MTB cone type wheel bearings.

wrote in message
...
On Monday, April 22, 2013 6:19:37 PM UTC-4, James wrote:
On 23/04/13 01:09, Ian Field wrote:



The ball to ball contact is very smooth & shiny - and hopefully well

lubricated.



In a bicycle hub or BB, not really well lubricated. I.e., not a

continuous flow of liquid lubricant.



The retainer is usually a stamped, formed and tempered bit of steel
with

sharp edges and burrs.



And smooth enough where it counts, and provides a lubricant reservoir,

I'm told.



And - apparently, some Muppets put them in back to front!



Anything is possible. Thankfully cartridge bearings with seals both

sides solve the problem.

With most sealed cartridge bearings I've examined, there's far more
friction from the seals than there would ever be from inter-ball contact.
Nonetheless, even that seal friction is negligible in any practical sense.

I've got sealed bearings on the hubs of three bikes and on the cranks of
at least two. (I'm not sure about some others.) Particularly with the
hubs when new, it was obvious when spinning the axles with my fingers
(wheels out of the bike) that there was a tiny bit more friction. But I
was happy to accept it in return for freedom from maintenance.

The manufacturer installed grease will eventually dry out and harden just
like any other grease.

Years ago, I vaguely remember seeing in a tool catalogue - a fixture for
forcing grease into sealed cartridge bearings, but they tend to drive the
seal out if you're not careful.

"J.B.Slocomb" wrote in message
news
On Mon, 22 Apr 2013 22:01:22 -0700 (PDT), Frank Krygowski
wrote:

On Apr 22, 8:56 pm, James wrote:
On 23/04/13 10:32, wrote:


With most sealed cartridge bearings I've examined, there's far more
friction from the seals than there would ever be from inter-ball
contact. Nonetheless, even that seal friction is negligible in any
practical sense.

Do you have any evidence?

You mean that I'm remembering correctly, or not lying? No, but I'd
prefer finding out whether anyone really doubts me before I go
downstairs, pull out a wheel, remove a quick release and take some
very fine torque readings.

I'd think it would be hard to believe that an elastomer seal rubbing
agains a bearing's race would not have a _little_ more friction than a
bearing lacking such a seal.

It is likely that it does have more drag than a seal less bearing but
like many things in the bicycle world there is probably so little
difference that no one can actually tell the difference.

There could be a cumulative effect on how knackered you feel after a 30 mile
ride.

On Tuesday, April 23, 2013 7:20:41 AM UTC-4, J. B. Slocomb wrote:

Who was it said, "A tempest in a tea pot"?

It may have been a psychic in days long ago, envisioning rec.bicycles.tech!

(Actually, this discussion is proceeding much more pleasantly than some. IOW, some things really do turn out to be negligible!)

- Frank Krygowski

On 4/23/2013 12:01 AM, Frank Krygowski wrote:
On Apr 22, 8:56 pm, James wrote:
On 23/04/13 10:32, wrote:


With most sealed cartridge bearings I've examined, there's far more
friction from the seals than there would ever be from inter-ball
contact. Nonetheless, even that seal friction is negligible in any
practical sense.

Do you have any evidence?

You mean that I'm remembering correctly, or not lying? No, but I'd
prefer finding out whether anyone really doubts me before I go
downstairs, pull out a wheel, remove a quick release and take some
very fine torque readings.

I'd think it would be hard to believe that an elastomer seal rubbing
agains a bearing's race would not have a _little_ more friction than a
bearing lacking such a seal.


And even so, I'd rather expend a little energy and keep good seals than
have bearings wear out fast.

Me too, as I said.



I've got sealed bearings on the hubs of three bikes and on the cranks
of at least two. (I'm not sure about some others.) Particularly
with the hubs when new, it was obvious when spinning the axles with
my fingers (wheels out of the bike) that there was a tiny bit more
friction. But I was happy to accept it in return for freedom from
maintenance.

And I bet the hub manufacturer didn't specify full complement bearings
either. Why is that? Don't you think they would want the best solution
for their product?

OK, that's an excellent question. Let's examine it! Does a
manufacturer want a "best solution"?

Well, I suppose that ideally, they want the solution that will best
influence their bottom line. When talking about things like hub
bearings (where sculptural art is not a factor), they want something
that will maximize profits. If they're smart, they'd be thinking
about long-term profits, which are influenced by market reviews and
their effect on the manufacturer's reputation. (Cheap junk might bump
profits up for a year, and lead to buyer abandonment.)

So the manufacturer has to take customer perception into account. He
also has to take his costs into account. In the case of hubs using
cartridge bearings, it's _much_ cheaper to install standard industrial
bearings rather than source custom ones from a bearing manufacturer.
(The hub guy won't have capability of making his own bearings.)

So what should he spec? Customers - aside from some geeks - don't
care much about friction torque measurements. So advertize
"maintenance free" and get significantly more sales.

But about full complement vs. cage bearings? Well, how would a
manufacturer even advertise that? Only a few super-geeks would know
the difference, and those guys would understand that the tradeoff
would be something like a few grams - i.e. that you could get the same
capacity by going to a much cheaper bearing one size up.

What a smart manufacturer would do would be to pick a standard sealed
cartridge bearing that would be expected to last 20 years with no
maintenance and advertise it as super-robust. That's what Phil Wood
did, and it worked well. Full complement? Seal friction? Who cares?
It's bulletproof! It's "best."


I never said I was concerned about *me* or anyone else overcoming the
bearing friction. I contend that there is more friction in a full
complement bearing than a caged type, that causes faster wear and
premature bearing failure.

I don't know what "premature bearing failure" is in a hub bearing.
The failures I've seen (with relatively inexpensive equipment) have
been almost imperceptible - a tiny rough spot on a cone or on a cup
after many, many years of service. The bearing balls get renewed when
overhauled anyway. Real failure is so rare it seems impossible to
determine if it's occurred ahead of schedule, whatever "schedule" may
be.

One place on a bike that does see bearing failure is a headset; but
that's a different, non-rotating animal, one that fails by a very
different mechanism, discussed here thoroughly in years past. But
interestingly, in that situation, a full complement of bearing balls
is thought to help.

The bearing losses has been done to death, and is about 0.5 Watts or so
per cartridge bearing at reasonable bike speed, and 0.4 Watts of that is
lost in the seals on a bearing with seals each side.

Really the question is of the suitability of full compliment bearings to
bicycles. And yes they "work" for some values of "work", but from what
I've seen some folks riding, "works" has a very wide range of acceptance.

That's true. Still, if a rider thinks any difference in friction
between caged vs. full complement bearings will affect his riding in
any measurable or perceptible way, he's a "princess and the pea"
guy.



Meanwhile back in the real world:
http://www.yellowjersey.org/photosfr...t/failfhub.jpg

Your average cyclist, for example this guy, who just wanted
us to stop the loud clanking noises, have other problems
than negligible losses from seal drag.

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

On Apr 22, 10:01*pm, Frank Krygowski wrote:
On Apr 22, 8:56*pm, James wrote:

On 23/04/13 10:32, wrote:

With most sealed cartridge bearings I've examined, there's far more
friction from the seals than there would ever be from inter-ball
contact. *Nonetheless, even that seal friction is negligible in any
practical sense.

Do you have any evidence?

You mean that I'm remembering correctly, or not lying? *No, but I'd
prefer finding out whether anyone really doubts me before I go
downstairs, pull out a wheel, remove a quick release and take some
very fine torque readings.

I'd think it would be hard to believe that an elastomer seal rubbing
agains a bearing's race would not have a _little_ more friction than a
bearing lacking such a seal.



And even so, I'd rather expend a little energy and keep good seals than
have bearings wear out fast.

Me too, as I said.



I've got sealed bearings on the hubs of three bikes and on the cranks
of at least two. *(I'm not sure about some others.) *Particularly
with the hubs when new, it was obvious when spinning the axles with
my fingers (wheels out of the bike) that there was a tiny bit more
friction. *But I was happy to accept it in return for freedom from
maintenance.

And I bet the hub manufacturer didn't specify full complement bearings
either. *Why is that? *Don't you think they would want the best solution
for their product?

OK, that's an excellent question. *Let's examine it! *Does a
manufacturer want a "best solution"?

Well, I suppose that ideally, they want the solution that will best
influence their bottom line. *When talking about things like hub
bearings (where sculptural art is not a factor), they want something
that will maximize profits. *If they're smart, they'd be thinking
about long-term profits, which are influenced by market reviews and
their effect on the manufacturer's reputation. *(Cheap junk might bump
profits up for a year, and lead to buyer abandonment.)

So the manufacturer has to take customer perception into account. He
also has to take his costs into account. *In the case of hubs using
cartridge bearings, it's _much_ cheaper to install standard industrial
bearings rather than source custom ones from a bearing manufacturer.
(The hub guy won't have capability of making his own bearings.)

So what should he spec? *Customers - aside from some geeks - don't
care much about friction torque measurements. *So advertize
"maintenance free" and get significantly more sales.

But about full complement vs. cage bearings? *Well, how would a
manufacturer even advertise that? *Only a few super-geeks would know
the difference, and those guys would understand that the tradeoff
would be something like a few grams - i.e. that you could get the same
capacity by going to a much cheaper bearing one size up.

What a smart manufacturer would do would be to pick a standard sealed
cartridge bearing that would be expected to last 20 years with no
maintenance and advertise it as super-robust. *That's what Phil Wood
did, and it worked well. *Full complement? *Seal friction? Who cares?
It's bulletproof! *It's "best."

In 1975, Phil was a lot better than Campy -- which had dust covers but
no real seals. But Phil hubs wore out in way less than 20 years.
Exposed wiper seals just weren't that good in wet weather, and Phil
didn't want you peeling back the seal to lubricate the bearings. When
the bearings in my various sets of Phil hubs went belly up, I'd just
tromp over to Phil's shop and get new ones (which used to be free --
not so anymore). Not so convenient after I moved to Oregon.

I still don't think exposed cartridge bearings are that good for use
in hubs in wet environments, at least not as good as fully serviceable
Shimano labyrinth/contact seals. My $35 sale table Deore disc hubs
have proven to be very durable, and even after riding through swamps
on them, they usually have a good amount of fresh grease in them when
I do my infrequent servicing.

-- Jay Beattie.

On Tue, 23 Apr 2013 15:42:53 +0100, "Ian Field"
wrote:



"J.B.Slocomb" wrote in message
news
On Mon, 22 Apr 2013 22:01:22 -0700 (PDT), Frank Krygowski
wrote:

On Apr 22, 8:56 pm, James wrote:
On 23/04/13 10:32, wrote:


With most sealed cartridge bearings I've examined, there's far more
friction from the seals than there would ever be from inter-ball
contact. Nonetheless, even that seal friction is negligible in any
practical sense.

Do you have any evidence?

You mean that I'm remembering correctly, or not lying? No, but I'd
prefer finding out whether anyone really doubts me before I go
downstairs, pull out a wheel, remove a quick release and take some
very fine torque readings.

I'd think it would be hard to believe that an elastomer seal rubbing
agains a bearing's race would not have a _little_ more friction than a
bearing lacking such a seal.

It is likely that it does have more drag than a seal less bearing but
like many things in the bicycle world there is probably so little
difference that no one can actually tell the difference.

There could be a cumulative effect on how knackered you feel after a 30 mile
ride.

In another post I mention that I recently changed wheels, from a set
using Shimano cage less bearing hubs to a set with cartridge bearings,
and really couldn't tell the difference. Which is true, I think, for
many of the "very important things" that we discuss here :-)
--
Cheers,

John B.

BULL TACO !


http://www.visualphotos.com/photo/1x...0043379 4.jpg

On Apr 23, 1:14*pm, Phil W Lee wrote:
James considered Tue, 23 Apr 2013 10:21:29
+1000 the perfect time to write:









On 23/04/13 09:38, J.B.Slocomb wrote:
On Tue, 23 Apr 2013 08:25:39 +1000,
wrote:

On 23/04/13 02:45, Phil W Lee wrote:
* considered Mon, 22 Apr 2013 08:27:11
+1000 the perfect time to write:

On 19/04/13 13:50, David Scheidt wrote:
Frank * *wrote:
:On Apr 18, 8:11 pm, * *wrote:
: * *On 19/04/13 08:38, Phil W Lee wrote:
:
: * * * * * *considered Thu, 18 Apr 2013 08:11:06
: * * * *+1000 the perfect time to write:
: * * * *The quote is not clear because it does not define what the increase is
: * * * *with respect to. *I have read other papers that say cageless bearings
: * * * *run hotter than caged bearings - thus more friction losses in the
: * * * *cageless variety.
:
: * * * *Unless you can explain how adding multiple points of sliding contact
: * * * *can reduce friction over a design with only rolling contact, that
: * * * *falls well short of sensible.
:
: * *What is a design with only rolling contact? *If you mean a cage less
: * *rolling bearing (ball or roller), then please explain how sliding
: * *contact is avoided when there is no cage to keep the rolling elements
: * *from touching?
:
: * *Please also explain to me, as I've obviously lost me bearings, why a
: * *google search yields results such as;
:
: * *"Ball Cage Effect
: * *The early forms of ball bearings were full-ball types without ball
: * *cages. Friction between balls caused loud
: * *noise, made high-speed rotation impossible and shortened the service
: * *life. Twenty years later, a Caged Ball
: * *design was developed for ball bearings. The new design enabled
: * *high-speed rotation at a low noise level,
: * *and extended the service life despite the reduced number of balls used.
: * *It marked a major development in
: * *the history of ball bearings.
: * *Similarly, the quality of needle bearings was significantly improved by
: * *the caged needle structure.
: * *With cage-less, full-ball types of ball bearings, balls make metallic
: * *contact with one another and
: * *produce loud noise. In addition, they rotate in opposite directions,
: * *causing the sliding contact between two
: * *adjacent balls to occur at a speed twice the ball-spinning rate. It
: * *results in severe wear and shortens the
: * *service life.
: * *In addition, without a cage, balls make point contact to increase
: * *bearing stress, thus facilitating
: * *breakage of the oil film. In contrast, each caged ball contacts the cage
: * *over a wide area. Therefore, the oil
: * *film does not break, the noise level is low and balls can rotate at a
: * *high speed, resulting in a long
: * *service life."
:
: * *(google "site:tech.thk.com Caged Ball SHS")

:Looks to me like they're advertising their design feature.

:I left all my bearing catalogs behind when I retired, but I know for
:sure that a bearing's load capacity is increased when the number of
:balls increases. For ordinary industrial ball bearings, the type with
:the cage is called a Conrad bearing; it's the basic type. *The type
:that crams an extra ball or two into the groove is called a slot-fill
:bearing, or full complement bearing. *Its radial load capacity is
:definitely higher, due to the higher ball count. *(Its axial load
:capacity is far lower, due to the groove.)

Conrad and slot fill bearings are relevant to bearings designed to
carry a purely radial load. *Remember, of course, that a cup and cone
bicycle bearing is an angular contact bearing, which can carry radial and
axial loads in different proportion by varying the angles of the races.
Importantly, they can be made as a full complement bearing without
needing the slot for assembly, since they come apart axially. *The Conrad
bearing solved James's objection that without a cage, the balls will
move, which does lead to bearing failure. *But with a full complement of
balls, you don't need a cage to maintain spacing, since there's no extra
space to dispalce into. *The cage is really just for easy of assembly
(and maybe stocking spares). *It's often used to reduce the number of
balls in the bearing, but it needn't be.

My objection was not only that the balls can move and not be evenly
spaced, but that they press against one another without a cage, and the
relative motion between the rubbing surfaces is in opposing directions.

But they won't press against each other any more than they would press
against a cage, and certainly not as continuously.

Where is your evidence?

The pressure and speed of relative motion is greater than that seen when
there is a cage to keep the balls separated.

The pressure is (at most) the same as for a cage, and is between two
hardened, curved, surfaces, and only intermittently.

Again, your evidence?

It has been noted that full compliment bearings run hotter, and need
better lubrication flow (than a caged bearing) to maintain a lubrication
film between the balls. *I.e. there is more friction and wear. *The type
of lubrication required for longevity is not so easy to achieve in a
bicycle hub or BB.

No more difficult than in a wheel bearing or suspension pivot of a
car, and much less heavily loaded.

Car wheel bearings (tapered roller) have a cage - at least all those
that I've worked on have. *And in fact wheel bearings on car trailers
are notorious for self destruction because the lubrication is
insufficient, they tend to not be used often, and the hub tends to let
moisture in - much like bicycle hubs.

I've yet to hear of any bicycle application where heat is a problem in
any (properly maintained) bearing.

Heat is indicative of friction. *The research papers and other sources
I've read say that full compliment bearings get hotter, therefore there
is more friction.

Friction usually leads to wear, therefore I conclude they wear out
faster too.

On the other hand the more balls the more the bearing will support,
attested by research papers also. So which is better, the pillar or
the post :-)

Yes, John, we've been over that before. *The more balls support more
load, and provided the speed is kept low (gee, think why that might be),
and the lubrication adequate (which is unlikely in a bicycle hub or BB),
they are ok.

Much better though to increase the size of the housing just a smidge to
allow for larger balls and bearing surfaces and a cage. *Then there is
less friction, less wear and a lubrication reservoir in the cage.

The lubrication reservoir is smaller if you fill part of it with a
cage.

Yeah, but I can see how a cage would sort of hold the bearing away and
allow more grease to coat it where it needs it in the cycle.

And adequate lubrication is just as easy in a bicycle hub or BB as it
is in a car wheel-hub or suspension pivot (probably easier, because of
the lower loads).
Of course, more balls don't just allow a greater overall load capacity
- at a given total load they reduce the individual load on each ball,
which reduces wear.

I like loose balls (had many bikes apart with trashed cages and balls
- just dont' care for scraps of metal in the moving bearing contact
space), although I'm coming around to thinking maybe the plastic cages
in my Shimano hubs aren't bad - the balls snap in and out - no metal
cage to scrape things up, get bent out of shape, break into metal
bits, etc.

On Apr 24, 8:05*am, Dan O wrote:
On Apr 23, 1:14*pm, Phil W Lee wrote:









James considered Tue, 23 Apr 2013 10:21:29
+1000 the perfect time to write:

On 23/04/13 09:38, J.B.Slocomb wrote:
On Tue, 23 Apr 2013 08:25:39 +1000,
wrote:

On 23/04/13 02:45, Phil W Lee wrote:
* considered Mon, 22 Apr 2013 08:27:11
+1000 the perfect time to write:

On 19/04/13 13:50, David Scheidt wrote:
Frank * *wrote:
:On Apr 18, 8:11 pm, * *wrote:
: * *On 19/04/13 08:38, Phil W Lee wrote:
:
: * * * * * *considered Thu, 18 Apr 2013 08:11:06
: * * * *+1000 the perfect time to write:
: * * * *The quote is not clear because it does not define what the increase is
: * * * *with respect to. *I have read other papers that say cageless bearings
: * * * *run hotter than caged bearings - thus more friction losses in the
: * * * *cageless variety.
:
: * * * *Unless you can explain how adding multiple points of sliding contact
: * * * *can reduce friction over a design with only rolling contact, that
: * * * *falls well short of sensible.
:
: * *What is a design with only rolling contact? *If you mean a cage less
: * *rolling bearing (ball or roller), then please explain how sliding
: * *contact is avoided when there is no cage to keep the rolling elements
: * *from touching?
:
: * *Please also explain to me, as I've obviously lost me bearings, why a
: * *google search yields results such as;
:
: * *"Ball Cage Effect
: * *The early forms of ball bearings were full-ball types without ball
: * *cages. Friction between balls caused loud
: * *noise, made high-speed rotation impossible and shortened the service
: * *life. Twenty years later, a Caged Ball
: * *design was developed for ball bearings. The new design enabled
: * *high-speed rotation at a low noise level,
: * *and extended the service life despite the reduced number of balls used.
: * *It marked a major development in
: * *the history of ball bearings.
: * *Similarly, the quality of needle bearings was significantly improved by
: * *the caged needle structure.
: * *With cage-less, full-ball types of ball bearings, balls make metallic
: * *contact with one another and
: * *produce loud noise. In addition, they rotate in opposite directions,
: * *causing the sliding contact between two
: * *adjacent balls to occur at a speed twice the ball-spinning rate. It
: * *results in severe wear and shortens the
: * *service life.
: * *In addition, without a cage, balls make point contact to increase
: * *bearing stress, thus facilitating
: * *breakage of the oil film. In contrast, each caged ball contacts the cage
: * *over a wide area. Therefore, the oil
: * *film does not break, the noise level is low and balls can rotate at a
: * *high speed, resulting in a long
: * *service life."
:
: * *(google "site:tech.thk.com Caged Ball SHS")

:Looks to me like they're advertising their design feature.

:I left all my bearing catalogs behind when I retired, but I know for
:sure that a bearing's load capacity is increased when the number of
:balls increases. For ordinary industrial ball bearings, the type with
:the cage is called a Conrad bearing; it's the basic type. *The type
:that crams an extra ball or two into the groove is called a slot-fill
:bearing, or full complement bearing. *Its radial load capacity is
:definitely higher, due to the higher ball count. *(Its axial load
:capacity is far lower, due to the groove.)

Conrad and slot fill bearings are relevant to bearings designed to
carry a purely radial load. *Remember, of course, that a cup and cone
bicycle bearing is an angular contact bearing, which can carry radial and
axial loads in different proportion by varying the angles of the races.
Importantly, they can be made as a full complement bearing without
needing the slot for assembly, since they come apart axially. *The Conrad
bearing solved James's objection that without a cage, the balls will
move, which does lead to bearing failure. *But with a full complement of
balls, you don't need a cage to maintain spacing, since there's no extra
space to dispalce into. *The cage is really just for easy of assembly
(and maybe stocking spares). *It's often used to reduce the number of
balls in the bearing, but it needn't be.

My objection was not only that the balls can move and not be evenly
spaced, but that they press against one another without a cage, and the
relative motion between the rubbing surfaces is in opposing directions.

But they won't press against each other any more than they would press
against a cage, and certainly not as continuously.

Where is your evidence?

The pressure and speed of relative motion is greater than that seen when
there is a cage to keep the balls separated.

The pressure is (at most) the same as for a cage, and is between two
hardened, curved, surfaces, and only intermittently.

Again, your evidence?

It has been noted that full compliment bearings run hotter, and need
better lubrication flow (than a caged bearing) to maintain a lubrication
film between the balls. *I.e. there is more friction and wear. *The type
of lubrication required for longevity is not so easy to achieve in a
bicycle hub or BB.

No more difficult than in a wheel bearing or suspension pivot of a
car, and much less heavily loaded.

Car wheel bearings (tapered roller) have a cage - at least all those
that I've worked on have. *And in fact wheel bearings on car trailers
are notorious for self destruction because the lubrication is
insufficient, they tend to not be used often, and the hub tends to let
moisture in - much like bicycle hubs.

I've yet to hear of any bicycle application where heat is a problem in
any (properly maintained) bearing.

Heat is indicative of friction. *The research papers and other sources
I've read say that full compliment bearings get hotter, therefore there
is more friction.

Friction usually leads to wear, therefore I conclude they wear out
faster too.

On the other hand the more balls the more the bearing will support,
attested by research papers also. So which is better, the pillar or
the post :-)

Yes, John, we've been over that before. *The more balls support more
load, and provided the speed is kept low (gee, think why that might be),
and the lubrication adequate (which is unlikely in a bicycle hub or BB),
they are ok.

Much better though to increase the size of the housing just a smidge to
allow for larger balls and bearing surfaces and a cage. *Then there is
less friction, less wear and a lubrication reservoir in the cage.

The lubrication reservoir is smaller if you fill part of it with a
cage.

Yeah, but I can see how a cage would sort of hold the bearing away and
allow more grease to coat it where it needs it in the cycle.

And adequate lubrication is just as easy in a bicycle hub or BB as it
is in a car wheel-hub or suspension pivot (probably easier, because of
the lower loads).
Of course, more balls don't just allow a greater overall load capacity
- at a given total load they reduce the individual load on each ball,
which reduces wear.

I like loose balls (had many bikes apart with trashed cages and balls
- just dont' care for scraps of metal in the moving bearing contact
space), although I'm coming around to thinking maybe the plastic cages
in my Shimano hubs aren't bad - the balls snap in and out - no metal
cage to scrape things up, get bent out of shape, break into metal
bits, etc.

.... Actually, what I really like are the cartridge bearings in my Phil
Wood and Mavic hubs - never need adjustment, ~never need service -
just Ride Bike!

On Apr 22, 10:01*pm, Frank Krygowski wrote:
On Apr 22, 8:56*pm, James wrote:


snip


I never said I was concerned about *me* or anyone else overcoming the
bearing friction. *I contend that there is more friction in a full
complement bearing than a caged type, that causes faster wear and
premature bearing failure.

I don't know what "premature bearing failure" is in a hub bearing.
The failures I've seen (with relatively inexpensive equipment) have
been almost imperceptible - a tiny rough spot on a cone or on a cup
after many, many years of service. *The bearing balls get renewed when
overhauled anyway. *Real failure is so rare it seems impossible to
determine if it's occurred ahead of schedule, whatever "schedule" may
be.


Try riding long, hard miles day-in-and-day-out (weather
notwithstanding), on and off road (not clinging stubbornly to the road
- and the cleanest part of the road at that) in a rainy, sedimentary
region.

One place on a bike that does see bearing failure is a headset; but
that's a different, non-rotating animal, one that fails by a very
different mechanism, discussed here thoroughly in years past. *But
interestingly, in that situation, a full complement of bearing balls
is thought to help.

The bearing losses has been done to death, and is about 0.5 Watts or so
per cartridge bearing at reasonable bike speed, and 0.4 Watts of that is
lost in the seals on a bearing with seals each side.

Really the question is of the suitability of full compliment bearings to
bicycles. *And yes they "work" for some values of "work", but from what
I've seen some folks riding, "works" has a very wide range of acceptance.

That's true. *Still, if a rider thinks any difference in friction
between caged vs. full complement bearings will affect his riding in
any measurable or perceptible way, he's a "princess and the pea"
guy.


Nobody (except you) is talking about the differences in that context.
(Although riding will suffer eventually if the bearings do.)