Double row angular contact ball bearings

[Frank]

Hi All,

I'm playing around with a Van Diemen RF91, and have a few questions about double row angular contact ball bearings (DRACBB's).

It's got DRACBB's in the rear hubs.

DRACBB's are 42x75x37mm, C3 internal clearance, steel inner cage, and a two-piece inner ring. The (pairs of) balls have about 18mm centreline spacing.

Do you find DRACBB's suffer from compliance due to the small ball centreline spacing?

I guess they are not equivalent to a pair of single row deep grooves with a spacer, because they contact differently when a moment is applied.

Do DRACBB's need to be a special construction (perhaps a certain inner ring design, or a particular contact angle) to resist a moment?

Are (large?) preload forces on the inner rings required to take up internal clearances?

Are split inner rings a bad idea? / good idea?

I have heard that people shim the inner rings to increase internal clearance (make them run colder, with less friction). Does this typically result in compliance?

Compliance that causes "pad knockoff" and poor camber and toe control?

Does anyone else have similar experiences with DRACBB's?

Are they reliable for racing (constantly undergoing a cornering moment) as opposed to OEM use (most of life spent driving in a straight line, with only a small moment)?

I see they save space axially, but I can't see that they result in a lighter design in typical FSAE rear ends.

Has careful detailing yielded lighter designs with these bearings?

Thanks in advance.

[jsmooz]

Frank - DRACBBs are the way to go, espically over DGBBs. Deep grooves are designed to take minimal axial load. Any type of road racing is going to exert axail loads, usually high ones. As with anything, the sizing of the bearing must be proportional to the load applied. Personally, I favor solid rings because the bearing is then assembled with the correct preload leaving out the human error.

[Brian Evans]

DRACBB's are common on smaller race cars up to about F3 and F. Atlantic. They are usually an automotive OEM application of some sort, VW Rabbit, Audi are common, and some are industrial. Speaking of Van Diemens, the RF-83 I am restoring uses a rear bearing that originated as a industrial bearing used in elevators, has been pretty much out of production for years, costs a silly amount, and is used in some Lamborgini's and if you get it from those guys costs double silly...

Larger cars and endurance cars tend to use double taper roller bearings with pre-load spacers.

If you install a DRACBB (I call them cartidge bearings, easier to say) with a split inner race you do need to watch how you install them, or you do lose the correct preload.

Brian

[Frank]

If you install a DRACBB (I call them cartidge bearings, easier to say) with a split inner race you do need to watch how you install them, or you do lose the correct preload.
Could you elaborate on this aspect a bit more?



Jim, you react a moment at your bearing compliment. This is the biggest source of (radial) load.



This is what the RF91 has at the moment
http://www.skf.com/images/cat/images/1/1_3/051_0210.jpg


Larger cars and endurance cars tend to use double taper roller bearings with pre-load spacers.

Is this what you mean Brian?
http://www.skf.com/images/cat/images/1/1_14/1000f05b.jpg

Or do you mean two bearings (could be a number of different bearings) separated by a spacer?

[Jersey Tom]

I wish I knew how to figure out how much to preload a DRACBB.

[Brian Evans]

Install DRACBB's with splt races - the only thing you need to be aware of is to press the bearing on to the hub so that the two inner races are pressed together during the installation. It's possible (people have done it so I know it can be done - to install the bearing so that the two inner races are pressed apart during the process, that ruins the design preload on the pair of bearings.

The correct preload on a DRACCB (in case it was a serious question) is zero, with no space between the two inner races. People have toyed with somehow installing a spacer so that the bearings run loose.

The taper roller setup that is more common on larger cars is shown in Prepare to Win, two separate taper roller bearings with a machined spacer to set the preload. It creates a very stiff assembly, high load potential. No particular reason why you couldn't use the one you show, but you lose the load potential from the larger space between the races so why not just use the ball bearings? hewland uses the double taper bearing as a pinion bearing in some gearboxes.

[Brian Evans]

To quote from Racecar Engineering March 2005, which I keep in my office as it is a good article on front uprights, ex-Team Lotus chief designer and designer of the Saxon car "Double taper rollers are good for larger cars but double ball bearings are okay if life is to be short and loads not too high" Peter Elleray on his design for the 2003 Bentley LeMans car "For a customer car you'd probably use a double taper roller, but we went for opposed angular contact ceramic bearings which are costly, light, but come with their own seals so they are easy to package. Although setup is critical, they were obviously reliable" In case you forget, the Bentley won...

Brian

[Analogue]

Back to back arrangement with angular contact bearings will have higher moment stiffness than deep groove ball bearings. The lines of action intersect further out on the shaft than the actual distance between the bearings.

Even deep groove ball bearings are angular contact if the races are displaced slightly. The trick is knowing what angle the balls are contacting at, and there's really no possible way for the DIYer.

We have found that the amount of press fit on the outer race is critical to proper fit. With VW rabbit bearings in an aluminum hub, we have a bit more clearnace than we would like. They have a split inner race, and the fit on the shaft is a fairly well done slip fit. We're talking maybe .0003 under.

Match ground angular contact bearings will have the proper amount of clearance when the edges of the inner and outer races are coincedent.

Large preload forces are never required, or desired, or even acceptable. The outer races must be restrained such that they are the proper distance from eachother-- must not be able to move together. The inner races are usually clamped together or against a spacer, but like typical tapered roller setups, it isn't required.

You would not use a front to front arrangment for high moment stiffness.

Keith

[Frank]

I should have probably not used the term "preload"

what I meant to say was...

"Do DRACCB's require a large axial clamping force on the inner rings to take up the radial clearance". Do you simply tighten until they "get a friction-y / sticky"?

I'm starting to seriously doubt that these bearings are suitable for reacting a moment.

Is it possible that people confuse these with double row taper rollers?

Is it possible these were only used in automotive applications where there was no reaction of a moment (solid rear axle arrangement)?

Is it possible that DRACCB's and double row angular contacts look identical from the outside, and possibly get interchanged by mistake? Perhaps this happened on the RF91 I'm looking at?

[jsmooz]

All I will say is F1 uses ball bearings, I know that for a fact.. That doesn't necessary mean it's gospel, but if any form of racing does a through job of engineering it's F1.

Frank, the way you are describing clamping force that is preload. If you take up radial clearance you are increasing preload. However, split inner ring usually rest against each other so inorder to over preload things you really have to tight the crap out of it.

[Analogue]

They are wonderful at taking a moment, don't worry about that.

They are very ideal for racing too. They are lower friction than rollers and have higher capacity than deep groove ball. Rollers might have a higher moment capacity for a given size, but they don't deal with thrust forces well due to the shallow angle on the rollers.

The industrial double row angular contact bearings are probably more intended for radial loads, such as machine spindles supported at both ends. Automotive double row angular contact bearings are a bit wider, and thus better for moments. If you are constrained on space though, the angular contact bearings will beat ball bearings every time.

If you are stuck using industrial bearings, get the highest contact angle possible, usually 45?.

All the angular contact bearings that I have seen used on cars have a split inner race and a one piece outer race. The inner races get clamped HARD, something like 180 ft-lb on around a 3/4" nut. You will not feel resistance increase as the nut is tightened. Proper housing diameter is critical for bearing clearance.

Keith

[Brian Evans]

All of the race car upright applications that I have seen using DRACBB's use a fairly stout press fit into the upright for the outer single race against a register, retain the outer race with a circlip, plate, threaded ring, or something else, and have a fairly stout press fit of the hub shaft into the split inner race. The hub shaft is somehow retained, usually a bolt arrangement, and the 3/4" 180 ft-lbs suggested is typical. the fit of the hub shaft into the inner hub has never been a slip fit in my experience, which is not to say it couldn't be.

Brian

[Jersey Tom]

The wheel assembly compression is the preload I'm talking about, ie '3/4" bolt at 180 ft-lbf'

Which I happen to feel is excessive. What is that, 6000-10000 lbf of preload?

[Brian Evans]

The bolt isn't really there to pre-load the bearing, it's really there to hold the hub on. The bearing pre-load is set by the accurate manufacture of the races, the amount of tension squishing the two inner races togther is somewhat irrelevant. The call-out by Holo-Krome for their 3/4" 16 tpi fastener is 6,750 inch-lbs, which induces 39,150 pounds of tension, and that is 75% of yield. From their catalog, I didn't calculate the numbers.

The torque on the bolt is usually less than that because the "bolt" is usually a spindle of some type, manufactured in a workshop, not by Holo-Krome, lathe cut threads, etc, and a grade 8 nut. All that's really required is to stretch the thing so that it doesn't auto-undo...

Don't forget that we started off by talking about a Formula Ford race car, not SAE. The bias is towards long life and not breaking down rather than light weight and innovative engineering. I just replace the wheel bearings in a RF83 Van Diemen for the first time since it was made, in 1983.. Yes it needed new ones, but 24 years ain't bad...and the wheels never fell off the car...

Brian

Brian

[Frank]

This is getting very interesting

Thanks for your patience Brian.

I usually think of interferance fits as being the way to take out radial clearance from a bearing.

Bearings are made with raial cleanance diliberately to allow for thermal expansion and interference fits. Hence why C3 (large clearance for heavy interference) and C2 (small clearance for light interferance or transitional fits) are made.

What I noticed on the DRACBB's is that the inner rings have a thin section (looks like an axial spacer with about 1.5mm wall. The bearing is machined down deliberately here). I'm of the opinion that this part of the bearing WILL compress with varying axial load (bolt tension). I'm guessing the force of tension is then somewhat responsible for the radial clearance.

Perhaps this shouldn't be a C3 bearing? (or it has not enough interferance.. common with ally uprights I guess... thermal expansion etc)


Automotive double row angular contact bearings are a bit wider, and thus better for moments.

Totally agree with this. The bearing is soooo narrow.

I also believe the contact angle must need to be different if the bearing is reacting a moment

Thanks everyone for your comments and help on this! Bearings are cool. Very fun topic.

Brian, how many different types of Formula Ford rear hubs have you seen disassembled? How many used DRACBB's? How many used DRTRB's? Other bearing compliments?

[Brian Evans]

Rear uprights on all of the race cars that I have seen were of three basic types. And I don't know the multi-letter acronyms...to be honest, I'm a guy who has been building race cars for 30 years and I find this forum to be both highly educational and entertaining...

The larger car uprights have dual taper roller bearings with a machined spacer. These are one things like Daytona Prototypes as a current example, and I have two cars in my shop with this bearing package using spindles out of a 1967 Corvair.

Many cars have dual ball bearings widely spaced with machined spacers to set preload. Completely adequate but hard to service - you have to press the bearing in before you can check the preload. Almost all modern cars use the cartidge bearings, and the difference is in the size - larger cars use Audi bearings, for example. 70 - 75 mm comes to mind as the larger size that I've seen as common.

Production cars, including race cars, need to finish races and be reliable first, fast second, maintainable third but still important. Always something to keep in the back of your mind. Club racers are going to ceramic bearings - club cars cost up to $100K these days.

Brian

[Brian Evans]

Correction: 90 - 100 mm OD on the bearing is the larger size that I have seen as common on cars like FFords.

Brian

[terra_dactile]

Hi guys,
I thought these documents might be of some help, or at least speak of the same subject being bearing in racing applications, the pdf documents come from SKF Racing on the www.SKF.Com (http://www.SKF.Com) website.

Im sure Jim having already worked at SKF must already be aware of these documents but for the others it might be something new.

Racing loads (http://www.skf.com/files/005030.pdf)

Racing Preload (http://www.skf.com/files/005031.pdf)

Jude Berthault
ETS FSAE
Vehicle Dynamics Leader

[Frank]

thanks for those Jude!

[jsmooz]

Jude - Nice find. I haven't seen those exact documents but I do have access to other documents with similar calculations and descriptions. And if I remember correctly you do also.

Brian - Your statement about almost all modern cars using cartridge bearings is incorrect. Most cars in developed countries use a hub unit. Hub units are bolt on bearings rather then press-in units. Descriptions here (http://www.skf.com/portal/skf/home/industries?contentId=0.000336.000728.000729.000952 ).

Frank - The contact angle would only change depending on the degree of loading, radial vs axial. Think of a deep groove, they handle very high radial loads but not axial. On the other hand, thrust bearings handle axial loads and not radial loads. The angle of contact in a ACBBs determines the degress of axial/radial loads that can be sustained. As the angle increase the axial load capacity increases, but the radial capacity decreases. It's a trade off.

[Brian Evans]

"Brian - Your statement about almost all modern cars using cartridge bearings is incorrect. Most cars in developed countries use a hub unit. Hub units are bolt on bearings rather then press-in units."

I was talking about small race cars, not production cars. I don't follow production cars at all, zero interest. Bolt in bearings sounds interesting from a production perspective.

Brian

[Chris Allbee]

how do you bolt in a bearing?

[kwancho]

The hub's pressed into the bearing, and the bearing's pressed into kinda a holder thing, which bolts to the upright. I worked with some Ford Five Hundred parts over the summer for a research vehicle, and they were like that. Kinda easy, I guess.

[Brian Evans]

According to the SKF reference link provided, in one example the inner race is pressed onto the upright (a stub, i guess), and the outer race rotates and has an integral flange to mount brake disc's and wheels. A second design has both the inner and outer race with integral flanges, the outer race bolts to the upright and the inner race directly mounts the brake disc and wheels. No machining of precise fits, no pressing, serviceable anywhere with a wrench, very expensive in small quantities (the flanges are custom per the manufacturers specification), but once you got going it would be great!

Brian

[jsmooz]

You got it Brian. Depending on the design the flange(s) that bolt to the rim or knuckle can have the raceway(s) machined directly onto them. Reduces problems from line workers assemblying the bearing incorrectly.

What's really cool is the Ferrari hub units. I've seen them when I was in Italy for work last year. The single center nut hub units are awesome. Got Ferrari lazer etched on em. I so wanted to walk off with one, but they are ungodly expensive.

[murpia]

Here (http://www.fiatforum.com/panda-guides/44993-front-wheel-hub-bearing-replacement.html?garpg=3) is a link to a page showing (possibly) the smallest / lightest version of these 'bolt in' bearing packages. I'd be interested to learn if any team has used such a product? Also I think SKF offer a 'racing' version called the HBU6 (http://vsmus-origin.skf.com/Files/Product_Flyers_PDFS/Hub_Unit_Broch_FINAL%20457616.pdf) but further info seems hard to come by...
Regards, Ian

[Brian Evans]

You know, this thread about bearing taught me that there are bearing/hub units, and that they combine the cartridge ball bearing assembly with the hub, and that a whole lot of cars have used them for years...who knew? not me...

So I looked into it, I found a street car (1990-ish Dodge Colt) that uses the same PCD bolt pattern as my vintage sports racer, and uses an integrated DRACBB bearing/hub unit for it's undriven rear wheels. So I found them for - get this - $20 each, new, from an on-line supplier. So I order a pair, and I'm going to see if I can adapt them to the front uprights on my car. Rather cool, I thought, and I thought I would say thanks...

Brian

[poweredbyvdub]

Originally posted by murpia:
Here (http://www.fiatforum.com/panda-guides/44993-front-wheel-hub-bearing-replacement.html?garpg=3) is a link to a page showing (possibly) the smallest / lightest version of these 'bolt in' bearing packages. I'd be interested to learn if any team has used such a product? Also I think SKF offer a 'racing' version called the HBU6 (http://vsmus-origin.skf.com/Files/Product_Flyers_PDFS/Hub_Unit_Broch_FINAL%20457616.pdf) but further info seems hard to come by...
Regards, Ian

do you know if this is the same as the Fiat X1-9 wheel bearing? thats what we are using this year, and they are still somewhat heavy in my opinion.

i was under the impression that even the smallest oem bearing will still be overbuilt for our application.... does anyone carry a 'bolt in style' DRACBB that is more tailored towards dsr/fsae applications?

[rjwoods77]

SKF worked on a ultra efficient bearing, seal hub setup that was advertised on the sae.org site in one of the automotive engineering magazines. They said it was designed to go into the new Caddy or something like that. Also timken is offering engineered taper roller and seal solutions on their website that alot of circle track stuff uses. I assume they go with this setup because of live rear axle spindles and "coleman"/"wide5" setup kingpin front spindles. I think the angular contact stuff is more for hubrights and variations of that design (something with a half shaft). Check it out but SKF and Timken are offering pre-engineered bearing packages for cars/race cars.

[rjwoods77]

Just looked at that skf link and i think that is what was on the sae site.

[jsmooz]

Rob - the article you are talking about was about the bearing for the STS-V. It's an asymmetrical ball bearing, ie the 2 rows of balls have different PCDs. This allows for increased stiffness, meaning less camber compliance in high g cornering.

Tapers are for muffler shop engineers....real engineers have balls. http://fsae.com/groupee_common/emoticons/icon_smile.gif

[rjwoods77]

Couple things that talk directly about bearing preload and bearing loads. No real answers but still somthign to look at.

http://www.skf.com/files/005030.pdf

http://www.skf.com/files/005031.pdf