accurate and round then there's no point with the higher grade bearings, its very easy to distort the races out of round, the second you do that you have lost the accuracy advantage!

  8. Likes thermite liked this post
  9. #88
    Join Date
    Sep 2011
    Country
    UNITED STATES
    State/Province
    Virginia
    Posts
    23,009
    Post Thanks / Like
    Likes (Given)
    7561
    Likes (Received)
    7133

    Default

    Quote Originally Posted by hanermo View Post
    I respectfully disagree on A.
    Both SKF and Timken (and others) are quite clear on this. The tolerance is about 3-4 microns, depending a bit on which table you pick.

    I respectfully disagree on B.
    Bearings are fit every day with great sucess. Likewise SKF, Timken and other bearing manufacturers make no mention of clean rooms being needed.
    A clean workspace, gloves and materials is quite sufficient.

    I disagree on C.

    No temperature control is needed at all.
    The delta in C between the 2 pieces must be small.
    4 degrees is about 1 micron.
    The pieces will be at the same temperature to much better than 1 degree, or 0.25 microns, as such it is a non-issue. IMO.

    The bearing, shaft and bore all expand or contract at the same rate, as all are made of steel.
    We are only interested in the relative size, not the absolute size.
    There is, I suppose, a measure of comfort in all that. The firms who have been actually SHIPPING precision spindles for the past hundred-odd years don't have a thing to worry about as to competition out of your corner.

    Not as if this is a NEW issue, after all...

    Bill

  10. #89
    Join Date
    Feb 2004
    Location
    peekskill, NY
    Posts
    23,923
    Post Thanks / Like
    Likes (Given)
    0
    Likes (Received)
    4088

    Default

    "I never said I want to make it to 1, note *1*, micron."

    No, the number originally mentioned by you was two microns.

    Two microns for the record is NOT ten thousanths of an inch. Two
    microns is 78.7 microinches. In inches: 0.0000787

    Two microns equals one micron for large values of one micron, as they say.

    I think what the comments are saying to you are:

    1) be sure of the bearing bore size tolerance requirments as specified by the bearing manufacturer.
    The two micron diameter tolerance on the bore should be double checked.

    2) be sure of the absolute position requirments for the spindle you are building, which means
    determine the net cylindricity requirments between the two bearing seats that are some distance
    apart. If you think getting a short bore to size +/1 two microns is exciting, try cylindricity at
    that level. This amounts to: Getting ALL points on two cylinders, say five cm apart, to lie
    inside a surface that is rms four microns. You can't do that with a surface plate and a good
    indicator alone.

    3) be sure of the assembly procedures you plan on using once the assembly is fabricated. You
    can NOT slide the bearing into a bore that is one micron oversized. Because you are using
    the same materials for bearings and spindle you can NOT use differential thermal expansion
    to do this.

    4) be sure of the axial preload requirments for your bearings. Because they will be locked
    in place in the axial direction you have to decide in advance, how this preload will be
    determined accurately. If you don't control this the bearings can be destroyed in use in
    short order if it is excessive.

    Look here's a bit of advice: go try to do this using your best available technology you have
    on hand RIGHT NOW. See how good it goes and see where you need to move up a notch
    in your metrology.

    Can you, RIGHT NOW do this job so the bores are stupid easy, +/- 0.0002 diameter tolerance
    and the spindle alignment to the assembly only has to be, say +/- 0.001 inch?

    Do that and then figure out how to squeeze another decimal place out of the sytem.

  11. Likes tdmidget, TeachMePlease liked this post
  12. #90
    Join Date
    Jul 2007
    Country
    UNITED KINGDOM
    Posts
    4,465
    Post Thanks / Like
    Likes (Given)
    2119
    Likes (Received)
    1957

    Default

    That's where grinding the spindle after mounting it comes in, isn't it. The fact that the arse end of the spindle is wallowing about by a few microns is irrelevant if the datum surfaces of the spindle have been ground to a gnat's whisker in the completed housing.

  13. Likes SAG 180 liked this post
  14. #91
    Join Date
    Aug 2005
    Location
    Toronto
    Posts
    4,986
    Post Thanks / Like
    Likes (Given)
    2939
    Likes (Received)
    3271

    Default

    Quote Originally Posted by jim rozen View Post
    "
    Two microns for the record is NOT ten thousanths of an inch. Two
    microns is 78.7 microinches. In inches: 0.0000787
    .
    I thought he first said 2 -3 microns, the midpoint of which is pretty much a tenth

  15. #92
    Join Date
    Sep 2009
    Location
    barcelona, spain
    Posts
    2,292
    Post Thanks / Like
    Likes (Given)
    464
    Likes (Received)
    1277

    Default

    Excellent advice Jim Rosen.
    And I agree with all of it.

    From 1. to 4. number 2. Cylindricity is the hard part.
    I have a plan (a,b, and c) on how to do this.

    I will post pics of running spindles.
    And sound.
    And TIR results.

    I took pics of the stuff, spent 2 hours getting them cleaned and set up on the surface plate, but cant find my Sony camera usb cable.
    Pics, when I get to the cable store.

    The far end of the shaft is 25.004 in one orientation.
    Shaft is less than 2 microns TIR at the moment. Just need to get it a tad smaller.

  16. #93
    Join Date
    Sep 2009
    Location
    barcelona, spain
    Posts
    2,292
    Post Thanks / Like
    Likes (Given)
    464
    Likes (Received)
    1277

    Default

    As the title says ..
    I am trying to hit a zone less than 4 microns in error.
    +0/-2 to -3. Between 0 and 1 preferably. um.

    A few microns extra would probably work. SKF or Timken says somewhere 4 um extra increases preload 80 kg.
    This indicates, to me, I dont have to have it exact.
    I am trying to get it exact.

    I really hope I succeed..

  17. #94
    Join Date
    Jul 2011
    Country
    UNITED STATES
    State/Province
    Oregon
    Posts
    462
    Post Thanks / Like
    Likes (Given)
    124
    Likes (Received)
    261

    Default

    I imagine that if your machines actually work out and you sell a few you will soon be using P4 bearings. The added expense of the P2 bearings will most likely not be worthwhile. You competition most likely uses P4 bearing.

    If you need more accuracy maybe you should look into air bearings or hydrostatic oil bearings. Not that they are going to be easier to make, but they are superior to roller element bearings in many ways.

  18. #95
    Join Date
    Sep 2006
    Location
    SE PA, Philly
    Posts
    5,200
    Post Thanks / Like
    Likes (Given)
    950
    Likes (Received)
    1653

    Default

    Read what Jim says, above.

    Hanermo, your emails seem a mix of extreme sophistication mixed with points that indicate either you, or many of us, are confused. It may be me, so if I'm wrong please educate me.

    I looked out ABEC specs, and at least one source (here) gives tables for inner- and outer-ring bore diameter tolerances. One of the points I could not find, to my satisfaction, was whether this was a bearing manufacturing tolerance (how big is the bearing), or an installation tolerance (how big is the counterbore). They SHOULD be different. Bearings of this class are manufactured in clean-room environments. Not all of them are installed, applied, or serviced in that same environment. So I think that that the ABEC specs are the manufacturer's tolerance. For ABEC-9 (ISO Class 2) bearings, the bore tolerances are given as minus 0, with the plus spec being from 0.0001 to 0.0003 inches, or, equivalently, 2.5 to 8 microns. In your bearing's size (25mm), the spec is 0.0001" and 2.5um. Agreed?

    Do we agree that this is the bearing mfr's spec, and not (necessarily) the installation spec for counterbore and spindle diameter?

    If so, the question becomes whether the bore and shaft diameters have to meet the same spec, for your spindle to be able to exploit the precision of your very expensive super-precision bearing. I would make the case that it's very important to have your two (or more) bearing bores to be precisely aligned, and for your spindle shoulders and the bottom of your counterbore to be flat, parallel, and perpendicular to the spindle axis. And you need some way of getting extremely precise preload. But I'm sure you are addressing those specs, so back to bore diameter.

    I understand precision and resolution. Precision has a repeatability component. If your DRO resolves to 1 micron, but you have a 10um backlash in the measurement system, or the tool deflects 5um under load, precision is degraded, and the resolution is distorted. From your note, I assume you are comfortable that your system not only resolves to 1um, but that it accurately reflects changes in the actual tool holder. So, if the DRO indicates 1um of motion, the tool moves 1um, and 1um of metal comes off. It sounds like you are saying that, with a 1um repeatable resolution, you can "sneak up on" the 2-3um tolerance you wish to hit, if you can get a cutting tool that can shave a 1um chip, and if you can measure the bore accurately. Yes?

    Here are some points, as I read them and process them
    1) Measuring 2um (less than 0.0001") with mechanical calipers or bore gages is going to be tough
    2) Air gaging might allow the resolution and accuracy, but this involves custom gages, and (in my estimation as an engineer) a pretty significant effort in calibrating those gages. You've indicated a resistance to spending $7000 for a precision grinding spindle. Because of the custom work involved, I think that the gaging would be around the same cost. Maybe less. But..
    3) Using 13E-6/°C coefficient of expansion, a 1°C change in the bearing or workpiece changes the bore 0.3 microns. So a 6°C elevation gives you a 2um bore size difference. Can you account for this?
    4) Those temperature differences affect the spindle in use. Assuming a mild temperature rise in operation (to say 37°C body temperature), both the outer ring and the bearing expand by about 5.5. micron. Does this mean your spindle will fail with a mild temperature rise? No! Why? Because the bearing axes are still aligned with the spindle axis.
    5) Is the 2 micron specification on counter bore necessary for the outer bearing ring to maintain positional accuracy? No! Most outer bearing races are slip-fit (think of overhauling an electric motor - the beaings slide out of the bell housing). This goes to your point about resolution and accuracy. Or perhaps I should say repeatabilty and accuracy. A standard sliip fit for the outer ring, with loctite, gives you repeatability. If the bore is displace by a micron, your grinder's positional feedback controls take that into account and correct.
    6) If friction fit is needed, there may be some spec on interference fit for these bearings. It is most certainly not 2 micron - the interference would be affected by everything from temperature to preload to service load. I was unable to find any manufacterer's recommendations for this, but there are sources for general interference fit discussion.

    The key points are that the bearing specs are not the hole size specs, and the hold need not be to the same spec as the bearing outer ring. The manufacterer's spec on bore size should be used, because the class of interference fit will change the bearing ring diameter. It would be interesting to hear back from you about what the bearing supplier recommends for interference in counterbore and spindle diameters.

  19. #96
    Join Date
    Feb 2010
    Location
    ch
    Posts
    2,055
    Post Thanks / Like
    Likes (Given)
    187
    Likes (Received)
    270

    Default

    why couldnt the bearings be put on the shaft, preloaded and glued into the bores? maybe youd have to correct some outside surface, if necessary.

  20. #97
    Join Date
    Aug 2005
    Location
    Tucson AZ
    Posts
    6,097
    Post Thanks / Like
    Likes (Given)
    8550
    Likes (Received)
    2806

    Default

    Quote Originally Posted by Mcgyver View Post
    I thought he first said 2 -3 microns, the midpoint of which is pretty much a tenth
    He did and now it's 4 microns. Maybe if this thread goes on long enough it will be +/- .010.

  21. Likes thermite liked this post
  22. #98
    Join Date
    Mar 2013
    Location
    Northern California, USA
    Posts
    60
    Post Thanks / Like
    Likes (Given)
    28
    Likes (Received)
    20

    Default

    Quote Originally Posted by hanermo View Post
    As the title says ..
    I am trying to hit a zone less than 4 microns in error.
    +0/-2 to -3. Between 0 and 1 preferably. um.

    A few microns extra would probably work. SKF or Timken says somewhere 4 um extra increases preload 80 kg.
    This indicates, to me, I dont have to have it exact.
    I am trying to get it exact.

    I really hope I succeed..
    Aside from the issue of what the necessary tolerance is; for making a super precision bore in a housing, would a jig grinder be a suitable machine? There are Moore #3 jig grinders available at pretty low cost, in well taken care of condition. They seem to be obsolete for most production purposes, but might suit this use. Probably not a suitable approach for boring a cylindrical part with very high concentric it's requirements, but maybe for other components?

  23. Likes tdmidget liked this post
  24. #99
    Join Date
    Sep 2007
    Location
    Cairns, Qld, Australia
    Posts
    2,651
    Post Thanks / Like
    Likes (Given)
    742
    Likes (Received)
    654

    Default

    Quote Originally Posted by adama View Post
    Actually pretty easy if you do it the way jones and shipman do there 540 spindle housing, a straight through accurate honed hole for the od fit of the bearing, then a large precision od ground and ends ground at the same time spacer tube.

    The housing is a simple id hone job, the spacer a simple grinding job.

    The spacer is actually split and has a bunch of springs, one end fixed by some bolts, other provides the pre load + copes with any thermal growth.

    If you can not get the bores true, accurate and round then there's no point with the higher grade bearings, its very easy to distort the races out of round, the second you do that you have lost the accuracy advantage!
    That's got to be the only reasonable way to do it: it divides the job into sub-assemblies so if one assembly surface fails, it doesn't ruin the whole job. It also makes some crucial surfaces externally accessible, not hidden inside the housing and finally it allows lapping the housing bore which will give an accurate cylinder shape when done right. You could even recess the bearings in from the openings to avoid any potential bell mouthing of the tube ends cause by the lapping process.

    One other potential way I'd attempt the job is to have a looser fit for the bearings and glue them in with alignment jigs to hold them in position. The spindle would then be a factory serviceable or disposable unit.

  25. #100
    Join Date
    Sep 2011
    Country
    UNITED STATES
    State/Province
    Virginia
    Posts
    23,009
    Post Thanks / Like
    Likes (Given)
    7561
    Likes (Received)
    7133

    Default

    Quote Originally Posted by SAG 180 View Post
    ...The spindle would then be a factory serviceable or disposable unit.
    Point. They long have been.

    Historically, when REALLY precise spindles have been wanted, that 'factory' was a specialist - one that had affordable rebuild services, even - in busier times, perhaps a quick-ship exchange program as you hint at - but one that could be amortized over a large customer base.

    EG: The "Pope" spindles fitted to MANY brands of surface-grinders, the "Precise" and Mac units for TP and T&C grinders. Old-line makers have vanished or downsized, but precision spindles are still the 'Day Job' of at least a few specialists who have made the investment in the requisite metrology, machinery, and tooling, hung-onto accumulated years of R&D and field experience.

    So, not-so-tongue-in-cheek, really ....the most pragmatic way to meet this challenge may very well be with .... a purchase-order.

    Newbie wants to enter the field?

    Going to need serious funding, more and better research.

    And more 'understanding', vs wish-for.

    Bill

  26. Likes tdmidget liked this post

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
  •  
2