What's new
What's new

Side-to-side play with new spindle bearings?

gargamel

Aluminum
Joined
Mar 22, 2012
Location
Boston, MA
Hi All,
This is part of an evolving series of posts related to new spindle bearings in an old Cincinnati Toomaster 1B.
The reason that I decided to change out the bearings was that the spindle had ~0.001" TIR side to side motion if i grabbed it and push pull back and forth. Not alot of force necessary, I don't need to mount a bar or anything for the extra torque. just a normal push pull with my feeble middle aged arms. Runout in the taper was less than that with no force, but there are lots of reasons for runout in the taper so i was not too hung up on that. Cheking the OD of the quill just above the spindle nut with the same test gives <0.0001" of motion for the same amount of applied force, so about 10X less motion on the quill than the spindle. This issue was leaving me with an inconsistent surface finish, and I surmised that the ~70 year old bearings were just shot.
Fast forward past the disassembly and reassembly hassles, I now have a new set of SKF bearings installed and am testing before finishing the full reassembly. Lo and behold... same problem. Spindle runout in the taper is much better with no force, a bit less than 0.00015". Dandy. However, the same push-pull test on the spindle gives roughly the same 0.001" TIR side to side motion. Quill OD for the same test is still ten times better, 0.0001" or less. Since there is no motor or drive parts on right now, I can also grab the top of the spindle and push pull on that to watch the indicator down at the taper - roughly same result with low force (spindle moves ~0.001" TIR). I would assume that cutting forces could well exceed what i am providing, so possibly worse motion in real operation.

So - with essentially no change for shiny new brand-name bearings we need a new plan.
Bearing preload? I guess its a possibility but all the spindle nuts are as tight as i can comfortably get them and the spacers looked good. Bearings are "medium" preload from the factory. Etched faces of the bearings are facing each other , and the etched arrows on the outside line up.

My suspicion is that OD of the outer bearing races are too loose against the ID of the quill. I'll try to get a decent measurement of the quill ID to confirm. I found it fishy that the spindle slides right into the quill, and back out again, very easily. It does not feel like any sort of interference fit between the outer race OD and the quill. I havent done this before, but I would assume that this is supposed to be a pretty tight fit.
Assuming that this is the problem, what are options to fix it? Somthing like a Cr plate and regrind of the quill bore seems excessive (and expensive).
Loctite 660 Bearing retainer or similar? It this stuff reasonable to use in this situation? If it works then it is cheap and easy, but i am concerned that if it does not fix the issue then i have probably made it impossible to get the spindle out again without trashing the bearings, and probably quite difficult to ever get it out again at all.

Ideas for other solutions?
Ideas for other causes of the problem?

As always, thanks for the help.
Mike
 
Can you make a video of the spindle movement? TIR is generally measured as a total runout from low to high as a spindle is rotated. What you're describing is more like free play of the spindle. Are you saying that the indicator stayed at "0" when pressed in one direction and then stayed at .001" displacement when pressed in the other direction? Or did it just flex and come back to the same number when pressed in both directions?
 
Thanks for the tip on Vibra-tite, I'll take a look. I am pleased to hear that you have been able to remove without too much force or heat. Loctite and Vibra-tite put the temperature range on their products to 300F, which is somewhat higher than the 220F SKF gives as maximum temperature for heating the bearings for install or removal, so I had a bit of concern that i would need to cook the assembly too hot if I ever need to take it apart again. Do you remember which Vibra-Tite you used? 530 looks like an low-strength oil-resistant option.

eKretz- apologies for my poor terminology. You are correct noting that i used TIR for describing this non-rotating measurement. I mean side to side free play as I push/pull the spindle without rotating. If the indicator is zeroed with no applied force it does come back to zero when the force is removed. For low-ish force it moves about 0.0005" + or - when I push/pull. If i just spin the spindle without pushing laterally then it runs true: TIR inside the taper during rotation is about 0.00015.
 
For removal, the higher the heat, the lower the required pressure so you go as hot as you feel is suitable and push the assembly out. A press-push is much gentler on the bearings than hammer blows, of course. The Vibra-Tite I use most is for 'close fits'. I've also used their 'gap fill' version which works up to .015" clearance.
 
Thanks gordon. McMaster lists Vibra-tite 530 for oily, close fitting parts max gap of 0.005". I'll probably order that and give it a shot.
thanks,
mike
 
Its likely one of a few potential issues. Lack of preload, small shaft od, or big housing id. We already know the housing is big if it slides in and out super easy, question is how big? Also how small is the bearing outer race od? It will say on the skf box (will be the big D spec) if the new bearings are say -6 and the housing is +15, you have a 21 micron slip fit, there could certainly be some movement there. Did you reset outer race capture? If so what did you set it at?

Also make sure when taking this radial play measurement that your indicator base is on the spindle housing, not the table or elsewhere, you want to isolate the spindle from the rest of the potential causes of play.
 
Thanks for the tip on Vibra-tite, I'll take a look. I am pleased to hear that you have been able to remove without too much force or heat. Loctite and Vibra-tite put the temperature range on their products to 300F, which is somewhat higher than the 220F SKF gives as maximum temperature for heating the bearings for install or removal, so I had a bit of concern that i would need to cook the assembly too hot if I ever need to take it apart again. Do you remember which Vibra-Tite you used? 530 looks like an low-strength oil-resistant option.

eKretz- apologies for my poor terminology. You are correct noting that i used TIR for describing this non-rotating measurement. I mean side to side free play as I push/pull the spindle without rotating. If the indicator is zeroed with no applied force it does come back to zero when the force is removed. For low-ish force it moves about 0.0005" + or - when I push/pull. If i just spin the spindle without pushing laterally then it runs true: TIR inside the taper during rotation is about 0.00015.
TIR stands for Total Indicator Reading and can therefore be used for any total measurement whether rotating or static.
 
001 total is only average 0005 clearance......so any bearing fit compound should be ideal.........the catch may be getting the shaft centred ,however surface tension of the liquid may do that automatically.
 
TIR stands for Total Indicator Reading and can therefore be used for any total measurement whether rotating or static.

I've never seen that done - in the U.S. at least - for a free play/deflection type of measurement. Using that term to describe this kind of movement will only lead to confusion, IMO.
 
I've never seen that done - in the U.S. at least - for a free play/deflection type of measurement. Using that term to describe this kind of movement will only lead to confusion, IMO.
He is absolutely correct about TIR being applicable to basically any indicated geometry, and I'm pretty sure that applies in the US too. You are also correct that TIR tends to be subjectively understood as a circular measurement, but that is incorrect.

This confusion is why FIM was presented as an alternative terminology, but I have seen that used very rarely in practice.
 
Ummm...

On my surface grinder (J&S 540) the spindle bearings are preloaded with a spring to take up expansion forces with a spring. If I push the spindle axially, it moves a bit, by design. Are you quite sure that your spindle is not meant to move like this?
 
Ummm...

On my surface grinder (J&S 540) the spindle bearings are preloaded with a spring to take up expansion forces with a spring. If I push the spindle axially, it moves a bit, by design. Are you quite sure that your spindle is not meant to move like this?
I’ve never seen that on any milling machine spindle I’ve worked on. They need to be rigid.

Regards Tyrone
 
Wow! Thanks for all of the additional feedback guys. I had thought this thread was fading into the background.
Re: type of bearings: A pair of angular contact bearings at the bottom and a single in up at the top. THe original bearings were from New Departure, replacements are a matched pair of 7007s from SKF and a single 7006 at the top.

Re: clearance in the housing. The two bearings bearings for the lower end of the spindle just above the taper were marked -5 in the box from SKF. I measured them around -4um as a sanity check but in a fight i would side with SKF on the true numbers. The bore in the quill measured about +5um, so we are somewhere in the ballpark of 10um clearance between the two. This was my virgin outing replacing spindle bearings and i bought them prior to disassembly hoping for a fast turnaround. If I were to do it again I would try to get a bearing set that more closely matched the bore.

I know some grumpy coot will say " You cant measure to 1um! You cant do that with your Harbor Freight calipers!" So the pre-emptive answer to how I measured the ID of the quill bore and OD of the bearings: Sideways on the surface plate in a V block. 0.00005" BesTest indicator swept for top dead center or bottom dead center. Starret Digi-Check gage block stack adjusted to zero the indicator on upper or lower gage block surface and then read the height directly from the DigiCheck. Subtract.

I did end up putting in some Vibra-Tite but it was a bit of a fiasco. The first time I tried it the Vibra-tite locked up while I was sliding the spindle w/bearings into the quill. The bottle says 20 minute working time so i was moving pretty slow, but i think i only paused for a few seconds for some reason and that was it - stuck.
Huge PIA to disassemble, clean and try again. Second time around I was deliberately light on applying the Vibra-Tite but the assembly went smoothly.
Side-to-side play is now +-0.0002" or so when I push and pull, so better than where I started but I would have hoped for a bit less than that. I think that the Vibra-tite application was light enough that I did not fill the gap as thoroughly as it could have been but i was gun-shy about it locking up again. Or maybe I am overestimating the bearing stiffness, I don't have too many data points to compare against.

cheers,
mike
 
Wow! Thanks for all of the additional feedback guys. I had thought this thread was fading into the background.
Re: type of bearings: A pair of angular contact bearings at the bottom and a single in up at the top. THe original bearings were from New Departure, replacements are a matched pair of 7007s from SKF and a single 7006 at the top.

Re: clearance in the housing. The two bearings bearings for the lower end of the spindle just above the taper were marked -5 in the box from SKF. I measured them around -4um as a sanity check but in a fight i would side with SKF on the true numbers. The bore in the quill measured about +5um, so we are somewhere in the ballpark of 10um clearance between the two. This was my virgin outing replacing spindle bearings and i bought them prior to disassembly hoping for a fast turnaround. If I were to do it again I would try to get a bearing set that more closely matched the bore.

I know some grumpy coot will say " You cant measure to 1um! You cant do that with your Harbor Freight calipers!" So the pre-emptive answer to how I measured the ID of the quill bore and OD of the bearings: Sideways on the surface plate in a V block. 0.00005" BesTest indicator swept for top dead center or bottom dead center. Starret Digi-Check gage block stack adjusted to zero the indicator on upper or lower gage block surface and then read the height directly from the DigiCheck. Subtract.

I did end up putting in some Vibra-Tite but it was a bit of a fiasco. The first time I tried it the Vibra-tite locked up while I was sliding the spindle w/bearings into the quill. The bottle says 20 minute working time so i was moving pretty slow, but i think i only paused for a few seconds for some reason and that was it - stuck
Huge PIA to disassemble, clean and try again. Second time around I was deliberately light on applying the Vibra-Tite but the assembly went smoothly.
Side-to-side play is now +-0.0002" or so when I push and pull, so better than where I started but I would have hoped for a bit less than that. I think that the Vibra-tite application was light enough that I did not fill the gap as thoroughly as it could have been but i was gun-shy about it locking up again. Or maybe I am overestimating the bearing stiffness, I don't have too many data points to compare against.

cheers,
mike
It’s a few years since I had to “ bodge “ up a spindle bearing assembly. Please don’t take “ bodge “ as an insult. “ There’s an old saying “ Needs must when the devil drives “. Sometimes you have to do these things.

I was working on a plastic extruder gearbox. Like a big lathe headstock but with a few less gears and lines. This thing had been run into the ground. All the bearings had “ tracked “ to some extent. Some had up to 0.005 “ or 0.006 “ clearance ! What we’d normally do is put it on a Hor bore, bore out all the bearing bores, and then re-sleeve them. That’s quite a long process and the customer was desperate to get the gearbox back into production. They said ” If you could just “ bodge “ it somehow whilst the Easter holidays you can have it back and do it properly. “ Easter was about 8 weeks away.

My manager made enquiries and came up with a new product at the time from “ Loc-tite “ called “ Loc-Metal “ if my memory is correct. It was a bit like grey toothpaste. It took something like 20 to 30 minutes to “ go off “. They reckoned that it was good for up to 0.010” clearance. Any way we fitted all the new bearings with this “ Loc-Metal “ and sent the machine back. When our sales people phoned the customer up with a view to having the job done properly near to Easter the customer said “ Err, it’s running great so we’re going to leave it for now “. We never heard from them again.

You had to fit the bearings with the gear box laid down to centralise them, otherwise the running clearance/backlash on the gears wouldn’t have been correct.
A pal of mine once fitted some bearings with the “ Loc-tite “ bearing fit liquid and he inadvertently allowed some of the liquid to get inside the bearing ! It had to come out and be replaced.

Regarding your existing running clearance of 0.0002”. I’d be happy with that. You need some clearance otherwise you’ll find that the bearings are running hot initially.

Edit - I just looked it up, it’s called “ Quick-metal “ now.

Regards Tyrone.
 
Last edited:
He is absolutely correct about TIR being applicable to basically any indicated geometry, and I'm pretty sure that applies in the US too. You are also correct that TIR tends to be subjectively understood as a circular measurement, but that is incorrect.

This confusion is why FIM was presented as an alternative terminology, but I have seen that used very rarely in practice.

I don't disagree, with the caveat that it's just never used that way in the shop. And it's not incorrect, it's just not the only way to use the term. It's just *practically* the only way it's used in basically every shop I have ever worked in or with. FIM, as I understand it, really means the same thing - it's just an updated, more defined term.
 
Ummm...

On my surface grinder (J&S 540) the spindle bearings are preloaded with a spring to take up expansion forces with a spring. If I push the spindle axially, it moves a bit, by design. Are you quite sure that your spindle is not meant to move like this?

You should not have play. Assuming you are not pushing from the back (or pulling from the front) and overcoming the spring. Axial play in the bearings on that spindle should have been taken up by the spring, if you can move it easily you dont have enough preload (or any). Picture attached should be the same or very similar to your spindle, springs preloading 2 tandem pairs, if you have play you have problems. In this case it could also be play from how the spacers pin to the housings even though your bearings are actually preloaded.
 

Attachments

  • DSC02336.JPG
    DSC02336.JPG
    4.9 MB · Views: 19








 
Back
Top