Re: Pedal-crank interface, fretting, and gouging.

In article <[email hidden]>,

Quoted message said:

So? As I said, the damage is not only the face of the crank, but
fretting in the threads. Besides, the washers on TA cranks I have
seen squish out into a slight cup shape beyond the diameter of the
spindle face and do not protect the crank face. The only difference
is that the crank face is already countersunk when new.

I do find myself wondering though, if the washers may still
have some marginal value to them. I would expect that having an aluminum
washer fretting against the crank face would be better for the crank
than a hard steel spindle face doing the fretting - the softer washer
would sacrificially take its share of erosion rather than let the
crank face take it all.

Also, I'd expect any cracks that intiated at the washer/spindle
interface, even if they propogate through the washer, to not make it
through the washer/crank interface - this reminiscent of the
difference in crack behaviour between fibreglass composites and solid glass.

Of course, as pointed out, washers still do absolutely nothing
to mitigate fretting in the threads.

That TA offers these washers, and the recess for them would
indicate they at least acknowledge there's an issue here. Perhaps they
would be more receptive to suggestions here than Shimano. Too bad they
don't have much market share here in the US.

-Luns

Quoted message said:

Chalo Colina said:

I have not tried pedal washers, because I have yet to crack a pedal
eye. Of course, I don't use feeble spindly cranks either.

Quoted message said:

In my opinion, a man your size who puts in as many miles as you do
should probably be riding tubular steel cranks.

From what you have written, you weigh a substantial bit more than I.

And from what you have written, you have cracked a lot of pedal eyes.
That seems to me like a strong indication that your cranks are a poor
choice for you. There are stronger cranks available that won't crack
so readily.

Chalo Colina

Dans le message de news:[email hidden],
[email hidden] <[email hidden]> a
réfléchi, et puis a déclaré :

Quoted message said:

I don't have much hope for that. Having discussed this with many
crank folks at InterBike and receiving either blank stares from
engineers who don't understand what I am demonstrating and explaining
to the ones who don't think anyone knows more about cranks than they
and have never seen a crank failure like this.

This is typical for the bicycle industry as I have experienced it over
the last decades. Shimano is an extreme case.

How much of a cone angle would it take to make the change you suggest ? Is
it radical, or mild, to achieve the effect you are aiming at ?
--
Bonne route,

Sandy
Verneuil-sur-Seine FR

Quoted message said:

Luns Tee said:

Quoted message said:

So? As I said, the damage is not only the face of the crank, but
fretting in the threads. Besides, the washers on TA cranks I have
seen squish out into a slight cup shape beyond the diameter of the
spindle face and do not protect the crank face. The only
difference is that the crank face is already countersunk when new.

Quoted message said:

I do find myself wondering though, if the washers may still have
some marginal value to them. I would expect that having an aluminum
washer fretting against the crank face would be better for the crank
than a hard steel spindle face doing the fretting - the softer
washer would sacrificially take its share of erosion rather than let
the crank face take it all.

Quoted message said:

Also, I'd expect any cracks that initiated at the washer/spindle
interface, even if they propagate through the washer, to not make it
through the washer/crank interface - this reminiscent of the
difference in crack behaviour between fibreglass composites and
solid glass.

Quoted message said:

Of course, as pointed out, washers still do absolutely nothing to
mitigate fretting in the threads.

Quoted message said:

That TA offers these washers, and the recess for them would indicate
they at least acknowledge there's an issue here. Perhaps they would
be more receptive to suggestions here than Shimano. Too bad they
don't have much market share here in the US.

I don't have much hope for that. Having discussed this with many
crank folks at InterBike and receiving either blank stares from
engineers who don't understand what I am demonstrating and explaining
to the ones who don't think anyone knows more about cranks than they
and have never seen a crank failure like this.

This is typical for the bicycle industry as I have experienced it over
the last decades. Shimano is an extreme case.

[email hidden]

what's this repeated assertion that fretting causes fatigue? you've
been saying that for years jobst, but just like "stress relief" it's
merely your assumption based on a somewhat limited understanding of
fatigue initiation. how on earth do you expect shimano to take you
seriously when, aside from the fact that you really do have some
substantial personality issues, you publish "engineering" material that
is so consistently & laughably incorrect? your "stress relief" theory
is based on assumptions about strain aging that don't exist in stainless
steels. such fundamental errors are just laughable. the fact that the
contradictions are published in your own book, and that you /still/
don't get the picture is just unbelievable. your only chance at any
credibility [if indeed that were to be possible] is to go back & correct
/all/ the serious errors in your publications, both dead tree & "faq",
then start being more civil to people. you could also try going to a
library occasionally.

your assumption that shimano are incompetent and incapable of employing
anything less than top flight engineers & metallurgists, ones that /do/
understand how to talk with each other [something clearly lacking in
certain quarters this side of the pacific] is simply ridiculous. the
fact that they spend their time addressing issues of actual concern
rather than getting side tracked with red herrings is a feather in their
cap, not a cause for your misplaced criticism.

Chalo Colina said:

Quoted message said:

Quoted message said:

I have not tried pedal washers, because I have yet to crack a
pedal eye. Of course, I don't use feeble spindly cranks either.

Quoted message said:

Quoted message said:

Quoted message said:

In my opinion, a man your size who puts in as many miles as you do
should probably be riding tubular steel cranks.

I suppose you haven't seen the failure pictures of such cranks but
they don't give me any hope for more durability. They look more like
tin cans that were torn apart above the pedal eye. I don't have any
pictures but have seen them in the past. What brand do you have in
mind?

Quoted message said:

Quoted message said:

From what you have written, you weigh a substantial bit more than I.

Quoted message said:

And from what you have written, you have cracked a lot of pedal
eyes. That seems to me like a strong indication that your cranks
are a poor choice for you. There are stronger cranks available that
won't crack so readily.

Please don't keep it a secret. I have used Campagnolo Record and
Shimano Dura Ace, the latter having cracked at the pedal eye and in
the square taper. What cranks do you believe are more durable?

Meanwhile let's not overlook the failure mode, that of bending off at
the bottom of the stroke. You'll note that most of the pictures we
have seen and all the cranks I have broken show failure wave
propagation across the crank from outside to inside in bending mode.

The latest Ritchey failure posted here was the same even though it was
one of the "modern" shaped extra wide semi-channel shaped profiles.

[email hidden]

Sandy Leurre said:

Quoted message said:

I don't have much hope for that. Having discussed this with many
crank folks at InterBike and receiving either blank stares from
engineers who don't understand what I am demonstrating and
explaining to the ones who don't think anyone knows more about
cranks than they and have never seen a crank failure like this.

Quoted message said:

Quoted message said:

This is typical for the bicycle industry as I have experienced it
over the last decades. Shimano is an extreme case.

Quoted message said:

How much of a cone angle would it take to make the change you
suggest? Is it radical, or mild, to achieve the effect you are
aiming at?

http://pardo.net/bike/pic/fail-001/FAIL-016.html

The effect is the same as that of conical lug nuts on car wheels, that
of centering and load carried in compression rather than shear. It
prevents fretting motions that cause wear and unscrewing of right hand
threads on left pedals or the converse for right pedals... that is to
say, left hand threads for left pedals would not be required.

[email hidden]

Quoted message said:

Chalo Colina said:

Quoted message said:

> I have not tried pedal washers, because I have yet to crack a
> pedal eye. Of course, I don't use feeble spindly cranks either.

Quoted message said:

Quoted message said:

> In my opinion, a man your size who puts in as many miles as you

do

Quoted message said:

Quoted message said:

Quoted message said:

> should probably be riding tubular steel cranks.

I suppose you haven't seen the failure pictures of such cranks but
they don't give me any hope for more durability. They look more like
tin cans that were torn apart above the pedal eye. I don't have any
pictures but have seen them in the past. What brand do you have in
mind?

I have put many hard miles on Bullseye cranks without the least
structural problem. I can't reasonably recommend them at this time,
though, as their bearings have undergone a design revision that i
suspect will not work in the long term.

Redline Flight cranks are impeccably made. They are offered in lengths
up to 190mm. To use them in a road bike, you'd have to fit it with a
threaded BB like the one from FSA.
http://www.danscomp.com/cgi-bin/hazel.cgi?action=DETAIL&item=451060
http://fullspeedahead.com/fly.aspx?layout=product&taxid=92&pid=117

FSA and Premium make nice sturdy tubular cranks that are not
offensively heavy, but they use an interference spline fit that I find
inferior to a pinch-bolted design.
http://fullspeedahead.com/fly.aspx?layout=product&taxid=88&pid=258
http://www.premiumbmx.com/cranks.html

Quoted message said:

Quoted message said:

There are stronger cranks available that
won't crack so readily.

Please don't keep it a secret. I have used Campagnolo Record and
Shimano Dura Ace, the latter having cracked at the pedal eye and in
the square taper. What cranks do you believe are more durable?

I use Primo Powerbite cranks on most of my bikes. They use cold forged
7075 aluminum arms pinch-bolted to a 22mm spindle made of heat treated
4140 chromoly. There is a lighter titanium spindle available for them,
and they are offered with a threaded BB.
http://www.danscomp.com/cgi-bin/hazel.cgi?action=DETAIL&item=451057

Quoted message said:

Meanwhile let's not overlook the failure mode, that of bending off at
the bottom of the stroke. You'll note that most of the pictures we
have seen and all the cranks I have broken show failure wave
propagation across the crank from outside to inside in bending mode

All the cranks I have mentioned are much thicker through the arm at the
pedal boss than typical road bike cranks. They are also made of higher
yield stress materials than road bike cranks. It's no mystery why they
hold up better with heavy use.

Chalo Colina

On Tue, 17 May 2005 16:37:18 GMT,
[email hidden] wrote:

[snip]

Quoted message said:

I suppose you haven't seen the failure pictures of such [tubular
steel] cranks but >they don't give me any hope for more durability.
They look more like >tin cans that were torn apart above the pedal
eye. I don't have any pictures but have seen them in the past.

[snip]

Dear Jobst,

While not pedal eye failures, these are pictures of tubular
cranks that failed:

A failed Bullseye crank:

http://pardo.net/bike/pic/fail-001/FAIL-002.html

Three pictures of a failed Tioga Revolver crank:

http://pardo.net/bike/pic/fail-001/FAIL-003.html

(The third picture on the page above shows the "tin cans
that were torn apart" notion, but that's merely what to
expect from such hollow failures--a solid crank looks like a
piece of snapped chalk when it breaks.)

Carl Fogel

Quoted message said:

Sandy Leurre said:

Quoted message said:

I don't have much hope for that. Having discussed this with many
crank folks at InterBike and receiving either blank stares from
engineers who don't understand what I am demonstrating and
explaining to the ones who don't think anyone knows more about
cranks than they and have never seen a crank failure like this.

Quoted message said:

Quoted message said:

This is typical for the bicycle industry as I have experienced it
over the last decades. Shimano is an extreme case.

Quoted message said:

How much of a cone angle would it take to make the change you
suggest? Is it radical, or mild, to achieve the effect you are
aiming at?

http://pardo.net/bike/pic/fail-001/FAIL-016.html

The effect is the same as that of conical lug nuts on car wheels, that
of centering and load carried in compression rather than shear. It
prevents fretting motions that cause wear and unscrewing of right hand
threads on left pedals or the converse for right pedals... that is to
say, left hand threads for left pedals would not be required.

[email hidden]

so, based on your research, what effect does the interface angle have on
the effect you're trying to achieve?

jim beam said:

[email hidden] wrote:

[...]

Quoted message said:

Quoted message said:

prevents fretting motions that cause wear and unscrewing of right hand
threads on left pedals or the converse for right pedals... that is to
say, left hand threads for left pedals would not be required.

so, based on your research, what effect does the interface angle have on
the effect you're trying to achieve?

For a given bolt tension, the torque required to unscrew the joint is

tau1 = mu*T*R/cos(theta)

where

T = bolt tension
tau = required torque to unscrew
mu = coefficient of friction (steel to aluminum)
R = average radius of taper
theta = angle of taper (0 is the usual)

With theta = pi/4 (Jobst's modification) the relative torque
is increased from 1 to sqrt(2) ~ 41%.

Suppose we model the interface as a ball joint and then compute the
torque required to make the joint rotate in the plane in which the
spindle lies. Note that this torque is orthogonal to the unscrewing
torque. The required torque is then

tau2 = 2*mu*T*R/sin(2*theta)

This is a minimum at theta = pi/4! However, this model is too
simplistic. It predicts an infinite torque is required with a
standard spindle (theta = 0) because the lever arm goes to infinity.

Joe Riel

Joe Riel said:

jim beam said:

[email hidden] wrote:

[...]

Quoted message said:

Quoted message said:

prevents fretting motions that cause wear and unscrewing of right hand
threads on left pedals or the converse for right pedals... that is to
say, left hand threads for left pedals would not be required.

so, based on your research, what effect does the interface angle have on
the effect you're trying to achieve?

For a given bolt tension, the torque required to unscrew the joint is

[snip unscrewing equations]

Dear Joe,

Is the torque required to unscrew the joint the same as
fretting?

That is, do the pedals actually unscrew from the crank in
this situation and then screw back in, remaining in the same
place forever?

I'm not arguing, just not clear on the motion.

Is it rotary from the point of view of the pedal shaft, like
a bolt being turned by a wrench, with an X on the bolt head
rotating?

Or is it more a bending motion, like a fence-post set in
concrete being pushed back and forth, with an X on the top
of the fence post not rotating when viewed from above, but
moving off-center to one side?

Carl Fogel

Carl Fogel said:

Is the torque required to unscrew the joint the same as fretting?

Quoted message said:

That is, do the pedals actually unscrew from the crank in this
situation and then screw back in, remaining in the same place
forever?

The pedal spindle is displaced as you would expect a steel pin in a
stiff rubber crank bore. It skews and precesses but cannot tighten
farther than it is because it is bearing on the crank face with a flat
shoulder. As the crank face erodes from fretting, the pedal screws in
as much as this loss of length predicts plus the erosion on the thread
faces that hold the pedal in place.

This moving thread interface is about as dirty as you can imagine. As
many of us have found, the wear debris makes the pedal almost
un-removable and about these we read here off and on. That is why the
most benign removal is done by heating the crank on a gas flame until
it sizzles to the wet finger.

Quoted message said:

Is it rotary from the point of view of the pedal shaft, like a bolt
being turned by a wrench, with an X on the bolt head rotating?

It would like to but because it is already hard up against the crank
face, it cannot advance farther. The main motion is a radial offset
from the geometric center of spindle and crank thread.

Quoted message said:

Or is it more a bending motion, like a fence-post set in concrete
being pushed back and forth, with an X on the top of the fence post
not rotating when viewed from above, but moving off-center to one
side?

Since these are metals and metals are elastic, there is always some
bending taking place, but in this interface the radial displacement is
the main culprit.

[email hidden]

Quoted message said:

Carl Fogel said:

Is the torque required to unscrew the joint the same as fretting?

Quoted message said:

That is, do the pedals actually unscrew from the crank in this
situation and then screw back in, remaining in the same place
forever?

The pedal spindle is displaced as you would expect a steel pin in a
stiff rubber crank bore. It skews and precesses but cannot tighten
farther than it is because it is bearing on the crank face with a flat
shoulder. As the crank face erodes from fretting, the pedal screws in
as much as this loss of length predicts plus the erosion on the thread
faces that hold the pedal in place.

This moving thread interface is about as dirty as you can imagine. As
many of us have found, the wear debris makes the pedal almost
un-removable and about these we read here off and on. That is why the
most benign removal is done by heating the crank on a gas flame until
it sizzles to the wet finger.

Quoted message said:

Is it rotary from the point of view of the pedal shaft, like a bolt
being turned by a wrench, with an X on the bolt head rotating?

It would like to but because it is already hard up against the crank
face, it cannot advance farther. The main motion is a radial offset
from the geometric center of spindle and crank thread.

Quoted message said:

Or is it more a bending motion, like a fence-post set in concrete
being pushed back and forth, with an X on the top of the fence post
not rotating when viewed from above, but moving off-center to one
side?

Since these are metals and metals are elastic, there is always some
bending taking place, but in this interface the radial displacement is
the main culprit.

[email hidden]

Dear Jobst,

So it's more two pieces rattling against each other than a
rotary motion between the two faces.

(Plus the tiny bit of tightening of the thread that you
point out will occur as the faces wear away.)

I'll try to think of it as a microscopic version of a
rattlesnake's rattle, where the buttons rock back and forth
at the joints, but don't twist relative to each other.

Thanks for explaining it (again),

Carl Fogel

Quoted message said:

Is the torque required to unscrew the joint the same as
fretting?

That is, do the pedals actually unscrew from the crank in
this situation and then screw back in, remaining in the same
place forever?

Good question. I don't claim to understand what is going on,
just making a few computations based on a simplistic model.

Given that with normal pedals there is no unscrewing torque
(the "precession" torque causes the spindles to tighten),
my first computation is pointless.

The more interesting computation is the second, where I make a
reckless attempt to estimate how Jobst's conical design may reduce the
presumed rocking/fretting motion caused by the cyclical load.
However, the model is clearly wrong, the fretting motion doesn't, I
think, cause a rocking motion that can be modeled as a ball joint, if
anything the the pivot is on the other side (i.e. through crankarm).
In which case the motion is more sliding/distorting. I'm not sure
that there is a simple way to model this.

Quoted message said:

Or is it more a bending motion, like a fence-post set in
concrete being pushed back and forth, with an X on the top
of the fence post not rotating when viewed from above, but
moving off-center to one side?

I believe that that is a reasonable approximation.

Joe

Quoted message said:

Since these are metals and metals are elastic, there is always some
bending taking place, but in this interface the radial displacement is
the main culprit.

So the idea behind the "chamfer" is to convert a radial displacement
into an axial displacement that increases the screw tension, hence
reduces the movement.

Joe

Joe Riel said:

Quoted message said:

Since these are metals and metals are elastic, there is always some
bending taking place, but in this interface the radial displacement
is the main culprit.

Quoted message said:

So the idea behind the "chamfer" is to convert a radial displacement
into an axial displacement that increases the screw tension, hence
reduces the movement.

Visualize the spindle as having a node in the middle of the tread
about which the spindle yaws in rotation. The outer end of the
spindle is not restrained from moving radially except by friction of
the plane shoulder of the spindle that is screwed tight against the
face of the crank.

With the conical seat, as with lug nuts on cars, radial loads can be
transmitted in compression between spindle and crank. A flat face
cannot do that nor can a cylindrical one without a radial preload
(press fit). That's why the square taper BB spindle is tapered. A
conical seat takes care of furnishing preload and carrying the canting
load in compression.

[email hidden]

Quoted message said:

Carl Fogel said:

Is the torque required to unscrew the joint the same as fretting?

Quoted message said:

That is, do the pedals actually unscrew from the crank in this
situation and then screw back in, remaining in the same place
forever?

The pedal spindle is displaced as you would expect a steel pin in a
stiff rubber crank bore. It skews and precesses but cannot tighten
farther than it is because it is bearing on the crank face with a flat
shoulder. As the crank face erodes from fretting, the pedal screws in
as much as this loss of length predicts plus the erosion on the thread
faces that hold the pedal in place.

How does the initial tightness of the pedal affect this situation?
From your description is sounds as if there will always be some wear
and further turning of the pedal in the crank. Does this lead all
pedal/cranks to eventually end up in the same state, or am I
misunderstanding something?

Joe Riel said:

Quoted message said:

Is the torque required to unscrew the joint the same as
fretting?

That is, do the pedals actually unscrew from the crank in
this situation and then screw back in, remaining in the same
place forever?

Good question. I don't claim to understand what is going on,
just making a few computations based on a simplistic model.

Given that with normal pedals there is no unscrewing torque
(the "precession" torque causes the spindles to tighten),
my first computation is pointless.

The more interesting computation is the second, where I make a
reckless attempt to estimate how Jobst's conical design may reduce the
presumed rocking/fretting motion caused by the cyclical load.
However, the model is clearly wrong, the fretting motion doesn't, I
think, cause a rocking motion that can be modeled as a ball joint, if
anything the the pivot is on the other side (i.e. through crankarm).
In which case the motion is more sliding/distorting. I'm not sure
that there is a simple way to model this.

the angled interface merely migrates the rocking pivot point from about
the middle of the spindle thread to the shamfer itself.

Quoted message said:

Quoted message said:

Or is it more a bending motion, like a fence-post set in
concrete being pushed back and forth, with an X on the top
of the fence post not rotating when viewed from above, but
moving off-center to one side?

I believe that that is a reasonable approximation.

Joe

Quoted message said:

The effect is the same as that of conical lug nuts on car wheels, that
of centering and load carried in compression rather than shear. It
prevents fretting motions that cause wear and unscrewing of right hand
threads on left pedals or the converse for right pedals... that is to
say, left hand threads for left pedals would not be required.

Would this then be a good mod to do on
regular-cranksets-converted-for-tandem-use?

Jasper