measuring runout - Page 2
Close
Login to Your Account
Page 2 of 3 FirstFirst 123 LastLast
Results 21 to 40 of 47
  1. #21
    Join Date
    Feb 2004
    Location
    peekskill, NY
    Posts
    27,595
    Post Thanks / Like
    Likes (Given)
    0
    Likes (Received)
    6487

    Default

    Strong, or Stong?

  2. #22
    Join Date
    Feb 2010
    Location
    ch
    Posts
    3,478
    Post Thanks / Like
    Likes (Given)
    287
    Likes (Received)
    484

    Default

    so i looked into this a little further. if the lathe has not been running (overnight) the "lobing" disappears on the 2µ test indicator. if i then rotate the chuck to different positions by hand the 1µ mitutoyo starts showing the same, a high and a low per rotation. if i wiggle the chuck in each position it also shows around 10µ runout.

    i did what 9100 said, blued up the piece and made a pass that nicely spread the blue. no axial marks, just circumferential stripes. the toop dug in in two spots.

    i probably knew this before but wont forget it now: dont trust a 1µ instrument without confirming what you see. there is friction/hysteresis at play.

    to investigate this a bit more i set up a 200 mm parallel good for 2µ (actually this is the deviation accross the narrow side) with a 10µ shim on one side.


    digital mitutoyo

    - sliding: first half no movement, catches up in second half showing 10µ. up/down the same
    - stopping in the middle and the end going up: first 100 mm 4-5µ, 10-12µ at the end
    - stopping in the middle and the end going down: 2-3µ/10-11µ
    - wiggling: improves down direction to 5µ (scatter of readings higher)

    tesa tast (2µ graduated)

    - sliding dragging it: shows perfectly im both directions
    - sliding pushing it: rises by 1µ (even when going down) and catches up after about 70 mm

    mitutoyo hicator (0.5µ graduated)

    - sliding up: shows 1,5µ total
    - sliding down: 1µ
    - wiggling up: 0-1.5/3-6µ
    - wiggling down: 0-0.5µ/3.5-4µ

    somet (graduated 1µ, 6 mm range)

    - sliding up: 11µ
    - sliding down: 2µ
    - wiggling down: 2µ/10µ
    - istalled additional spring. read worse, its quite stiff as it comes already.

    so what happens when measuring runout is pretty much open except for the test indicator of course. i have always benn pulling it and thats good. will probably get a 1µ graduated one soon. i didnt include the mikrokator (0.25µ) because i only use it on the measuring stand (very narrow range).
    Attached Thumbnails Attached Thumbnails 0-253.jpg   0-256.jpg  

  3. Likes Trueturning liked this post
  4. #23
    Join Date
    Nov 2004
    Location
    Webster Groves, MO
    Posts
    7,985
    Post Thanks / Like
    Likes (Given)
    2077
    Likes (Received)
    3909

    Default

    Here is an interesting (to me) article. Scroll down to page 29 for some comments about bearings.

    https://wp.optics.arizona.edu/optome...irror-Full.pdf

    Bill

  5. #24
    Join Date
    Nov 2004
    Location
    Webster Groves, MO
    Posts
    7,985
    Post Thanks / Like
    Likes (Given)
    2077
    Likes (Received)
    3909

    Default

    Quote Originally Posted by jim rozen View Post
    Strong, or Stong?
    stong(Noun) An area of land equivalent to a quarter of an acre.

    John Strong (noun) a PHD who wrote the book on experimental physics.

    I have seen his name given as Stong, probably just a typo.

    strong-1.jpg

  6. #25
    Join Date
    Feb 2004
    Location
    peekskill, NY
    Posts
    27,595
    Post Thanks / Like
    Likes (Given)
    0
    Likes (Received)
    6487

  7. #26
    Join Date
    Feb 2010
    Location
    ch
    Posts
    3,478
    Post Thanks / Like
    Likes (Given)
    287
    Likes (Received)
    484

    Default

    Quote Originally Posted by 9100 View Post
    Here is an interesting (to me) article. Scroll down to page 29 for some comments about bearings.

    https://wp.optics.arizona.edu/optome...irror-Full.pdf

    Bill
    so i should get a hydrostatic oil spindle?

  8. #27
    Join Date
    Nov 2004
    Location
    Webster Groves, MO
    Posts
    7,985
    Post Thanks / Like
    Likes (Given)
    2077
    Likes (Received)
    3909

    Default

    Quote Originally Posted by dian View Post
    so i should get a hydrostatic oil spindle?
    That is a decision you can make for yourself, but this link lists applications where the spindle is held centered when it is motionless, not depending on the oil wedge. It will still move off center with side loads, where the term "stiffness" comes in. I suppose if you wanted the closest to perfection possible you would have the spindle vertical and driven by a coupling that did not produce any side loads.

    https://rotorlab.tamu.edu/me626/Note...20Bearings.pdf

    Bill

  9. #28
    Join Date
    Jul 2019
    Country
    UNITED STATES
    State/Province
    Florida
    Posts
    5,202
    Post Thanks / Like
    Likes (Given)
    4872
    Likes (Received)
    972

    Default

    Quote Originally Posted by 9100 View Post
    I'm the one getting lobes. The bar is solid and it is left in the chuck without changing anything, just making another pass at the same settings. The type of chuck doesn't matter except possibly contributing to resonance of the spindle.

    Bill
    Usually for critical runout I will take a rough/rough then two finish finish/semifinish and oversized that equal to allow for my two final finish passes. I check runout before bringing it to the finished size. If it is expensive material or even in practice it has served me well minimizing uncorrectable idiotic errors. Many times I am given only enough material to complete a complex part on the lathe and then to goes often to the mill.

    That is where one understands it becomes even more important not to make a error considering all the wok put into the lathe work and no extra material unless ordered after a mistake to start over.

    I take all of my parts completing the work on both machines myself mostly yet I like watching others when they get a finished lathe part and watch how they handle it. Most of the time they can not handle any helpful feedback and so I consider their demeanor and personality before offering anything they might consider doing to get a good part off most likely. It is a lot of pressure in the trade under these circumstances.

  10. #29
    Join Date
    Sep 2009
    Location
    barcelona, spain
    Posts
    2,625
    Post Thanks / Like
    Likes (Given)
    643
    Likes (Received)
    1578

    Default

    Typical lathe tolerances in terms of TIR are about 1-5 microns, trending about 2 um.

    The spindle bearing accuracies may be less, but this tends to cancel out when turning.
    See Timken et al lit regarding same.

    I believe runout cannot be measured under power, or above say 0.5 revs per second.

    I know that polished workpieces I have made have 0 microns runout, aka less than 1 micron, and can be accurate to 6.000 mm +/- 1.

    Of course they are not *really* accurate to that size, nor really accurate to that runout, but are probably within 1 micron of it.

  11. Likes Trueturning liked this post
  12. #30
    Join Date
    Jun 2012
    Country
    UNITED STATES
    State/Province
    Michigan
    Posts
    14,392
    Post Thanks / Like
    Likes (Given)
    4902
    Likes (Received)
    5204

    Default

    Prety small run out .000039 - .000078 for lathe work/machine.

    Could be weight out of balance of the chuck if turning at a higher speed, some could the pressure of one's hand pushing the chuck to turn, could be the bearings eccentric or eccentric mismatch or berinq quality or breaking wear.

    The under power RO should not even be considered because nobody does that.

    REF: https://www.kalsi.com/handbook/D04_S...ial_motion.pdf

  13. #31
    Join Date
    Jul 2019
    Country
    UNITED STATES
    State/Province
    Florida
    Posts
    5,202
    Post Thanks / Like
    Likes (Given)
    4872
    Likes (Received)
    972

    Default

    Quote Originally Posted by hanermo View Post
    Typical lathe tolerances in terms of TIR are about 1-5 microns, trending about 2 um.

    The spindle bearing accuracies may be less, but this tends to cancel out when turning.
    See Timken et al lit regarding same.

    I believe runout cannot be measured under power, or above say 0.5 revs per second.

    I know that polished workpieces I have made have 0 microns runout, aka less than 1 micron, and can be accurate to 6.000 mm +/- 1.

    Of course they are not *really* accurate to that size, nor really accurate to that runout, but are probably within 1 micron of it.
    Are you saying “under power” is placing the indicator on the od and engaging the spindle?

  14. #32
    Join Date
    Oct 2011
    Location
    West-Central Illinois, USA
    Posts
    1,427
    Post Thanks / Like
    Likes (Given)
    2306
    Likes (Received)
    674

    Default

    Quote Originally Posted by Trueturning View Post
    Are you saying “under power” is placing the indicator on the od and engaging the spindle?
    I believe that’s exactly what hanermo means, at ˝ RPM an indicator might tell you something. At higher speeds you need non-contact solutions to a surface. An poor-man's alternate that is used for running large long shaft compressors is to cast a strong light at the shaft and observe the shadow seen on a white backgound at 20-20ft & do the arithmatic. (recip I know, but a wobbling rotating shaft will cast a fuzzy edge that can be measured at a distance, even @ 3600rpm you see it)

    Good luck,
    Matt

  15. Likes gbent, pavt liked this post
  16. #33
    Join Date
    Sep 2011
    Country
    UNITED STATES
    State/Province
    Virginia
    Posts
    22,149
    Post Thanks / Like
    Likes (Given)
    7561
    Likes (Received)
    12732

    Default

    Quote Originally Posted by dian View Post
    the runout i see is aperiodic from what i can tell (like moving lobes). to put this into perspective im seeing an abec3 runout: abec9=2.5µ, abec5=5µ, abec3=10µ (kia for 50-80 mm id). bad news. from what i understand the runout in a ball bearing will be aperiodic to rotation with alower frequency and can be calculated for a specific case (no idea how realistic that would be).
    Knew I'd encounter it. Didn't BOTHER to do the "math". If you have about 12 hours to WASTE, you can do it ..on a ROLLER BEARING spindle.

    Prior to start, you'll need

    A) "whatever it takes" to establish a surface with no bumps. My case it was a sore tedious task to get dings and burrs stoned, 1942 10EE spindle bore, for arrival at a blued fit-up of one of Brian Miller's superb test bars.

    B) A "real" support for your DI. Noga & such just will NOT "cut it" when the DI is a ten millionths per division unit - better if yah have better.

    Otherwise? Holder flex be easily greater than what you are trying to assess! World is FULL of optimists who lova a Noga and think otherwise. "Optimist" being the polite term for "naif", "fool", or both.. in this case?

    Not Noga's fault. A mono arm is a mono. S**t moves.

    "To the rescue"? Raiding my extensive firearms cleaning goods for Dural cleaning rods, tubular. The fat 12-bore shotgun ones made up a decently stable three-point mount. Ok . it looked like metallic mutant spiders or king crabs shagging.. but it worked.



    C) A "degree wheel". Not that hard. It CAN be paper if you don't happen to already own one from back in the day work on IC engine valve & ignition timing, Old Skewl means.

    D) Polar-coordinate paper.

    METHOD:

    - Log the deviation every "so may" degrees to find and highlight the postion of the low and high spot at their "O'Clock" position. For MANY hand-turns. You'll really NEED about four iterations. Not of the spindle. Of the whole SET!

    That's over a dozen revolutions until it repeats. IF even it DOES exaclty repeat. Because it SHOULD NOT.

    Same as designing-in "hunting tooth" to an axle's differential gear set ratio.

    IF the bearing specifier COULD do, (s)he selected the same sort of "hunting" off-by (n) the rollers as is sought for gears. It greatly extends the LIFE.. and 10EE spindle bearings ARE legendary for long life.

    Multiple sets of many-turn data is so you can discard truly-off outliers, then average the apparently decent numbers. Still from more than two data records.


    CAVEAT: NONE of this stuff was/is calibrated. NONE had ANY "traceability".

    So WHY TF did I even BOTHER?

    Simple enough.

    10EE spindle bearings are rather dear.



    Could I count on the existing bearings to still support good work, and if so, just about HOW good?

    Bottom line?

    At 130 millionths, "net-net" and a sort of "DEE" shaped" plot as had an uncanny resemblance to test of studio-grade audio microphone directional field response curve shapes? See "cardioid microphone"

    I have a no-longer in-spec rolling element in the bearing-pair.

    Does that meant work @ 130 millionths out or WORSE?

    Positively NOT! Enter the "saving grace" of countervailing error!

    I did say "roller bearings"? Cutting tool tip position, nor TP grinder axis are NOT synced to that rolling element as is doing the "precessing" at a periodic rate produced by rolling element count, element diameter, hence inner and outer race differentiating ratio.

    The tool or grinding wheel will be skimming the high lobe second and subsequent round.. because it has MOVED!

    Is the actual net-net RESULT only 70 millionths "effective" TIR?

    I don't actually know. And care even less.

    Brand new, summer of 1942 the spec was either of 50 millionths, standard, or 35 millionths extra cost. And I hadn't even asked Terrie which it was shipped with.

    Nor whether shipped to Ford Motor Co.

    Or Ford Instrument Company. Look-up whom THEY were. They put these bearings-as-uber-fine differentiating and integrating devices to deadly WORK. On warships. Main battery pointing. Iowa-class and immediate predecessors.

    So THEY knew all those "bearing maths" I didn't BOTHER with.

    There's a possible source if you need one? See-also "Vannevar Bush" and HIS "lab work" at OSRD.

    Now . . I - not being "they" - no longer had the need to even split ONE thou but ever so rarely. Let alone try to split a "tenth".

    Avoiding roughly two "thou", but as-in SPEND, not "TIR" .. of two thousand US dollars was my REAL goal!

    I'm good with that. Penurious Iyam. Also lazy!

    NB: Plain bearing lobing is someone else's rice-bowl. I don't have any that are involved in anything all that close to "high precision".

  17. #34
    Join Date
    Jan 2007
    Location
    Flushing/Flint, Michigan
    Posts
    11,118
    Post Thanks / Like
    Likes (Given)
    666
    Likes (Received)
    8891

    Default

    The do not measure runout check under power brings up questions.
    Is the end part made when under power?

  18. #35
    Join Date
    Sep 2011
    Country
    UNITED STATES
    State/Province
    Virginia
    Posts
    22,149
    Post Thanks / Like
    Likes (Given)
    7561
    Likes (Received)
    12732

    Default

    Quote Originally Posted by CarbideBob View Post
    The do not measure runout check under power brings up questions.
    Is the end part made when under power?
    Ultimately, it is measurements of the "end parts" that determine the pragmatic net-net of the whole shebang, is it not?



    To the core, though.. a mechanical indicator is applying VERY low loading. and must, of itself, do exactly that to even serve it's OWN needfuls accurately. Bending the holder is not helpful

    Better example might be dynamic NON-contact monitoring, as has become practical- even come within the "budget" of OCD "hobbyists" off the back of surplused prior generations of technology, etc.

    Even so.. any time the work is not being loaded with cutting-forces - or a GOOD simulation of same?

    Neither are the spindle's bearings! "(Captain Obvious, etc...)

    Ultimately, it is measurements of the "end parts".....


    NB; W/R "OCD"? My OTHER 10EE.. nor the HBX-360-BC have not and WILL NOT get days of tedious pre-prep and any sort of "all day, sucker! " close look at wot's wot.. as described for the FIRST 10EE.

    I now know "about" what to expect. And how much .. or how little.. it actually matters.

    Like the rest of "we chikn's" .... so long as the work comes good enough?

    Need of new bearings can "self-ID" when they are ready!
    Meanwhile?

    "Run what you got, while you got it!"

  19. #36
    Join Date
    Jun 2013
    Country
    UNITED STATES
    State/Province
    New York
    Posts
    1,080
    Post Thanks / Like
    Likes (Given)
    734
    Likes (Received)
    609

    Default

    Quote Originally Posted by Matt_Maguire View Post
    An poor-man's alternate that is used for running large long shaft compressors is to cast a strong light at the shaft and observe the shadow seen on a white backgound at 20-20ft & do the arithmatic. (recip I know, but a wobbling rotating shaft will cast a fuzzy edge that can be measured at a distance, even @ 3600rpm you see it)

    Good luck,
    Matt
    That's really interesting to know, thanks! As an aside, a similar technique is used when manufacturing shotgun barrels. Any bulging (in a damaged barrel) or any bending is immediately found by looking thru the barrel while it is pointed at something like the sun, or at least a bright patch of sky. Indoor lighting not so good for this technique. Error shows up as a dark area inside the barrel which is otherwise mirror polished.

    I wouldn't hesitate to use a chunk of old shotgun barrel for a metrology tool.

  20. Likes Trueturning, Matt_Maguire liked this post
  21. #37
    Join Date
    Jul 2019
    Country
    UNITED STATES
    State/Province
    Florida
    Posts
    5,202
    Post Thanks / Like
    Likes (Given)
    4872
    Likes (Received)
    972

    Default

    Quote Originally Posted by Matt_Maguire View Post
    I believe that’s exactly what hanermo means, at ˝ RPM an indicator might tell you something. At higher speeds you need non-contact solutions to a surface. An poor-man's alternate that is used for running large long shaft compressors is to cast a strong light at the shaft and observe the shadow seen on a white backgound at 20-20ft & do the arithmatic. (recip I know, but a wobbling rotating shaft will cast a fuzzy edge that can be measured at a distance, even @ 3600rpm you see it)

    Good luck,
    Matt
    That is good. I do like to turn the part by hand preferably. Gear driven lathes I would neutral it out also not that it would make a huge difference. Polishing and finish are factors also. I think sometimes it is nothing much which might make us think something is wrong when it is our process which overlooks something.

  22. #38
    Join Date
    Feb 2010
    Location
    ch
    Posts
    3,478
    Post Thanks / Like
    Likes (Given)
    287
    Likes (Received)
    484

    Default

    "Brand new, summer of 1942 the spec was either of 50 millionths, standard, or 35 millionths extra cost":

    sorry, you lost me somewhere. what device are you talking about? btw, you forgot the hertzian stuff.

    background to all this:

    if you want to do precise work, they tell you to turn between centers. unfortunately, unlike on a grinder, the live center rotates. in my case it has 20µ tir if stuck in the spindle. din 8606/8605 specifies inside taper as 10µ/7µ, chuck taper as 10µ/7µ and (surprisingly) tir of work as 10µ/7µ. i suspect schlesinger is the same. so i have a slight problem there, spindle might be bent etc. so i came up with the idea to hardturn a center i would place in the chuck (like a sacrificial center) that, if in the same orientation (i was thinking a pin) would hopefully reduce tir and save messing with the chuck. so i did and i saw tir of 10µ without rechucking and thought i could be better. so i started investigating and it never went away.

    after i hardturned the inside jaws i get a tir of 10µ on the same diameter and of 20µ elsewhere. there is a chucking "procedure" involved. that is on a h4 ground parts which themselves have a tolerance of 5-10µ, depending on size. so the whole between centers idea becomes pointless.

    has anybody measured tir on his live center recenly? (manual lathe in good condition.)

  23. #39
    Join Date
    Sep 2011
    Country
    UNITED STATES
    State/Province
    Virginia
    Posts
    22,149
    Post Thanks / Like
    Likes (Given)
    7561
    Likes (Received)
    12732

    Default

    Quote Originally Posted by dian View Post
    sorry, you lost me somewhere.
    Always did appreciate how predictably consistent the residents of Schweiz try to be...

    So "that's a feature, not a bug"


  24. #40
    Join Date
    Jun 2012
    Country
    UNITED STATES
    State/Province
    Michigan
    Posts
    14,392
    Post Thanks / Like
    Likes (Given)
    4902
    Likes (Received)
    5204

    Default

    QT: Has anybody measured tir on his live center recently? (manual lathe in good condition.)

    One of my tasks was grinding live centers. I would set them up on a B&S 13 grinder and grind the 60* while the shank was dead(stopped) and the point turning about 60 to 100 RPM on its won bearing set with pressure to the foont to out some load in the right direction.. They were inspected with a tenth indicator for no noticeable needle movement so about less than .0000050 max error.

    Some just off the machine and some with having new bearings installed.

    20um is a mile for a live center...even 10um is still a mile.
    A solid dead center or a center turned in the lathe/grinder spindle and then allowed to turn with the part is a very good method.

    Some China live centers turning at a tenth or two could not be trued because the main shaft had wobble.


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
  •