I am a little confused. How could i appropriate call when i have related movement between outer ring bearing and housing???

clearance and looseness??

Very thanks.

Fabián.

i think the right description is looseness because "clearance" is a definintion for space between two surface and able to decrease and and increase. in the bearing ring and housing surface no place to the "clearance".

You can have a clearance fit or an interference fit or a transition fit (in between).

Generally outer rings have more of a clearance fits to housing and inner rings have more of an interference fit to the shaft.

As a result, some movement of outer ring in the housing is normal and expected. But if you think there is more movement than normal, or more clearance fit than appropriate, then you can say: "The shaft/bearing housing fit is too loose" or "The outer ring is spinning within the housing". Spinning within the housing implies more movement than the normal expected slow creep of the outer ring within the housing.

It's not clear from your post what is the basis for this type of diagnosis. It would be easy to say after you remove the bearing for inspection (measure as-found outer ring and housing dimensions and look for evidence of movement). But I'm not sure how you can detect/prove this before. There may well be some recognizeable pattern that I'm not familar with. I'd be curious to learn about how you reached the conclusion.

I didn't mean to steal this post, just want to ask forum's opinion on the following related subject.

Will losseness in bearing-shaft fit (interference fit by design )
exhibit itself differently in the spectrum from excessive looseness in bearing-housing fit ( loose fit by design) or excessive internal antifriction bearing clearance?

David

It seems that when looking at this looseness strictly from the spectrum as in 1X, 2X, 3X RPM or some what higher multiple of RPM’s, looseness would show itself regardless of where it would be on the component. Typically vibration amplitude is highest at the location where most of the looseness can be found. But, to go further and look for harmonics of components then something different is seen. Say it is a fit issue with the shaft to inner race then harmonics of the inner race might be seen and the same can be said for the outer race fit in the housing. Higher frequencies could also show themselves in the spectrum as in metal-to-metal contact, sliding of the two surfaces as they move over each other. The HFD or gSE readings would probably be higher than normal. It would depend on how far a person would want to break it down as what could be discovered.

hello

losseness are 2 types:
1- Structral losseness : losseness between 2 stationary parts .
diagnoses : High 1X radial vibration . 1X H >1X V.(example : machine and base , bearing housing and machine body)

2- Rotating looseness : looseness between rotating part and stationary part (rotor and bearing).
Charactrized by : High vibration 1X and its harmonics 2X,3x,...etc may reatch 10X

bye

quote:

Originally posted by David_G:
I didn't mean to steal this post, just want to ask forum's opinion on the following related subject.

Will losseness in bearing-shaft fit (interference fit by design )
exhibit itself differently in the spectrum from excessive looseness in bearing-housing fit ( loose fit by design) or excessive internal antifriction bearing clearance?
David

Just to interject a bit of 'reality', here, while it is quite true what electricpete said that generally bearings are a snug fit on the shaft and a clearance fit in the bore, the plain fact of the matter is that things have to be assembled in the first place, and if you press two bearings together on a shaft to 'sandwich' a machine of some sort, you often end up having to whale on the end of the shaft with a hammer to get it to free up enough to rotate without heavy friction in the grooves of the bearings. Even after you whack it with a hammer you're in a worrying place because heat cycling is going to stress the bearings as the shaft expands and contracts. It is far easier, and just more sane to have one bearing tight to the shaft and the other bearing either tight to the bore or even better loose to both but held in place by a flange bolted to the outside of the bore, so that it assembles easy, floats on the shaft, but can be fixed in place easily. To pull the shaft you just pull it in the right direction and out it comes with one bearing attached.

Good comments. There may be more than one way to look at things and more than one way to do things. I'll tell you how I look at it and how we do it.

Recommended bearing shaft and housing fits for motors are given in EASA AR100 page 15/33 (for ball bearings) here:
http://www.easa.com/indus/AR100_0406.pdf
As described above, shaft fits are tighter than housing fits.

These recommendations are similar to what is given by the bearing manufacturers and by AFBMA.

More importantly, this is what you get when you buy a motor from an OEM and this is what you will normally find when you take measurements of shaft and housing during bearing replacement (if it's not, then you need to evaluate repairs). If you ever end up with a loose fit between inner ring/shaft, then it's because of wear, or worse yet you have intentionally machined the shaft.

A machine might operate satisfactorily that way (loose shaft fit), but it is not the ideal approach. A loose inner ring can sometimes accelerate to a problem a lot easier/faster than a loose outer ring.

We have no problems replacing motor bearings on-site with the recommended tight fit to the shaft on both beairngs. A typical sequence to replace bearings on a 100hp TEFC horizontal motor might be something like: uncouple, remove couping hub, remove fan shroud/fan/both endbells, remove bearings with puller. Now reassemble (rotor never comes out of stator). For larger bearings heat replacement bearings with a heater immediately before installing onto shaft, press endbells on to bearings, reinstall fan, shround, hub, recouple. (normally we wouldn't realign if all we have done is replace bearings). We also include in our motor repair spec for work done off-site, that the EASA bearings fits are to be used.

I'm not sure what you mean by "if you press two bearings together on a shaft to 'sandwich' a machine of some sort, you often end up having to whale on the end of the shaft with a hammer to get it to free up enough to rotate without heavy friction in the grooves of the bearings"

This message has been edited. Last edited by: electricpete,

quote:

Originally posted by electricpete:
Recommended bearing shaft and housing fits for motors are given in EASA AR100 page 15/33 (for ball bearings) here:
http://www.easa.com/indus/AR100_0406.pdf
As described above, shaft fits are tighter than housing fits.

I only checked two at semi-random (two from the top line), but the fits seem pretty close to me.
The tightest shaft fit for that 10mm ID is 10.00505mm, or about two ten-thousandths of an inch of interference. the tightest bore fit for a 30mm OD is 29.99994, which is size-for-size. While being a bit looser, two tenths is not a great deal looser. Honestly, in a shop or plant environment I doubt many people could reliably measure within two tenths. It only takes a couple of degrees of ambient temperature change to throw you out, or holding the micrometer in your hand for a minute can change the reading that much.
That having been said, I'm splitting hairs, here. You said the bore fits were tighter, and they are (according to the tiny sampling that I took). My comments were about bearings in general. Assembling machines is what I did for a living for years before I switched to maintenance, but I didn't assemble electric motors. Maybe they're different. If I were building an electric motor (a good-sized one) I'd have the drive-end bearing tight to the shaft and locked in the frame by a flange, but allow the NDE bearing to float.

quote:

A machine might operate satisfactorily that way (loose shaft fit), but it is not the ideal approach. A loose inner ring can sometimes accelerate to a problem a lot easier/faster than a loose outer ring.

Keep in mind that by 'loose' I'm talking about half a thou loose, not rattling around. Unless the surface finish is precision ground and highly, highly polished the fit will degrade under any kind of duress as the high points get worn down. With a turned finish I'd go a lot tighter than a ground finish.

quote:

We also include in our motor repair spec for work done off-site, that the EASA bearings fits are to be used.

I honestly don't check the sizes. If the old bearing came off properly, then I figure the new one will go on properly. Bearings are very reliably sized.

quote:

I'm not sure what you mean by "if you press two bearings together on a shaft to 'sandwich' a machine of some sort, you often end up having to whale on the end of the shaft with a hammer to get it to free up enough to rotate without heavy friction in the grooves of the bearings"

When you have a gearbox or some other machine that had bearings pressed to a shaft from both ends, usually you end up driving them on using a sleeve over the shaft to push on the inner race of the bearing. You put one on the shaft in the right spot, then install the shaft (with bearing) into the machine. Now you have to drive the other bearing on while the shaft is in the machine. You drive it in as far as it will go, but when you lock the bearing in place with the flange you find that the outer race was a bit farther out than the inner race (understandable, because you were pressing by the inner race) and now that the flange has been tightened to lock the outer race in place you have a discrepancy between the location of the outer and inner races, so the grooves bind on the balls. Grab a big dead-blow hammer and beat the ends of the shaft, and everything sorts itself out. Yeah, it weirded me out, too, the first time I saw my old boss take a sledge-hammer to a newly-built high-speed spindle with four $900 (each) bearings in it, but that's how it's done. If one bearing can have a bit of clearance on the shaft, or if the outer diameter can float a bit, then that issue is eliminated.

By the way... I have never once managed to pull a DE bearing from an electric motor without having to remove either it, or the machine it was coupled to, from the base plate. How do you do it when the shafts of the two machines are only separated by an inch or so?

Good comments.

quote:

By the way... I have never once managed to pull a DE bearing from an electric motor without having to remove either it, or the machine it was coupled to, from the base plate. How do you do it when the shafts of the two machines are only separated by an inch or so?

We have some machines that have enough clearance to pull the hub, enbdbell, and the bearing without moving the motor and some that don't. I was certainly incorrect to imply it could be done on all machines.

One of ours that doesn't require moving the motor is a 100hp 3600rpm TEFC motor driving pump thru Thomas Shim pack coupling. One that does require moving the motor is a 75 hp 3600 rpm TEFC motor machine driving pump thru gear coupling.

To gain a little extra room to work at the inboard end, you can remove outboard endbell, remove inboard endbell from frame so it is hanging on the shaft, and move the whole rotor outboard, but don't do this unless you are confident you won't cause damage while moving the rotor. You can help assure that by lifting the outboard rotor extension to slide a protective sheet into bottom airgap between rotor and stator from the outboard end, and/or by supporting the rotor from both ends while moving it).

quote:

When you have a gearbox or some other machine that had bearings pressed to a shaft from both ends, usually you end up driving them on using a sleeve over the shaft to push on the inner race of the bearing. You put one on the shaft in the right spot, then install the shaft (with bearing) into the machine. Now you have to drive the other bearing on while the shaft is in the machine. You drive it in as far as it will go, but when you lock the bearing in place with the flange you find that the outer race was a bit farther out than the inner race (understandable, because you were pressing by the inner race) and now that the flange has been tightened to lock the outer race in place you have a discrepancy between the location of the outer and inner races, so the grooves bind on the balls. Grab a big dead-blow hammer and beat the ends of the shaft, and everything sorts itself out. Yeah, it weirded me out, too, the first time I saw my old boss take a sledge-hammer to a newly-built high-speed spindle with four $900 (each) bearings in it, but that's how it's done. If one bearing can have a bit of clearance on the shaft, or if the outer diameter can float a bit, then that issue is eliminated.

Your binding scenario does not seem applicable for the assembly sequence I describe. Both bearings are put onto the shaft before either endbell/housing is assembled onto the bearing. The endbells housings should slide onto the bearings fairly easily because of that looser fit. I have not heard of binding like that but I don't work much with other types of equipment than motors.

quote:

I honestly don't check the sizes. If the old bearing came off properly, then I figure the new one will go on properly.

You can probably get away with not checking bearing fits in most cases. As you say, if there is no sign of significant sliding, spinning, heating or distress, you're probably ok. BUT, every facility has a different balance between reliability and maintenance effort. For highest reliability, you should absolutely invest the 10 or 15 extra minutes to check those fits during every bearing replacement imo. It is also a standard practice in my industry.

quote:

Bearings are very reliably sized.

We find more problems with the shaft and housing dimensions than the bearings.

quote:

Keep in mind that by 'loose' I'm talking about half a thou loose, not rattling around. Unless the surface finish is precision ground and highly, highly polished the fit will degrade under any kind of duress as the high points get worn down. With a turned finish I'd go a lot tighter than a ground finish.

Interesting comments about finish. You're right that it plays an important role in allowing/preventing slipping. The new bearing of course has a specified finish. The shaft and housing seats are supplied with a certain speicified smooth finish and we're supposed to inspect them during bearing replacement to ensure they remain in good condition. That is usually a pretty subjective process (eyeball and fingernail). More exocitc measurements are possible but usually not used during maintenance. I guess the main danger we expose ourselves to is reassembling a machine with rough housing finish so that a designed floating bearing cannot slide within the bearing as intended. Arie mol has also pointed out a lot of other factors that can get in the way of floating bearings moving as designed in another thread this past spring.

quote:

Originally posted by electricpete:
One of ours that doesn't require moving the motor is a 100hp 3600rpm TEFC motor driving pump thru Thomas Shim pack coupling. One that does require moving the motor is a 75 hp 3600 rpm TEFC motor machine driving pump thru gear coupling.

Almost every motor we have that doesn't drive through belts or chains uses a jaw coupling for drive. The only exceptions I can think of off the top of my head are two DC motors with 'rubber tire' elastomeric couplings, but they still don't have enough room to pull the bearings. I've never worked with those shim pack couplings.

quote:

To gain a little extra room to work at the inboard end, you can remove outboard endbell, remove inboard endbell from frame so it is haing on the shaft, and move the whole rotor outboard, but don't do this unless you are confident you won't cause damage while moving the rotor. You can help assure that by lifting the outboard rotor extension to slide a protective sheet into bottom airgap between rotor and stator from the outboard end, and/or by supporting the rotor from both ends while moving it).

I think I'll just stick with removing the motor and taking it to the bench. I hardly ever have to replace bearings in electric motors, so I don't mind the time taken to re-align them. Plus at least I know then that the alignment is right. The previous regime here was... untrained.

quote:

You can probably get away with not checking bearing fits in most cases. As you say, if there is no sign of significant sliding, spinning, heating or distress, you're probably ok. BUT, every facility has a different balance between reliability and maintenance effort. For highest reliability, you should absolutely invest the 10 or 15 extra minutes to check those fits during every bearing replacement imo. It is also a standard practice in my industry.

What instrument(s) do you use to measure the bore?

quote:

Plus at least I know then that the alignment is right. The previous regime here was... untrained.

Good point. There are some schools of thoughts that alignment should be checked periodically anyway (not that we do it). So the extra time to realign is not necessarily wasted and you can certainly gain back time on some other parts of the job if you can pull the motor out to the shop.

quote:

What instrument(s) do you use to measure the bore?

I don't do the measurements myself. You probably know more about it than I do. I know the technique is very important and our guys practice their technique to be able to reproduce known measurements. I have seen our guys use snap gages to capture the inside distance of the housing bore. Then they remove the snap gage and measure it with an outside mic.

We ask them to take three measurements 60 degrees apart and if we're being really precise, two different axial locations in the bearing. (this is the type of detail that comes out of lots specifications and proceduresf... the way we do business).

This message has been edited. Last edited by: electricpete,

By the way, this is just me venting spleen about the ridiculous expectations placed on metrology in an industrial environment. This shouldn't be taken as being directed at any particular person.

quote:

Originally posted by electricpete:
I don't do the measurements myself. You probably know more about it than I do. I know the technique is very important and our guys practice their technique to be able to reproduce known measurements. I have seen our guys use snap gages to capture the inside distance of the housing bore. Then they remove the snap gage and measure it with an outside mic.

That sounds like a telescopic guage (shaped like a 'T'?). I wouldn't expect better than half a thou (on a really good day) accuracy with those things. Mind you, if all you're looking for is large variances then they'll work just fine. If you really want to measure _accurately_ to tenths in a bore, though, then you need specialized metrology, and you need to do it in a controlled environment. Probably a bore guage with a calibrated ring gage of a close size. You can't use go/no-go gages because they won't detect an oval bore... anyway, these problems are the primary reason why I usually don't bother measuring bores - because I know that the number I get will not be accurate enough to be worth doing. So many precision measuring/reading tools are sold with resolution far beyond the accuracy that it's downright disturbing. You can buy DROs, digital slide calipers and other tools that will display inches to five or even six decimal places. Not just ten-thousandths of an inch but hundred-thousandths of an inch. Take a look at the manufacturer's specification and you'll see that the guaranteed accuracy is plus or minus one whole thousandth. That's a 2 thou spread when it resolves to half a percent of that. Makes you think. Anyway, that's a pet peeve of mine. To me, a precision measuring tool should have at least five times more accuracy than resolution. If you take a reading and it says 'n' then you should be able to be confident that it really is 'n'. If you buy, for instance, a Fowler electronic micrometer which resolves to half a tenth... for $112... do you think it has accuracy to a hundredth of a thou? I sure don't. If I had to use one I'd just put tape over the last two or three digits and hope for the best. This is turning into a diatribe so I should just stop now.

quote:

We ask them to take three measurements 60 degrees apart and if we're being really precise, two different axial locations in the bearing. (this is the type of detail that comes out of lots specifications and proceduresf... the way we do business).

That's a good procedure, but the two axial locations isn't very feasible in a short bore. Telescopic guages (if, indeed, I'm not mistaken in thinking that's what they are) need a reasonable amount of room to 'swing' in use. Less constricted are tubular inside micrometers (a far better way to measure a bore). Neither of these, though, are well equipped to measure two axial locations on a short bore like that for a bearing of less than perhaps 3/4" (20mm)-ish thickness, which is a pretty reasonably sized bearing in the grand scheme of electric motors. Anyway, you have a process, it's approved, so you use it and it no doubt meets your 'due diligence' requirements. It's better than no process at all, and if the bore is banged up enough to be bad it will _look_ bad, and no doubt the technician would see it was messed up and just fudge the numbers to make it fail. That's what _I'd_ do.

Yes, it's a telescopic T-shaped instrument with rounded ends (maybe snap gage wasn't the right word).

To read between the lines, I suspect you are pointing out that due to limited resolution of measurements, the measurements don't buy us a lot. Along with the previous comment: if there is an gross problem with the shaft/housing dimensions, it should show up in the as-found condition of the machine/bearings.

It could very well be that we are wasting our time measuring bearing fits. I can tell you that a few times it changes how we do business... sometimes we go back to the warehouse to see if we have another bearing perhaps on the high-end of the O.D. tolerance to help compensate for loose housing fit. On occasions we sleeve a housing. What would have happened to the machines if we didn't do this... no-one really knows.

It is very common industry practice to do those checks and not accept fits that deviate signfiicantly. Where does the practice originate from? Bearing manufacturer recommendations for one... but we all know how conservative they are. And it also originates from repair shop practice... but if you're a repair shop and you find a fit out of tolerance, you just created more business for yourself. So while I feel very confident to say it's a standard practice to measure bearing fits during bearing replacement, it is certainly worthwhile to question how much we are really gaining from that practice.

quote:

Originally posted by electricpete:
Yes, it's a telescopic T-shaped instrument with rounded ends (maybe snap gage wasn't the right word).

A snap gauge is something else, but the telescopic gauge is the most common method out there for general use, because it's dirt cheap by comparison. It's not really any different than the time-honoured method of using firm-joint inside calipers (with two legs like dividers). So, common it is, but high-accuracy it ain't. It relies heavily (!) on feel and technique. Accuracy to a thou - not a real problem. Decent technique will get you there in a polished bore. To better than a thou... well, I'd rather not. In the shop I worked in if you needed to use a telescopic gauge to better than a thou you'd take readings until you got three the same and that would be the number, and you might need to take nine readings to get you the three the same. But! They're cheap. 8-) The biggest drawback to them is the extensive need for 'feel' which to my mind invalidates them as a precision measuring device. The second drawback is that they will develop flats on the ends of the rounded tips that add a bit of extra randomness to the readings because bores vary in size but the micrometer anvils are always flat.

quote:

To read between the lines, I suspect you are pointing out that due to limited resolution of measurements, the measurements don't buy us a lot. Along with the previous comment: if there is an gross problem with the shaft/housing dimensions, it should show up in the as-found condition of the machine/bearings.

Yes, that's pretty much it. With bearing bores and journals that have been in use for some time, you can usually see if anything is amiss. When they are brand new, of course, you need the measure the bejeebers out of them, but for that we always used a tubular inside micrometer.

quote:

It could very well be that we are wasting our time measuring bearing fits.

I wouldn't say that it's a waste... if it satisfies your due-diligence for ISO or warranty or whatever then it is worth doing. Also, it makes it more likely that a bored tech will at least _look_ at the journal and bore before pressing a new bearing in, so that alone can make it worthwhile. I may think that it is gilding the lily somewhat, particularly if you do a vibe check on the motor after assembly, but that doesn't mean it's not worth doing.

quote:

I can tell you that a few times it changes how we do business... sometimes we go back to the warehouse to see if we have another bearing perhaps on the high-end of the O.D. tolerance to help compensate for loose housing fit.

If the bore was clean and tidy but slightly oversized, then I'd just put the new bearing in. I might use a bit of retaining compound on it, but I'd figure that if the previous bearing fit in there nicely and didn't spin in the housing then the next bearing will be just fine, too. Now, if it was sloppy enough that bearing retaining compound wouldn't do it then it would need to be sleeved, for sure, or if the meat is there you can switch to the next series and just bore out the existing bore. Like replace a 6208 with a 6308. If, as mentioned, there is enough meat in the bore walls to do it and the extra width isn't a problem.

Still, if you want accurate results from your inside bore measurement, I'd suggest that your shop get themselves a tubular inside micrometer. There are other ways that are even more accurate (and expensive) but this would be a big step forward for a few hundred bucks.

quote:

It is very common industry practice to do those checks and not accept fits that deviate signfiicantly.

Ah, but how much is significant? Humans being what we are, we tend to see what we want to see, or what we expect to see. If the bore looks good and you want to see good numbers, you will see good numbers. This is because of the heavy reliance on _feel_ in measuring using this kind of technique. Books have been written on the various kinds of inaccuracy in measuring, and pretty much every variety of inaccuracy can rear its ugly head when measuring with a telescopic gauge.
It's easy to test... just get a polished bore from somewhere and have every one of the guys who normally measure these things give it a try without knowing ahead of time what the bore is supposed to be. Have everyone measure it once without having seen it measured by his buddies, and see what the results are. If you have a range of less than two thou over five guys, then I will be quite surprised.

Clearance is intentionally provided by the Manufacturer and users.

But looseness never be in the machine.