measuring runout

dian · Mar 10, 2021

i turned a cylindrical part and measured runout in the chuck using a 1µ mitutoyo with a large radius tip and turning chuck by hand. i saw 1-2µ. great. then i measured it with spindle running at 25 rpm. 11µ. hm. thinking vibrations. disengaged the gear and let the machine run. no movement. turned spindle by hand and got 1-2µ again. got a 2µ graduated test indicator with a 3 mm tip and it shows 12µ.

1. whats happening? do fine vibrations transmitted to the mitutoyo make it show correctly by eliminating hysteresis?

2. guessing that 11-12µ was what it actually is i tried adjusting the plain bearings wihout any success (turned another piece). how bad do you thing this is? (lathe has 370 mmm swing.)

crossthread · Mar 10, 2021

I don't think it is vibrations. The only thing that vibrations should effect is stiction in your gauge and I doubt that is an issue. How does your finish look? I think you are correct in looking at your main spindle bearings. You say they are plain bearings like the SB used. I think that what is happening is that the drive mechanism is pulling the spindle to the side and it is oscillating slightly. Could be wrong of course but to me it's the only thing that makes sense. Try taking a shim or two out of the head stock and see if it goes away. I know you said you adjusted the bearings but they are iffy to adjust. On my old SB I took shims out until the bearing caps would just be a bit warm at high speed. I tried Plastiguage with no real success there. I don't know exactly why except the tolerances on a lathe spindle are probably tighter then an old Ford tractor.

eKretz · Mar 10, 2021

Turn a diameter at lower RPM and another at higher RPM. Check both and see if the runout is the same as what you got at low speed by hand or better yet between dead centers or out of the lathe in v-blocks.

jim rozen · Mar 10, 2021

STOP measuring your workpiece with *any* indicator when the part is under power. Those results are meaningless - throw those numbers away.

mnl · Mar 10, 2021

The gauge has a spring constant, so if there is any surface texture, that texture will drive the system. When the rpm hits the magic number you will get a resonance and potentially a big oscillation on the gauge.

Conrad Hoffman · Mar 10, 2021

Runout is usually measured on an air bearing rotary fixture. Could probably use a Zero Spindle, though they can be a bit stiff. All sorts of thermal and speed-dependent stuff probably happens on the lathe spindle.

michiganbuck · Mar 10, 2021

I agree with all the above posts…
Perhaps an air bearing with pressure to radial and axial directions would give near-absolute precision. The part finish need be very good and the indicator point need be air also. The driving force needs to be remote (away from the spindle.) The whole shebang in a salt mine and a sensor to watch for earth vibrations.

The company used to calibrate gryows in Michigan salt mines, ops: keep that on the QT.

Not related to the ops hand/power motion question but a good experiment is to fixture or chuck a part under a micron indicator and then give it a push with one finger, two fingers, one hand, two hands and note the readings, also try this in different directions.

edwin dirnbeck · Mar 10, 2021

It is difficult to GRIND something TRULY ROUND within a millionth of a meter,this is about 1/3 of a tenth of a thousandth of an inch.I find it hard to believe that you TURNED something round to this level of accuracy.There is the possibility that all of the variables that occur to make something OUT OF ROUND,could align themselves and make something round.Edwin Dirnbeck

eKretz · Mar 10, 2021

Yeah the spindle itself almost certainly doesn't turn that round. But the change in runout by 11 microns or nearly .0005" may warrant some investigation to insure that it's just a false indication.

dian · Mar 10, 2021

i usually dont do any measurement under power and thats probably why i havent noticed this behaviour of the mitutoyo before. however, as this corresponds to what i see on the test indicator im led to believe the 11µ are real.

i can adjust "preload" with a screw, now im at the point where the chuck makes one turn with one hand, maybe 1-1/4 turns if i try hard (fingers hurt). if i grab the chuck with two hands i can move it by exactly 2µ, which seems o.k. if i press on the part there is no movement. i tried the same procedure on the small lathe and im seeing 2-3µ (even when running). thats to be expected, i think. interestingly the small, 100 mm chuck doesnt move at all.
then i tried again on the big one and now im seeing 9µ.

i agree, its not too bad, but im almost sure it was not that much before, i would have noticed. lathe doesnt get used that much, so not a wear issue. i dont have any "air stuff" at my disposal. was turning 30 mm diameter, btw.

i guess i cant do much about it, but kind of disappointed. im getting about 20µ runout when carefully re-chucking (was always proud of that) and half of it is in the bearings?

edit: of course i have no idea how round the part really is. this is just from as turned to as measured.

michiganbuck · Mar 10, 2021

Not uncommon to run gauges at 6 to 12 millionths of an inch..about .000039" is a millionth of a meter.

Mostly just need a couple of thousandths or tenths on lathe work.
Most lathes that have run tenths (.000x) are very hard to find...even with polishing to finish the heat makes tenths difficult.

jim rozen · Mar 10, 2021

Quickest trip out the door in a commercial shop: take their high end indicator and slap it on a part rotating under power.

9100 · Mar 10, 2021

Depending on speed, my 14 1/2" south Bend spindle will move in a slightly triangular pattern. If I turn a piece of hard aluminum, chosen for needing low force to cut it, getting the best finish the lathe will produce, then shut it off and blue the part and make the same cut again without changing the settings, the part will have three bright lines. Apparently, the spindle is moving in a slight triangular motion and stopping and restarting moves the triangle to a different rotational orientation. Putting an indicator on a part and prying the chuck up with a 2 X 4 gave .002" when it was new, now .0025" A journal bearing must have some clearance. They work better when running at a moderate speed than they do at a very slow one because the oil wraps around the spindle and centers it. When indicating a part in by hand turning, I have to turn an increment and wait for the spindle to push oil out of the bottom of the bearing and rest against against the bearing bottom.

Bill

dian · Mar 11, 2021

thanks for the data. south bend has plain bearings, righ? so i was thinking, in a plain bearing the spindle will "climb" with rotation at low speed at settle on the opposite side onece the hydrodynamic regime is established. where does the tri-lobing come from?

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).

so questions on the ones in the know:

1. im seeing "lobing" (several deviations per rotation) on the smal lathe with roller bearings as well. where does it come from?
2. where does the "lobing" with the plain bearings come from?

toolsteel · Mar 11, 2021

Couple thoughts they may have little or no bearing on your specific issue....
Is it the original tip that came with the indicator...or at least one of the exact same length as original?
I assume you have the tip at the correct angle to the part....12 degrees I think is what Interapid recommends. I have witnessed guys complain about indicator readings with the tip at damn near 90 degrees to the part.
I am not a fan of measuring with an indicator....when I do I try to use it to compare to a known dimension...gage blocks etc. I have been told fwiw a Test indicators can be used to measure.
The lobing you are seeing....is it in any way relatable to the chuck you are using? Assuming a 3 jaw...do you see 3 similar lobes either in line with the jaws or between the jaws equally? obviously part material, features and chuck pressure would make a difference with this
Is the tip on one of the indicators sticking??? any flat spots on the tip???? was one of them ever repaired by someone not qualified to do so????

9100 · Mar 11, 2021

toolsteel said:
Couple thoughts they may have little or no bearing on your specific issue....
Is it the original tip that came with the indicator...or at least one of the exact same length as original?
I assume you have the tip at the correct angle to the part....12 degrees I think is what Interapid recommends. I have witnessed guys complain about indicator readings with the tip at damn near 90 degrees to the part.
I am not a fan of measuring with an indicator....when I do I try to use it to compare to a known dimension...gage blocks etc. I have been told fwiw a Test indicators can be used to measure.
The lobing you are seeing....is it in any way relatable to the chuck you are using? Assuming a 3 jaw...do you see 3 similar lobes either in line with the jaws or between the jaws equally? obviously part material, features and chuck pressure would make a difference with this
Is the tip on one of the indicators sticking??? any flat spots on the tip???? was one of them ever repaired by someone not qualified to do so????

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

pavt · Mar 11, 2021

I would not have any confidence in a measurement that small, when it is in the chuck. I think a better test is to use vee blocks and rotate your part on those. I know my spindle bearings are good, but any measurement I took with a 1-micron indicator would get lost in the "noise" that comes from all other factors. When I am measuring to "tenths" I do it on the surface plate, very carefully --

jim rozen · Mar 11, 2021

9100 said:
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

Probably you should try measuring it while it's spinning at 100 rpm, might make it different....

Matt_Maguire · Mar 11, 2021

9100 said:
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

With a fluid bearing the oil wedge can move around for any number of reasons. If the shiny lobes are straight along the shaft and regular spacing it’s more likely something in the drive or load doing it.

Used to look at shafts with magnetic (proximity) probes using IRD equipment & O’scopes trying to figure out things happening… I agree with all above comments to not use a DI or DTI on a rotating shaft under power.

On the bench a 60 & 90 v-block can be used for sorting lobes (some).

Good luck,
Matt

9100 · Mar 12, 2021

jim rozen said:
Probably you should try measuring it while it's spinning at 100 rpm, might make it different....

I haven't tried that. I am not using an indicator at all in this test. I think it is a resonance between the bearing clearance, belt tension, spindle mass and speed. It may be akin to the triangle pattern a drill makes when started without a center drill first.

Yes, the lines are straight and parallel to the axis.

In one of my incarnations I made concave aluminum mirrors for experimental spectrophotometers. I cut the concave on a 9" Logan lathe using a tool bit on a rod pivoted by a leaf spring on the tailstock and fed by resting on a notch held by the cross slide. I lapped them on a pitch or honeycomb foundation lap, first by hand just like a telescope maker, then on a machine much like Strong describes in his book "Procedures in Experimental Physics". After a few minutes, they showed a sunburst pattern, clearly a result of having high and low spots in the surface. I never measured them but they were quit small, of the order of millionths. Before long, they polished out. The lathe was new, bought for our R&D operation. The spindle bearings were ordinary ball bearings, I'm sure they were not very high quality ones, which would have cost more than the purchase price of the lathe. The sunburst lines were straight, implying something like each ball passing a loading slot.

Bill

measuring runout

Titanium

Titanium

Diamond; Mod Squad

Diamond

Hot Rolled

Diamond

Diamond

Hot Rolled

Diamond; Mod Squad

Titanium

Diamond

Diamond

Diamond

Titanium

Titanium

Diamond

Stainless

Diamond

Stainless

Diamond

Similar threads