robertj298 

Ok I have my 1985 Sequoia's bottom bracket apart and I'm just waiting for bearings to
come in the mail. There seems to be dispute over which is best, loose or caged bearings.
The bearings that came out of the bike were caged and both cages are in good shape.
So should I use the cages or put them in loose? If I put them in loose do I add additional
bearings? The ones that came out have 11 1/4 in bearings in them.

Wileyone 

I like my balls loose.

The cages help to speed up the production process.

dddd

If it's not too difficult to get the existing cages clean, then for sure the 11-ball cage arrangement was only ever found on high-end assemblies and should be re-used.

9-ball cages were the cheaper way, and I replace these with 11 loose balls for the greater durability against fatigue of the race surfaces.

Really cheap bikes, some of them, had 7-ball cages.

79pmooney

In general, loose is considered better because you can install more and reduce the load on each bearing. The old fill the race and remove one adage of years ago still holds true and will almost always result in one or two more bearings in there then caged. (I'd do this first with the spindle and left side cup dry. Load the bearings, then push the spindle in firmly and rotate. Easy way to ensure you don't have too many. (Oh, yeah, you can skip this and just do it for real, but 1) it will be far harder to feel anything and 2) if you do need to remove one, fishing it out is a messy bear.

Edit: a little fun. Caged vs loose bearings? (You'll find those free roaming critters on any shop floor.)

2cam16

Loose for me.

bikemig 

If you have cages that take 11 1 and 1/4, then it's a no brainer. Use the cage. Loose is better than caged if the cage holds 9 or 7.

BFisher

Both. Go caged on one side and loose on the other.

Walk on the wild side.

robertj298 

Guess there is no final verdict lol

3alarmer 

...me too. I always wear boxers, not briefs.

brian3069

I have a good friend that is a mechanical engineer. He told me that 9 caged bearings is better than 11 loose bearings because the predictable position of the bearing balls reduces wear on the race. He also said that because of the light load and loose tolerances of a bicycle bottom bracket, it really doesn't matter if they're caged or loose, lubrication is the most important thing.

3alarmer 

...the one advantage of switching from caged in a headset to loose is that you can usually space the bearings differently by adding one or two. So if it's a little indexed, you can usually extend the life through some more service cycles. In a BB, any additional bearing you can get to run in the cups is another contact point, in an area that sees some heavy loading. Having said all that. if I take something apart and it seems to be working OK with the current caged bearings, I might swap out the bearings in the cages, but I don't usually switch. Either way, not that big a deal.

There's probably someone who feels strongly about it, though. They'll be along shortly.

jethin

This is one of my pet issues, probably because it’s so minor.

I generally use loose mostly out of superstition. I believe I’ve felt slightly increased friction in some caged applications, but it could’ve been lack of caffeine/beer.

Why don’t you try both assemblies and report back? I’m sure that will end this argument once and for all.

I respect and defer to engineers, but some of them will tell you anything.

wrk101

I've had several cages disintegrate and damage the spindle and cups.

Loose balls and NEW ones every time. Are you reusing? Better get a microscope to inspect the surface for wear.

Agree lubrication is most important. I rarely/never find an old bike where the bottom bracket has been maintained. People ride bikes to failure.

Cycle Tourist

If the cages have disintegrated you've got bigger problems than deciding which way to go. Take em out. If they're good, clean em, lube em and put em back. I suppose more is better and for headsets if they're notchy or hard to adjust loose balls can help.

Reynolds 

Loose for headsets, except if it's a new high end one that came with caged bearings. For BBs, 11 bearings, either caged or loose.

100bikes

IMO, it is more about the quality of the surfaces and tolerances of the bearings themselves than it is about having
a retainer around the bearings.

A secondary point is the application. A OPC(one piece crank) is a bear to load loose ball bearings into.You generally
do not see precision ground surfaces in this application.

Steel on steel is the most energy efficient mechanical set up - that is the closest to pure, near 100% energy transfer.
Think of a locomotive's steel wheels on a steel rail.

The more bearings, as long as they are not crowded into space, the more points in contact and less pressure on
each one.

With chrome or zinc finish, the surface is relatively "rough" as compared to a machined surface on on a high grade
(and generally better quality) steel race. The bearing surfaces benefits from more points in contact(ie more ball bearings).
Retainer and loose ball of the same count would perform equally.

Grade of bearing more influences the performance - you can get retainer bearings with grade 25 balls as well
as grade 300 balls(less precise and "softer" steel).Hardware store, bulk bearings are generally grade 300 and
Campagnolo bearings are generally grade 25.

You can get Campagnolo grade 25 bearings either loose or in a retainer.

A precision ground surface using lower grade 300 ball bearings(in a retainer or not) will be more prone to wear/pitting than
when using grade 25 balls.

A chrome or zinc, unfinished surface will not benefit in any significant way from using higher grade 25 bearings.

Note: the grade of the bearings reflects not only the quality and precision of their manufacture, but the close tolerances
between each bearing.

In the case of performance, you get what you pay for.
rusty

dddd

Not quite the same as in a bottom bracket, but when I replaced the tiny 1/8" ball bearings in my Huret derailer pulleys with Gr25 balls, the pulleys turned quite a bit smoother, making their correct adjustment more discernible and easy.

One more thing about the caged vs. loose argument. When the bearing turns, the balls move in and out of the loaded zone along the races, and the balls actually change shape cyclicly as they do this.
In a loose ball bearing assembly carrying significant load, the balls are free to move toward and away from each other as they pass through the unloaded zone along the races.
They will randomly come in contact and then have their positions next to each other locked in as they roll into the loaded zone of the races.
Under load, the balls then expand in the circumferential direction along the races, which forces the balls against each other .
At the same time, the contact surfaces between adjacent balls are moving in opposite directions, producing friction and heat!

Ball cages address the above concerns by allowing some flex to absorb that fluctuating interference between balls. Part of this yielding to adjacent-ball interference can be the oil layer that gets squished away as the balls expand towards each other (the cage tines providing twice as many oil layers as exist between any pair of loose balls). Also, the contact angles of the cage tines with the balls is angled well away from the axis of greatest expansion of the balls, helping further to absorb the physical ball-cage-ball interference before much contact force develops.
All part of good bearing design and why Sugino and Campagnolo always used the finely-made 11-ball retainers in their finest bottom brackets.

Lastly, when dealing with lesser bearing surface quality, it seems to me that improving the finish and roundness of the ball surfaces will certainly reduce the contact friction between the balls (where sliding contact, not rolling contact is occurring). I believe that just might be the difference I felt after replacing the balls in the Huret pulleys with Gr25 balls.

SurferRosa

They're usually found that way (caged).

It's up to you. Either is fine. I always use them loose in the bb. I think a count of 11 (caged or loose) in each cup spreads the load better than 9 caged.

No. Do not add.

But if the cages held 9 each, you would add 2 (for a total of 11) if you wanted to use them loose.

/thread