[sendkeys]

Does anyone know at full heat what a 1 to 1.5 in.(ID) tapered roller bearing should be preloaded too(how many pounds)? I can't find it on any bearing website. only the basic directions like you use on a car.

What i was hoping to do is preload them with a spring washer, to so many pounds with a little slack. so low rpm dont have any play and high rpm doesnt overheat. bytheway im going to be pushing them right around there max rpm(5k+-). I would have went with angular contact bearing but didnt have the room. so my only opinion was highend tappers.

I would like to use the highest preload i could but also dont want to overheat. Also if no really knows opinions on a starting point for testing would be nice too

Thanks for the help.

[HuFlungDung]

I'm doubtful of the success you'll have running tapered roller bearings at high speed for very long. I find it strange that you could not find angular contact bearings that were more compact than tapered roller bearings.

Are you running your two bearings back to back, or how far apart are they?

The preload is quite delicate to set for high speed. It should almost be zero.

If we could see a sketch of how you have the bearings laid out, we might be able to give you some ideas. Ideally, you'd have the bearings back to back or something, to prevent having to deal with heat growth of the spindle. Then, at the other end, you'd have a simple ball bearing steady, which is a slide fit in the housing bore. This one would be the one that deals with the heat expansion of the spindle.

[sendkeys]

It's one bearing on the shaft far apart on each end like a 9x12 lathe or benchtop mills.

well problem with the angualar bearings is i would need 4 of them 2 top 2 bottom (left right in this case) but only had room for 1 on each end. I'm hoping to be able to seal the bearings in and use a oil bath. not in crapy sketch but there are two caps that go over each end that i think i can fit with seals.

It just sounded better to me if i loaded it with say 100 pounds of springwashers with 1/4 turn back that i could have it all :P low rpm with no play and high rpm with less heat :P

[marx911]

well normally tapered bearings are set with an internal precision spacer that holds the 2 bearings apart at some predetermined amount of endplay... after you set the endplay you would run the spindle up to speed is steps while measuring the temp on the bearing housing.... what you would want is to see is the bearing heat up then stabilize, if it never heats up then your endplay is too much and you would make the spacer smaller..If it never stabilizes and keeps getting hotter then you have too little endplay and you need a bigger spacer ... these are just general guidelines and I agree with HuFlungDung "The preload is quite delicate to set for high speed. It should almost be zero”.... the problem with using a spring washer as I see it is that as things heat up and start to grow things are going to get tighter which is going to create more heat which is going to make it grow and you will end up with thermal runaway
The other thought is that if you are using a spring in a spindle you will have a springy spindle it might tend to move around under load
Just my thoughts
Mark

[HuFlungDung]

It's one bearing on the shaft far apart on each end like a 9x12 lathe or benchtop mills.

well problem with the angualar bearings is i would need 4 of them 2 top 2 bottom (left right in this case) but only had room for 1 on each end. I'm hoping to be able to seal the bearings in and use a oil bath. not in crapy sketch but there are two caps that go over each end that i think i can fit with seals.

It just sounded better to me if i loaded it with say 100 pounds of springwashers with 1/4 turn back that i could have it all :P low rpm with no play and high rpm with less heat :P

Sendkeys,
If you are building this spindle housing yourself, its still not too late
You said you needed 4 angular contact bearings: the reason being that you would run them in adjacent pairs. True, but the preload method is the same, whether using tapered roller or angular contact styles, so you could also run pairs of tapered roller bearings back to back. One pair at the front end of the spindle would be best, and then a plain ball bearing at the rear end to steady the spindle and allow heat growth would also be best. It must not be captive, but must be able to float endwise in the housing.

I'm looking in my FAG bearing catalogue (can I say FAG in here? ), and it says the limiting speed is about 5000 rpm in grease, and maybe 6500 with oil.

Be aware that your shaft seals are going to cause a significant rise in temp just from doing what seals do best: rubbing on the spindle. A bearing running in oil is not necessarily better cooled, either. If the oil bath is captive, the bearing actually generates heat from stirring the oil. Some kind of a stream of oil is the most superior method of applying the oil to the bearing. But that requires the added complexity of an oil pump.

If your duty cycle with this machine is short, say 10 minutes on, and 10 minutes rest, you might not have to worry about overheating. But, the long distance between your bearings is going to give you grief when you run at high speed, IMO. It will just be difficult to make it right at both low and high temperatures. That's why I recommend you change your construction plan before its too late

[sendkeys]

this is for a old horizontal mill that im sure im asking to much out of it hehe. main reason i need 4 angular bearings was 2 didnt have the load rating i needed. But that could have worked guess if this doesnt work i might end up doing that way ,anyway.
By the way duty time is around 4 hours on, 1 off.

ok well thank you both very much for the help, Guess the idea wasnt really a good one. Seeing i already have the bearings i will give a it a shot doing it the normal way. if it doesnt work will have to try somthing else or live with slower rpm,or lower load rating.

Thanks again

[IndHobby]

If posible assemble the spindle hot, a pinch hotter than it will see in normal operation. This way expansion (preload change) will already be handled in the process and you don't need to worry about changes quite so much.

[TPPJR]

[QUOTE=HuFlungDung]Sendkeys,
If you are building this spindle housing yourself, its still not too late
You said you needed 4 angular contact bearings: the reason being that you would run them in adjacent pairs. True, but the preload method is the same, whether using tapered roller or angular contact styles, so you could also run pairs of tapered roller bearings back to back. One pair at the front end of the spindle would be best, and then a plain ball bearing at the rear end to steady the spindle and allow heat growth would also be best. It must not be captive, but must be able to float endwise in the housing.


Is it possible to convert a taper bearing spindle to angular contact? I have a RF-45 benchtop that I am running at 3k rpm with tapered bearings. I have been fighting an heating problem with the gearbox and noticed the lower bearing was hot. Upon removing the spindle both bearings had been pretty hot burning the grease I was using. Going to go with high speed grease this time to see if it helps but angular bearings caught my attention.

Thomas

[wizard]

Interesting comments but I'm troubled by one. Spring loading of spindle bearings is done to prevent losening of the preload as the spindle heats up when mounted as you describe. At least that is the way it has always been explained to me. This seems rational as the spindle itself will elongate with teperature rise, a rise that should happen faster than the surrounding bearing housing.

To be honest the few spindles I've torn apart with spring loading have been high speed units with angular contact type spindle bearings. Almost all of the machines I've seen with roller bearings had the preload set with an adjustable nut. As far as the precision spacers, as described above, those are almost always in spindles using angular contact spindle bearings. Of course you have to realize that the vast majority of my spindle work was done 25 some odd years ago on machinery that was old at the time. I don't think the technology has changed that much since then.

HuFlungDung seems to be on the right track here with bearing arraingements. The issue is do you have the room to do such a retro fit on that old machine. I can't answer that but it might pay to consult with a bearing manufacture for non standard product that might do the trick. As to load rating you are likely to loose some on the taper rollers at high speed so maybe angular contacts are not all that bad.

By the way what are you expecting to do with the machine anyways?

Thanks
dave

[NC Cams]

Tapered rollers are designed to run with 0.000 to 0.005" endplay.

The "preload" deal is a quick-a$$ way to try to obtain this figure with a quick-a$$ one shot tightening approach. If you literally run the tapers with preload, you run the risk of having problems as the edges of the rollers start to rub the flanges of the raceway and overheating and/or flange/roller scuffing develops.

The above clearance achieved by shoving the shaft in one direction while rotating with a dial indicator on the end of shaft until it quits moving axially, Then continue rotating while you shove it axially in the other direction. Keep tightening the bearing end clearance until you get above clearance.

The longer the shaft/ergo span between the bearings, the more you have to compensate for thermal growth of shaft and housing.

If you're doing a lathe spindle, you can consider tapers. If you're doing a mill spindle, stick with A/C's.

Wave washers are good to load the end support bearing opposite the workend of the spindle. On the work end, you want a postion preloaded A/C bearing.

Example: if you use a 206 size bearing on work end, use a 7206CTYDULP4 or DUM or DUH. Use either a 6205 ot 6206 with 1%-2% preload washer on drive end.

You can't maintain bearing rigidity if you spread the bearings out like you're suggesting with a mill spindle. Lathe spindle (due to preominant radial load application) will work OK with tapers, providing you get machine tool grade.

Generic "automotive tapers"or the like typically won't have the tighter runout needed for M/T used.

[WhiteTiger]

A lot of the problem with tapered rollers goes away if you get away from static lube (grease). If you go to oil seals and a central sump between your bearings you can avoid the pump question the way you have your bearings illustrated. As long as your oil level is up high enough to feed the open small end of the bearings and you have an oil return channel, the bearings themselves will sling the oil to the larger diameter and suck new in from the sump.

Some old babbit bearing precision spindles used just that sort of arrangement, with slinger discs on the spindle shaft running in the sump to auto circulate the oil.

Has the benefit of conducting the heat away from the running surfaces and into the housing. Down side is you have to design for seals, and ideally you would have a closed sump on each end to supply the bearings independently, which means 4 seals and seal surfaces.

Just a side note: when I was doing a lot of high speed, small diameter work on my lathe I had a bit of heat problem despite sump type oil lubrication. I drained all the oil, stuck slick 50 in and let the thing idle at high speed long enough to heat up, then ran it a few minutes on lowest speed to give the ptfe time to start to cure on the race and roller surfaces without building up. Cured the heat problem, and it's something I re-did periodically all the years I actively used the lathe.


Tiger

[Stevie]

I use FAG tapered roller brgs in my small lathe (actually the big one too)

I have no problems with setting perload and running the spindle at 3600 all day; course you need to cool the head if it's Alum like mine or expansion creates more preload and so on and so on

[sendkeys]

man i been around some time post that over year ago :P hehe As i remember,The old idea was to over drive my mill to a high speed but i ended up just keeping the slow rpm and building a new spindle for the side.

TPPJR you can convert it over without much problem but just remember that the angular contact bearing are rated much lower than the taper bearing. That's why i was going to have to run two on the bottom and two on the top. With your mill this might not be a problem but you will have to do the math

[Stevie]

Ok here's the scope
My Fag are 30mm bore taper rollers; max with oil is over 10,000rpm
I run about 1/3 of that but sometime more; the odd time only 15%
So if we take 33% of the max for the brg; and take the number of hours it runs
5 nights a week for min 2.5 hours 780hours
plus 8 hours every weekend day 820hours (sometimes more than 8 hours)
Lathe has been online for 16months at least so lets add 1/3 more = 2340hours
now we need mins so x60 = 1404480mins
say 3450rpm = 4,845,456,000 times my spindle has revolved since it came on line with the same set of 2 taper brgs and seals (CR brand)
I have busted 2 Polycord belts (1/4" round) by my stupidity; they did not wear out

It's actually running now while I type this
I set a small preload at first; maybe 1/8th turn on 14tpi threads for the locking nuts from the no movement setting 0.000; not 0.005 as NC cams said
so I have 14tpi divided by 8th turn = 0.0089 preload roughly


yes this might be an old thread; but it's providing good info
From the numbers here of mine;you can see taper bearings are good for high speeds
My product is so small i need to control the spindle to the max possible accuracy; try to make a brass part 0.020 diameter with 22 grooves 0.002 deep 0.002 wide spaced 0.005 apart with a crappy spindle; you'll never do it; the 0.020 is the biggest diameter on the part; the smallest is 0.009

[NC Cams]

There is a bit of an issue of symantics involved in this discussion:

In the bearing industry, it is well known that tapers will give the ultimated in life if it they are run with a TRUE slight negative clearance (IE: preload). HOWEVER, the key word is SLIGHT.

This method loads all the elements AT THE RISK of bearing damage should load and/or thermal stresses cause the lube film to rupture and/or the raceways or rolling elements yield.

Failure will occur shortly thereafter if "good/slight" negative clearance progresses to "oops, too much" and this transition has a VERY low threshold. This is why the 0.000 to 0.005" TOTAL axial end play spec is so commonly used with tapers. NOTE: this can be hard to measure properly unless one uses the proper equipment and procedures.

If you "snug up" a taper as a lot of procudures indicate, you do "preload" them. HOWEVER, once they start to rotate, the rollers move around and settle or "bed-in"in.

It can take 5 to 15 COMPLETE revolutions for the rollers to "settle" in to the rolling path that will result from the internal geometry of the parts. It is during this time that "preload" hopefully/typically settles out to the minimal axial clearance one hopes to achieve.

The 1/8th turn at 15tpi may have been the intitial "preload" set in the example. However, I wonder what it is now that the spindle has rotated and the elements have "bedded themselves"??? (see my prior post for axial clearance checking procedure).

I'd be very surprised if there was anywhere's near 0.008" negative clearance when the spindle is at operating temp.

As indicated in my prior posts, the 0.000 to 0.005" end play (as checked in the prior posted method) is the industry standard for axial clearance for tapers. One is free to follow or ignore it.

My point about "preload" is this: if you're going to use it in concert with bearings, one had better know how to properly calculate AND apply it... It is NOT a "universal, one procedure fits-all, do it the same way all the time" sort of deal.

Finally, it does help to know the ramifications of too much and/or too little if you do start to play with the process of arbitrarily preloading tapered roller bearings.

[Stevie]

I rotated the brgs with the drive motor for a few mins; so that argument is out the window; then pre-loaded them
I have no play; I knowthis because I regul chk it; as I said with the tiny parts i make; if the spindle was lifting or being pushed off under cut loads my parts would be scrapped
I don't follow why your still saying what you are; the proof is right here; and has been for over a yr
BTW
I use chainsaw bar oil for lube

I know reading your posts and what you do with race car stuff; you feel you are the primo expert here; well I grind 36" diameter 317 stainless steel Ball Valves with a Tungsten Carbide coating (0.01") and we get them dead round and a bubble tight fit to the seats with no CNC or CMM equipment; just experiance

Which is what I use to make my small cnc machines; which also seem to work just fine for long periods

[NC Cams]

My post was meant to be read by someone who was asking how to preload tapers (see post #1) who clearly didn't know how to do so.

Imagine someone inexperienced simply reefing down on a preload nut while NOT rotating the spindle while tightening it noror rechecking it after rotation as you did (seen it done).

Moreover, had you included the procedural qualification in your original post that you included in the later pose, I wouldn't have chimed in with my offerings - I could not and would not disagree with your amended post, especially if the procedure netted a 0.000" to VERY SLIGHT preload as is truly optimum for tapers.

We're both seasoned veterans at what we do. Thus, we can easily and perhaps often do tend to forget to include some of the intuitively obvious details that we've learned via our life experiences. Novices often take what we say literally and therein lies a potential for misinterpretation.

During my career in the auto industry, I saw guys preload pinion bearings numerous times by running them down with impact wrenches to obtained the specified "preload" rotating torque. I also saw guys use breaker bars to creep up on the same spec all the while rotating while they tightened (which is how they are done at the factory by the way).

More often than not, the flat rate beater, impact wrench jockeys had comebacks due to noise than the guy who used the breaker bar....Same "preload", different attainment procedure, drastically different results...

Point being, the devil was in the details on HOW to get to the results in the proper way.

Bearing expert??? Hardly. I just spent more time performing bearing applications engineering for a living than the average DIY'er who frequents the site.

I'm only trying to share what I learned via conveying the methods that would be recommended to someone who took the time to contact a bearing engineer at a major bearing company....

[Stevie]

I think you missread that last post
I preloaded the 1/8th turn after a run with the motor; so i do have 0.009 preload plus my headstock runs about 40dgr C adding more preload than the 0.009 as my spindle is steel (4140) and the head is 6061 T651
The expansion must increase the preload by at least another 0.0015
In summer I've seen my head temp hit 60drg; then i used to shut it down
Now I have a compumotor full size heatsink (about 12" by 6" fully finned with PC PS fan)
It now stays at a max of 45dgr even in summer; winter about 35dgr
strangely enough I found the head stays a lot cooler with hardly any oil; as it only holds about a shot glass anyways; min must be about 2 TeaSpoon's worth of bar oil
I tried full Syn/ic oil; just made it hotter; 5W30 was close to the bar oil; but this chainsaw oil seems the best up to now
I change it about every 2 months

[TPPJR]

As indicated in my prior posts, the 0.000 to 0.005" end play (as checked in the prior posted method) is the industry standard for axial clearance for tapers. One is free to follow or ignore it.

Question:
How do you set the preload for a set of taper bearings when both sides are press fit onto the shaft?

Thomas

[NC Cams]

Two replies - 1 post

Setting preload per post #19: This is tough to do as you are dealing with both a position preload combined with having to deal with the inner ring growth due to the press fit (ring growth does affect clearance). Each situation is different so it is impossible to give a fits-all method.

One method involves doing a trial fit where you press the parts together and check for desired/prescribed clearance/preload. You then R&R the bearings and set the clearance/preload with shims. This method is done with ring and pinions all the time.

There are other methods s but they involve a lot of tricky measuring procedures which are to complicated to deal with in a forum like this. (see following for more details)

re; Stevie's post #19: Depending on how the bearings are mounted in your machinery, you could either ADD or LOSE preload.

If the cones are pressed onto the shaft with large OD facing each other, the housing that holds the cups would tend to expand AWAY from the cones as the aluminum housing expands faster than the steel shaft.

If the cups are mounted in the housing, and the cones are slid onto the shaft so that the SMALL ends face each other, your preload should increase as thermal growth takes place.

The important thing to emphasize is that the amount of preload that an S5 automotive wheel bearing set would tolerate is NOT the same as what a 94649/94112XD cup/cone set would tolerate simply due to the physical size difference.

There is also a difference in cone/cup roller angle. A shallow angle bearing is designed to absorb more radial thrust. Whereas a steeper roller angle is used where higher axial thrust is involved. Thus, the same size bearing will NOT necessarily tolerate as much preload as the next depending on internal geometry.

Moreover, housing and shaft fits (pressed versus loose), thermal growth, lube type and method, et al come into play to determine the optimum clearance/preload for a particular situation.

This is why Timken has three factors that they consider when establishing bearing clearance/preload:

a3k = life adjustment factor for bearing's internal load zone
a3/ = life adjustment factor for lubrication
a3m = life adjustment factor for alignment.

these factors are multiplied together to comput "a3" which is an "environmental" or "load zone factor". This effective is a determination of how many rollers share the applied load.

The load zone is definitely a function of endplay (internal clearance) or preload within the bearing system. Ultimately, load zone is affected by the initial settings, bearing geometry, load application and shaft/housing deformation and alignment.

The rated L10 life of the bearing merely assumes that the load zone factor where a3=1. If this condition exists, there is a 180 deg load zone where the for the bearing rollers are loaded.

If the load zone factor is LESS than 1 (>180 deg), Timken recommends that a computer analysis be performed to determine life potential, regardless if there is clearance or preload.

The above explanation was obtained from Timken's 2002 "The tapered roller bearing guide".

It is important to note that a SPECIFIC clearance spec is NOT published for any bearing in the guide. Clearance/preload tend to be empiricaly derived/verified values that are machine and/or application specific in nature.

In conclusion, what works for Stevie's application might/might not cause problems for someone else with a different loading and/or useage cycle.