What's new
What's new

Rules of thumb for setting spindle bearing preload

rhb

Aluminum
Joined
Apr 27, 2019
Location
A small town in central Arkansas
Due to some "cleverness" on my part I had to remove the spindle nuts on a new lathe I'm setting up. Prior to that the spindle bore clocked at less than a tenth TIR. I have no preload data from the OEM yet. I sent an email, but I expect it will take a few days to get a reply.

I've never worked in a big manufacturing plant, but I assume that the repair crew has rules of thumb for resetting preload when they don't have OEM data. Can anyone who has such experience comment?

I've read through the "Tips for adjusting spindle bearing preload on a lathe" thread. It generally describes what I have done, but I'm nervous. I'm OK with making mistakes, but not with doing damage.

What I did was clock the spindle bore and when I got to about a tenth TIR, stopped tightening the nut. I'm familiar with adjusting bearings based on both axial end play and friction and have the kit to do either. But I'm used to having a specification from the OEM.

Tearing it apart to find out who made the bearings and then track down the bearing OEM specs is not very attractive. That takes time and I want to get this in service.

The mechanical engineers who design machine tools have to determine required preload for whatever bearing system they decide to use. I've got a very large technical library, but I could find nothing on the topic other than in a Torrington catalog which did not list the size taper bearing bearing I'm dealing with. There was no discussion on how this is determined by the bearing OEM.

Ideally I'd like to find a mechanical engineer level discussion of how to determine the required preload for a bearing given a set of operating constraints for any type of bearing, taper, angular contact, deep groove, etc. It seems to me in my ignorance that a small number of people who have to do this daily know the answers.

Thanks,
Reg
 
does it spin freely? to loose.
does it bind? something is cockeyed, rotate bearing in housing and or rotate shaft in bearing. Voodoo ritual and ratio until order is close enough to restored.
does it not turn? to tight.

let sleeping dogs sleep - and do not disturb good bearings.
 
The bearing type might help with providing an answer.

I am used to Hardinge lathes that use angular contact ball bearings on the spindle. This type of bearing has the preload specified by the machine designer ("light" is typical) and built into the bearings during final bearing manufacture. The machine parts just have to clamp the bearing races tightly together to produce the specified preload. The machine assembler had no means of changing the preload; just tighten the nut.

I think tapered roller bearings are different. I recall back in the 1960's that heavy truck rear axle pinion assemblies had the preload set during final axle assembly. The assembler wrapped a string around the bearing housing and pulled on the string with a spring scale calibrated in ounces of force. He adjusted the preload to get a certain string pull force and then bent a toothed washer to lock in the nut position.

Back then, I was a mechanical engineer in the axle test laboratory of a major truck manufacturer. We built engines, transmissions and axles for our trucks, so there was a lot of stuff to test. I saw the assembly process and wrote the reports on the tests, but had nothing to do with bearing design. Hypoid pinions produced simultaneous axial and radial loads on the supporting bearings when driving the ring gear. We could measure actual pinion input torque while operating test vehicles. The axle dynamometer (photo below) could run steady state endurance tests at various speeds and torques to monitor quality. Timken provided excellent bearing design handbooks and their engineers worked with our design engineers to select bearings and specify preload and assembly methods. We then published truck maintenance manuals that described the assembly methods for repair shops. Those manuals were also used in the test labs.

Larry

Scan_20201110 cropped.jpg
 
Last edited:
IIRC air cooled VW front wheel bearings are 0.001-0.002" end play. My Toyota truck is 5 lbs of force on the wheel lugs to rotate the bearing. In both cases you then lock the castle nut and insert a cotter pin. If I had to retighten a front wheel bearing today I'd look up the factory spec and do exactly that.
I did not want to loosen the spindle nuts, but I didn't have a choice. I had thoroughly stuck an MT taper ER collet chuck in the spindle. A fairly sharp blow from a 4 oz hammer on the drawbar would not dislodge it. I did not want to risk spanking the bearings by hitting it harder so I removed the spindle nuts and the leadscrew drive gear, set up a gear puller on a heavy washer and used that to free the collet chuck by pressing on the loosened drawbar.

That left me with the task of resetting the bearing preload on a pair of taper roller bearings without any OEM data. I'm engaged in helping a friend with an old Nardini he's been rebuilding with the same issue. How much preload and how should it be measured?

I've got a roughly double the TIR on the inside of the spindle relative to when I clocked it prior to getting the collet chuck arbor stuck. I feel reasonably sure I'm not too tight. I don't want to overtighten the spindle nut and risk damage to the bearings, but I'd like to cut the TIR to the half tenth it was before I got clever.

I'm probably being OCD about it, but I've always taken pride that I did things to the highest standards. When I bought my first dial indicator circa early 70's to set things to VW factory spec, I bought a 0.0005" ENCO branded unit made in Japan which I still have and use. IIRC it cost roughly half my monthly allotment for living expenses at university. So no "fun" money for several months after.

Even at my age I make mistakes, but I try to learn something from it if I do. I foolishly used the drawbar to hold the spindle while tightening the collet chuck nut instead of fetching a larger wrench to fit the flats on the chuck. Color my face *very* red.
 
Check and make sure you have no endplay to start. From there tighten it up a little at a time and run it at a decent speed and check temperature. If they're still the original bearings you don't need to worry about break-in so just make sure they don't get too hot. You don't want to be much above 120° - 130° max. after extended runs at high speed. Too cold means probably too loose. You don't want that either or you'll have no end of chatter problems on ops like parting or form tool work.
 
Last edited:
I am sensing one of those 'pronoun' situations. You know, you meet someone and they keep mentioning their 'partner' and how 'they' did this or that and it dawns on you that they are purposely avoiding mentioning the gender of their partner.....

Is this a Grizzly lathe?
 
Old rule of thumb regarding the heat of headstock bearings.

If you can keep your hand on the housing for a while - that should be ok.
If you can’t keep your hand on the housing for more than 5 seconds - that’s a cause for concern.
If when you spit on the housing your spit sizzles - Stop right away.

I always used a 3ft long crowbar between the spindle flange and the headstock casting. With a DTI plunger on the face of the flange and using moderately firm force on the bar I’d expect to see movement in the range of 0.0005” to 0.00075”. I always insisted on seeing some movement. You don’t want the assembly to be solid.

Regards Tyrone.
 
Thanks.That reminded me I have a thermal camera. So I'm going to run it and monitor the heating. Should have thought of that earlier.

A bit too focused on the indicator using it to measure end play and runout.
 
If I recall from a post here, SIP jig borer spindles had preset set by giving the loading nut a twirl and having it stopped by the bearings. Some super magical procedure performed by wizards or some such.
 
I am sensing one of those 'pronoun' situations. You know, you meet someone and they keep mentioning their 'partner' and how 'they' did this or that and it dawns on you that they are purposely avoiding mentioning the gender of their partner.....

Is this a Grizzly lathe?

Doesn't really matter in this situation, he's within the rules even if it is. Questions about procedures applicable to any lathe are just fine.
 
In general use such as wheel bearings or gearbox shafts I'd agree John; but on a lathe, my experience is that they do need to be preloaded. Not heavily, but just enough. And yes, there is reading material on spindle bearings abound. Good suggestion for the OP to do some more reading of the right stuff if he hasn't already.
 
LoL! Busted!

It's a 6" x 10" G0937 combo lathe/mill. Basically an improved (dovetail ways) version of the EMCOs I lusted after in the 90's but couldn't afford. I bought it specifically to make gun screws and parts which doesn't require great precision, but when I started clocking it and getting errors in the 1-2 tenths range I started getting ambitious. The machine is a lot stiffer than anything it can hold. The lathe appears to be the Grizzly M1015 with a milling head bolted on. Normally I'd frown on a vertical milling using a lathe, but the dovetail ways prevent the carriage lifting.

I have a Clausing 4902, but that's not really a convenient machine for making #6-48 screws. I plan to use a crank on the spindle a la Sherline since the screws are so short.

After many cycles of prying, adjusting and testing it's now much quieter. I've let it run for 15-20 minutes at 2000 rpm since my last tweak of the spindle nut. It's at about 110 F as a guess based on sticking my finger in the spindle. I've got 1-2 tenths TIR on the spindle face and it moves in and out about 2 tenths when I reverse the direction. I think it's time to stop. The spindle bore is less than a tenth TIR. The needle barely twitches running a low speed. The temperature and fact that the spindle face consistently moves in and out by 2 tenths when I change direction makes me comfortable it is not too tight. And the sub tenth TIR on the spindle bore says it's "good enough".

After I got it, but before I had gotten serious about clocking the spindle I holed up with a paper copy of the Torrington catalog for days thinking I should get higher grade bearings for it. Nothing I read addressed the logic of how to adjust bearings with basic tools. Bearing maker docs on basic procedures are the sort of references I am looking for. I prefer paper, but I'm going to download all the major maker publicartions to my iPad.

Many thanks to eKretz and Tyrone for their suggestions. That definitely got me on the right track. It wasn't until I started using a prybar that I got proper endplay readings. But that combined with the temperature monitoring makes me pretty confident I've got it set properly. And I've gained a much better understanding of how to adjust rolling contact bearings. In the past I did what I was told rather than figure it out.
 
In general use such as wheel bearings or gearbox shafts I'd agree John; but on a lathe, my experience is that they do need to be preloaded. Not heavily, but just enough. And yes, there is reading material on spindle bearings abound. Good suggestion for the OP to do some more reading of the right stuff if he hasn't already.
Another great pub is the two volume Timken Engineering Journal
 








 
Back
Top