What's new
What's new

Cat40 spindle rebuild: Thoughts on floating bearings outer races spinning in housing

nmorong

Plastic
Joined
Jul 1, 2022
Hi everyone,

I've been working on rebuilding a cat40 spindle from a Milltronics VMC. This morning I assembled the spindle and noticed the upper bearing set outer races are spinning in the bearing bore. This is the upper set of angular contact bearings that is not restrained in order to allow axial floating with thermal expansion. I had heated the housing for assembly, so I figured there was just too much clearance because it was still hot but after 8 hours of sitting in a cold shop the outer races still spin in the bore.

I am not new to precision work, but this is my first rebuild of this kind of spindle. I used NSK P4Y bearings, and treated the NSK technical guide (linked below) as the bible. On page 235 it gives the recommended fits for housings. I did measure everything, carefully, using mics and a dial bore gauge, verifying each measurement by a comparison to a gauge block stack equaling the measured value. I was even more anal for the flatness and parallelism measurements. Although now I wish I had written down all the measurements... Anyways, it recommends a clearance of .007mm to .010mm (~.0003-.0004"). After assembly I can't really re-measure this without destroying the brand new bearings taking things apart, so I tried to improvise a sanity check. Using an indicator mounted on the outside of the housing with the needle on the shaft just above the locknut, I can deflect the spindle enough to make the bearings make contact with the bore. Going around I get a consistent reading of .0002" at the point where the outer race contacts the bore. See attached image for setup. There is no detectable shaft runout just above the locknut with a .0001" indicator without applying a side force.

I know measuring tenths withs mics and a dial bore gauge I could have been off by a tenth or 2, but I'd be surprised if it were 2 with the procedure I was using. It must be close to the recommended fit, but it just doesn't seem right to have the outer races spinning with the shaft. Is this normal? Is .0002" clearance between a bearing and the bore enough for the bearing to spin freely? It doesn't seem to make contact at all with no side load applied. I lubed all outer surfaces of the bearings very lightly with spindle oil after greasing to prevent rust. Right or wrong, I'm absolutely certain I never would have noticed anything if I had used a thick lubricant for installation. It just doesn't seem right to me, but I would hate to buy new bearings and tear it down again if this is normal and OK.

-Nate

 

Attachments

  • IMG_2891.jpeg
    IMG_2891.jpeg
    2 MB · Views: 43
A spot of l;octite will cure it
Have you ever been inside of a precision spindle? Wouldn't you be more au fait with bulldozers. Yet your out here embarrassing us Australians with your lack of knowlege. Its why many of us don't participate any more. The reason the upper bearings are a sliding or transition fit, is that they can transittion axially with thermal growth. Bottom bearings are retained axially, top bearings can float, to compensate thermal growth of the spindle. To suggest loctiting / glueing them in is just moronic. Floating rear bearings goes back a 80 years, from spring loaded grinder spindles, to current best practice, state of the art, hydraulic porportional valves that apply different pre-loads depending on speed.

Your an embarresment to this nation. Our version of Thermite.
 
This morning I assembled the spindle and noticed the upper bearing set outer races are spinning in the bearing bore.
Hi Nate
Unless you could detect any bore wear when you did all that good work with a dial bore gauge. Thats perfectly normal. As you are already aware, that bearing to housing fit is a transition fit. Were I grew up, we'd call that a two thumb fit. If you could push the bearing in by two thumbs, it was a transition /sliding fit. It has to be a sliding fit, or it defeats the purpose. 8 hours of equallising temerature wont be enough.
I'd note you have P4Y bearings. P4 being the running accuracy. Most Americans would associate that class with ABEC7. Your P4Y "Y' meaning they are in the middle of the I.D and O.D size range. Its a NSK tolerance tighter than the JIS standard.

Not uncommon to see the outer race spin in the housing. Its on the open side of a transition fit. I'd assume Milltronics arn't doing anything out of the ordinary for lubrication. So these are dragging grease. What happens in practice, it will spit that excess grease out. running torque will drop way down, inner ring being the much smaller,torque arm / lever, will spin, leaving the outer to do as it should do. Just creep axially up and down as the spindle expands. Heat only occurs at the bearings. Few degree's expands inner and outer rings.

You never specified bearings sizes. But Im going to guess your in the 70 or 72 series, bore code of 12-15?

Regards Phil.
 
Last edited:
The OP says its spinning,which is a long way from a transition fit ,or light push fit ..............if the bearing is loose ,it will wear the housing and heat in operation.
 
The OP says its spinning,which is a long way from a transition fit ,or light push fit ..............if the bearing is loose ,it will wear the housing and heat in operation.
He further went on to report that he could detect 0.0002" of deflection. Thats 2 ten thouandths of an inch deflection. Appox 5 microns. Let me guess your solution would be to weld it up and line bore it like an excavator pin? Wouldn't the fact be, you have no fucking idea what your talking about here?
 
The OP says its spinning,which is a long way from a transition fit ,or light push fit .............
No its exactly how it should react when its full of grease. Your thinking 3600 rpm or what ever pole motor your thinking of. 3,600 was it? He's only just hand assembled it. The outers will creap, just the same as they are designed too axially and radially. The fits are that good, you dont need a 14lb sledge hammer to fit. They'e not loose, the entire spindle is designed so the upper bearings can float with thermal growth. You seemed to have missed the fact that you suggested gluing them in.

I maintain, your still an embarrassment to the Australian community, Our Thermite.
 
Its beyond a joke trying to have a true techical conversation here. Bearing sizes of 0 - 8 um, and housing sizes of plus / minus 0.0025mm. I wish this bloke good luck, but I'm going back to the ignore list.
 
Time to talk to the bearing maker methinks.

Concerning Loctite I do wonder how much actual differential thermal growth / shift under load occurs during operation and whether Loctite stopping any mutual self adjustments really would prevent things behaving.

Way back when Loctite was a rather new thing I raised similar questions when faced with a similar bearing and spindle set-up where fit had degenerated from light push to slides out under gravity with similar just detectable deflections. The rep assured me that as the designed light fit was purely to enable the bearing to take up its correct position under assembly with no possibility of unwanted loads from pressing in loctite would be fine. He suggested a lighter grade than the usual bearing fit tho'. I think I used the one intended for small screws.

It certainly worked for me. The bearing seated properly and the spindle ran true. To tenth TIR or so as I recall it. Probably overall a less loaded and less demanding set-up than the spindle the OP is dealing with but it does indicate that correctly used in appropriate circumstances loctite can be successful.

Whether the OPs situation is appropriate being a whole n'other matter of course.

Clive
 
I appreciate all the responses.

I am not about to loctite it. Even if I thought it was a good idea, I couldn't clean inside the gap for proper application without taking things apart. I can't get the assembly apart again without destroying the bearings. The lower triplex set is too tight a fit to drop out. The housing is too massive to heat fast enough to get a temperature differential between the housing and these bearings. Loctite is something that I might try if I KNEW it was wrong, and was hoping to salvage the situation. That said, I am aware of enough people that have done it successfully that if I run it as is and it fails and I have to do the job again I might try it. This is my spindle on my machine and I have no one but myself to answer to if it goes wrong. And if it worked fine, maybe I would admit it years later. The spindle bearings are being replaced because of how visually rusty they were. It was still working fine, and the outer races of the old bearings were seized to the housing with a layer of rust and certainly weren't moving as designed.

Phil, I appreciate the input from someone who is familiar with doing this. I now wish I had just used a thicker lubricant on the outer races. Maybe I can make a tool that seals against the outer race and hydraulically forces some thick lubricant into the gap? Would you recommend trying, or just installing it and not worrying so much?

-Nate
 
Questions from a guy who has almost 0 experience with spindle bearings (did rebuild a nikken speeder once):

Does it literally spin freely inside the bore, where the inner and outer race spin at the same rate (ie, the balls in the bearing don't actually rotate)?

2nd question: is the spindle belt driven, and if so, does the side load from the belt push the outer race against the bore, causing it to not spin?

I don't know if either of those questions will change any opinions.

My assumption? the bore is probably a bit big, and whoever bored it out or assembled it at milltronics probably said "oh well, just ship it". If it was initially like that from the factory, sounds like running it for years will "fix" it anyway.
Having re-read machtool's 2nd post, sounds somewhat normal. I mean, i still feel like I personally would be surprised it it completely freely spun in that bore, but I totally get the 2 thumb type fit. And again, with belt tension, I would imagine you won't see it freely spinning any more.
 
Last edited:
Standing vertically in a free state, yes it does spin totally free in the bore. With just a tiny bit of side pressure, it makes enough contact with the bore to mostly stop the outer race from spinning. This is probably a small enough amount of pressure that if I set the spindle on its side, its weight would be enough. It's currently standing on end in a fixture I made for assembly. With firm side pressure, it contacts the bore with enough force that the outer race can't be made to spin at all, even with abrupt changes in direction or by poking at it with a rubber tipped tool. (This is at the .0002" deflection I mentioned previously).

It is a belt driven spindle, but there is a 6th bearing above these resisting the radial load from the belt. This 6th bearing is just a 6008 that sits in a housing that is bolted onto the top of the head casting, and adjusted to position with set screws. At the cost of a slight bend in the shaft, I could certainly adjust the 6th bearing housing position to load the shaft so that the angular contact bearing outer races are always in contact. Maybe if the slight bend induced by doing so is a small fraction of the bend in the shaft caused by the radial load from the belt, this wouldn't be an issue.

I attached a little hand drawn diagram to the initial post showing the bearing sizes and positions.

-Nate
 
Questions from a guy who has almost 0 experience with spindle bearings (did rebuild a nikken speeder once):

Does it literally spin freely inside the bore, where the inner and outer race spin at the same rate (ie, the balls in the bearing don't actually rotate)?

2nd question: is the spindle belt driven, and if so, does the side load from the belt push the outer race against the bore, causing it to not spin?

I don't know if either of those questions will change any opinions.

My assumption? the bore is probably a bit big, and whoever bored it out or assembled it at milltronics probably said "oh well, just ship it". If it was initially like that from the factory, sounds like running it for years will "fix" it anyway.
Having re-read machtool's 2nd post, sounds somewhat normal. I mean, i still feel like I personally would be surprised it it completely freely spun in that bore, but I totally get the 2 thumb type fit. And again, with belt tension, I would imagine you won't see it freely spinning any more.

No.
You know yourself how tight a fit .0002" clearance is, and the whole purpose of the balls is to have lower friction than a plain bearing.

Any bit of assembly oil or dust will be enough friction to hold the outer race, and even without that, as soon as you tighten the belt it will deflect enough to pull the bearing tight into the bore.

If it had ultra precision bearings on either side that kept it a perfect .0001" gap, and you had it jam packed with grease, it MIGHT spin, but if you have that much grease it in the bearing will fail reasonably soon anyways.
 
The .0002" is a radial clearance, so .0004" diameter clearance.

I greased the bearings with Kluber NBU15 at 1.5cc/bearing for the 7009 and 1.6cc/bearing for the 7010 bearings (15% free space). I erred on the side of too much grease, as getting precisely 1.5 or 1.6cc from a 50cc syringe is difficult. NSK recommends up to twice as much grease as I was shooting for (30% free space), and I'm sure I didn't come close to that. I am aware that the bearings will require careful run in to expel extra grease. Once greased the bearings did feel fairly stiff, but not more than I would expect from a new factory greased bearing.

-Nate
 
Phil, I appreciate the input from someone who is familiar with doing this. I now wish I had just used a thicker lubricant on the outer races. Maybe I can make a tool that seals against the outer race and hydraulically forces some thick lubricant into the gap? Would you recommend trying, or just installing it and not worrying so much?

-Nate
I think what you are seeing if perfectly normal. At the fit its specified at, it must be free to axially float. Were they P4Y's you removed? That might be the difference, being tighter controlled on OD. Your Ying and Yang may not have collided. If you couldn't decern any wear of the housing bore. I'd say you are good to go. A tenth bigger bearing, and a tenth smaller bore, you wouldn't be seeing the race spin. That's how close to transition that fit is.

I wouldn't be trying to offset that 6008. I'd assume that's just a good commercial grade deep groove, it will have way more clearance and inaccuracy than a P4 angular contact. You did measurements up to and including dial bore gauges confirmed via Gauge / gage / Jo blocks. (I'm multi lingual when it comes to precision). If you didn't find an error in the housing bore due to wear. Your good to go.

Ed made a great point, about rolling verses sliding friction. It was passed down to me, Great grandfather used to say, a horse could pull a 1 tonne log at of the forest, but put it on wheels, a dray or cart and the horse could pull 4 tonnes. Love to have that argument with him, as I think the factor of rolling to sliding friction is more in the order of 7 : 1. But he died 40 years before I was born.
You noticed the stiffness with new grease. Once run in, that will expel from the running grooves and the rolling resistance will drop dramatically. Now that we know that the upper bearings are 7009's 45 x 75mm Id to Od ratio. That equates to 166% driven to drive ratio. Your driven / locked - spindle mounted bearings have to produce 166% more torque to make the outers spin in the housing. Its not like they are going to spin at 6k rpm chewing out the housing. A certain amount of radial creep is to be expected. Thats never going to happen on a belt driven spindle. I worry about it on integral motor spindles.

In certain circumstances, I've Jerry rigged a bearing that has chewed out a housing by a tenth or two, from spining by modifing the upper retainer plate, and incorporating an O-Ring that rides on the outer. Normally because the customer, doesnt have the time or money for a grind / chrome / final grind. Or we are waiting for Christmas / Easter or the Melbourne cup long weekend. Tempory thing. I'm sure some of those tempory things are out there in the wild for 20 years. Trick is remembering that the bearings have to float up and down. So only set the O-Ring with about 5% compression. For a 2mm cord, set it at 1.9mm compressed. My best guees. I must have been in the pub the day they taught about Shore hardness and compressive strenghts v friction.

This very same science is employed in every electrical motor you have ever touched. How many of you have pulled apart an electric motor and found a wave / floppy / spring washer one end? Up to and including a car altinator.

And just because its Sunday afternoon down here, the thermal expansion. Consider the lower and upper bearings are 152mm apart (6"). Spindle might get delta 20 deg C. Anyone that has ever dry cut knows the cutter body is warm to the touch. That heat all sinks into the spindle. Upper bearings must float.

At a coefficient of 0.000012 for steel. mm / C. It equates to 0.0356mm, 36 microns expansion at just 20 degrees centergrade. Kelvin is the same unit. From my bearing tables a 7009 15 degree moves from light preload into heavy with a shift of 26 micron. Good luck gluing them in.

I wouldn't normaly be here. I black banned this place 2 years ago. But some black smith, big hammer swinger, tractor driver, Steptoe & Son. Dirty Hoarder character, Got to open their mouth at every oportuniy Yoo Hoo, embarrassing my country, whilst diseminating close to legaly curupt information. That got my attention.
I'm going to fade out again for the next few years. You might find me on an other forum.

Best regards Phil.
 
Last edited:








 
Back
Top