Tapered roller bearing preload or clearance?

[strokersix]

Hello,

Automotive tapered roller wheel bearings are often recommended to run with a little clearance. Other applications, machine tool spindles for example, are preloaded. I would like to know what the issues and trade-offs are when designing a tapered roller bearing spindle.

Clearance- cooler running, maximum load capacity, maximum life

Preload- maximum rigidity

Looking forward to responses because I expect it is more complicated than this.

Thanks, Mike

 

[evelrod]

Before it gets too complicated, the "rule of thumb" has always been to 'never' preload roller bearings and 'always' preload ball bearings. As this relates to automotive use you will read things like "tighten to 'X' ft.lb. and back off until the washer is free" or "tighten to...and back off to next..." typically yielding some lash in roller wheel bearings. On the other hand, "after all the free play is removed, tighten to next slot"--- giving prealoaded wheel bearings in ball bearing applications.
In early GM cars using ball bearings, failures were very common. Probably caused by setting them up as you would a Ford (roller bearings) with lash. Working in a auto repair garage while I was in college, this was one of the most common repairs.

Since this is a narrow application and "bearings" uncorporate some pretty advanced theory that's my best.
After 50 years of fooling with cars, I tend to stick with what works.

Rod

 

[GregLocock]

In my thankfully short job as a chassis design engineer I was responsible for wheel bearings, which were taper rollers at the time.

Facory spec was 1.4 Nm (finger tight on a 1 1/4" nut) torque.

We would have preferred to lower that but that was the setting on the machine that gave the most reliable assembly.

TRBs hate having low preload, but, the expansion of the hub under braking would be enough to cause direct damage to the races, so we had to accept it.



Cheers

Greg Locock

 

[strokersix]

Greg,

"TRB hate having low preload"

Please elaborate on why you say this.

'60s and '70s vintage General Motors service manuals specify clearance in the wheel bearing assembly. Is the design intent to prevent too much preload as the hub expands with braking heat?

A rough way to set machine tool spindle preload is to run the spindle toward the high end of it's speed range and monitor the temperature for an hour or so. If it's barely warm then a little more preload can be used. If it's too hot to touch then back off the preload. I wouldn't do this with a new high dollar machine but I've had good luck with this method for simple machinery.

Mike

 

[Metalguy]

Strokersix,

Yes, it's the heat, as GregLocock already pointed out.

 

[GregLocock]

GM were ball bearings, so what they used does not apply to TRBs.

I've dug out my notes - optimum is zero preload, but no slop. Slop is worse than slight preload.

So, I was wrong, TRBs do like light preloads, use just enough to ensure that you don't run into end play.

However, if you are running gears then more preload is a good idea since Timken claim this will help stabilise the geometry. They suggest that a preload of half the end thrust generated by the gear set is ideal, if I'm reading this correctly.

Here's the general recommendations for typical automotive installations:

passenger car front wheel - slight preload to slight freeplay (big help!)

engine idler gear train 15 lb in torque (these are very short spacings, and the gears and the shaft are the same material, and are probably all running at roughly the same temperature - around 100 deg C)

Now the question I ask myself is why we didn't reverse the bearings, so that as the wheel expands due to brake heat freeplay develops, but that most of the time we could have run with a bit of preload (that 1.4 Nm is just enough to pinch the bearing, it is not what I would call a real preload).



Cheers

Greg Locock

 

[Metalguy]

"Now the question I ask myself is why we didn't reverse the bearings, so that as the wheel expands due to brake heat freeplay develops, but that most of the time we could have run with a bit of preload (that 1.4 Nm is just enough to pinch the bearing, it is not what I would call a real preload)."
Greg Locock


I think this one has some simple answers. It is obvious that having a high preload when cold will give an even higher one when hot. But the temp. range isn't very large, and the distance between the races is fairly short. Only if we start with a high cold preload do we have a problem. Also, the balls/rollers on car/light truck front wheels are small, so the skidding that will destroy a larger roller/ball, because of its inertia, doesn't have any effect at the slow wheel speeds-even at 100+ MPH.

I was involved in one case which lasted a few years. We had 4 large vertical-shaft fans in a $$$-critical environment (nuclear power). We only needed 1 of the 4, or we had to shut down (at ~$1 million /day). The motor mfg. had screwed up big-time, using back-to-back angular-contact ball bearings without much thought to preload. The bearings were about 8" OD, with balls ~1"+. We had many problems, always with the unloaded bearing.

We called in the bearing mfg. (one of the largest and best, IHO), and after studying the drawings they immediately saw the problem-no preload. But some of the dimwits decided that a *major* electric motor mfg. couldn't really be that stupid, so there must be something else. Yeah, there *was* something else-they had used bearings with NO seals, so when the grease got hot and thin it just ran down out of the bearings! That part was easy to fix, but getting some accurate preload wasn't. We also discovered that the bearings were approaching 300+ deg. F, so I came up with a fluoro-silicon grease which could take 450 Deg. F (at $100+ per tube, back in 1985!)-it's used in expensive "sealed-for life" bearings.

About 2 weeks later one of the bearings with *my* grease failed-and I mean completely. Looked like it had caught on fire, and someone tried to put it out with gasoline and a large grinder! I was dumbfounded. Much later I discovered the mechanics had packed it *full*, thinking that would help things.

It gets worse-when one of these bearings failed, it usually vibrated so much that a large ring of bolts holding the fan shroud together would fail via fatigue cracking, and we actually dropped one section of shroud right on the floor-which was a ways below, and I'll not elaborate on just what was struck on the way down. So part of the interim fix was (I am NOT making this up!) was to weld a piece of chain on the outside of the shroud to catch the lower piece on the way down! Yes, we also did the usual bolting things-went from the orig. soft bolts to grade 5, then grade 8 and raised the torques to match, but nothing was able to withstand the level of vibration when the bearing failed.

Naturally the fix that worked was to change the motors to a design which would allow preload to be set and maintained. At one point of this whole episode an engineer wrote an "official" report which blamed the *new* grease failure on the "fact" that "the new grease was so slippery it allowed the balls to skid instead of rotating". He was lucky I didn't see that until years later!

 

[GregLocock]

I don't understand what you mean about my idea, the idea is to have the outer getting hot and expand. With the TRBs reversed then they will get less preload when hot. The main disadvantage I can see is that the you have to have a nut bearing on the outer race to apply preload.

Anyway here's NTN's take on preloads for a few different bearings

Precision angular contact bearings, TRBs -to increase rigidity - fixed position preload. (eg lathes milling machines, differentials and axles)

Angular contact bearings, deep grooves, trbs- prevent vibration etc- constant force preloads (eg small machines, electric motors)

Various thrust applications - sprung preload so that unloaded side does not smear. (sounds like your problem)

Cheers

Greg Locock

 

[Metalguy]

Greg,
"I don't understand what you mean about my idea, the idea is to have the outer getting hot and expand. With the TRBs reversed then they will get less preload when hot. The main disadvantage I can see is that the you have to have a nut bearing on the outer race to apply preload."

What I meant was that yes, your idea would work, but I think it's unnecessary for cars and light trucks, because of the small roller diameters and low RPM involved.