Bicycle Mechanics - Bearing retainers. Engineering or convenience?

Previous experience with headsets has suggested that the use of bearing retainers may not be the best practice. Whilst these retainers make it easier to assemble the parts they don't allow free running of the bearings and restrict the number of balls one can include between the races, thus increasing the load (and wear) at the remaining contact points.

So, on installing my new headset I left the retainers out and fitted loose bearings. With the retainers, the parts would not move freely against each other when tried by hand - everything was binding :( The plus side was that it was possible to increase the bearing count from 16 to 24 balls in both the top and bottom races, whilst still allowing some freedom of movement.

This must be a better way... The retainers seem to do NOTHING for the mechanics of the system, except aid in rapid assembly & servicing.

Cheers,

Ed

you MUST be an engineer. If not, I suggest that you become one. :)

No doubt, reatainers are a convenience factor, especially for assembly at the factory, where reducing time to perform an operation increases the profits. Loose bearings are better for load carrying capability in that they permit the use the use of more bearings in the same space. There may also be less friction, but it is not significant. A retainer does not increase friction enough to cause binding, unless it was installed upside down or damaged. Installing retainers upside down is a common error among novice mechanics.

So, on installing my new headset I left the retainers out and fitted loose bearings.

You have hit upon a very common tried-and-true practice by experienced cyclists and mechanics. Many people yank out the retainer in order to place more balls in the bearing for the very same reasons you mentioned.

...all of the above figures. In fact I have chosen to disregard retainers for some time now - but was reminded about their problems during installation of a new headset.

T-Mar - I agree with your analysis. However the bearing/retainer combination in my headset kit was a cause friction despite correct assembly. I suspect the retainers supplied did not allow enough freedom for the balls to move independantly. Before fitting to the frame, I was familiarising myself (mucking about with...;)) the components and was posible to see the bearings skidding, rather than rolling between the raceways. At this point I really started to question the retainer's function, though I was going to omit them anyway.

Phatman: I'm not an engineer, its worse than that, I'm an experimental Physicist :)

Cheers,

Ed

Phatman: I'm not an engineer, its worse than that, I'm an experimental Physicist :)


It could be even worse, you could be theoretician.

:beer:

If the bearings with the retainers were binding, something was definitely wrong. A headset with caged bearings shouldn't spin noticably different than a headset with loose bearings. I've never been able to casually tell if any pivot used caged or free bearings without taking it apart.

Sheldon Brown sez, "Don't use retainers on bottom brackets, do use retainers on headsets." (See here (http://www.sheldonbrown.com/gloss_r.html#retainer) for the brief summary, also a longer article on bottom brackets that I'm too lazy to find)

sealed bearings are the only way to go IMO.

Orientation of the carrier is important. If there is binding after reassembly, likely the cage/ball unit has been inserted the wrong way round, and the cage is rubbing on the race. I speak from regular personal experience (darn it). I am very practised in assembling, disassembling and reassembling headsets, BBs and hubs.

sealed bearings are the only way to go IMO.sure.

I concur with Sheldon's statement that retainers speed and simplify headset assembly, but I strongly advocate using loose ball bearings, at least on the lower bearing set, to reduce spot pressure on the race and cone.

...It is odd that there should have been friction when the bearing retainer was used. I assure you that assembly was correct, incorrect orientation was obvious. The bearing retainers are very crude/ cheap items and can easily be deformed, which could contribute to problems.

I can see no harm, only potential benefit in using loose balls - it just requires more care in assembly. Sealed bearings are great, I just don't have them for this project.

Cheers,

Ed

...It is odd that there should have been friction when the bearing retainer was used. I assure you that assembly was correct, incorrect orientation was obvious. The bearing retainers are very crude/ cheap items and can easily be deformed, which could contribute to problems.

The thinness or "delicacy" of the retainer is an issue with headset bearings. Less so with BB and hub bearings. But we've concluded the additional balls are a better option.

Loose bearings are OK but you don't want there to be so many bearings that they don't all ride in their proper track. This *will* cause binding. Leave one out and reassemble; the results may surprise. The small gap between a few balls won't matter.

Without a retainer, and assuming that the ball count is max, the relative motion (and therefore friction) between the balls themselves (not between the ball and the race) is 2x the speed of the ball, at both sides of the balls! Ball cages eliminate this source of friction. FWIW, industrial recirculating ball bearing applications use alternating size bearings (large-small-large-small, with the small balls only a few thousandths smaller than the large) because it significantly reduces this friction. Adjacent balls work with each other, not against.

Many, many high precision, heavily loaded bearings are spec'd with cages - granted, they're not the cheap pieces of tin we often see. Cages allow much more room for lubricant between the balls, good for lubricant distribution, diluting contaminants, and removing heat!

IMO, more balls = more friction, up to the point that the balls and races are being significantly deformed due to excess loading. I don't think we see that extreme on bicycles, and I beleive that the friction difference between properly applied caged and uncaged bearings is neglible for us as well.

sure.

Dunno. I've just replaced a freehub whose outer cartridge bearing was so badly worn out that the cassette was wobbling.

Bike mounts Chorus, about 10'000 miles old. It's the outermost bearing on the freehub, cassette side. Campy checked the piece and replaced it at less than 30% of normal cost (v. cheap). Apparently the issue is not so unusual.

I doubt this would happen with loose bearings..

How often do you rebuild your bearings? I have had a few caged BBs wear out, but it was after much use in short period with grit and no maintenance. In the bike context I think uncaged is better for duty rating, since I have seen more than a few brinelled and pitted races from freeriders. 5 balls is a lot less contact area than 9 or 11. Thats also why I have had some cheaper sealed cartridges pit out too(though the industrial grade replacements rock). The uncaged ones, though harder to work on, seem to have a much higher tolerance for abuse.