Measure runout on shaft while rotating? 1

[cnuk]

I am looking for a method of measuring runout on a shaft while it is rotating. The shaft will be approx 2" diameter, rotating 500RPM, and there may be a film of grease on OD of shaft. There is a small 1/4" dia hole providing access to the shaft from a housing that covers it. For a "static" measurement we are simply putting a dial indicator on the shaft and rotating the shaft by hand to get a reading. The problem is that I think the readings get worse under operating conditions but I have no way of measuring it. I am trying to resolve whether the shaft is running out, for example, .005" or .010", not down to 1 tenth.

Any ideas? The only link I can find is at

http://www.lionprecision.com/targa/index.html

but it looks expensive and not really shop suitable.

Thank You

 

[sbozy25]

So if I understand this correctly, you think that as the shaft rotates it becomes "out of round" kind of like it is off balance?

I would think that if you got a good digital indicator gauge you could stick it in the hold and measure. Out side of that I think your best bet would be to re-produce operating conditions on a shaft that is not incased and try to measure off that.

 

[eromlignod]

You'd better put a hell of a spring on it at that rpm or you're going to have a terrible time tracking your reading. The speed that the pushrod and the needle of your indicator moves will have a lot of inertia at that rpm and frequency, not to mention it will be hard on the gauge.

You might want to look into some optical/laser measurement methods.

If the runout is enough to see visually with the naked eye, you might want to try shining a strobe on it. By varying the frequency of the strobe, you can make the shaft appear to move in slow-motion. Alternately you could also strobe your laser sensor for the same effect.

By the way you can pick up a cheap stroboscope on Ebay pretty easily. They come along almost every day.

Don
Kansas City

 

[Ataloss]

For an optical/laser measurement device, I assume the oil/grease film would affect the measurements (depending on the thickness and how well the laser passes through the oil). Could some sort of scraper be used while the measurement is being taken to clean the oil off the shaft?

 

[electricpete]

Sounds like a job for inductive proximity probes such as sold by Bently Nevada

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[Heckler]

Here is a company that does balancing. They balance a small shaft for us to 90K RPM. Ask for Chuck....he an old timer that knows a ton about high/low speed....just don't get him on the subject of China.

http://www.polynetics.com/

Heckler
Sr. Mechanical Engineer
SWx 2007 SP 3.0 & Pro/E 2001
XP Pro SP2.0 P4 3.6 GHz, 1GB RAM
NVIDIA Quadro FX 1400
o
_`\(,_
(_)/ (_)

(In reference to David Beckham) "He can't kick with his left foot, he can't tackle, he can't head the ball and he doesn't score many goals. Apart from that, he's all right." -- George Best

 

[LarryEF]

This is out of my field, but could you use an LVDT or resistance type linear measuring instrument with a fairly high sampling rate?

Larry

 

[MikeHalloran]

Jam a 1/4" wooden dowel against the shaft, hard, using something compliant, like a Pink Pearl, to push against its end. Use a crowbar or something like that, not your thumb, to back up the eraser. The idea is to get the dowel to follow the shaft surface for a few revolutions.

Then measure the scar that the hole leaves on the side of the dowel.




Mike Halloran
Pembroke Pines, FL, USA

 

[pmover]

although i do not recommend this procedure unless due caution and thought is use - a responsible person too.

the human finger is capable of detecting movement as little as one-thousandth of an inch. i will let you improvise a device that allows the human sensor to determine and quantify movement.

another method is to use proximity probes mfg by BN.

an aside, if you suspect the runout is greater while operating than while safely and properly measuring, then other problems exist with the rotating assembly. that is unless the shaft is poorly designed.

i would definitely check the balance of the rotating assembly and inspect the bearings (support structure). any imbalance may very well cause shaft bending and/or premature bearing failure. if bearings have failed, then carefully inspect the rotating assembly and have the bearings return to mfg for analysis and cause of failure.

good luck!
-pmover

 

[tygerdawg]

I second electricpete's suggestion.

If you only have a 1/4" diameter hole, then you are limited. The solution depends upon how deep is that hole.

Keyence, Pepperl+Fuchs, maybe Turck & others make inductive analog position sensors, but they're not necessarily cheap. But it would certainly have the frequency response and dimensional range to give you the info you want without having to touch the shaft. The grease shouldn't affect the readings. You would also need something to scale/display the readings, and maybe something to log it.

If you want to go cheap, perhaps you could buy some sensor hardware and partner with your local university to use their student labor and laboratory reading gizmos and software.

The next alternative would be laser triangulation sensors from Keyence, Banner, many others. And still not really cheap. But the same readout requirements apply, and then the grease would be part of the reading.

And ultimately you have to ask the question "how much is downtime and repair worth if the shaft fails, versus the cost of finding out the answer quickly" ?

TygerDawg

 

[ScottyUK]

electricPete and Tygerdawg are on the right lines:

Bently Nevada, now devoured by GE, and CSI, swallowed up by Emerson a while ago, are two of the bigger players in the vibration analysis market. They have equipment designed specifically for this type of application. The processing of the signal is important, especially bandwidth. Many of the common proximity sensors are bandwidth limited by the electronic signal conditioning they incorporate.

http://www.ge-energy.com/prod_serv/products/oc/en/bently_nevada.htm

http://www.compsys.com/

----------------------------------

Sometimes I only open my mouth to swap feet...

 

[GregLocock]

I'd still stick a screwdriver down the hole before getting too analytical- 500 rpm is very slow

Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[Tmoose]

I've used a regular vibration analyzer set to MILS with the pickup (either accelerometer or velomiter) on a shaft rider as an electronic dial indicator. At 500 rpm I'd expect to be able to simply poke a metal rod as a shaft rider down the hole, after I tested for shaft keys etc with a coffee stirrer

 

[zekeman]

I would do the following:
1) juryrig a wire linkage to the shaft which will be acting like a cam to multiply the radial motion by a factor of say 5. The low speed 500RPM shouldn't bend the linkage.
2) Get a stobe light to "freeze" the motion so that you can easily detect 5x the runout or in your case, if you are looking for .005, then .025 should be easily seen.

 

[BobM3]

If the shaft deflects worse while running (dynamic problem), are you sure that the maximum deflection is inline with your 1/4" hole?

 

[byrdj]

two inductive prox probes
one to read displacement, the other set to "see" a once per rev notch in the shaft for phase (notch say 0.030 X 0.030 cut into shaft)

a dual channel o-scope with a sweep of 0.012 sec/div (10 division axis)

channel a being displacement with scale set per 100mv/div (0.001" for prox probe)

channel b being once per rev with scale set at 500mv/civ

 

[cnuk]

Thanks to all those who replied. I probably owe a little better explanation since there was initially some confusion. My 2" diameter shaft rotates inside an outer steel housing that is approx 3/8" thick so the depth of the hole is approx that deep. It is currently 1/4" dia but could be made up to 1/2" dia if that helps. The shaft runs on bearing that are a sliding (clearance) fit into the housing. The shaft has load transferred to it by some eccentric movements (by design) elsewhere in the system. The load transferred will go up with speed (thus the runout will increase) which is why I believe the reading taken turning the shaft by hand and using a dial are unrealistically low.

I will look into some of the sensors suggested. I don't claim to be an expert in proximity sensors but it sounds like I need to read up.

Thank You All

 

[automatic2]

What are your objectives in considering this runout? Are you concerned that the shaft is deflecting under load. Gauging the shafts torsional strain, or load force applied, may be a more 'telling' measurement. Otherwise, your runout will be a function of bearing clearance or stability of axial support.

 

[IRstuff]

Are you really trying to measure elongation or wobble? A 0.2% elongation just doesn't seem possible at normal operating temperatures.

A wobble, because the shaft is bent would be different. You could measure it with a phonograph needle mounted on a micrometer.

TTFN

Eng-Tips Policies FAQ731-376

 

[cnuk]

I am trying to measure the runout of the shaft because also located on the shaft is a radial lip seal. The lip seal life is largely affected by runout and I think some of the failures we are having are because the runout increases with RPM. The increase in runout is due only to clearance in the bearings and elsewhere in the system and the radial load on the shaft, not things like the shaft going out of round.