WGB 

I have a League Fuji road bike I am rebuilding.

Previously I asked about using caged bearings in wheels and members here were against that, pointing out that with loose bearings I'd get more bearings on each hub side.

When I removed the headset for this Fuji I found a set of caged bearings - 5/32 - (front and bottom)

When I removed the bottom bracket (Sugino) I again found a set of caged bearings, (haven't measured but either 1/4 or 7/32) one for each side.

Can I switch to loose bearings for these places??

Bill Kapaun

I keep the retainer for the head set only.
BB should have 1/4" bearings. 11/side loose.

fietsbob

In the case of one piece, 'ashtabula' cranks you cannot put it together without the retainer.. I Tried, just once.


But in most bikes 3 piece cranks , sticking the balls in the grease in the cup race is OK...






....

superstring

For the headset, for sure. You should be able to get another ball or two in there without the cage(s).

AnkleWork

So called "full compliment" caged bearings have the maximum number of balls and lower friction than loose. They are also easier to work with than loose balls.

sweeks

I'll agree that caged bearings are easier to work with; that's one reason they are used in bicycle manufacture, where "time is money".

What is the source of your first statement? According to SMB Bearings (A Guide to Bearing Retainers or Bearing Cages) "A full complement (or full ball) bearing contains extra balls and has no retainer. It is used for its greater radial load capacity although axial load capacity is very small due to the risk of the balls fouling the filling slot during rotation. These bearings can only be used at low speeds and bearing torque is increased due to ball to ball friction."

These are different than the cup-and-cone bearings used in bicycles. Of course, they are operating at low speeds, but bike bearings don't have "filling slots". In bicycle bearings with the sort of cages that are commonly used, it is possible to add at least one or two additional balls and *still* have enough space between the balls that there is minimal contact force. If this were a problem, then higher-end components would always use cages. This is not usually the case.

For example, my mountain bike (low-end Shimano hub) had 8 caged balls per side in the front wheel. My road bike (Ultegra hub) has 10 uncaged balls per side in its front hub. I replaced the caged balls in the mountain bike wheel with 10 balls per side. There is plenty of space, and when the balls contact each other they do so at a single point rather than all the places a cage touches each ball.

My old 10-speed had 7 (!) caged balls on each side of the bottom bracket. I replaced them with 11 balls per side, which represents a significant improvement in load-bearing capacity.

Finally, here's what Jobst Brandt had to say about cages (Source: Ball bearings (Jobst Brandt)):

"The bearings of which we speak in bicycles are primarily cup and cone
bearings in which the cage serves merely as a convenience for
assembly. It has no function other than that. Cheap bearings, that
we needn't consider, often use cages to reduce the number of balls,
cages being cheaper than balls.

"The contact of one ball on another is not the reason for cages
although this is a problem with needle or roller bearings that have a
long contact line."

Steve

AnkleWork

NASA would be one source I've read, among many others all of which indicate that ball retainers serve substantially greater purposes than convenience. What was Jobst's source (apart from his own rather expansive sense of authority)?

" . . . bearing torque is increased due to ball to ball friction." Loose ball friction is higher than in caged bearings due to skin friction between the balls. Skin friction is proportional to the square of the surface speed. Ball-to-ball surface speed is double ball-to-cage surface speed and thus skin friction is four times as high.

JohnDThompson 

FWIW, you can get 11-ball retainers to use in your bottom bracket:

SquidPuppet

Could you provide a few links to these sources? I'd like to read more about full complement bearings.

sweeks

Citation needed (see below). I can't speak for Jobst, and can't ask him since he's dead. He's generally regarded as a credible source, though not completely without controversy.

Here are two:
https://ntrs.nasa.gov/archive/nasa/c...9830018943.pdf
https://ntrs.nasa.gov/archive/nasa/c...9810009866.pdf

The first link contains the following (p. 17):
"Ball and roller bearing separators, sometimes called
cages or retainers, are bearing components that, although
never carrying load, are capable of exerting a vital
influence on the efficiency of the bearing. In a bearing
without a separator, the rolling elements contact each
other during operation and in so doing experience severe
sliding and friction. The primary function of a separator
is to maintain the proper distance between the rolling
elements and to ensure proper load distribution and
balance within the bearing. Another function of the
separator is to maintain control of the rolling elements in
such a manner as to produce the least possible friction
through sliding contact."

I stand corrected on the friction issue. Still, there must be some reason separators are not used in many bicycle applications.

The second link is for historical interest, but since it's NASA it must be good!
Steve

fietsbob

They fall out as you put the crank arm in.

SquidPuppet

Thanks for the links. But... something doesn't add up here.

No retainer + more balls = higher material cost
No retainer + more balls = higher labor cost

A Shimano CB-E110 coaster brake hub comes with retainer bearings and can be had for $20.00

A Shimano Dura-Ace 7710 rear track hub comes with loose balls and runs about $130.00

Why is an inexpensive hub that is designed for low speed clunking around given the "superior" retainer bearing system, and the hub designed specifically for racing given the "inferior" loose ball bearing system that has greater friction?

The CB-E110 comes with grade 300 balls, while the Dura-Ace 7710 comes with grade 25 balls.

Why would Shimano spend more money on materials and labor to provide racers with hub bearing performance and service life that is inferior to a beach cruiser or child's bike?

Something doesn't make sense here.

robert schlatte

Regarding headsets, Sheldon Brown says to leave the race one or two balls short of full. Apparently, this allows the balls to move around a little bit to decrease the chance of "brinelling"- the little dimples in the race made by the balls.

For what its worth, I have always read that loose balls are better than caged balls in all bearing applications at least where bikes are concerned.

rydabent

With loose balls you can get more in the race and spread the load. IMO caged bearings are just a convenience when building on an assembly line.

79pmooney

"Uncaged" bearings aka "free range" bearings. I have a few roaming under my work bench. (The caged ones are a lot easier to look after.)

Ben

79pmooney

The "rule" from many years ago - fill the race and remove one. Headsets, two is probably a better bet.

Ben

sweeks

Grade 25 balls are higher quality than grade 300... I read that somewhere here.
Here's a reference 'cause I know you like them!
https://en.wikipedia.org/wiki/Ball_(bearing)
Steve

79pmooney

The reason we were all told that loose bearings were better than caged bearings wasn't friction, it was wear and impact (headsets) to the races. With (say) 11 balls instead of 10, that would be ~9% less weight and force transmitted by each ball, making life a little easier for the races (where we have money invested).

Before I sweat the friction from the contact of loose bearings, someone is going to have to show me real numbers on just how high this friction really is. The bearings may be touching, but there is very little force being transmitted between them. NASA being able to detect some drag there and finding it is more than the drag of the retainers isn't going to sway me. I'm guessing these are very small numbers.

Ben

Racing Dan

Without backtracking the quoted text, Im betting it only applies to cartridge bearings that cant be assembled without some space between the balls. The cage then spaces the balls out evenly in the race. In an open cup and cone bearing like shimano uses, you can use loose balls and just fill it up, mitigating the need for a ball cage.
Imo, on a bike its not the number of balls that determine the longevity of the bearing. Its the sealing. A clean bearing could last "forever" and a dirty one will rust and pit very fast.

SquidPuppet

Yeah, I know. That's why it doesn't make sense. Superior materials used in an "inferior" system.

wphamilton

I think the spacing is wider between balls than in other, high performance applications. I've also seen it claimed that the balls naturally spread themselves out in the available space, which means that sliding friction between the balls would be negligible. In that case, the primary advantage noted by NASA would not be applicable.

Personally, I see no downside to leaving the retainers out. Retainer in, I have seen (poorly maintained) retainers disintegrated with pieces in amongst the balls, which has to be degrading efficiency and possibly damaging to the races over time.

sweeks

Engineers also work with manufacturing... caged bearings are faster to install than loose balls and also cost less. In some low-cost hubs, the speed of assembly and material costs may be significant in terms of competitive pricing. This doesn't apply to home mechanics adding more balls and pitching the cages. As noted, some higher-end hubs (Ultegra in my case) have loose balls.


Steve

Ghrumpy

It is the case that caged bearings allow for fewer balls safely to be used in a given assembly, and that it can speed assembly.

Admitting that, however, I seriously doubt that Campagnolo, for one example, had either cost or speed of assembly in mind when they decided to use caged bearings in their Record bottom brackets and headsets way back in the 1950s. (It is true that their hubs have always used loose balls, as do most hubs.) And Campagnolo BBs and headsets use a full complement of bearings. Removing the retainers will not increase the number of balls you can fit in them.

So I'd submit that it doesn't have to be one reason OR another; there can be multiple reasons for retainers. The friction reason for them is, in fact, to keep them from touching each other. That's because sliding friction between lubricated steel parts is literally 100x more than rolling friction. (see https://www.tribology-abc.com/abc/cof.htm. The coefficient of sliding friction for lubricated steel is 0.15. The coefficient of rolling friction for steel balls is 0.0015) Friction reduction is the entire reason for ball bearings in the first place. Now, since the balls themselves are rolling in the same direction, the points where they touch will not just be sliding against each other, but also at twice the speed they are rolling. If the sliding friction becomes high enough, which wouldn't take much, it will stop the balls from rolling in the races momentarily, which will quickly damage the bearings and the races.

FWIW, the mangled retainers I've seen appear mostly to have been damaged by the failure of one or more balls, not the other way around. Balls are much harder than retainers, which are quite soft as steel goes. Or, just as often, they were installed the wrong way around in the assembly.

sweeks

Yet this sort of damage seems not to occur in bicycle bearings. I think it is because when the balls contact each other it is not with great pressure. Sliding friction is proportional to the "normal" force, which in a bearing is the pressure of a ball against another at its point of contact. If this pressure is low, so is the friction. Also, these are surfaces lubricated with grease that typically has pretty good film strength.
Steve

Lakerat

I'll surmise that bearing cost vs retainer cost isn't much of a factor. Both are cheap on an industrial manufacturing scale. I doubt a retainer saves money compared to adding a ball or two.

I also doubt that factory assembly convenience is much of a factor either. I assume automated systems for assembling with or without retainers is easily doable. Retainers are certainly convenient for repair during use, and may have some perception of value or preference to most users. Freewheels never seem to have retainers and apparently efficient methods of manufacturing have been utilized.

When retainers are employed by manufacturers such as low end vs high end hubs, does little to explain the benefits of either system. Weight could be a bigger factor than friction. The diameter of the bearing track seems to be more of a determinant of whether retainers are used than anything. Larger diameter tracks having retainers, and smaller ones without, excepting freewheels.

The OS system used by Campagnolo and Fulcrum, replaced common loose ball cup and cone hub designs used by Campagnolo for about 70 years. If you put any validity to the: "low end = retainers, high end without" theory, Campagnolo's redesign doesn't support it. This redesign increased the bearing track diameter and added plastic retainers.

Plastic retainers add another factor in that they may have different friction characteristics than metal retainers. They may also be more consistent in contact points with the bearing ball.

It's also possible that bearing contact without retainers is more of a problem with less precise cones and races. I can imagine balls bouncing off each other happily in a Dura Ace hub.