Re: summary - the disk brake debate

Kinky Cowboy said:

On Sun, 09 Oct 2005 14:33:25 -0700, jim beam <[email hidden]>

Quoted message said:

even with a shear yield at only 100Nmm^-2 for a really soft magnesium
alloy fork end, shearing through 100mm^2 of enmeshed serrations takes
10,000N, order of magnitude. annan's theory resolves only 1000N, order
of magnitude. be my guest, re-work it. annan won't.

If it's so hard to shear through the material, why bother? Climbing
over the slopes of the serrations requires the QR skewer to stretch
only fractions of a millimetre. A full tenth of a millimetre of skewer
stretch only adds about 4kN to the tension, i.e the clamping force is
not quite doubled even riding over a 0.1mm serration engagement depth.

you talk like 4kN is trivial.

Quoted message said:

If you disregard indentations into the paint, which is as likely to
lubricate the interface as restrain it, 0.1mm seems like some
seriously deep grooving; my Fox certainly doesn't seem anywhere near
that depth. So, in the best case, serrations might add less than 100%
to the clamping force, and therefore pullout force. For steel
dropouts, expect much less.

how do you go from trivializing the increase in clamping force to
dismissing shear??? all that matters is that a serration engages with
an indentation. when it does, you suddenly have to shear through
significant amounts of material if you want to drag it loose as gross
deformation is the only way. depth of shear isn't relevant, it's area.

Quoted message said:

Shear yield looks like a red herring, but given their ubiquity
serrations should not be ignored completely as a contributor to the
disparity between Annan's prediction and the low reported in service
failure rate

ya think??? sure, reality's got "herring" written all over it.

Quoted message said:

Kinky Cowboy*

*Batteries not included
May contain traces of nuts
Your milage may vary

On Mon, 10 Oct 2005 20:15:20 -0700, jim beam <[email hidden]>

Quoted message said:

Kinky Cowboy said:

On Sun, 09 Oct 2005 14:33:25 -0700, jim beam <[email hidden]>

Quoted message said:

even with a shear yield at only 100Nmm^-2 for a really soft magnesium
alloy fork end, shearing through 100mm^2 of enmeshed serrations takes
10,000N, order of magnitude. annan's theory resolves only 1000N, order
of magnitude. be my guest, re-work it. annan won't.

If it's so hard to shear through the material, why bother? Climbing
over the slopes of the serrations requires the QR skewer to stretch
only fractions of a millimetre. A full tenth of a millimetre of skewer
stretch only adds about 4kN to the tension, i.e the clamping force is
not quite doubled even riding over a 0.1mm serration engagement depth.

you talk like 4kN is trivial.

No, it's almost doubling the original clamping force; but only if the
serrations are engaged to a depth of 0.1mm, which doesn't seem at all
likely, even for steel serrations biting into Mg dropouts. If the
skewer only has to stretch <0.01mm to let the serrations ratchet past
one another, which is not altogether unreasonable for a steel dropout
with Al QR nuts and axle (the worst case, admittedly, but we have to
consider it) then we only have <400N of extra clamping load, which
really is trivial.

Quoted message said:

Quoted message said:

If you disregard indentations into the paint, which is as likely to
lubricate the interface as restrain it, 0.1mm seems like some
seriously deep grooving; my Fox certainly doesn't seem anywhere near
that depth. So, in the best case, serrations might add less than 100%
to the clamping force, and therefore pullout force. For steel
dropouts, expect much less.

how do you go from trivializing the increase in clamping force to
dismissing shear??? all that matters is that a serration engages with
an indentation. when it does, you suddenly have to shear through
significant amounts of material if you want to drag it loose as gross
deformation is the only way. depth of shear isn't relevant, it's area.

Path of least resistance; if it's easier to go over it than through
it, shear is irrelevant. That makes the depth of engagement (and the
slope angle) very relevant. Remember, all that's stopping the
serration from lifting out of the groove is 100mm of 5mm dia. tie rod
under about 5kN of preload. Just like a chain jumping over sprocket
teeth instead of shearing them off. And just like your jumping chain,
it's probably going to shear off very small areas of the peaks; as it
climbs, the load goes up and the area needing to be sheared off goes
down, and some equilibrium point will be reached where it's easier to
flatten off the top of the hill than to go all the way over it. That
pretty much defines the limit of how significant shear is; it happens
when it's providing less resistance than the friction is. At the other
end of the scale, and at the other end of your bike, there's a set of
indentations where you hope the springy bit will ALWAYS deform enough
to stop the pointy bits from wiping the tops off the serrations; it's
called a freewheel. My final paragraph stands; on the numbers you give
shear is irrelevant because there's an easier path, but indentations
can help to increase clamping load in the event of movement, just like
any other kind of wedge.

Quoted message said:

Quoted message said:

Shear yield looks like a red herring, but given their ubiquity
serrations should not be ignored completely as a contributor to the
disparity between Annan's prediction and the low reported in service
failure rate

ya think??? sure, reality's got "herring" written all over it.

Quoted message said:

Kinky Cowboy*

*Batteries not included
May contain traces of nuts
Your milage may vary

jim beam said:

how do you go from trivializing the increase in clamping force to
dismissing shear??? all that matters is that a serration engages with
an indentation. when it does, you suddenly have to shear through
significant amounts of material if you want to drag it loose as gross
deformation is the only way. depth of shear isn't relevant, it's area.

That's not true, you can also skate right over it if the skewer elongates
enough to do so. Which is what he said. Claiming that the *only* option is
shearing through that material is either blinkered or disingenuous. I'm
also extremely suspicious of your 100 mm^2 figure, incidentally.

Jasper

Kinky Cowboy said:

Path of least resistance; if it's easier to go over it than through
it, shear is irrelevant. That makes the depth of engagement (and the
slope angle) very relevant. Remember, all that's stopping the
serration from lifting out of the groove is 100mm of 5mm dia. tie rod
under about 5kN of preload. Just like a chain jumping over sprocket
teeth instead of shearing them off. And just like your jumping chain,
it's probably going to shear off very small areas of the peaks; as it
climbs, the load goes up and the area needing to be sheared off goes
down, and some equilibrium point will be reached where it's easier to
flatten off the top of the hill than to go all the way over it. That
pretty much defines the limit of how significant shear is; it happens
when it's providing less resistance than the friction is. At the other
end of the scale, and at the other end of your bike, there's a set of
indentations where you hope the springy bit will ALWAYS deform enough
to stop the pointy bits from wiping the tops off the serrations; it's
called a freewheel. My final paragraph stands; on the numbers you give
shear is irrelevant because there's an easier path, but indentations
can help to increase clamping load in the event of movement, just like
any other kind of wedge.

It occurs to me, there's another factor we've been leaving out.

Jim Beam is convinced that the serrations on a steel QR nut will bite
into the softer dropout, giving some degree of mechanical interlock.

If that's true just one time, then what we have is a set of concave
indentations in the dropout, matching the convex protrusions in the QR
nut.
IOW, matching pegs and holes.

Matching _that time_, that is. There is NO guarantee that the second
installation of the front wheel will put the pegs into the holes. IOW,
a perfectly competent cyclist could put the axle in exactly the same
location, but have the circumferential position of the "pegs" slightly
different than before. He clamps the quick release exactly correctly -
but the "pegs" can be biting on the side slope of the "holes."

If this happens - and, statistically, it must - then any microscopic
motion at that interface would cause the "pegs" to slide deeper into
the "holes," and cause some wear at that slope while doing so.

The result would be a loss in QR skewer tension, equal to the amount of
force corresponding to the amount the "stretching" is reduced when the
pegs slide into the holes. For a 5mm steel skewer, 0.1 mm settling
into previous indentations would reduce skewer tension by about 4 kN =
900+ pounds.

Again, ISTM that at the same time this is happening, it's probably
making the side slope of those indentations shallower.

Now for a conventional braked wheel with conventional dropouts, this
hardly matters. The metal to metal interface of the dropout above the
axle prevents almost all relative motion, because the force on the axle
pushes the axle against dropout metal. And even if the QR loses
tension, it needs to retain only enough force to prevent gravity from
pulling the wheel down when, say, the rider lifts the front wheel.

Not so with a common disk brake setup. The force diagram clearly shows
that hard braking tries to lever the axle down and out. A loss in
skewer tension from this mechanism I described might be significant.

Furthermore, as with all threaded fasteners, loss in tension makes
further loosening by vibration easier. It lessens the normal force at
all contacting surfaces, thus lessening the friction force, thus
increasing the chance that microscopic motion within the clearance
interface of the male and female threads will lead to the thread's
sliding down their helix angle. A self-perpetuating situation.

Idealistic calculations of dropout clamping force and metal shear
stress thus need to be modified, because a certain percentage of the
time, the serrations will NOT match, and a QR will soon be looser than
the operator - or the calculators - think.

By the way, a question for the "It never happens" crew: Did you never
ride a bike with horizontal rear dropouts? Did you never experience
your chain pulling your wheel sideways, causing the tire to rub on the
left chainstay, even though you thought the axle was properly clamped?

ISTM that one experience with that would tell a person "Hmmm. Axles
_can_ slip. That's why I'll make sure my next bike design has the
front axle slots facing _away_ from the direction of the ejection
force."

Maybe the guys who designe the first bike disk brakes always rode with
vertical rear dropouts, and never learned that simple lesson?

- Frank Krygowski

In article <[email hidden]>,

Quoted message said:

Maybe the guys who designe the first bike disk brakes always rode with
vertical rear dropouts, and never learned that simple lesson?

Maybe the folks who designed the first bike disk brakes did so with
mountain biking as the only application in mind.

That would make sense, since: (1) All the mtb forks that I have have
dropouts that are in-line with the fork stanchions; they aren't parallel
to the braking force direction, based on my eye-balling the pad locations
on the disk, once installed; (2) All the mtb forks that I've seen have
some sort of lawyer lips-style construction in the dropouts; (3) The
calipers need to be rear mounted to enable protection from damage by the
fork legs; (4) The max braking force that's achieveable is limited by
front wheel frictional forces, or reactions to in-plane (endo) overturning
moment; since the application involves dirt and rocks and such, this is
less force than one would experience on dry roads in roadie applications.

john

Quoted message said:

Kinky Cowboy said:

Path of least resistance; if it's easier to go over it than through
it, shear is irrelevant. That makes the depth of engagement (and the
slope angle) very relevant. Remember, all that's stopping the
serration from lifting out of the groove is 100mm of 5mm dia. tie rod
under about 5kN of preload. Just like a chain jumping over sprocket
teeth instead of shearing them off. And just like your jumping chain,
it's probably going to shear off very small areas of the peaks; as it
climbs, the load goes up and the area needing to be sheared off goes
down, and some equilibrium point will be reached where it's easier to
flatten off the top of the hill than to go all the way over it. That
pretty much defines the limit of how significant shear is; it happens
when it's providing less resistance than the friction is. At the
other end of the scale, and at the other end of your bike, there's a
set of indentations where you hope the springy bit will ALWAYS
deform enough to stop the pointy bits from wiping the tops off the
serrations; it's called a freewheel. My final paragraph stands; on
the numbers you give shear is irrelevant because there's an easier
path, but indentations can help to increase clamping load in the
event of movement, just like any other kind of wedge.

It occurs to me, there's another factor we've been leaving out.

Jim Beam is convinced that the serrations on a steel QR nut will bite
into the softer dropout, giving some degree of mechanical interlock.

If that's true just one time, then what we have is a set of concave
indentations in the dropout, matching the convex protrusions in the QR
nut.
IOW, matching pegs and holes.

Matching _that time_, that is. There is NO guarantee that the second
installation of the front wheel will put the pegs into the holes.
IOW, a perfectly competent cyclist could put the axle in exactly the
same location, but have the circumferential position of the "pegs"
slightly different than before. He clamps the quick release exactly
correctly - but the "pegs" can be biting on the side slope of the
"holes."

If this happens - and, statistically, it must - then any microscopic
motion at that interface would cause the "pegs" to slide deeper into
the "holes," and cause some wear at that slope while doing so.

The result would be a loss in QR skewer tension, equal to the amount
of force corresponding to the amount the "stretching" is reduced when
the pegs slide into the holes. For a 5mm steel skewer, 0.1 mm
settling into previous indentations would reduce skewer tension by
about 4 kN = 900+ pounds.

Again, ISTM that at the same time this is happening, it's probably
making the side slope of those indentations shallower.

Now for a conventional braked wheel with conventional dropouts, this
hardly matters. The metal to metal interface of the dropout above the
axle prevents almost all relative motion, because the force on the
axle pushes the axle against dropout metal. And even if the QR loses
tension, it needs to retain only enough force to prevent gravity from
pulling the wheel down when, say, the rider lifts the front wheel.

Not so with a common disk brake setup. The force diagram clearly
shows that hard braking tries to lever the axle down and out. A loss
in skewer tension from this mechanism I described might be
significant.

Furthermore, as with all threaded fasteners, loss in tension makes
further loosening by vibration easier. It lessens the normal force at
all contacting surfaces, thus lessening the friction force, thus
increasing the chance that microscopic motion within the clearance
interface of the male and female threads will lead to the thread's
sliding down their helix angle. A self-perpetuating situation.

Idealistic calculations of dropout clamping force and metal shear
stress thus need to be modified, because a certain percentage of the
time, the serrations will NOT match, and a QR will soon be looser than
the operator - or the calculators - think.

By the way, a question for the "It never happens" crew: Did you never
ride a bike with horizontal rear dropouts? Did you never experience
your chain pulling your wheel sideways, causing the tire to rub on the
left chainstay, even though you thought the axle was properly clamped?

Yes. I skidded across the road bleeding profusely. My QR was not
sufficiently tightened. I loosened up the cones in the hub, then I properly
tightened my steel QR onto my steel frame and steel locknuts, and have had
no problems since.

--
Phil, Squid-in-Training

Squid-in-Training said:

Quoted message said:

By the way, a question for the "It never happens" crew: Did you never
ride a bike with horizontal rear dropouts? Did you never experience
your chain pulling your wheel sideways, causing the tire to rub on the
left chainstay, even though you thought the axle was properly clamped?

Yes. I skidded across the road bleeding profusely. My QR was not
sufficiently tightened. I loosened up the cones in the hub, then I properly
tightened my steel QR onto my steel frame and steel locknuts, and have had
no problems since.

So would you have preferred a front wheel ejection for exactly the same
mistake?

Jasper

Jasper Janssen said:

Squid-in-Training said:

Quoted message said:

By the way, a question for the "It never happens" crew: Did you
never ride a bike with horizontal rear dropouts? Did you never
experience your chain pulling your wheel sideways, causing the tire
to rub on the left chainstay, even though you thought the axle was
properly clamped?

Yes. I skidded across the road bleeding profusely. My QR was not
sufficiently tightened. I loosened up the cones in the hub, then I
properly tightened my steel QR onto my steel frame and steel
locknuts, and have had no problems since.

So would you have preferred a front wheel ejection for exactly the
same mistake?

My dropouts don't have lawyer lips.

--
Phil, Squid-in-Training

Jasper Janssen said:

jim beam said:

how do you go from trivializing the increase in clamping force to
dismissing shear??? all that matters is that a serration engages with
an indentation. when it does, you suddenly have to shear through
significant amounts of material if you want to drag it loose as gross
deformation is the only way. depth of shear isn't relevant, it's area.

That's not true, you can also skate right over it if the skewer elongates
enough to do so.

"if" it elongates enough??? a rough calc shows the clamping force for a
normal skewer is ~>20kN. you're going to "skate" over a set of
indentions being held together like that???

Quoted message said:

Which is what he said. Claiming that the *only* option is
shearing through that material is either blinkered or disingenuous.

like being "supicious" about an area calc you can easily check ourself?

Quoted message said:

I'm
also extremely suspicious of your 100 mm^2 figure, incidentally.

get your calipers out and measure it. area = pi.r^2

Quoted message said:

Jasper

Quoted message said:

Kinky Cowboy said:

Path of least resistance; if it's easier to go over it than through
it, shear is irrelevant. That makes the depth of engagement (and the
slope angle) very relevant. Remember, all that's stopping the
serration from lifting out of the groove is 100mm of 5mm dia. tie rod
under about 5kN of preload. Just like a chain jumping over sprocket
teeth instead of shearing them off. And just like your jumping chain,
it's probably going to shear off very small areas of the peaks; as it
climbs, the load goes up and the area needing to be sheared off goes
down, and some equilibrium point will be reached where it's easier to
flatten off the top of the hill than to go all the way over it. That
pretty much defines the limit of how significant shear is; it happens
when it's providing less resistance than the friction is. At the other
end of the scale, and at the other end of your bike, there's a set of
indentations where you hope the springy bit will ALWAYS deform enough
to stop the pointy bits from wiping the tops off the serrations; it's
called a freewheel. My final paragraph stands; on the numbers you give
shear is irrelevant because there's an easier path, but indentations
can help to increase clamping load in the event of movement, just like
any other kind of wedge.

It occurs to me, there's another factor we've been leaving out.

Jim Beam is convinced that the serrations on a steel QR nut will bite
into the softer dropout, giving some degree of mechanical interlock.

If that's true just one time, then what we have is a set of concave
indentations in the dropout, matching the convex protrusions in the QR
nut.
IOW, matching pegs and holes.

Matching _that time_, that is. There is NO guarantee that the second
installation of the front wheel will put the pegs into the holes. IOW,
a perfectly competent cyclist could put the axle in exactly the same
location, but have the circumferential position of the "pegs" slightly
different than before. He clamps the quick release exactly correctly -
but the "pegs" can be biting on the side slope of the "holes."

what, like this you mean?

http://home.comcast.net/~carlfogel/download/Img_3199.jpg

that wheel had been inserted & reinserted approximately 100 times. the
indentaion pattern looks remarkably persistent to me. but maybe that's
why shimano makes the serrations with sloping faces to ensure they
always mesh. just like a dog clutch in a gear box.

Quoted message said:

If this happens - and, statistically, it must

"must" it? it's not evidenced above.

Quoted message said:

- then any microscopic
motion at that interface would cause the "pegs" to slide deeper into
the "holes," and cause some wear at that slope while doing so.

The result would be a loss in QR skewer tension, equal to the amount of
force corresponding to the amount the "stretching" is reduced when the
pegs slide into the holes. For a 5mm steel skewer, 0.1 mm settling
into previous indentations would reduce skewer tension by about 4 kN =
900+ pounds.

that's a real sqirm. any regular user of qr skewers can feel when it's
tightening solidly or not. pictorial evidence shows no attrition that
would be caused by the description you're desperately trying to
artificially manufacture.

Quoted message said:

Again, ISTM that at the same time this is happening, it's probably
making the side slope of those indentations shallower.

Now for a conventional braked wheel with conventional dropouts, this
hardly matters. The metal to metal interface of the dropout above the
axle prevents almost all relative motion, because the force on the axle
pushes the axle against dropout metal. And even if the QR loses
tension, it needs to retain only enough force to prevent gravity from
pulling the wheel down when, say, the rider lifts the front wheel.

Not so with a common disk brake setup. The force diagram clearly shows
that hard braking tries to lever the axle down and out. A loss in
skewer tension from this mechanism I described might be significant.

Furthermore, as with all threaded fasteners, loss in tension makes
further loosening by vibration easier. It lessens the normal force at
all contacting surfaces, thus lessening the friction force, thus
increasing the chance that microscopic motion within the clearance
interface of the male and female threads will lead to the thread's
sliding down their helix angle. A self-perpetuating situation.

Idealistic calculations of dropout clamping force and metal shear
stress thus need to be modified, because a certain percentage of the
time, the serrations will NOT match, and a QR will soon be looser than
the operator - or the calculators - think.

go ahead and modify them frank - you're supposed to be the engineering
professor. get with the math.

Quoted message said:

By the way, a question for the "It never happens" crew: Did you never
ride a bike with horizontal rear dropouts? Did you never experience
your chain pulling your wheel sideways, causing the tire to rub on the
left chainstay, even though you thought the axle was properly clamped?

ISTM that one experience with that would tell a person "Hmmm. Axles
_can_ slip. That's why I'll make sure my next bike design has the
front axle slots facing _away_ from the direction of the ejection
force."

Maybe the guys who designe the first bike disk brakes always rode with
vertical rear dropouts, and never learned that simple lesson?

- Frank Krygowski

jim beam said:

"if" it elongates enough??? a rough calc shows the clamping force for a
normal skewer is ~>20kN.

I think you should check your working on that "rough calc".

Or do you imagine that just tossing out random (and wildly inaccurate)
numbers will lend some spurious air of authority to your guesses?

James

James Annan said:

jim beam said:

"if" it elongates enough??? a rough calc shows the clamping force for a
normal skewer is ~>20kN.

I think you should check your working on that "rough calc".

Or do you imagine that just tossing out random (and wildly inaccurate)
numbers will lend some spurious air of authority to your guesses?

James

oh, james, /please/ educate me; /please/ do show off your superior
intellect...

jim beam said:

Quoted message said:

It occurs to me, there's another factor we've been leaving out.

Jim Beam is convinced that the serrations on a steel QR nut will bite
into the softer dropout, giving some degree of mechanical interlock.

If that's true just one time, then what we have is a set of concave
indentations in the dropout, matching the convex protrusions in the QR
nut.
IOW, matching pegs and holes.

Matching _that time_, that is. There is NO guarantee that the second
installation of the front wheel will put the pegs into the holes. IOW,
a perfectly competent cyclist could put the axle in exactly the same
location, but have the circumferential position of the "pegs" slightly
different than before. He clamps the quick release exactly correctly -
but the "pegs" can be biting on the side slope of the "holes."

what, like this you mean?

http://home.comcast.net/~carlfogel/download/Img_3199.jpg

that wheel had been inserted & reinserted approximately 100 times. the
indentaion pattern looks remarkably persistent to me. but maybe that's
why shimano makes the serrations with sloping faces to ensure they
always mesh. just like a dog clutch in a gear box.

I see no particular way sloping faces on the QR serrations are going to
guarantee locking into those same serrations. In fact, it looks like
the flat spot between serrations is roughly twice as wide as the
serration grooves themselves. That means on a random insertion,
there's only a one-out-of-three chance of mesh.

I admit, I'm surprised at the apparent consistency of the serrations.
I just checked my Cannondale rear dropouts and I saw no such grooves.

:-) Of course, if I were "gmschemist" or whatever his name is, I'd
simply say "We don't _know_ what you say is true" and avoid the issue.
But instead I'll just ask, do you do something special to locate the
same circumferential position each time? Again, odds are two to one
that it wouldn't happen by random chance for a given instance.

Quoted message said:

Quoted message said:

If this happens - and, statistically, it must - then any microscopic
motion at that interface would cause the "pegs" to slide deeper into
the "holes," and cause some wear at that slope while doing so.

The result would be a loss in QR skewer tension, equal to the amount of
force corresponding to the amount the "stretching" is reduced when the
pegs slide into the holes. For a 5mm steel skewer, 0.1 mm settling
into previous indentations would reduce skewer tension by about 4 kN =
900+ pounds.

that's a real sqirm. any regular user of qr skewers can feel when it's
tightening solidly or not.

Pay attention, please. If the serrations were right on the edge of
meshing, but not quite meshing, they _would_ clamp tightly. The
operator could feel it.

But if there were some slight movement, the nearly-meshed serrations
could come into mesh, with the "pegs" settling into the "holes." That
would cause a loss of tension.

Quoted message said:

pictorial evidence shows no attrition that
would be caused by the description you're desperately trying to
artificially manufacture.

One picture shows something different, and we don't know what makes it
different. My dropouts don't show any serrations at all. (Sorry, no
digital camera lives here.) By my count, you're one-for-two.

Now I'd like an explanation of how your QR beats the obvious two-to-one
odds against landing precisely in the same grooves. Seems to me the
chance of that happening every time by random chance over 100
installations is (1/3)^100 or about 2E-48.

- Frank Krygowski

Phil said:

Jasper Janssen said:

So would you have preferred a front wheel ejection for exactly the
same mistake?

My dropouts don't have lawyer lips.

Phil, that did not answer the question.

Let me ask another. What would _you_ think about riding a
conventional-brake road bike whose front fork dropout slots faced
upward? Would you call that a good design?

None of your teammates seem willing to answer.

- Frank Krygowski

jim beam said:

James Annan said:

jim beam said:

"if" it elongates enough??? a rough calc shows the clamping force for a
normal skewer is ~>20kN.

I think you should check your working on that "rough calc".

Or do you imagine that just tossing out random (and wildly inaccurate)
numbers will lend some spurious air of authority to your guesses?

James

oh, james, /please/ educate me; /please/ do show off your superior
intellect...

"A man convinced against his will,
Is of the same opinon still".

I'll let you go through your own working, maybe do a little googling,
and perhaps even read that document you cited a few days ago...

One thing that has been interesting to me over the last couple of years
is to see how people have adopted this idea, pretty much at their own
speed, as an inreasing proportion have come to see how it fits with
their (and their friends'😉 experiences. Whenever someone posts (on a
web forum or here) about their difficulty with a slipping or loosening
QR, someone quickly pops up with the answer and a link to my pages.

I don't really see that there is much point in _my_ trying to argue the
case in any detail, especially with anonymous usenet trolls such as
yourself. But you can hardly have avoided noticing how your supposed
"summary" brought out criticism from even the likes of "gcmschemist"
(hardly my strongest supporter!) who seems to recognise your
smokescreen and bluster for what it is.

James

Quoted message said:

Phil said:

Jasper Janssen said:

So would you have preferred a front wheel ejection for exactly the
same mistake?

My dropouts don't have lawyer lips.

Phil, that did not answer the question.

Let me ask another. What would _you_ think about riding a
conventional-brake road bike whose front fork dropout slots faced
upward? Would you call that a good design?

None of your teammates seem willing to answer.

- Frank Krygowski

wow - you're just obsessed with this aren't you. no one answers because
it's so dumb. but since being dumb seems to be the order of the day for
this debate - the answer, professor, is that the g's of an impact can
easily exceed anything seen in in any other mode of service. how about
you do a little math and demonstrate the problem since you seem so
damned fixed on posing such a defeatingly stupid question?

Quoted message said:

jim beam said:

Quoted message said:

It occurs to me, there's another factor we've been leaving out.

Jim Beam is convinced that the serrations on a steel QR nut will bite
into the softer dropout, giving some degree of mechanical interlock.

If that's true just one time, then what we have is a set of concave
indentations in the dropout, matching the convex protrusions in the QR
nut.
IOW, matching pegs and holes.

Matching _that time_, that is. There is NO guarantee that the second
installation of the front wheel will put the pegs into the holes. IOW,
a perfectly competent cyclist could put the axle in exactly the same
location, but have the circumferential position of the "pegs" slightly
different than before. He clamps the quick release exactly correctly -
but the "pegs" can be biting on the side slope of the "holes."

what, like this you mean?

http://home.comcast.net/~carlfogel/download/Img_3199.jpg

that wheel had been inserted & reinserted approximately 100 times. the
indentaion pattern looks remarkably persistent to me. but maybe that's
why shimano makes the serrations with sloping faces to ensure they
always mesh. just like a dog clutch in a gear box.

I see no particular way sloping faces on the QR serrations are going to
guarantee locking into those same serrations. In fact, it looks like
the flat spot between serrations is roughly twice as wide as the
serration grooves themselves. That means on a random insertion,
there's only a one-out-of-three chance of mesh.

I admit, I'm surprised at the apparent consistency of the serrations.
I just checked my Cannondale rear dropouts and I saw no such grooves.

:-) Of course, if I were "gmschemist" or whatever his name is, I'd
simply say "We don't _know_ what you say is true" and avoid the issue.
But instead I'll just ask, do you do something special to locate the
same circumferential position each time? Again, odds are two to one
that it wouldn't happen by random chance for a given instance.

Quoted message said:

Quoted message said:

If this happens - and, statistically, it must - then any microscopic
motion at that interface would cause the "pegs" to slide deeper into
the "holes," and cause some wear at that slope while doing so.

The result would be a loss in QR skewer tension, equal to the amount of
force corresponding to the amount the "stretching" is reduced when the
pegs slide into the holes. For a 5mm steel skewer, 0.1 mm settling
into previous indentations would reduce skewer tension by about 4 kN =
900+ pounds.

that's a real sqirm. any regular user of qr skewers can feel when it's
tightening solidly or not.

Pay attention, please. If the serrations were right on the edge of
meshing, but not quite meshing, they _would_ clamp tightly.

"not quite meshing"??? so how come i can so unerringly and completely
manage to get such consistently comlete repetition over 100
insertions??? am i secretly using some precision alignment device that
micro-rotates my axle to an exactly pre-determined position as i throw
it in from the trunk of my car?

Quoted message said:

The
operator could feel it.

But if there were some slight movement, the nearly-meshed serrations
could come into mesh, with the "pegs" settling into the "holes." That
would cause a loss of tension.

absolutely there's movement. that's how it shows up in the same indent
pattern every time!!! but there's no logical connection with your
assumption that the operator can't feel whether it's clamping tight from
the lever resistance. and that's why you re-tighten if you can feel
it's not clamping tight enough.

Quoted message said:

Quoted message said:

pictorial evidence shows no attrition that
would be caused by the description you're desperately trying to
artificially manufacture.

One picture shows something different, and we don't know what makes it
different. My dropouts don't show any serrations at all. (Sorry, no
digital camera lives here.) By my count, you're one-for-two.

how damned convenient. and you /dare/ you trot out a smooth faced
non-disk axle in this debate? you got chutzpah frank.

Quoted message said:

Now I'd like an explanation of how your QR beats the obvious two-to-one
odds against landing precisely in the same grooves.

"obvious" assumptions trying to stack the odds against the photographic
evidence you can't/[won't] try to refute with your own? yeah, right.

Quoted message said:

Seems to me the
chance of that happening every time by random chance over 100
installations is (1/3)^100 or about 2E-48.

- Frank Krygowski

you're really grasping for those straws there frank - keep it up! if
one set of facts don't fit, why, just make some others up! or maybe i'm
just very very good at photoshop. why don't you try that red herring
for a lark and make that accusation? i'm sure you'll get someone to
corroborate that fabrication of that photo is technically feasible.

Quoted message said:

Phil said:

Jasper Janssen said:

So would you have preferred a front wheel ejection for exactly the
same mistake?

My dropouts don't have lawyer lips.

Phil, that did not answer the question.

He didn't ask a relevant question.

Quoted message said:

Let me ask another. What would _you_ think about riding a
conventional-brake road bike whose front fork dropout slots faced
upward? Would you call that a good design?

Sure, if I knew it was time-tested, worked well in my experience, and didn't
have glaringly obvious problems that affected the vast majority of riders.
Sounds like QR disc brakes are a good design. A running design mod is okay
by me, too.

If it were up to me, I would redesign the quick-release to make it
idiot-proof, not the disc brake. But it looks like we'll have to settle for
forward-facing dropouts.

--
Phil, Squid-in-Training

Squid-in-Training said:

If it were up to me, I would redesign the quick-release to make it
idiot-proof, not the disc brake. But it looks like we'll have to settle for
forward-facing dropouts.

Feel free to elaborate on *how* to idiotproof QRs.

Jasper