Bearings = Better Loose or Caged?

686 views
Skip to first unread message

Sir Ridesalot

unread,
Aug 22, 2009, 10:48:29 PM8/22/09
to
Hi there.

Once upon a time bicycles came with non-cartridge bearings. Some of
these bearings were loose ball bearings whilst some had the ball
bearings captured in cages to make servicing easier. What I am
thinking is that the loose ball bearings, because there would be more
of them and thus more load bearing surface than if they were caged,
would be better than the caged bearings. Is this true?

I know that an advantage to the caged bearing design was that you
could prolong the life of a bearing surface by removing the caged
bearings and inserting loose bearings.

I usually place loose bearings in all of my cup-and-cone bearing
applications when I am servicing my vintage bicycles and have the
headset, bottom bracket and/or wheel apart.

Is it worth the effort to replace caged bearings with loose ball
bearings in these areas?

Thanks from Peter

Mark J.

unread,
Aug 22, 2009, 11:10:51 PM8/22/09
to
Sir Ridesalot wrote:
> Once upon a time bicycles came with non-cartridge bearings. Some of
> these bearings were loose ball bearings whilst some had the ball
> bearings captured in cages to make servicing easier. What I am
> thinking is that the loose ball bearings, because there would be more
> of them and thus more load bearing surface than if they were caged,
> would be better than the caged bearings. Is this true?

Most cages did not hold as many balls as a loose-bearing installation
would allow, that's true, though many Campagnolo cages held a "full
complement."

> I know that an advantage to the caged bearing design was that you
> could prolong the life of a bearing surface by removing the caged
> bearings and inserting loose bearings.
>
> I usually place loose bearings in all of my cup-and-cone bearing
> applications when I am servicing my vintage bicycles and have the
> headset, bottom bracket and/or wheel apart.
>
> Is it worth the effort to replace caged bearings with loose ball
> bearings in these areas?

If you're gonna replace the bearings, the difference in effort is almost
nil, may as well use loose bearings. Will it make a significant
improvement in bearing life? Dunno, but cost is so small, value
calculations are silly.

Mark J.

someone

unread,
Aug 22, 2009, 11:15:30 PM8/22/09
to
On Aug 23, 3:48 am, Sir Ridesalot <i_am_cycle_pat...@yahoo.ca> wrote:
> Hi there.
>
> Once upon a time bicycles came with non-cartridge bearings. Some of
> these bearings were loose ball bearings whilst some had the ball
> bearings captured in cages to make servicing easier. What I am
> thinking is that the loose ball bearings, because there would be more
> of them and thus more load bearing surface than if they were caged,
> would be better than the caged bearings. Is this true?
>
No.

> I know that an advantage to the caged bearing design was that you
> could prolong the life of a bearing surface by removing the caged
> bearings and inserting loose bearings.

Only by renewing lubricant.


>
> I usually place loose bearings in all of my cup-and-cone bearing
> applications when I am servicing my vintage bicycles and have the
> headset, bottom bracket and/or wheel apart.
>
> Is it worth the effort to replace caged bearings with loose ball
> bearings in these areas?

No. The space between the balls means there is a good reservoir of
lubricant, most important if using grease, it also prevents ball to
ball contact which ovalises the balls and can result in their early
demise due to fatigue caused by continually running on the same axis.
The design of the cup and cone cycle bearing encourages spin on the
balls which enhances their life expectancy. If lubrication is kept up
(and the bearing tracks remain smooth) you should never need to
replace the balls in a quality assembly Well perhaps if you ride
20,000 miles a year for 20 years it might be a good time to replace
the balls.

Sir Ridesalot

unread,
Aug 23, 2009, 1:02:25 AM8/23/09
to

Hi there.

Thanks for your input.

I for one really do appreciate the advice and help that so many of you
on these groups provide.

Cheers from Peter

notme

unread,
Aug 23, 2009, 1:33:52 AM8/23/09
to
> No. The space between the balls means there is a good reservoir of
> lubricant, most important if using grease, it also prevents ball to
> ball contact which ovalises the balls and can result in their early
> demise due to fatigue caused by continually running on the same axis.
> The design of the cup and cone cycle bearing encourages spin on the
> balls which enhances their life expectancy.

This doesn't sound right. More balls = less load for each. This can only be a
good thing.

Good lubricant will result in minimal effect between the balls.

"ovalize"? Can you cite some references? I'd be interested.

Thanks.


Jobst Brandt

unread,
Aug 23, 2009, 2:21:30 AM8/23/09
to
a shy person wrote:

> No.

> Only by renewing lubricant.

Cages are used to make assembly easier. Typically gaged 9-0.25dia
ball complements in cages in hubs are as close together as without a
cage and similarly 11-0.25 dia ball complements of BB's are also close
packed. Leaving out the cage is for weight weenies and believers in
better rolling. Bearing balls spin individually in non parallel
rotation that does not cause wear, the inter-ball contact being
light on an oil film that is greater than between ball and race.

I suppose it should be apparent that bearing balls running in curved
races skate backward in the center of the contact ellipse and forward
at the edges. Diagrams of this are found in bearing analyses of which
"Rolling Bearings Analysis" by Tedric Harris (SKF) is the best
reference:

http://tinyurl.com/y38had
http://tinyurl.com/nnc2xn

Cages are used to use fewer bearing balls where there are more than
enough balls for the load, typically in RECORD Campagnolo head
bearings where 22 bearing balls would fit but with a cage, 20 can be
equally spaced. Inexpensive head bearings have used curved wire
spacers to reduce the number of bearing balls to 16 without cages that
cost more than bent wire and menial labor to install them.

Jobst Brandt

Chalo

unread,
Aug 23, 2009, 2:39:44 AM8/23/09
to
Sir Ridesalot wrote:
>
> Once upon a time bicycles came with non-cartridge bearings. Some of
> these bearings were loose ball bearings whilst some had the ball
> bearings captured in cages to make servicing easier. What I am
> thinking is that the loose ball bearings, because there would be more
> of them and thus more load bearing surface than if they were caged,
> would be better than the caged bearings. Is this true?

More balls can support higher loads. But full-complement bearing
balls rub against the faces of loaded rolling balls rather than
relatively stationary unloaded cages, and thus they incur higher
bearing drag. You can feel the difference even while turning the
races of unmounted full-complement bearing cartridges versus Conrad
type caged ball bearing cartridges.

> Is it worth the effort to replace caged bearings with loose ball
> bearings in these areas?

It doesn't hurt. It probably doesn't help, either.

Chalo

Ben C

unread,
Aug 23, 2009, 4:17:41 AM8/23/09
to

Putting loose balls in a headset that's starting to index a bit can
solve the indexing problem.

The other problem with a cage is that sometimes when things go
pear-shaped the cage breaks up and gets dragged around scouring out the
races and things.

But I don't go around taking cages out of working stuff (if it ain't
broke).

someone

unread,
Aug 23, 2009, 8:39:44 AM8/23/09
to
On 23 Aug, 06:33, notme <no...@notme.org> wrote:
> > No.  The space between the balls means there is a good reservoir of
> > lubricant, most important if using grease, it also prevents ball to
> > ball contact which ovalises the balls and can result in their early
> > demise due to fatigue caused by continually running on the same axis.
> > The design of the cup and cone cycle bearing encourages spin on the
> > balls which enhances their life expectancy.  
>
> This doesn't sound right. More balls = less load for each. This can only be a
> good thing.

Separating the balls with a spacer encourages spin which evens out the
wear and and loading axiii. The larger grease reservoir close to the
balls also permits a dumping ground for wear particles.

> .
> Good lubricant will result in minimal effect between the balls.
>
> "ovalize"? Can you cite some references? I'd be interested.
>
> Thanks.

http://www.colorsofthecity.org/allfiles2/645_Ewec2006fullpaper.pdf

From a google search.

someone

unread,
Aug 23, 2009, 9:13:46 AM8/23/09
to

one aspect.

Typically gaged 9-0.25dia
> ball complements in cages in hubs are as close together as without a
> cage and similarly 11-0.25 dia ball complements of BB's are also close
> packed.

got that.

 Leaving out the cage is for weight weenies and believers in
> better rolling.

I suppose those bearing cages can be weighed. To reduce rolling
resistance in a ball bearing assembly it must be kept clean and
lubricated. This will maintain its low rolling resistance. Lubricant
is required to enable the low resistance sliding of the ball at the
edges of its contact area with the tracks, continual wetting of the
ball is essential to prevent drag and wear when the wheel is used.
'Better rolling' is best effected (in life terms) by better
lubrication which is to use ball spacing cages when using a grease.
For instantaneous lower resistance grease should be replaced with
oil. Oils benefit , that is gradually seeps out of the bearing taking
away wear debris is also its downfall in the hands of the incompetent
home mechanic who fails to replenish the lubricant at the regular
intervals required. Grease which has been slackened by the addition
of supplementary oil seem most satisfactory for bicycle bearings. It
results in excellent lubricant supply to the bearing surfaces and yet
is still mostly retained within the bearing housing. Wear seems
minimal despite extended intervals between oilings.

 Bearing balls spin individually in non parallel
> rotation that does not cause wear, the inter-ball contact being
> light on an oil film that is greater than between ball and race.


Without a spacer the balls are pushed tightly together and the weight
is carried on the two lowest balls only, which means early fatigue
failure of the cones bearing track or balls

>
> I suppose it should be apparent that bearing balls running in curved
> races skate backward in the center of the contact ellipse and forward
> at the edges.  Diagrams of this are found in bearing analyses of which
> "Rolling Bearings Analysis" by Tedric Harris (SKF) is the best
> reference:
>
>  http://tinyurl.com/y38had
>  http://tinyurl.com/nnc2xn
>
> Cages are used to use fewer bearing balls where there are more than
> enough balls for the load, typically in RECORD Campagnolo head
> bearings where 22 bearing balls would fit but with a cage, 20 can be
> equally spaced.  Inexpensive head bearings have used curved wire
> spacers to reduce the number of bearing balls to 16 without cages that
> cost more than bent wire and menial labor to install them.
>


A space keeps the grease reservoir intact, pumping occurs of this
reservoir as the relative position of the balls change. Without the
cage the grease is pushed away from the balls to outside their track,
only positive pressure from outside the bearing track can keep the
balls adequeately lubricated. This why it is impoortant to use a
fluid lubricant when there is no cage installed.

Message has been deleted

someone

unread,
Aug 23, 2009, 11:35:36 AM8/23/09
to
On 23 Aug, 16:15, "Still Just Me -<"
<stillnoEmail...@stillnodomainer.com> wrote:
> On Sun, 23 Aug 2009 06:13:46 -0700 (PDT), someone

>
> <thirty-...@live.co.uk> wrote:
> >Without a spacer the balls are pushed tightly together and the weight
> >is carried on the two lowest balls only, which means early fatigue
> >failure of the cones bearing  track or balls
>
> You'd really need to cite some references on something like that.
> Sounds like serious conjecture to me.

Early failure of hub cones is commonplace when a wheel is used without
refreshing a grease lubricant. By early I mean less than 20 000
miles. The cones pit at the lowest point. Periodic rotation of the
axle will maximise its fatigue life.

>
> I always preferred loose balls - mostly because the only ones that
> came caged in the early days were those in cheap components. Serious
> hardware always came with loose bearings. That changed in later years
> as even quality component manufacturers wanted to save money.

The cage made up for the bearing's other deficiencies such as soft
cones and aspherical balls. Quality manufacturers likely thought they
would attract more ball sales if they were caged for a specific duty
along with special bicycle value.
>

> That said, a higher number of bearings will spread the load among more
> bearings, resulting in a lighter load per bearing. Wear would be lower
> per bearing.

cramming in an extra one or two balls gains no advantage.
>
> The only caveat would be if the bearings were such that they'd never
> wear under the loads applied in use - I don't think that's true with
> headset bearings - they take a lot of shock for a bearing. More is
> better.  

False. Brunel took six months to launch the Great Eastern because he
missed the spring tide.

Dan O

unread,
Aug 23, 2009, 12:05:17 PM8/23/09
to
On Aug 23, 1:17 am, Ben C <spams...@spam.eggs> wrote:

I just had a (Shimano XT M770) rear hub apart and discovered a caged
bearing on the non-drive side behind a seal that was pressed into the
hub. Loose balls I could have fished out and more easily cleaned and
overhauled the hub without disturbing the seal. (The other side -
inside the freehub - has loose bearings behind a seal that Shimano
explicitly says should not be "disassembled".)

I guess I'm going to investigate my options for either replacing the
caged bearings with loose, or routinely removing and replacing the
seal (which I guess I'll have to do at least one for the first option
anyway) for hub and wheel bearing service. (it looks from the service
diagram like the M770 front has caged bearings behind a seal on both
sides.)

I guess you use some kind of puller to get such a seal out, and then
like a socket and hammer to put it back in (or the new replacement if
you munged it).

_

unread,
Aug 23, 2009, 12:05:47 PM8/23/09
to
On Sun, 23 Aug 2009 11:15:05 -0400, Still Just Me -< wrote:

> On Sun, 23 Aug 2009 06:13:46 -0700 (PDT), someone

> <thirt...@live.co.uk> wrote:
>
>>Without a spacer the balls are pushed tightly together and the weight
>>is carried on the two lowest balls only, which means early fatigue
>>failure of the cones bearing track or balls
>

> You'd really need to cite some references on something like that.
> Sounds like serious conjecture to me.
>

Not serious, laughable.

The spacer makes no difference to the clearance between the balls and the
races. Any assymetry of load due to excessive clearance (due to variations
in ball size) will occur regardless of spacers; and of course, all bearings
where the use of spacers is optional are adjustable for clearance.

This claim of spacers preventing failure by such a method is just as
reliable as the assertion that brake bolts are "hardened chromoly".

someone

unread,
Aug 23, 2009, 12:15:09 PM8/23/09
to
On 23 Aug, 17:05, _ <jtayNOSPAM...@hfDONTSENDMESPAMx.andara.com>

Try to comprehend before responding, pillock.

someone

unread,
Aug 23, 2009, 12:23:52 PM8/23/09
to

Some sheilds can be tight and removing them by force will distort
them. I'd be tempted to drill an oiling port in it as it stands.
Flush everything out using a pipe brush (bottle brush or tooth brush
etc if you want) and kerosene. Oil it through the oiling port in
future using a thick oil from a squirt can.

_

unread,
Aug 23, 2009, 1:07:26 PM8/23/09
to

Insults are a sign that the user has no other support for his claims.

Jobst Brandt

unread,
Aug 23, 2009, 1:35:41 PM8/23/09
to
Chalo Colina wrote:

Full complement ball bearings, as I mentioned, can also have cages,
whose purpose is ease and reliability of assembly rather than to alter
friction or wear characteristics. Full complement (no empty ball
spaces) bearings are not press fit bearing ball to bearing ball but
rather an arrangement where the race circumferences are designed to
accept a number of balls with small spacing that is not accidentally
just enough to have a reasonable oil film between them.

Two hardened steel spheres, even when touching, have a no-load point
contact in such a bearing. The reason for ball or roller cages in
bearings is for assembly rather than a load related one except in ones
where the cage allows using fewer balls for economic effect.

Jobst Brandt

Jobst Brandt

unread,
Aug 23, 2009, 1:45:46 PM8/23/09
to
Ben C? (in umbra) wrote:

>>> Once upon a time bicycles came with non-cartridge bearings. Some
>>> of these bearings were loose ball bearings whilst some had the
>>> ball bearings captured in cages to make servicing easier. What I
>>> am thinking is that the loose ball bearings, because there would
>>> be more of them and thus more load bearing surface than if they
>>> were caged, would be better than the caged bearings. Is this
>>> true?

>> More balls can support higher loads. But full-complement bearing
>> balls rub against the faces of loaded rolling balls rather than
>> relatively stationary unloaded cages, and thus they incur higher
>> bearing drag. You can feel the difference even while turning the
>> races of unmounted full-complement bearing cartridges versus Conrad
>> type caged ball bearing cartridges.

>>> Is it worth the effort to replace caged bearings with loose ball
>>> bearings in these areas?

>> It doesn't hurt. It probably doesn't help, either.

> Putting loose balls in a headset that's starting to index a bit can
> solve the indexing problem.

I take it you didn't try this in the day when dimpled head bearings
were common and removing the cage and adding two balls produced the
same effect. Fretting, once microscopically started continues to work
the bearing balls in the same place, the bearing being a classic
steering environment where the home position of straight ahead is held
for the life of the bearing. For this reason, automobile "king pin"
bearings are bushings rather than rolling element bearings. Steering
gears used to wear a notch in the straight ahead position giving the
steering wheel a certain dead space in the middle position.

> The other problem with a cage is that sometimes when things go
> pear-shaped the cage breaks up and gets dragged around scouring out
> the races and things.

What causes things to go "pear-shaped" and what is the result.

> But I don't go around taking cages out of working stuff (if it ain't
> broke).

Why not if you believe it has benefit?

Jobst Brandt

Jobst Brandt

unread,
Aug 23, 2009, 1:54:38 PM8/23/09
to
SJM secretively snipes:

>> Without a spacer the balls are pushed tightly together and the
>> weight is carried on the two lowest balls only, which means early
>> fatigue failure of the cones bearing track or balls

> You'd really need to cite some references on something like that.


> Sounds like serious conjecture to me.

> I always preferred loose balls - mostly because the only ones that


> came caged in the early days were those in cheap components.
> Serious hardware always came with loose bearings. That changed in
> later years as even quality component manufacturers wanted to save
> money.

> That said, a higher number of bearings will spread the load among


> more bearings, resulting in a lighter load per bearing. Wear would
> be lower per bearing.

The only bearings that need more bearing balls to support the load are
ones in which the contact surfaces spall and of these I don't find any
in a bicycle except in wheel bearings when a so called perfectly
adjusted hub gets its bearings hugely overloaded by closing the QR
tightly. More bearing balls is not the proper solution to this.

> The only caveat would be if the bearings were such that they'd never
> wear under the loads applied in use - I don't think that's true with
> headset bearings - they take a lot of shock for a bearing. More is
> better.

"You'd really need to cite some references on something like that.


Sounds like serious conjecture to me."

Show me a non round bearing ball that does not have part of its
surface spalled.

Jobst Brandt

Jobst Brandt

unread,
Aug 23, 2009, 2:10:33 PM8/23/09
to
Dan Overman wrote:

I think you'll find that assembling that hub without the cage will be
difficult and can lead to a single ball can slip out of its place and
jam the assembly. Such an assembly must often be done with stiff
grease that will hole bearing balls in place while the parts are
assembled. However, freewheels with large-heel-radius ratchet pawls
should not be assembled with grease. That can make them rebound too
slowly making poor engagement while trying to engage a grease filled
ratchet ramp.

> I guess you use some kind of puller to get such a seal out, and then
> like a socket and hammer to put it back in (or the new replacement if
> you munged it).

Removal may be destructive. I'm not clear on why you want to remove
these parts rather than rinse them and re-lubricate. There should be
no grease in there so cleaning and oiling is easy. Use 10-20W oil
unless you are operating in sub freezing weather where 10W oil is
best. Don't use spray lubes (aka solvents).

Jobst Brandt

Message has been deleted

Ben C

unread,
Aug 23, 2009, 3:24:14 PM8/23/09
to
On 2009-08-23, Jobst Brandt <jbr...@sonic.net> wrote:
> Ben C? (in umbra) wrote:
[...]

>> Putting loose balls in a headset that's starting to index a bit can
>> solve the indexing problem.
>
> I take it you didn't try this in the day when dimpled head bearings
> were common and removing the cage and adding two balls produced the
> same effect.

Well, I had some dimpled bearings, took out the cage and put lots of
balls in without one, and it stopped indexing and has been fine now for
about 3 years and quite a few thousand km.

[...]


>> The other problem with a cage is that sometimes when things go
>> pear-shaped the cage breaks up and gets dragged around scouring out
>> the races and things.
>
> What causes things to go "pear-shaped" and what is the result.

My guess is the whole thing being far too loose (due to neglected
maintenance).

("Pear-shaped" is just a British expression for when things have gone
wrong somehow. Nothing in the bearings is actually shaped like a pear.)

>> But I don't go around taking cages out of working stuff (if it ain't
>> broke).
>
> Why not if you believe it has benefit?

In case I take it all apart and put it back together wrong.

DaveC

unread,
Aug 23, 2009, 3:28:00 PM8/23/09
to
> ("Pear-shaped" is just a British expression for when things have gone
> wrong somehow. Nothing in the bearings is actually shaped like a pear.)

On the Yank side of the Pond we say things "have gone south" (with due
respect to those living South of the Mason-Dixon line...) or "sideways".

Michael Press

unread,
Aug 23, 2009, 3:36:52 PM8/23/09
to
In article
<d1196f5d-b6fb-4bc4...@w6g2000yqw.googlegroups.com>,
Sir Ridesalot <i_am_cyc...@yahoo.ca> wrote:

> Hi there.
>
> Once upon a time bicycles came with non-cartridge bearings. Some of
> these bearings were loose ball bearings whilst some had the ball
> bearings captured in cages to make servicing easier. What I am
> thinking is that the loose ball bearings, because there would be more
> of them and thus more load bearing surface than if they were caged,
> would be better than the caged bearings. Is this true?

Yes.

> I know that an advantage to the caged bearing design was that you
> could prolong the life of a bearing surface by removing the caged
> bearings and inserting loose bearings.

I do not understand what you say to be an advantage of cages.
Cages are used by manufacturers to reduce assembly cost.

> I usually place loose bearings in all of my cup-and-cone bearing
> applications when I am servicing my vintage bicycles and have the
> headset, bottom bracket and/or wheel apart.
>
> Is it worth the effort to replace caged bearings with loose ball
> bearings in these areas?

It is for me. Do you see an advantage in caged
bearing balls for yourself?

--
Michael Press

Dan O

unread,
Aug 23, 2009, 5:03:20 PM8/23/09
to

Yep.

> I'm not clear on why you want to remove
> these parts rather than rinse them and re-lubricate. There should be
> no grease in there so cleaning and oiling is easy. Use 10-20W oil
> unless you are operating in sub freezing weather where 10W oil is
> best. Don't use spray lubes (aka solvents).
>

The benefit of your knowledge and experience is sincerely
appreciated. Maybe I wasn't clear, though - I'm talking about the
wheel (axle) bearings. I would prefer to be able to remove the
bearing balls when servicing - in order to more thoroughly and easily
clean and inspect the bearings and races.

http://techdocs.shimano.com/media/techdocs/content/cycle/SI/Deore-XT/SI_3CZ0A_002/SI_3CZ0A_002_EN_v1_m56577569830621150.PDF

Chalo

unread,
Aug 23, 2009, 5:09:25 PM8/23/09
to
Michael Press wrote:

>
>  Sir Ridesalot wrote:
> >
> > Is it worth the effort to replace caged bearings with loose ball
> > bearings in these areas?
>
> It is for me. Do you see an advantage in caged
> bearing balls for yourself?

That's spoken like a man who's never lain on the floor fishing cat
hair and dead bugs out from underneath the fridge in the pursuit of
that stray ball bearing.

Chalo

Tom Sherman °_°

unread,
Aug 23, 2009, 5:14:12 PM8/23/09
to
Not that a cat would ever help the bearings to go astray!

--
Tom Sherman - 42.435731,-83.985007
Celebrity culture is an opposite of community, informing us
that these few nonsense-heads matter but that the rest of
us do not. - Jay Griffiths

Dan O

unread,
Aug 23, 2009, 5:23:23 PM8/23/09
to
On Aug 23, 2:14 pm, Tom Sherman °_°

<twshermanREM...@THISsouthslope.net> wrote:
> Chalo Colina wrote:
> > Michael Press wrote:
> >> Sir Ridesalot wrote:
> >>> Is it worth the effort to replace caged bearings with loose ball
> >>> bearings in these areas?
> >> It is for me. Do you see an advantage in caged
> >> bearing balls for yourself?
>
> > That's spoken like a man who's never lain on the floor fishing cat
> > hair and dead bugs out from underneath the fridge in the pursuit of
> > that stray ball bearing.
>
> Not that a cat would ever help the bearings to go astray!
>

I was laying on my back underneath a van, working on a roadside heater
core bypass for a broken heater hose. I had parts laying on a rag so
as not to get lost in the gravel. A car pulls up, the door opens, and
a big dog comes bounding out. Do you suppose he goes about whatever
it is they stopped for? No, he makes a beeline for the rag that my
parts were on, snatches it up, and runs away.

AMuzi

unread,
Aug 23, 2009, 5:53:36 PM8/23/09
to
Sir Ridesalot wrote:
> Hi there.

>
> Once upon a time bicycles came with non-cartridge bearings. Some of
> these bearings were loose ball bearings whilst some had the ball
> bearings captured in cages to make servicing easier. What I am
> thinking is that the loose ball bearings, because there would be more
> of them and thus more load bearing surface than if they were caged,
> would be better than the caged bearings. Is this true?
>
> I know that an advantage to the caged bearing design was that you
> could prolong the life of a bearing surface by removing the caged
> bearings and inserting loose bearings.
>
> I usually place loose bearings in all of my cup-and-cone bearing
> applications when I am servicing my vintage bicycles and have the
> headset, bottom bracket and/or wheel apart.
>
> Is it worth the effort to replace caged bearings with loose ball
> bearings in these areas?

"loose ball bearings, because there would be more of them"

Not necessarily true. Yes, bearing retainers were invented
in Chicago during WWII to use a smaller count and speed
assembly. In many bicycle applications such as crank
bearings, there are/were indeed 5, 7 and 9 ball retainers
intended to save balls with a smaller count.

Campagnolo developed the 11 ball retainer shape 2
generations ago and that design has passed out of patent
protection. Now, premium quality Japanese (Tange, Sugino)
11-ball retainers are readily available at a low price.
There's no difference in use from loose balls and assembly
is much quicker.

In most headsets the same is true although headset design
varies a lot. Usually you will use the same count loose or
in retainers for many relatively modern headset.

Traditional design hubs do not lend themselves to retainers
due to lack of space so most quality systems (current
Shimano) are loose, Campagnolo "big bearing" hubs such as
current Record use a retainer. The very bottom of the heap
(steel Chinese hubs) have 5 or 7 ball retainers so a small
improvement may be achieved, theoretically, with loose balls.

--
Andrew Muzi
<www.yellowjersey.org/>
Open every day since 1 April, 1971

someone

unread,
Aug 23, 2009, 6:09:25 PM8/23/09
to
> http://techdocs.shimano.com/media/techdocs/content/cycle/SI/Deore-XT/...

Keep them wet with oil, visually inspect when you trash the rim.
Turning the axle in your fingers with correct cone adjustment is
adequate in the meantime. Listen with a stethescope if you have not
got safecracker fingers.

Michael Press

unread,
Aug 23, 2009, 6:15:15 PM8/23/09
to
In article
<f22838f2-d5d1-4f1f...@c2g2000yqi.googlegroups.com>,
Chalo <chalo....@gmail.com> wrote:

Too true. My jar of bearing balls has more than I need
at any one time, plus the balls and jar interior are
coated with grease.

--
Michael Press

someone

unread,
Aug 23, 2009, 6:18:58 PM8/23/09
to
FFS the axle load parts the balls at the bottom pushing the balls all
together around the rest of the bearing. Use some pennies inside a
pastry cutter and use a larger coin to act as the con/axle and apply
the radial load, the pennies spread in the load direction and are
pushed tight together around the ring and against it.

someone

unread,
Aug 23, 2009, 6:23:36 PM8/23/09
to

So there really is too cheap a hub.

Jobst Brandt

unread,
Aug 23, 2009, 7:36:27 PM8/23/09
to
Dan Overman wrote:

>>>>> Is it worth the effort to replace caged bearings with loose ball
>>>>> bearings in these areas?

>>>> It is for me. Do you see an advantage in caged bearing balls for
>>>> yourself?

>>> That's spoken like a man who's never lain on the floor fishing cat
>>> hair and dead bugs out from underneath the fridge in the pursuit
>>> of that stray ball bearing.

>> Not that a cat would ever help the bearings to go astray!

> I was laying on my back underneath a van, working on a roadside
> heater core bypass for a broken heater hose. I had parts laying on
> a rag so as not to get lost in the gravel. A car pulls up, the door
> opens, and a big dog comes bounding out. Do you suppose he goes
> about whatever it is they stopped for? No, he makes a beeline for
> the rag that my parts were on, snatches it up, and runs away.

What sort of eggs were you "laying" and what were the parts "laying" I
can imagine lying under a car with parts lying all around, but what
you describe defies sense.

Jobst Brandt

Dan O

unread,
Aug 23, 2009, 7:58:39 PM8/23/09
to

"Common" sense - get some.

jim beam

unread,
Aug 24, 2009, 1:38:45 PM8/24/09
to
AMuzi wrote:
> Sir Ridesalot wrote:
>> Hi there.
>>
>> Once upon a time bicycles came with non-cartridge bearings. Some of
>> these bearings were loose ball bearings whilst some had the ball
>> bearings captured in cages to make servicing easier. What I am
>> thinking is that the loose ball bearings, because there would be more
>> of them and thus more load bearing surface than if they were caged,
>> would be better than the caged bearings. Is this true?
>>
>> I know that an advantage to the caged bearing design was that you
>> could prolong the life of a bearing surface by removing the caged
>> bearings and inserting loose bearings.
>>
>> I usually place loose bearings in all of my cup-and-cone bearing
>> applications when I am servicing my vintage bicycles and have the
>> headset, bottom bracket and/or wheel apart.
>>
>> Is it worth the effort to replace caged bearings with loose ball
>> bearings in these areas?
>
> "loose ball bearings, because there would be more of them"
>
> Not necessarily true. Yes, bearing retainers were invented in Chicago
> during WWII to use a smaller count and speed assembly.

mmm, well, not that simple. you can't assemble a deep groove cartridge
bearing with a full compliment - you have to use a reduced count. then,
to ensure the inner race stays centered with the outer, you have to
evenly distribute them. the only way to keep them distributed is with a
cage.

so, while cages do indeed convenience assembly and reduce count for
cheap bike-type cup and cone bearings, their origin derives from the
necessities of captivation in cartridges.

http://www.gizmology.net/bearings.htm - scroll to the bottom.

jim beam

unread,
Aug 24, 2009, 1:46:31 PM8/24/09
to

er, modern "king pin" vehicle ball joints are used because

1. they don't need micron precision.

2. they have a much wider range of motion available.

3. they're cheap.

and you don't even know what a journal bearing /is/ based on some of
your previous utterings on this subject.


> Steering
> gears used to wear a notch in the straight ahead position giving the
> steering wheel a certain dead space in the middle position.

1. bearings both true brinell and false brinell jobst. you appear to be
unaware of the former.

2. false brinelling has been discovered to be primarily a function of
oxidation, not "fretting".

you really do need to crack open a book occasionally. and quit
bullshitting your same old spew.


>
>> The other problem with a cage is that sometimes when things go
>> pear-shaped the cage breaks up and gets dragged around scouring out
>> the races and things.
>
> What causes things to go "pear-shaped" and what is the result.
>
>> But I don't go around taking cages out of working stuff (if it ain't
>> broke).
>
> Why not if you believe it has benefit?
>

why don't you open a book? do you not believe doing so would have benefit?

jim beam

unread,
Aug 24, 2009, 1:48:36 PM8/24/09
to
Ben C wrote:
> On 2009-08-23, Jobst Brandt <jbr...@sonic.net> wrote:
>> Ben C? (in umbra) wrote:
> [...]
>>> Putting loose balls in a headset that's starting to index a bit can
>>> solve the indexing problem.
>> I take it you didn't try this in the day when dimpled head bearings
>> were common and removing the cage and adding two balls produced the
>> same effect.
>
> Well, I had some dimpled bearings, took out the cage and put lots of
> balls in without one, and it stopped indexing and has been fine now for
> about 3 years and quite a few thousand km.

classic true brinelling.

modern boat bearing grease can be effective in mitigating against false
brinelling - it can reduce oxygen activity.

jim beam

unread,
Aug 24, 2009, 1:50:10 PM8/24/09
to

defies sense? like the "engineer" that defies elasticity in brake
cables by failure to observe or do math?

jim beam

unread,
Aug 24, 2009, 2:08:16 PM8/24/09
to
Sir Ridesalot wrote:
> Hi there.
>
> Once upon a time bicycles came with non-cartridge bearings. Some of
> these bearings were loose ball bearings whilst some had the ball
> bearings captured in cages to make servicing easier. What I am
> thinking is that the loose ball bearings, because there would be more
> of them and thus more load bearing surface than if they were caged,
> would be better than the caged bearings. Is this true?
>
> I know that an advantage to the caged bearing design was that you
> could prolong the life of a bearing surface by removing the caged
> bearings and inserting loose bearings.
>
> I usually place loose bearings in all of my cup-and-cone bearing
> applications when I am servicing my vintage bicycles and have the
> headset, bottom bracket and/or wheel apart.
>
> Is it worth the effort to replace caged bearings with loose ball
> bearings in these areas?
>
> Thanks from Peter

higher ball bearing count means higher load capacity. thus, loose is
good for things like head sets. use of a bearing with proper cup and
cone hardness/precision and appropriate lubricants will however have
more benefit.

Ben C

unread,
Aug 24, 2009, 2:43:54 PM8/24/09
to
On 2009-08-24, jim beam <m...@privacy.net> wrote:
> Ben C wrote:
>> On 2009-08-23, Jobst Brandt <jbr...@sonic.net> wrote:
>>> Ben C? (in umbra) wrote:
>> [...]
>>>> Putting loose balls in a headset that's starting to index a bit can
>>>> solve the indexing problem.
>>> I take it you didn't try this in the day when dimpled head bearings
>>> were common and removing the cage and adding two balls produced the
>>> same effect.
>>
>> Well, I had some dimpled bearings, took out the cage and put lots of
>> balls in without one, and it stopped indexing and has been fine now for
>> about 3 years and quite a few thousand km.
>
> classic true brinelling.
>
> modern boat bearing grease can be effective in mitigating against false
> brinelling - it can reduce oxygen activity.

Actually I finally got the headset replaced on one of my other bikes the
other day because the threads in the top cup had gone south.

Long story short, but I have the old one, and it's a bit dimpled (and
was indexing a bit) on the original caged bearings. What should I look
for with my magnifying glass to determine whether it was true or false
brinelling?

jim beam

unread,
Aug 24, 2009, 2:49:58 PM8/24/09
to

with visual microscopy, it's very hard if not impossible to tell. with
electron, you might be able to find some sub-surface plastic deformation
under true brinell. i'd speculate about finding evidence of oxygenation
with false, but that's after my time at the screen, so i don't know for
sure.

Jobst Brandt

unread,
Aug 24, 2009, 10:46:05 PM8/24/09
to
Ben C? who wrote:

>>>>> Putting loose balls in a headset that's starting to index a bit
>>>>> can solve the indexing problem.

>>>> I take it you didn't try this in the day when dimpled head
>>>> bearings were common and removing the cage and adding two balls
>>>> produced the same effect.

>>> Well, I had some dimpled bearings, took out the cage and put lots
>>> of balls in without one, and it stopped indexing and has been fine
>>> now for about 3 years and quite a few thousand km.

>> classic true brinelling.

>> modern boat bearing grease can be effective in mitigating against
>> false brinelling - it can reduce oxygen activity.

> Actually I finally got the headset replaced on one of my other bikes
> the other day because the threads in the top cup had gone south.

> Long story short, but I have the old one, and it's a bit dimpled
> (and was indexing a bit) on the original caged bearings. What should
> I look for with my magnifying glass to determine whether it was true

> or false Brinelling?

See whether the dimples are mainly in the fore and aft quadrants of
the bearing races. This shows that the damage is by fretting as the
fork crown rocks fore and aft as the steer tube bends. This also
causes dimples in the upper bearing where no impact, even through a
rubber tire can land, shows that it is a fretting effect from a
bearing rocking out of its normal rotational plane.

Brinell indentations are as shiny as the bearing ball that made them.
Fretting dimples are matte from millions of tiny weld breaks. Fork
vibration makes bearing balls penetrate the oil film that otherwise
separates balls from bearing races. How ball bearings roll in use is
described in "Rolling Aliment Bearings":

http://tinyurl.com/y38had
http://tinyurl.com/nnc2xn

Jobst Brandt

jim beam

unread,
Aug 24, 2009, 11:09:26 PM8/24/09
to
Jobst Brandt wrote:
> Ben C? who wrote:
>
>>>>>> Putting loose balls in a headset that's starting to index a bit
>>>>>> can solve the indexing problem.
>
>>>>> I take it you didn't try this in the day when dimpled head
>>>>> bearings were common and removing the cage and adding two balls
>>>>> produced the same effect.
>
>>>> Well, I had some dimpled bearings, took out the cage and put lots
>>>> of balls in without one, and it stopped indexing and has been fine
>>>> now for about 3 years and quite a few thousand km.
>
>>> classic true brinelling.
>
>>> modern boat bearing grease can be effective in mitigating against
>>> false brinelling - it can reduce oxygen activity.
>
>> Actually I finally got the headset replaced on one of my other bikes
>> the other day because the threads in the top cup had gone south.
>
>> Long story short, but I have the old one, and it's a bit dimpled
>> (and was indexing a bit) on the original caged bearings. What should
>> I look for with my magnifying glass to determine whether it was true
>> or false Brinelling?
>
> See whether the dimples are mainly in the fore and aft quadrants of
> the bearing races. This shows that the damage is by fretting as the
> fork crown rocks fore and aft as the steer tube bends.

no, exactly the same thing exists with true brinelling!!! jeepers
jobst, that is /real/ basic!


> This also
> causes dimples in the upper bearing where no impact, even through a
> rubber tire can land, shows that it is a fretting effect from a
> bearing rocking out of its normal rotational plane.

er, see above for loading caused by tilting...


>
> Brinell indentations are as shiny as the bearing ball that made them.

not if they've been used! apparently you've never understood any
bearing failure you've ever looked at.


> Fretting dimples are matte from millions of tiny weld breaks.

false brinelling is an oxidation process jobst. go to the library.


> Fork
> vibration makes bearing balls penetrate the oil film that otherwise
> separates balls from bearing races.

which doesn't mean a thing. most bike bearings operate in "full
contact" mode all the time since they're not rotating fast enough for
elasto-hydrodynamic separation. you /do/ know that don't you jobst?


> How ball bearings roll in use is
> described in "Rolling Aliment Bearings":
>
> http://tinyurl.com/y38had
> http://tinyurl.com/nnc2xn

jobst, seriously, why don't you actually /read/ that freakin' book???
you cite it frequently, but apparently you're incapable of actually
cracking it open or you wouldn't be so ridiculously, er, "confused".

Ben C

unread,
Aug 25, 2009, 3:37:47 AM8/25/09
to

The crown race has shiny dimples all the way around. So does the
corresponding cup (from the bottom end of the head tube). So maybe that
was a true brinnel.

The top lower race (the one that's drifted into the head tube) has matte
dimples in a continuous arc spanning about 200 degrees, and the top cup
has corresponding shiny dimples.

All the races seem to be made out of the same kind of steel (the top and
bottom cups are aluminium but with steel races pressed into them).

jim beam

unread,
Aug 25, 2009, 9:09:37 AM8/25/09
to

fact 1. i, and several others on this group, have personal experience of
single event impact causing true brinelling in bicycle headsets. in my
case, on two bikes no less - and one was about a week old.

fact 2. "chas" here on this group performed rockwell hardness testing on
headsets he found to have brinelled - they were significantly softer
than what would be considered acceptable for bearing steels.

fact 3. myself and carl fogel have deliberately brinelled headsets.
myself with a hand hammer. fogel with an air hammer against a straw
mat. we posted pics.

fact 4. jobst doesn't believe in true brinelling as he says the bearing
steel is too brittle and will, quote, "shatter". needless to say,
bearing steels do have ductility, they do not shatter and they do indeed
true brinell.

fact 5. modern headset bearings almost never brinell, despite the fact
that they're operated under exactly the same conditions and subject to
exactly the same loads!

bottom line - jobst will never admit he's wrong on this because he's got
no interest in learning and is too stubborn to read beyond some chapter
heading he read 50 years ago. meanwhile, he continues to piss in the
knowledge pool with false and/or outdated information, and looks like a
complete ass for doing so.

Ben C

unread,
Aug 25, 2009, 9:13:55 AM8/25/09
to

Thanks for the summary. One unresolved question is this: you said that a
matte dimple doesn't imply false brinelling (I guess because it can get
like that afterwards whatever caused the original dimple). But does a
smooth shiny dimple imply true brinelling?

jim beam

unread,
Aug 25, 2009, 9:24:03 AM8/25/09
to

more typically, yes, but that is not a 100% rule. you have to look at
the circumstances. with single impact event it's easy to say true
brinelling, but you can have multiple smaller impact events whose sum is
the same as the larger event, as excellently illustrated by fogel's air
hammer. this would still be true brinelling, not false brinelling.
false brinelling happens under low/no load conditions in the presence of
oxygen.

someone

unread,
Aug 25, 2009, 4:08:55 PM8/25/09
to
On 25 Aug, 14:24, jim beam <m...@privacy.net> wrote:
> Ben C wrote:
> > On 2009-08-25, jim beam <m...@privacy.net> wrote:
> >> Ben C wrote:

Another part of headset damage is due to environmental contamination
of the lower bearing. More wear will occur facing the front and the
rear (without mudguard) because the lack of articulation prevents the
spread of the dirt. F.W.Evans fitted their frames with lubrication
ports so that oil could be supplied direct to the headset bearing.
Thiis would wash out contaminated lubricant leaving the bearing clean
with fresh oil. Pumping of the bearing will cause grease to be
expelled and dirt to be drawn in. Continuing use wears the bearing
unless the contaminated grease is removed and the bearing
relubricated.

Message has been deleted

Jobst Brandt

unread,
Aug 25, 2009, 5:58:46 PM8/25/09
to
"SJM who? wrote:

>> Thanks for the summary. One unresolved question is this: you said
>> that a matte dimple doesn't imply false brinelling (I guess because
>> it can get like that afterwards whatever caused the original
>> dimple). But does a smooth shiny dimple imply true brinelling?

> This whole argument begs the question "WTF difference does it
> make?". True brinell, false brinell, whatever. This isn't a gyro
> in the space shuttle. If the headset or bearings are worn, try
> replacing the bearings and see if it's serviceable. If so, continue
> to use it. If not, replace it. Repack your headset often enough to
> keep the bearings clean.

Fretting is not cause by lack of clean grease (or oil) which furnishes
lubrication to the interface of bearing balls to races. In normal use,
bearing balls remain separated from their race by a mono-molecular oil
films. A great example that such fluid films are not readilydisplaced
from smooth interfaces is loss of traction of railway wheels on wet
rails. To prevent this, all locomotives have sanding tubes that
spread special hard sand ahead of their drive wheels. Oil being more
viscous and tenacious, separates bearing balls from races in spite of
what is commonly believed. This is also what makes a sharp edged
squeegee glide over a windshield or a razor blade to glide over wet
skin. I suppose the picture of a snail climbing over the edge of a
razor has been seen by all.

Fretting damage to races is causes by vibrational pressure that
enables bearing balls to penetrate the oil film by rocking back and
forth on the Hertzian contact patch until the oil film is penetrated.
As explained in "Rolling Element Bearings", bearing balls skate when
they roll and cause scuffing welding that subsequently tears out tiny
bits of races when the ball rotates far enough to lift off.

In steering gears, fretting while driving straight ahead for longer
high speed stretches that cause vibration, gave rise to many patented
mechanisms to prevent dimpling in steering gears. Fortunately a far
simpler design has replace these mechanisms that woks well today with
power steering.

> There, all done.

I'm not sure to what you are referring. It does make a difference and
knowing what the difference is enables current head bearings to avoid
fretting damage by taking rocking motion in a spherical plain bearing
on which a radially preloaded angular contact ball bearing rides while
making its rotary steering motions.

Jobst Brandt

someone

unread,
Aug 25, 2009, 9:09:57 PM8/25/09
to
On 25 Aug, 22:58, Jobst Brandt <jbra...@sonic.net> wrote:

> Fretting is not cause by lack of clean grease (or oil) which furnishes
> lubrication to the interface of bearing balls to races. In normal use,

dirt enters the heaset at the front and behind if no mudguard is in
use. The contamination causes rapid oxidation,wear and subseque=nt
corrosion. Bearing quality is variable as is any sealing or lubricant
properties.

> bearing balls remain separated from their race by a mono-molecular oil
> films.  A great example that such fluid films are not readilydisplaced
> from smooth interfaces is loss of traction of railway wheels on wet
> rails.  To prevent this, all locomotives have sanding tubes that
> spread special hard sand ahead of their drive wheels.  Oil being more
> viscous and tenacious, separates bearing balls from races in spite of
> what is commonly believed.  

That contamination can penetrate a single molecule of oil.

> Fretting damage to races is causes by vibrational pressure that
> enables bearing balls to penetrate the oil film by rocking back and
> forth on the Hertzian contact patch until the oil film is penetrated.

Except bicycles are operated in a dirty environment so damage due to
contamination is highly likely and fretting corrosion is not.

> As explained in "Rolling Element Bearings", bearing balls skate when
> they roll and cause scuffing welding that subsequently tears out tiny
> bits of races when the ball rotates far enough to lift off.

Capillary action maintains lubrication, when movement is micrscopic,
with the correct oil.

Message has been deleted

jim beam

unread,
Aug 26, 2009, 1:15:19 AM8/26/09
to
Jobst Brandt, the "engineer" who simply never gets it right wrote:
> "SJM who? wrote:
>
>>> Thanks for the summary. One unresolved question is this: you said
>>> that a matte dimple doesn't imply false brinelling (I guess because
>>> it can get like that afterwards whatever caused the original
>>> dimple). But does a smooth shiny dimple imply true brinelling?
>
>> This whole argument begs the question "WTF difference does it
>> make?". True brinell, false brinell, whatever. This isn't a gyro
>> in the space shuttle. If the headset or bearings are worn, try
>> replacing the bearings and see if it's serviceable. If so, continue
>> to use it. If not, replace it. Repack your headset often enough to
>> keep the bearings clean.
>
> Fretting is not cause by lack of clean grease (or oil) which furnishes
> lubrication to the interface of bearing balls to races. In normal use,
> bearing balls remain separated from their race by a mono-molecular oil
> films.

untrue.

1. it's not monomolecular.

2. separation does not occur unless rotation exceeds a certain speed.
and that itself is a function of load. bike bearings rarely, if ever,
reach sufficient speed or have low enough loads to achieve separation.

of course, you'd know both these things if you'd ever bothered to read
your own cites or had read any other material post 1950.


> A great example that such fluid films are not readilydisplaced
> from smooth interfaces is loss of traction of railway wheels on wet
> rails.

eh? jobst, that's because they're not actually smooth and they have low
contact pressure!


> To prevent this, all locomotives have sanding tubes that
> spread special hard sand ahead of their drive wheels.

don't you see how this contradicts your previous statement?


> Oil being more
> viscous and tenacious, separates bearing balls from races in spite of
> what is commonly believed.

not at low speeds it doesn't!


> This is also what makes a sharp edged
> squeegee glide over a windshield or a razor blade to glide over wet
> skin. I suppose the picture of a snail climbing over the edge of a
> razor has been seen by all.

jobst, now you're taking the piss - these are deliberately misleading
"analogies".


>
> Fretting damage to races is causes by vibrational pressure that
> enables bearing balls to penetrate the oil film by rocking back and
> forth on the Hertzian contact patch until the oil film is penetrated.

you got that wrong jobst.

1. you don't get hertzian contact unless you have load. false
brinelling doesn't occur under load.

2. you can have a bearing subject to those conditions without presence
of oxygen, and it doesn't happen. introduce oxygen under exactly the
same conditions and false brinelling occurs.

even a stanford grad should be able to exercise a little logical
analysis of this data.


> As explained in "Rolling Element Bearings", bearing balls skate when
> they roll and cause scuffing welding that subsequently tears out tiny
> bits of races when the ball rotates far enough to lift off.

indeed they do, but that's not false brinelling, that's "wear".
[ignoring of course the fact that you're contradicting your own
statements about oil films...]


>
> In steering gears, fretting while driving straight ahead for longer
> high speed stretches that cause vibration, gave rise to many patented
> mechanisms to prevent dimpling in steering gears. Fortunately a far
> simpler design has replace these mechanisms that woks well today with
> power steering.

jobst, have you any idea just how ridiculous that self-contradiction is?
power steering simply amplifies steering force - it has /ZERO/ effect
on steering gear surface contact!!!


>
>> There, all done.
>
> I'm not sure to what you are referring. It does make a difference and
> knowing what the difference is enables current head bearings to avoid
> fretting damage by taking rocking motion in a spherical plain bearing
> on which a radially preloaded angular contact ball bearing rides while
> making its rotary steering motions.

jobst, if only you bothered to read your own cites and intake any
information newer than 1950...

jim beam

unread,
Aug 26, 2009, 1:16:05 AM8/26/09
to
Still Just Me -< wrote:

> On 25 Aug 2009 21:58:46 GMT, Jobst Brandt <jbr...@sonic.net> wrote:
>
>>> There, all done.
>> I'm not sure to what you are referring. It does make a difference and
>> knowing what the difference is enables current head bearings to avoid
>> fretting damage by taking rocking motion in a spherical plain bearing
>> on which a radially preloaded angular contact ball bearing rides while
>> making its rotary steering motions.
>>
>> Jobst Brandt
>
> Maybe it makes a difference to a headset engineer. To anyone here
> riding a bike with an installed headset, all that matters is how you
> maintain and/or replace that headset.
>
> There might be a useful discussion of the relative advantages of ball
> vs. roller bearings and wear/lifetime, but that's another thread.

indeed.

the thing that really totally amazes me about all jobst's blatherings is
his failure to recognize the fact that modern cartridge bearing headsets
hardly ever brinell. exactly the same loading conditions as the old
days, so if he were right, the same failures would still be occurring!

Ben C

unread,
Aug 26, 2009, 4:11:28 AM8/26/09
to
On 2009-08-25, Still Just Me -< <stillno...@stillnodomainer.com> wrote:

> On Tue, 25 Aug 2009 08:13:55 -0500, Ben C <spam...@spam.eggs> wrote:
>
>>
>>Thanks for the summary. One unresolved question is this: you said that a
>>matte dimple doesn't imply false brinelling (I guess because it can get
>>like that afterwards whatever caused the original dimple). But does a
>>smooth shiny dimple imply true brinelling?
>
> This whole argument begs the question "WTF difference does it make?".
> True brinell, false brinell, whatever. This isn't a gyro in the space
> shuttle.

By your logic, why have a space shuttle? I mean, WTF difference does it
make?

someone

unread,
Aug 26, 2009, 9:27:56 PM8/26/09
to
On 26 Aug, 09:11, Ben C <spams...@spam.eggs> wrote:
> On 2009-08-25, Still Just Me  -< <stillnoEmail...@stillnodomainer.com> wrote:

>
> > On Tue, 25 Aug 2009 08:13:55 -0500, Ben C <spams...@spam.eggs> wrote:
>
> >>Thanks for the summary. One unresolved question is this: you said that a
> >>matte dimple doesn't imply false brinelling (I guess because it can get
> >>like that afterwards whatever caused the original dimple). But does a
> >>smooth shiny dimple imply true brinelling?
>
> > This whole argument begs the question "WTF difference does it make?".
> > True brinell, false brinell, whatever. This isn't a gyro in the space
> > shuttle.
>
> By your logic, why have a space shuttle? I mean, WTF difference does it
> make?

not another political thread.

someone

unread,
Aug 26, 2009, 9:30:29 PM8/26/09
to
On 23 Aug, 23:18, someone <thirty-...@live.co.uk> wrote:
> On 23 Aug, 18:35, Jobst Brandt <jbra...@sonic.net> wrote:

>
> > Chalo Colina wrote:
> > >> Once upon a time bicycles came with non-cartridge bearings.  Some
> > >> of these bearings were loose ball bearings whilst some had the ball
> > >> bearings captured in cages to make servicing easier.  What I am
> > >> thinking is that the loose ball bearings, because there would be
> > >> more of them and thus more load bearing surface than if they were
> > >> caged, would be better than the caged bearings.  Is this true?
> > > More balls can support higher loads.  But full-complement bearing
> > > balls rub against the faces of loaded rolling balls rather than
> > > relatively stationary unloaded cages, and thus they incur higher
> > > bearing drag.  You can feel the difference even while turning the
> > > races of unmounted full-complement bearing cartridges versus Conrad
> > > type caged ball bearing cartridges.
> > >> Is it worth the effort to replace caged bearings with loose ball
> > >> bearings in these areas?
> > > It doesn't hurt.  It probably doesn't help, either.
>
> > Full complement ball bearings, as I mentioned, can also have cages,
> > whose purpose is ease and reliability of assembly rather than to alter
> > friction or wear characteristics.  Full complement (no empty ball
> > spaces) bearings are not press fit bearing ball to bearing ball but
> > rather an arrangement where the race circumferences are designed to
> > accept a number of balls with small spacing that is not accidentally
> > just enough to have a reasonable oil film between them.
>
> > Two hardened steel spheres, even when touching, have a no-load point
> > contact in such a bearing.  The reason for ball or roller cages in
> > bearings is for assembly rather than a load related one except in ones
> > where the cage allows using fewer balls for economic effect.
>

> FFS [(exasperated)], the axle load parts the balls at the bottom pushing the balls all
> together around the rest of the bearing.  Use some pennies inside a
> pastry cutter and use a larger coin to act as the con[e]/axle and apply
> the radial load, the pennies spread in the load direction and are
> pushed tight together around the ring and against it.

speeling correxion.

Jobst Brandt

unread,
Aug 28, 2009, 8:32:55 PM8/28/09
to
SJM wrote secretively:

>>> There, all done.

>> I'm not sure to what you are referring. It does make a difference
>> and knowing what the difference is enables current head bearings to
>> avoid fretting damage by taking rocking motion in a spherical plain
>> bearing on which a radially preloaded angular contact ball bearing
>> rides while making its rotary steering motions.

> Maybe it makes a difference to a headset engineer. To anyone here


> riding a bike with an installed headset, all that matters is how you
> maintain and/or replace that headset.

That's my point. If you know what causes these failures and give it
some thought, you can buy a suitable replacement.

> There might be a useful discussion of the relative advantages of ball
> vs. roller bearings and wear/lifetime, but that's another thread.

I see you are new to RBT and didn't read of the failures with head set
roller bearings ans why they fail and have more drag than ball
bearings. They also suffer from fretting and do so more easily than
balls because the rollers slice axially on a rocking fork crown and
they get off axis and skate with more drag than balls.

Jobst Brandt

Jobst Brandt

unread,
Aug 28, 2009, 8:42:52 PM8/28/09
to
SJM wrote secretively:

>>> There, all done.

>> I'm not sure to what you are referring. It does make a difference
>> and knowing what the difference is enables current head bearings to
>> avoid fretting damage by taking rocking motion in a spherical plain
>> bearing on which a radially preloaded angular contact ball bearing
>> rides while making its rotary steering motions.

> Maybe it makes a difference to a headset engineer. To anyone here


> riding a bike with an installed headset, all that matters is how you
> maintain and/or replace that headset.

That's my point. If you know what causes these failures and give it
some thought, you can buy a suitable replacement.

> There might be a useful discussion of the relative advantages of ball
> vs. roller bearings and wear/lifetime, but that's another thread.

I see you are new to RBT and didn't read of the failures with head set

roller bearings and why they fail and have more drag than ball

jim beam

unread,
Aug 29, 2009, 9:41:58 AM8/29/09
to
Jobst Brandt wrote:
> SJM wrote secretively:
>
>>>> There, all done.
>
>>> I'm not sure to what you are referring. It does make a difference
>>> and knowing what the difference is enables current head bearings to
>>> avoid fretting damage by taking rocking motion in a spherical plain
>>> bearing on which a radially preloaded angular contact ball bearing
>>> rides while making its rotary steering motions.
>
>> Maybe it makes a difference to a headset engineer. To anyone here
>> riding a bike with an installed headset, all that matters is how you
>> maintain and/or replace that headset.
>
> That's my point.

no it's not - you've been barking up entirely the wrong tree for
decades. the only "point" you've had is to pose as an "expert" and
muddy the water with bullshit about fretting. if you'd known what you
were talking about, or even done some rudimentary experimentation, you'd
have identified the problem and publicized it in a manner appropriate
to getting it cured. instead, being as you've been wrong, nobody in the
industry would listen to you, and thus the problem persisted for decades
longer than it need have.


> If you know what causes these failures and give it
> some thought, you can buy a suitable replacement.

not if an effective replacement is not available. do /you/ know how to
prevent false brinelling jobst? if you think you do, you should spell
it out! [and provide entertainment!]


>
>> There might be a useful discussion of the relative advantages of ball
>> vs. roller bearings and wear/lifetime, but that's another thread.
>
> I see you are new to RBT

i see you don't learn a damned thing, even when the information is
served you on a plate. you don't even read your own cites.


> and didn't read of the failures with head set
> roller bearings ans why they fail and have more drag than ball
> bearings.

er, jobst, the roller bearings in my bike are about 20 years old. no
true or false brinelling. total failure from you to explain this
because you haven't identified the correct problem.


> They also suffer from fretting and do so more easily than
> balls because the rollers slice axially on a rocking fork crown and
> they get off axis and skate with more drag than balls.

the only thing off axis around here is your understanding of either
bearings or tribology.

corrections of jobstian myth:

1. bike bearings are not elasto-hydrodynamically separated.

2. bike bearings do true brinell.

3. old bike bearings were insufficiently hard. [see #2]

4. false brinelling is an oxidation process.

Reply all
Reply to author
Forward
0 new messages