Hot Spindle Bearings

[Anorak Bob]

I've been fooling around with a pair of high speed drive pulleys on my little 9" Hercus. Because of a small difference between the diameters of the counter shaft V pulley and the headstock V cone pulley on the standard lathe (mine) and the high speed version, my lathe runs at slightly higher speeds. The one exception is the top speed, mine runs 70 rpm slower than the stated 2600 rpm. Why, I don't know.

At top speed, the spindle roller bearings become hot, enough for the headstock casting around the bearings, to be uncomfortable to touch.

The preload on the Timken tapered roller spindle bearings is 2 inch pounds as per the manufacturer' s recommendation. This is the preload suggested for the standard lathe, with a maximum spindle speed of 700 rpm ( 1050 with a 2 speed motor ) and the beautiful Craftsman lathe with it's maximum of 1840 rpm. I'm wondering if preload is contributing to the heating. Obviously, a bearing is going to generate more heat running at high speed than if it is idling along at 700rpm.

The simplest way of avoiding the heat would be to not run it flat out but I need to know how hot's too hot. Any suggestions?

Bob.

[.RC.]

Are the bearings grease or oil lubed.... If they have too much lube or the wrong type of lube they can get hot... But 2600rpm is not a fast RPM for a small bearing

[Dave J]

I have never heard anyone give a figure for preload in all my research about bearings.
I would say you have to much preload on them and as it gets hotter it gets worse because tolerances tighten up even more with everything expanding with the heat.

Getting bearing preload is tricky between being to tight and getting hot or having movement in the spindle, it is something you have to play with. The bearing place told me that wheel bearing grease is fine for something you can have your hand on, after that you need a HT bearing grease.

Dave

[Anorak Bob]

Oil R.C.

Mobil DTE Heavy Medium circulating oil, ISO 68.

The Mobil oil recommended for the old 9 inch and newer 260 Hercus lathe was Vactra Oil Heavy. DTE Heavy Medium is it's modern day replacement, according to Mobil's technical department.

On Hercus roller bearing equipped 9" lathes, there are oil cups at each bearing location. Oil enters the bearing from the cup then drains out through a hole in the bottom of the bearing cap. There is no ponding of oil in the bearing, the only pond being the one in the chip tray.

Bob.

[Anorak Bob]

Here you go Dave.

Page nine of Peter's book.

[Dave J]

I have seen that way of testing torque before but I was thinking you where meaning tightening torque on the nut.
What I did with my mill spindle which has tapered bearings as well, was to tighten it up just until there was not side ways movement with a dial indicator resting on the side of it. Then monitor the spindle by the heat and it should run warm but not hot. If it is running cold after 15 minutes it needs nipping up a bit and if it is too hot give loosen it slightly and give the end of the spindle a bump with a soft hammer to release the preload.

Dave

[Dave J]

With your oilers do they have some felt or something in them to release the oil over time? If not, when it's running a these higher speeds I would be putting a drop in every 5 - 10 minutes of running.

Dave

[Anorak Bob]

Dave,

There is no felt and given that I have 20 litres of DTE, I'm liberal with the oiling.

I will check the preload I have tomorrow.

Bob.

[Dave J]

It would not take long for the bearing to push the oil out a high speed.
If you find you are using these higher speeds a lot, it might be worth while getting some drip oilers to save the hassle and time of having to oil the lathe through the job.
I have seen some really nice brass and glass/plastic oilers around.
Here are a few I have seen
RDG Tools - Online Engineering Tools Drip Feed Oilers (http://www.rdgtools.co.uk/cgi-bin/sh000001.pl?REFPAGE=http%3a%2f%2fwww%2erdgtools%2eco%2euk%2facatalog%2fACME_TAPS__LEFT_HAND_%2ehtml&WD=oiler&PN=Drip_Feed_Oilers%2ehtml%23a1117546#a1117546)

Dave

[RayG]

I have seen that way of testing torque before but I was thinking you where meaning tightening torque on the nut.
What I did with my mill spindle which has tapered bearings as well, was to tighten it up just until there was not side ways movement with a dial indicator resting on the side of it. Then monitor the spindle by the heat and it should run warm but not hot. If it is running cold after 15 minutes it needs nipping up a bit and if it is too hot give loosen it slightly and give the end of the spindle a bump with a soft hammer to release the preload.

Dave

Hi Dave,

What would you think is too hot in your view?

You'll remember I did the same tapered bearing mods that you did on your mill.

Anyway, I've been running my mill for longish periods today, (cutting a new idler gear for the lathe) and the spindle bearings are fairly warm to touch, probably 40-50 degrees C. My question is, how hot is too hot? I notice the belts and pulleys are probably about the same temperature.

Regards
Ray

[Dave J]

Hi Ray,
That sounds about spot on after running for a while. The bearings will cope with the heat it's the grease that fails.

Dave

[Anorak Bob]

I reduced the preload on the bearings yesterday to about 1 1/2 pounds. At 2530 rpm they are still running hot after a couple of minutes. I checked the preload with the hot bearings and it was about 2 pounds.

I'm using a spring balance that could very well be an inaccurate POS. Something made to puff up a fisherman's ego.

I'll contact Steve Durden at Hercus and see what he reckons.

Bob.

[Greg Q]

I'm going to look for the reference I have, but I am sure that I have read 65º C is an acceptable temperature for a roller bearing with grease lube, higher with oil.

Greg

[Anorak Bob]

I did a bit more experimenting tonight. The most sensitive indicator that I own measures in 1um increments. As per Dave's suggestion, I loosened the nut on the spindle to the point that there was play then carefully adjusted the nut to remove the play. At that point the bearings still ran hot at 2500 rpm but nowhere near as hot as yesterday. The fish scale may have been more of a hindrance than a help.

I'm interested to read Greg's findings on bearing running temperatures.

I need to buy a cheapish non contact infra red thermometer. There is an abundance of cheap ones on Ebay. Any suggestions?

Bob.

[Greg Q]

Bob, I just checked my source books to find the following:

Industrial Machine Maintenance quotes 150º C

SKF Bearing Catalogue only mentions plastic cages, and states 120ºC continuous, 140ºC part time and 180ºC peak

I guess with oil lube you are limited by your paint job and spindle growth with temperature.

FWIW my 4000 rpm lathe has grease lube, and the books recommend grease change after 3000+ hours of continuous use.

I think my 65ºC figure came from the maximum safe touch temperature, which could be lower depending on time of exposure...as low as 50ºC for continuous contact.

Greg

[Dave J]

Hi Bob,
When I check spindles with a indicator I put it on the side of the spindle and adjust until there is no side ways movement. I think it is easier to detect movement that way. I also use a long tool in the spindle taper so I can get leverage to pull and push on it.

Dave

[Anorak Bob]

Thank you Greg,

I need to establish just how hot the bearings become. I fiddled around with a jam thermometer last night. A complete waste of time.


When you have some leisure time, I would be very interested to find out more about your German lathe. Maybe some nice photos if that's not pushing things too far.:2tsup:

Bob.

[Greg Q]

Bob, it too is in the re-building process. It's a Dutch made Hembrug DR-1, pretty much like the one on lathe.co.uk. I am doing it very slowly as I had to detour around the drill press and the mill, and learn scraping and re-engineering the defunct drive system. All good fun.

Greg

[Anorak Bob]

Greg,

I thought for some reason it was German. I hope It's green like this.:D

Bob.

[eskimo]

IMHO it would be too hot if you could not leave your hand on it?

even at these temps you would need high temp grease though...thats only to stop the grease from melting and dribbling all over the place...mucs the sime as mee wen I've hvae a fwe 2too many...hic!

[Pete F]

I have a plain bearing Hercus and it's actually the countershaft bearings that are more of a limit regarding the temperature. I understand yours probably has the roller bearings in this area, but just a heads up to keep an eye on them while you're doing all this.

My non-scientific, largely worthless, 2 cents worth is that if you can't touch the bearing area without burning yourself it's too hot. My Hercus mill has, I think, a similar spindle setup to the roller lathe and that doesn't get hot at all. In contrast the vertical head is a real problem.

You could try a different weight oil and see if that helps. The oil you're using is as specified by Hercus but at the end of the day there's nothing special about these bearings (assuming they're the same ones as the mill). I thought they were something special, looked up the bearing number, and found they are sold as trailer bearings! Good ol' Freddy!!

If you want to set up the pre-load I'd be using a weight of 2 lbs and hanging that from the 2" pulley if that's possible? Personally I've just finished stripping my mill (somebody greased it) and still need to fine tune the pre-load. It's not as critical as the lathe as it doesn't affect the finish as much as a lathe, but if I get fussy I'll be doing it by the end play method, which is indeed basically what I've done just by feel.

Pete

[Bryan]

I just happened to be reading a lathe manual that reinforces the correct temp idea. The short version of the spindle preload adjustment is to tighten the nut just till all end float is gone, then a further .13" along the cirumference of the nut. The test is to run it for 20 mins at high speed and feel the housing.



— if the inboard end of the spindle tube is
slightly warm to the touch, you have the
correct bearing preload.

— if the inboard end of the spindle tube is
hotter than you can comfortably keep your
hand on, the preload is too tight and you
must repeat the spindle bearing preload
adjustment procedure.


Anyone who wants to read the full version will find it here (http://grizimages.s3.amazonaws.com/manuals/g9249_m.pdf), on pages 58 - 60. The file is about 17 meg.

[Anorak Bob]

Pete,

The roller bearing lathes have ball bearings supporting the countershaft. They hardly become warm when running the spindle at 2530 rpm. Nothing much gets warm on my 1959 Model A when running flat out at 700rpm.

There has been considerable discussion on this forum in regards to the class of bearing used in the 9 inch and the 260 lathe headstock.

My 1969 ARL has Timken class 3 bearings ( max.0.0003" runout) in the headstock. Statewide Bearings quoted me $650 to replace the pair 6 years ago.

Bob.

[Anorak Bob]

Thank you Bryan.
BT

[Pete F]

Oh they used Class 3 bearings? I had read that the standard Hercus was just a schmuck bearing and the higher grade was an option. I didn't read the bearing number on the mill when I had the chance, but it's no big deal to whip the plate off. However it's quite possible they would use a lower grade of bearing in the mill anyway.

The mills bearing number according to the manual is Timken 18690 18690/18620 Timken Trailer Bearings Taper Roller - Bearing King (http://www.bearing-king.co.uk/bearing/18690-18620-timken/5084)

Incidentally there's a good discussion on bearings from Timken here http://www.timken.com/en-us/products/Documents/5918_09-09-29.pdf

[Anorak Bob]

Pete,

If you delve through the Hercus threads you should find a scanned price list for the 260. Hercus did offer a higher precision upgrade for that lathe.

Maybe my bearings are not standard. The story goes that the original owner of the ARL was a Queen's Counsel. An extra few hundred for bearings would have been play money. He had been using the lathe for wood turning hence the high speed pulleys.

Bob.

[Pete F]

Ah, ok, ordering precision bearings for a wood lathe. No wonder lawyers charge like they do! :D

Another alternative is that I'm full of it, and they were in fact all fitted with class 3 bearings. I was just repeating what I'd read on a BB search. I recall the discussion at the time, but not having a roller model myself the actual discussion was simply "of interest". Next time I have the end caps off the mill I'll have a look at the bearing numbers on the bearings that's fitted with.

Pete

[Anorak Bob]

I placed an inch round bar about 8 inches long in the chuck and set up an indicator at the outward end of the bar. Winding the preload nut up to the point that the bearings are hot when running did not prevent me being able to detect movement sideways on the bar when pushed without much force. The bar will obviously flex to some extent and maybe there is some movement in the chuck.

I replaced the chuck and bar with the 3 Morse taper mill arbor in an attempt to reduce the variables. Same result. It's pretty easy to move the end of the arbor 0.015". Even at the spindle nose there is movement if side load is applied to the arbor. The end of the arbor is roughly 9 inches from the spindle nose. The runout measured at the end with a tenth indicator is 0.0003". Could be just a lucky coincidence where everything has lined up.

When you blokes say take up the play so there is no radial movement, by what method are you measuring things?

Bob

[RayG]

Hi Bob,

Wouldn't it make more sense to have the DTI as close to the spindle as possible, otherwise you are just measuring the flexing in the bar?

Regards
Ray

[Bryan]

Bob, the manual I referred to talks about axial play, not radial.

[Anorak Bob]

Ray,

I did measure the play at the spindle nose. The photo is of the indicators positioned for checking the runout.

Bryan,

I was pursuing Dave's suggestion of checking the radial play with a spindle extension.
I've got bruises on my hand from pushing the spindle backwards and forwards to check the axial play.

So if you had no axial play should that suggest you would have no radial play also? What would be the cause of the radial play in my case? Spindle flex, roller deformation?

I'm sort of stumped. Maybe the bearings are rooted. I've run out of ideas.:(

Bob

[Pete F]

Bob, sorry I'm away at the moment and struggling through with an iPhone so it's a bit tricky to jump forward and backward through threads, so this may already be pointed out. ALL bearings need some level of clearance in order to operate and some machinery uses the humble plain bearing since it's easier to get a bearing with small clearance in that configuration versus a roller bearing. Even with the pre-load adjusted it will still have some clearance if not it would simply seize. The higher the quality of the bearing the less this clearance need be. I believe there are figures available for the Hercus with regards the amount of movement you're measuring. If you haven't already done so, you can compare what you're measuring with the factory spec to get an idea of where you stand. Personally I would doubt your bearings are faulty if they are running quietly, feel smooth to turn, give a good finish, yet you're measuring excessive lateral play. That should be purely an adjustment issue.

Pete

[Pete F]

Bob, I just got this darn phone to look at you images. I'd suggest you're measuring incorrectly for bearing adjustment. Take everything off the spindle and place your indicator on the spindle. You then lever the spindle and check the play meets factory spec. Personally I use a length of broom handle as it will pass right through the bore and won't damage it while providing plenty of leverage. The setup you show would be used to indicate runout at a specific distance from the nose with the lathe running. While it's true excessive bearing clearance may give excessive runout, the two really are 2 completely different animals.

[Dave J]

I check mine with a dial indicator on the side of the end of the spindle, but with something longer in it to get some leverage to pull and push.

You can use the setup you show in the last picture, just move the indicators down to the end of the spindle and use that arbour to pull and push on.

Dave

[.RC.]

FWIW there is an interesting discussion here on bearings for machine tools.. And what you get with cheap versus very expensive... All boils down to your needs..

Timken tapered roller bearings – truck bearings in my spindle? - Practical Machinist - Largest Manufacturing Technology Forum on the Web (http://www.practicalmachinist.com/vb/general/timken-tapered-roller-bearings-truck-bearings-my-spindle-218992/#post1509374)

[Metmachmad]

ALL bearings need some level of clearance in order to operate and some machinery uses the humble plain bearing since it's easier to get a bearing with small clearance in that configuration versus a roller bearing. Even with the pre-load adjusted it will still have some clearance if not it would simply seize.

Pete
Sorry Pete F, I can't understand your reasoning. Preload on a tapered roller bearing means that all clearence between rollers and the races is negative, in other words the bearing is adjusted in a controlled manner to a loaded condition when it is installed.

To counter preloaded roller bearings from seizing and depending on the amount of preload they have when installed they are only run up to certain speeds for certain amounts of time, the amount of load is known in their intended operation, the cooling effect of the oil (usually an oil bath) and the addition of EP (extreme pressure) additives in the oil.

[Anorak Bob]

I decided to check the play again. With the indicator contact point touching the face of the spindle behind the threaded nose and pushing the spindle from both ends, there was no discernible play. I then mounted the indicator to read the radial play at the same location. Applying a fairly heavy load on the arbor about 6-7 inches from the spindle end resulted in a reading of about 0.0004".

I ran the spindle at 2530 rpm for 5 minutes and whilst the bearings were quite warm, I could comfortably leave my hand on the headstock casting. After 8 minutes of continuous running, the bearings were considerably hotter. The occasions that I might need to run the machine at that speed and for that length of time will most probably be few and far between. I might just leave it as it is.

RC. That was a very interesting and informative discussion on PM. Thank you.

Bob.

[Pete F]

Sorry Pete F, I can't understand your reasoning. Preload on a tapered roller bearing means that all clearence between rollers and the races is negative, in other words the bearing is adjusted in a controlled manner to a loaded condition when it is installed.

To counter preloaded roller bearings from seizing and depending on the amount of preload they have when installed they are only run up to certain speeds for certain amounts of time, the amount of load is known in their intended operation, the cooling effect of the oil (usually an oil bath) and the addition of EP (extreme pressure) additives in the oil.

G'day, from what you've said, and I highlighted above, it would mean that the races and/or the rollers would actually distort under the pre-load. I'm afraid I'm a little sceptical that occurs in hardened races at the sort of pre-load levels being discussed here, so would be very interested if you could provide more information if you know that to indeed be the case. One of my precision spindles is pre-loaded with a simple spring, and again I'd be very interested to find that spring had enough power to distort 4 precision bearings and achieve negative clearance.

I'm keen to be corrected if wrong, but my understanding is that the pre-load removes unnecessary clearance, to effectively "zero", along the axis that they are pre-loaded. However I believe there is still an extremely small amount of clearance still there due to the physical imperfections of the bearing and the requirement to function without overheating.

Again, I'd be keen to hear more if you believe that I'm mistaken.

Pete

[RayG]

Hi Pete, Metmachmad,

I got interested in this topic, when doing the preload on the AC bearings for the mill ballscrews. Always happy to learn more.

I came to the conclusion that preload is just the amount of force applied, not a dimensional measurement as such. Although you can calculate the preload as movement if you know the elasticity of the parts involved. For very non-elastic parts like ball bearings, a tiny movement makes a big difference. Further, I suspect that at least some of the heat generated by an AC bearing with too much preload is caused by the deformation of the balls and races.

So, you are both correct (in my view) the negative clearance metmachmad refers to is the amount of compression that the components are under, which is directly related to the force applied and the elasticity of the parts.

I have heard of instances where bearings were dimpled (crenellated sp?) by excessive preload.

Regards
Ray.

[Pete F]

Ray, I agree, excess heat will be created by excessive pre-loading of the bearings, indeed it could lead to a situation where the bearing heats, reducing the clearance still further, which heats it more, etc until the bearing is destroyed. In principle I understand what is being alleged with regards negative clearance but I'm afraid I'm still very sceptical that this is in fact the case in these situations. I raise the point as I'm curious to know whether that is in fact what happens. I have the manual for my surface grinder on my desk, and the configuration for its spindle is a very common configuration of a spring pre-loading 4 back-to-back bearings. Precisely the same configuration was used in a spindle I was looking at making for a toolpost grinder. I find it very difficult to believe that spring has enough power to actually distort those 4 bearings such that the elasticity of the bearing material itself was coming into play. I haven't had that spindle apart yet (but soon need to), so don't know precisely how strong it is, but the one in the toolpost grinder spindle wasn't especially strong, and as mentioned above I believe simple reduces the excessive clearance.

That's my story and so far I'm sticking to it. Ha ha :U

[RayG]

Hi PeteF

You ask a very good question, do the balls and races actually distort under preload?

I don't know the answer with any degree of certainty, (I think yes they do) but here is my logic..

1. Where does the heat come from?
I think it comes from the balls and races compressing under the preload.
More force (preload) == more heat.

One way of thinking about it is to imagine a truck, with no load on it, drive around for a while and check the tyre temperatures.. Then load it up to maximum and drive the same distance. The tyres will be a lot hotter when loaded.

The extra heat in the tyres is caused by the tyres compressing more each revolution.
I think that something similar happens in preloaded bearings.

2. Can you even apply enough force to deform the balls into the races?
The answer must be yes, if you can actually dimple bearings with too much preload.

Interesting topic, it would be good to find out for sure, if what I'm thinking is correct..:)

Regards
Ray

[Stustoys]

I have the manual for my surface grinder on my desk, and the configuration for its spindle is a very common configuration of a spring pre-loading 4 back-to-back bearings.

Pete
Are the "4 back-to-back bearings" tapered roller? I doubt it. A surface grinder wouldn't have much axial load compared to a lathe.
As I understand it, preload on tapered roller bearings is to ensure the bearing doesn't lose preload under load. The higher the load the higher the preload and the shorter the bearing life. Once preload is lost the load is carried by one or two rollers.
As always I could be wrong. I'll try and find a graph of bearing life V preload.

Stuart

http://www.skf.com/files/005031.pdf

[Pete F]

Stu, I'd suggest it doesn't matter what the bearings are, the principles of the pre-load are the same, to remove excess clearance.

Ray, I love cycling! I say that because I've just come back from a ride and thought about this very question for much of the ride. Here's my conclusion, for what it's worth.

I'd suggest the heat is coming from 2 sources, one is the hydraulic action of the lubricant being squashed between the rollers/balls and the races. Forget that, it's another issue for the moment.

The other is indeed a distortion of the surface of the bearing. Perhaps it seems like I'm being pedantic, but I see a major difference between this situation and deforming the whole bearing itself, which is what "negative clearance" implied to myself. What I'm supposing is that the bearing surface is not perfectly smooth, but if magnified greatly can be thought of as a series of "mountains and valleys". Everything put in contact with another surface will deform somewhat and it is these "mountains" that are being squashed; the very peaks on the surface of the bearing. That process generates heat. However the overall shape of the bearing isn't really being changed much at all. As greater load (whether pre-load or actual load) is placed on the bearing more "peaks" on the surface are being "squashed", hence more heat is generated. Eventually the point would come where the overall shape of the bearing is being changed by the load, but I'd expect that to be way beyond a pre-load state.

What I expect is that under no load at all there is a physical "gap" between the bearing components. As a load is applied (eg pre-load) that gap (clearance is reduced) to the point where the surface of the bearing begins to be affected by the load. This is the point at which we pre-load a bearing.

Now I need somebody to tell me if I was close ... or do I need to stick to cycling :doh:

[Anorak Bob]

A clear, analogous decription of the events that create bearing heat. Thank you Pete.

[Pete F]

A clear, analogous decription of the events that create bearing heat. Thank you Pete.

Maybe, but is it correct? :U The link Stu posted implies the whole bearing yields, thus creating the heat, yet I've read other papers (also by SKF interestingly enough) that imply something different and, sadly without saying as much one way or the other, more along the lines I've what I suggested. Mind you, that SKF paper linked to above talks about tapered bearings, the diagram isn't displaying well on my monitor, but it looks like a plain roller bearing from what I can see. All very strange! It may come down to semantics as to precisely what is defined as "clearance" in this situation.

Pete

[Stustoys]

Stu, I'd suggest it doesn't matter what the bearings are, the principles of the pre-load are the same, to remove excess clearance.
Yes the principles would be the same but the pressure needed to preload of angular contact bearings would be lower than tapered roller. The preload on deep groove ball bearings would be set on assembly.
Do you do your own wheel bearing work on your bike?

As for bearings or races being dimpled by preload. Try hitting one with a hammer. I've never seen one "dint". There is a problem with vibration of a bearing that is not rotating which will dimple the races.

Stuart

[Pete F]

Do you do your own wheel bearing work on your bike?

Stuart

Yep, do everything on my bikes, from the initial build onwards. All of the different types of wheel bearings use pre-load.

Pete

[Stustoys]

Pete
I have some maths on the preload of cup and cone bearings, but I cant remember where. I'll see if I can dig it up.
Stuart

[Metmachmad]

G'day, from what you've said, and I highlighted above, it would mean that the races and/or the rollers would actually distort under the pre-load. I'm afraid I'm a little sceptical that occurs in hardened races at the sort of pre-load levels being discussed here, so would be very interested if you could provide more information if you know that to indeed be the case. One of my precision spindles is pre-loaded with a simple spring, and again I'd be very interested to find that spring had enough power to distort 4 precision bearings and achieve negative clearance.

I'm keen to be corrected if wrong, but my understanding is that the pre-load removes unnecessary clearance, to effectively "zero", along the axis that they are pre-loaded. However I believe there is still an extremely small amount of clearance still there due to the physical imperfections of the bearing and the requirement to function without overheating.

Again, I'd be keen to hear more if you believe that I'm mistaken.

Pete
Yes Pete your are correct when you say that preload (negative clearence) distorts the bearings, and maybe some of the other components also, depending on the setup used.

Have a good look around this SKF site, it has plenty if info on this topic-
Application of bearings - SKF.com / Products / Interactive Engineering Catalogue/Rolling bearings (http://www.skf.com/portal/skf/home/products?maincatalogue=1&lang=en&newlink=1_0_85)

[RayG]

Hi Metmachmad,

Thanks for posting that link, that all makes perfect sense now. The bearings compress elastically by an amount given by the bearing stiffness ( measured in kN/um) and the more compression (more preload force applied) means the heat generated by the compression increases. (same as the truck tyre example I gave earlier in this thread)

Here is an example of pre-load calculation, refer to the graphs for bearing stiffness ( elasticity)
http://www.dynaroll.com/preload-calculation.asp

Stuart,
The word I was looking for was not dimpling, but brinelling, which happens when
the preload gets too high and the bearings are deformed beyond the elastic limit.
And yes they do make little dents in the races, after which you chuck em out. :)

http://www.dynaroll.com/excessive-force.asp

There is another similar effect which can be caused by vibration, rather than excessive preload.

Regards
Ray

[eskimo]

There is another similar effect which can be caused by vibration, rather than excessive preload.

Regards
Ray

I think that it is called False Brinelling

you mainly get it in bearings that are, say in a electric motor, in transit via a steel boat.

they used to rotate the shafts of motors to prevent it happening while in transit

Apparently it doesnt happen on a truck due to the tyres (rubber)
They say it can also happen on a train

[Pete F]

Thanks guys, yes I think Ray's analogy of a truck tyre is the most accurate from what I've now read, it's the contact points on the bearing that are distorted much like the bottom of a truck tyre as Ray said. I couldn't believe the whole bearing was "flattened", and indeed that's not the case. Also because there are different levels of pre-load, there will be greater or lesser yielding at the contact points, with virtually nothing on the spring preloaded surface grinder I mentioned (and I'd expect behaving much more like I mentioned by simply "flattening the peaks"), but considerably more on something like a lathe spindle. All very interesting, though I don't know if it was any assistance to the OP :doh:

Pete

[Stustoys]

Ray I think its a little rough to call the load of incorrect press fitting of the bearing "preload", but I see your point. Would be interesting to know that the max preload would be on a bearing used in a lathe headstock.

eskimo As I understand it False Brinelling first became a problem on the wheel bearings of new cars moved by rail(bikes had been putting up with it for years by then;)). Improved surface finish on the bear shells fixed the problem. (So the story goes)

Stuart

p.s. lol wiki backs me up. (I really added this for the second paragraph about "back up" motors/pumps, I thought might be interesting to anyone that has a system like that).

"The discovery of false brinelling is unclear but one story describes how, in the 1930s, new automobiles were loaded on to trains for delivery; when they were unloaded, some would show severe wheel bearing damage. On further inspection, it turned out that many wheel bearings were slightly damaged. The damage was eventually traced to rocking of the autos and the regular impact every time a railroad car wheel passed a track joint. These conditions led to false brinelling.
Although the auto-delivery problem has been solved, there are many modern examples. For example, generators (http://www.woodworkforums.com/wiki/Electrical_generator) or pumps (http://www.woodworkforums.com/wiki/Pump) may fail or need service, so it is common to have a nearby spare unit which is left off most of the time but brought in to service when needed. Surprisingly, however, vibration from the operating unit can cause bearing failure in the unit which is switched off. When that unit is turned on, the bearings may be noisy due to damage, and may fail completely within a few days or weeks even though the unit and its bearings are otherwise new. Common solutions include: keeping the spare unit at a distance from the one which is on and vibrating; manually rotating shafts of the spare units on a regular (for example, weekly) basis; or regularly switching between the units so that both are in regular (for example, weekly) operation."

[Anorak Bob]

I imagine that the discussion generated by my original post has been beneficial to more than just me, that in it's self makes it worthwhile.

Bob T

[RayG]

Hi All,

Interesting discussion, I'm not sure it has solved Bob's problem, but it seems to me that measuring the temperature is a simple way of checking if you've got the pre-load correct.

Stuart,
Further down that page, is the bit about pre-loaded bearings. Sorry, I should have highlighted it a bit better...

"Bearings are susceptible to shock or impact loads, especially when the axial play has been removed through preload. Dropping a preloaded assembly, or even placing it on a hard surface, can cause brinelling. It is recommended that assembly work surfaces are covered with a layer of shock-absorbing material."

Delicate little buggers.... :)

Regards
Ray

[Stustoys]

It has been interesting to me Bob(sorry if we haven't fixed your problem yet)
I find bearings interesting because the worse ones you can buy(from good suppliers) are out of round by 0.0002".
One thing I just thought of which maybe confusing people. "preload" isn't metal to metal -.XXXmm there is still a film of oil between the metal parts you are squeezing it thinner. (or maybe I am the one that is confused)

Ray they are still talking about extra loads doing the damage not the preload. Also it would be nice if they mentioned what they consider a "small bearing". They also seem to be talking about balls, I assume rollers would have the same problem but the loads would need to be much higher. The loads a "normal" sized bearing can take must be pretty high. Think about the load that can be placed on the bottom head stem bearing or the tiny little suckers in clipless pedal spindles. As someone once said "if you think impact load is the cause of brinelling on a head stem put one together and start hitting it with a hammer"

Stuart

[Pete F]

One thing I just thought of which maybe confusing people. "preload" isn't metal to metal -.XXXmm there is still a film of oil between the metal parts you are squeezing it thinner. (or maybe I am the one that is confused)



I can only speak for myself, but personally I found the term "negative clearance" somewhat confusing as it implied (to me at least) the entire bearing was deformed in the process of pre-loading. I couldn't believe there was enough force being generated to do that, and indeed there isn't. Instead it is the very small contact point that is being deformed within the elastic limits of the material concerned. The analogy of the flat part of a tyre is an extremely good one, as the tyre as a whole remains essentially "round" however it is deformed where it contacts the road surface.

Pete