I fixed a broken spoke!

jim beam said:

spokes angle towards a rim at say 6 degrees both sides front, and say 6
+ 3 for rears. and that is what hub flanges are canted to. in
addition, hub holes are angled perpendicular to the flange, so they too
are aligned for optimal spoke position.

see:
http://www.flickr.com/photos/38636024@N00/316202144/
http://www.flickr.com/photos/38636024@N00/316202143/

I find your "pencil test" a bit unconvincing.

I have a new, never spoked, LX rear hub (Shimano SH-FHM570-32). I found
that a 8d finish nail is almost an interference fit in the spoke holes.
I can state with absolute certainty that the holes are drilled parallel
to the axle, not perpendicular to the flange.

Quoted message said:

After all, the most likely explanation for an occasional spoke
breaking is variation in the spoke material. The bending and
tensioning and so forth are quite regular. If squeezing spoke pairs
was solving some residual stress problem by raising tension, the same
problem would be causing all the other spokes to fail.

I see Jobst's comment in 1993 that spokes somehow became much more
durable as strong evidence that he was mistaken about how efffective
spoke-sqeezing is.

I think you're missing the point. Spoke failure is from fatigue. Fatigue
life is shortened both by poor material quality and residual stress.
Removing residual stress will make poor spokes and good spokes last
longer. That even modern, presumably better made, spokes can benefit
from stress relief is supported by the observation that spoke failures
still occur.

Quoted message said:

This illustrates how easily we fool ourselves and how easy it is to
make huge mistakes. I have never doubted the good faith of the last
such calculation that appeared on RBT. A 30 pound side force moved a
spoke a certain distance and bent the spoke at such and so an angle,
which in turn allowed calculations for steel of specifc thickness to
indicate that the spoke tension would rise approximately 150 pounds, a
5 to 1 ratio.

I think you're mis-remembering that thread. Phil Holman reported
measuring that 5:1 ratio on his wheel. After much posting, all concluded
that little could be improved in Jobst's description in his book. The
reason you found such a surprising (to yourself) ratio of 1:1 was
because your rims were not very stiff.

Quoted message said:

The next thing to test was the effect of tension on residual stress.

Quoted message said:

So I cobbled together a crude vise rig and started bending spokes into
U shapes.

A spoke bent but not stretched expands considerably when heated.

A spoke bent and tensioned slightly in the vise will expand noticeably
less.

A spoke bent and tensioned more will expand even less.

A spoke bent and tensioned to normal bicycle wheel levels (~200 lbs)
will hardly expand at all.

A spoke bent and tensioned to ~400 pounds will hardly expand at all,
just like the spoke tensioned to ~200 pounds.

I had trouble following this "experiment". I assume that after
tensioning the spokes had very different "starting angles" before heat
was applied? It's very difficult to know exactly what was going on,
specifically if the various changes in angle with heating have anything
to do with residual stresses at all.

Quoted message said:

My problem with the spoke-squeezing theory is that it seems to be a
lot of theory developed to support preconceived and self-flattering
conclusions about how we can cure the King's Evil by the laying on of
hands.

This method of relieving residual stress is well known and used in
various industrial processes.

On Wed, 20 Dec 2006 16:59:12 -0500, Peter Cole
<[email hidden]> wrote:

[snip]

Quoted message said:

I think you're mis-remembering that thread. Phil Holman reported
measuring that 5:1 ratio on his wheel. After much posting, all concluded
that little could be improved in Jobst's description in his book. The
reason you found such a surprising (to yourself) ratio of 1:1 was
because your rims were not very stiff.

[snip]

Dear Peter,

No, I tested various wheels, including a box-section Mavic MA3, and
found the same results. You must have missed those details.

Apparently, you also missed my more recent test, in which I hung
weights on a horizontal spoke tensioned in a massive pipe-clamp rig,
far stiffer than any bicycle wheel:

http://groups.google.com/group/rec.bicycles.tech/msg/bce241a75595fbed

A 60-pound weight produced only a 90-pound tension rise when hung from
the middle of horizontal spoke whose original tension was 235 pounds.

Of course, no bicycle rim will ever approach the stiffness of a pair
of steel clamps on a 3/4" steel pipe whose walls are thicker than the
spoke itself.

Again, the impressive bend angles seen when pairs of bicycle spokes
are squeezed are not signs of huge tension increases. The bends are
the result of the slack gained when the rim distorts into a faint N or
Z shape and the slack gained when the squeezed spoke presses down
against the unsqueezed spoke in a 3-cross patter.

Feel free to do some actual testing instead of just asserting that
stiffer rims will produce the results that you want.

The pipe clamp cost eight bucks, the spoke was about fifty cents. A
tension gauge costs about $50, but no one really interested in this
kind of thing can consider that a barrier.

Cheers,

Carl Fogel

Quoted message said:

On Wed, 20 Dec 2006 16:59:12 -0500, Peter Cole
<[email hidden]> wrote:

[snip]

Quoted message said:

I think you're mis-remembering that thread. Phil Holman reported
measuring that 5:1 ratio on his wheel. After much posting, all concluded
that little could be improved in Jobst's description in his book. The
reason you found such a surprising (to yourself) ratio of 1:1 was
because your rims were not very stiff.

[snip]

Dear Peter,

No, I tested various wheels, including a box-section Mavic MA3, and
found the same results. You must have missed those details.

You apparently assume the MA3 is a stiff rim. A box section does not
improve stiffness, as it adds material in the neutral plane (something
Jobst has pointed out before). Rim stiffness is pretty much a function
of cross section height.

Quoted message said:

Apparently, you also missed my more recent test, in which I hung
weights on a horizontal spoke tensioned in a massive pipe-clamp rig,
far stiffer than any bicycle wheel:

http://groups.google.com/group/rec.bicycles.tech/msg/bce241a75595fbed

A 60-pound weight produced only a 90-pound tension rise when hung from
the middle of horizontal spoke whose original tension was 235 pounds.

Of course, no bicycle rim will ever approach the stiffness of a pair
of steel clamps on a 3/4" steel pipe whose walls are thicker than the
spoke itself.

How can you be sure of the stiffness of your fixture without measuring
it? How can you be sure that your spoke mounting/support exactly
replicates a hub and rim? There are many ways for errors to creep in.

Quoted message said:

Again, the impressive bend angles seen when pairs of bicycle spokes
are squeezed are not signs of huge tension increases. The bends are
the result of the slack gained when the rim distorts into a faint N or
Z shape and the slack gained when the squeezed spoke presses down
against the unsqueezed spoke in a 3-cross patter.

As said so many times before, the spoke angle only reveals the ratio of
forces. To say you get low ratios and large angles is just repeating
yourself. The real question is why the angles are large. This can only
be from something in the path that is less stiff than the spoke itself.
Likely candidates are the rim and the spoke path/end support. You
generalize from your measurements on your wheels to predict that all of
us get less tension increases than you, despite at least one report to
the contrary.

This is, of course, a tempest in a teapot, since it makes little
difference what the precise amount of spoke overload is as long as it
falls within the rough bounds of being enough to improve the residual
stresses but not so much as to damage the rim.

Quoted message said:

Feel free to do some actual testing instead of just asserting that
stiffer rims will produce the results that you want.

I would be perfectly happy with 60 lbf that your worst case tests seem
to predict.

Peter Cole said:

jim beam said:

spokes angle towards a rim at say 6 degrees both sides front, and say
6 + 3 for rears. and that is what hub flanges are canted to. in
addition, hub holes are angled perpendicular to the flange, so they
too are aligned for optimal spoke position.

see:
http://www.flickr.com/photos/38636024@N00/316202144/
http://www.flickr.com/photos/38636024@N00/316202143/

I find your "pencil test" a bit unconvincing.

I have a new, never spoked, LX rear hub (Shimano SH-FHM570-32). I found
that a 8d finish nail is almost an interference fit in the spoke holes.
I can state with absolute certainty that the holes are drilled parallel
to the axle, not perpendicular to the flange.

ok, i accept that. maybe it's a high flange vs. low flange thing.
maybe it's a quality level thing - mine was xt. but my hub /definitely/
has holes perpendicular to the canted flange.

but all this is dodging the fundamental questions:

1. does or does not hub hole conformation lessen the required spoke angle?
2. does or does not "correcting the spoke line" /prior/ to hub hole
conformation exaggerate the "correction", i.e. make it excessive?
3. how does elastic stress from an undeformed spoke bending toward the
rim compare to that where the spoke bends around its interleaving
partner? from what i can see, the angle is very similar so skin stress
has to be of similar magnitude.

Peter Cole said:

Quoted message said:

After all, the most likely explanation for an occasional spoke
breaking is variation in the spoke material. The bending and
tensioning and so forth are quite regular. If squeezing spoke pairs
was solving some residual stress problem by raising tension, the same
problem would be causing all the other spokes to fail.

I see Jobst's comment in 1993 that spokes somehow became much more
durable as strong evidence that he was mistaken about how efffective
spoke-sqeezing is.

I think you're missing the point. Spoke failure is from fatigue. Fatigue
life is shortened both by poor material quality and residual stress.
Removing residual stress will make poor spokes and good spokes last
longer. That even modern, presumably better made, spokes can benefit
from stress relief is supported by the observation that spoke failures
still occur.

sure they do. and it's almost always inferior no-name spokes. one
of my coworkers was breaking spokes every week earlier this year and
gave me the broken ones to examine. sure enough, no-name and serious
surface quality issues. all he needed was a rebuild with a quality
brand spoke and his breakage issues have been eliminated.

Quoted message said:

Quoted message said:

This illustrates how easily we fool ourselves and how easy it is to
make huge mistakes. I have never doubted the good faith of the last
such calculation that appeared on RBT. A 30 pound side force moved a
spoke a certain distance and bent the spoke at such and so an angle,
which in turn allowed calculations for steel of specifc thickness to
indicate that the spoke tension would rise approximately 150 pounds, a
5 to 1 ratio.

I think you're mis-remembering that thread. Phil Holman reported
measuring that 5:1 ratio on his wheel. After much posting, all concluded
that little could be improved in Jobst's description in his book. The
reason you found such a surprising (to yourself) ratio of 1:1 was
because your rims were not very stiff.

Quoted message said:

The next thing to test was the effect of tension on residual stress.

Quoted message said:

So I cobbled together a crude vise rig and started bending spokes into
U shapes.

A spoke bent but not stretched expands considerably when heated.

A spoke bent and tensioned slightly in the vise will expand noticeably
less.

A spoke bent and tensioned more will expand even less.

A spoke bent and tensioned to normal bicycle wheel levels (~200 lbs)
will hardly expand at all.

A spoke bent and tensioned to ~400 pounds will hardly expand at all,
just like the spoke tensioned to ~200 pounds.

I had trouble following this "experiment". I assume that after
tensioning the spokes had very different "starting angles" before heat
was applied? It's very difficult to know exactly what was going on,
specifically if the various changes in angle with heating have anything
to do with residual stresses at all.

absolutely they do! that's why frames need to be re-set after heat
treatment for example.

Quoted message said:

Quoted message said:

My problem with the spoke-squeezing theory is that it seems to be a
lot of theory developed to support preconceived and self-flattering
conclusions about how we can cure the King's Evil by the laying on of
hands.

This method of relieving residual stress is well known and used in
various industrial processes.

indeed. but with a quality material without residual stress, it's
pointless and can initiate fatigue, not eliminate it. when's the last
time you saw a bridge being "stress relieved"?

jim beam said:

Peter Cole wrote:

Quoted message said:

Quoted message said:

I find your "pencil test" a bit unconvincing.

I have a new, never spoked, LX rear hub (Shimano SH-FHM570-32). I
found that a 8d finish nail is almost an interference fit in the spoke
holes. I can state with absolute certainty that the holes are drilled
parallel to the axle, not perpendicular to the flange.

ok, i accept that. maybe it's a high flange vs. low flange thing.
maybe it's a quality level thing - mine was xt. but my hub /definitely/
has holes perpendicular to the canted flange.

but all this is dodging the fundamental questions:

If you say so, but I'd still like to see evidence of a hub drilled
perpendicular to the flange.

Quoted message said:

1. does or does not hub hole conformation lessen the required spoke angle?
2. does or does not "correcting the spoke line" /prior/ to hub hole
conformation exaggerate the "correction", i.e. make it excessive?
3. how does elastic stress from an undeformed spoke bending toward the
rim compare to that where the spoke bends around its interleaving
partner? from what i can see, the angle is very similar so skin stress
has to be of similar magnitude.

First, I'm not aware of anyone (other than Sheldon, & I don't agree with
him) correcting the spoke line before the wheel is tensioned, so I'm not
sure who you're arguing with.

In all of the above cases, the only important thing is that the spoke
take as straight a line as possible. If the spoke has a bend in it, the
spoke will flex under change of tension, accelerating fatigue. If, after
the wheel is tensioned, the spokes have a bow in them where they exit
the flange, it is necessary to correct the line to prevent flex in
operation. If they don't, it isn't. The elbow is a critical fatigue
location, the crossover is not. The only spokes I have ever seen break
at places other than the elbow or threads were chain-nicked.

jim beam said:

Peter Cole wrote:

Quoted message said:

Quoted message said:

I think you're missing the point. Spoke failure is from fatigue.
Fatigue life is shortened both by poor material quality and residual
stress. Removing residual stress will make poor spokes and good spokes
last longer. That even modern, presumably better made, spokes can
benefit from stress relief is supported by the observation that spoke
failures still occur.

sure they do. and it's almost always inferior no-name spokes. one
of my coworkers was breaking spokes every week earlier this year and
gave me the broken ones to examine. sure enough, no-name and serious
surface quality issues. all he needed was a rebuild with a quality
brand spoke and his breakage issues have been eliminated.

I have had a few DT spokes fail (at elbows) before I began stress
relieving, none since.

Quoted message said:

Quoted message said:

I had trouble following this "experiment". I assume that after
tensioning the spokes had very different "starting angles" before heat
was applied? It's very difficult to know exactly what was going on,
specifically if the various changes in angle with heating have
anything to do with residual stresses at all.

absolutely they do! that's why frames need to be re-set after heat
treatment for example.

Carl didn't "heat treat". Thermal warping has many causes. If even the
direction (never mind the magnitude) revealed residual stresses, why did
carbon steel spokes bend the other way?

Quoted message said:

Quoted message said:

This method of relieving residual stress is well known and used in
various industrial processes.

indeed. but with a quality material without residual stress, it's
pointless and can initiate fatigue, not eliminate it.

You assume that spokes are delivered w/o residual stresses and that none
are introduced in wheel building. The only way the first could be true
is if the manufacturer annealed the spokes -- I have seen no claims of
this -- you? If, after unlacing a tensioned wheel, spoke elbow angles
are no longer equal, there must be residual stresses introduced.

Quoted message said:

when's the last
time you saw a bridge being "stress relieved"?

Perhaps not bridges, but I posted a link to an overload stress relief
procedure required by the Coast Guard for large pressure tanks. At least
some companies market equipment for vibrational stress relief of large
structures. The principle is the same. I don't know what parts of a
bridge you think might need stress relief, but for known problems like
welds, mitigation is done.

Peter Cole said:

jim beam said:

spokes angle towards a rim at say 6 degrees both sides front, and say 6
+ 3 for rears. and that is what hub flanges are canted to. in
addition, hub holes are angled perpendicular to the flange, so they too
are aligned for optimal spoke position.

see:
http://www.flickr.com/photos/38636024@N00/316202144/
http://www.flickr.com/photos/38636024@N00/316202143/

I find your "pencil test" a bit unconvincing.

I have a new, never spoked, LX rear hub (Shimano SH-FHM570-32). I found
that a 8d finish nail is almost an interference fit in the spoke holes.
I can state with absolute certainty that the holes are drilled parallel
to the axle, not perpendicular to the flange.

Peter,
On your hub, are the inner faces of the flanges perpendicular to the axle,
or are they canted at the same angle as the outer faces (that is, are they
parallel to the outer faces)?

On my XT rear hub, they are perpendicular to the axle. I suspect that's a
compromise made because of the difficulty of making a concave surface in a
forged hub, at least without subsequent machining.

The reason I ask is that if some of the commonest hubs out there don't
have optimized inner faces, then jim beam's theory that spoke lines don't
need adjusted doesn't hold water.

Gary Young said:

The reason I ask is that if some of the commonest hubs out there don't
have optimized inner faces, then jim beam's theory that spoke lines don't
need adjusted doesn't hold water.

jim beam has admitted that his theory is limited to special cases. It's
so limited as to have little application in real life.

--
Dave
dvt at psu dot edu

Peter Cole said:

jim beam said:

Peter Cole wrote:

Quoted message said:

Quoted message said:

I find your "pencil test" a bit unconvincing.

I have a new, never spoked, LX rear hub (Shimano SH-FHM570-32). I
found that a 8d finish nail is almost an interference fit in the
spoke holes. I can state with absolute certainty that the holes are
drilled parallel to the axle, not perpendicular to the flange.

ok, i accept that. maybe it's a high flange vs. low flange thing.
maybe it's a quality level thing - mine was xt. but my hub /definitely/
has holes perpendicular to the canted flange.

but all this is dodging the fundamental questions:

If you say so, but I'd still like to see evidence of a hub drilled
perpendicular to the flange.

i gave you a pic peter! the pencil may not have been a wise choice of
indicator for a hardened disbeliever like you, but i assure you, it's
jammed in so it's perpendicular to the hole - the wood deforms to a
cylinder that perfectly fits and is square to the pencil axis on removal.

Quoted message said:

Quoted message said:

1. does or does not hub hole conformation lessen the required spoke
angle?
2. does or does not "correcting the spoke line" /prior/ to hub hole
conformation exaggerate the "correction", i.e. make it excessive?
3. how does elastic stress from an undeformed spoke bending toward the
rim compare to that where the spoke bends around its interleaving
partner? from what i can see, the angle is very similar so skin stress
has to be of similar magnitude.

First, I'm not aware of anyone (other than Sheldon, & I don't agree with
him) correcting the spoke line before the wheel is tensioned, so I'm not
sure who you're arguing with.

jobst brandt!!! its specifically stated in "the book"!

Quoted message said:

In all of the above cases, the only important thing is that the spoke
take as straight a line as possible. If the spoke has a bend in it, the
spoke will flex under change of tension, accelerating fatigue.

no, the spoke elbow is fundamentally going to flex regardless. the
elbow's offset means load is non-axial, hence the elbow /will/ bend.
and it is /that/ which causes fatigue, not the ethereal dragon that is
alleged to be "stress relief".

Quoted message said:

If, after
the wheel is tensioned, the spokes have a bow in them where they exit
the flange, it is necessary to correct the line to prevent flex in
operation.

but if that's true, it should also include the necessity to "correct"
the bow at interleaving crossover!

Quoted message said:

If they don't, it isn't. The elbow is a critical fatigue
location, the crossover is not.

agreed. but we may have different reasoning for that. to me, the
crossover is subcritical because the straightening of the loading cycle
runs /with/ it's natural shape and is /not/ fighting an offset like the
elbow. additionally, it's not subject to the surface defects that
inevitably accompany a bending operation like elbow forming. "orange
peel effect".

Quoted message said:

The only spokes I have ever seen break
at places other than the elbow or threads were chain-nicked.

indeed. and they are surface defects.

Gary Young said:

Peter Cole said:

jim beam said:

spokes angle towards a rim at say 6 degrees both sides front, and say 6
+ 3 for rears. and that is what hub flanges are canted to. in
addition, hub holes are angled perpendicular to the flange, so they too
are aligned for optimal spoke position.

see:
http://www.flickr.com/photos/38636024@N00/316202144/
http://www.flickr.com/photos/38636024@N00/316202143/

I find your "pencil test" a bit unconvincing.

I have a new, never spoked, LX rear hub (Shimano SH-FHM570-32). I found
that a 8d finish nail is almost an interference fit in the spoke holes.
I can state with absolute certainty that the holes are drilled parallel
to the axle, not perpendicular to the flange.

Peter,
On your hub, are the inner faces of the flanges perpendicular to the axle,
or are they canted at the same angle as the outer faces (that is, are they
parallel to the outer faces)?

On my XT rear hub, they are perpendicular to the axle. I suspect that's a
compromise made because of the difficulty of making a concave surface in a
forged hub, at least without subsequent machining.

The reason I ask is that if some of the commonest hubs out there don't
have optimized inner faces, then jim beam's theory that spoke lines don't
need adjusted doesn't hold water.

The inner faces of the flanges on my LX hub are perpendicular to the
axle, yes.

Peter Cole said:

jim beam said:

Peter Cole wrote:

Quoted message said:

Quoted message said:

I think you're missing the point. Spoke failure is from fatigue.
Fatigue life is shortened both by poor material quality and residual
stress. Removing residual stress will make poor spokes and good
spokes last longer. That even modern, presumably better made, spokes
can benefit from stress relief is supported by the observation that
spoke failures still occur.

sure they do. and it's almost always inferior no-name spokes. one
of my coworkers was breaking spokes every week earlier this year and
gave me the broken ones to examine. sure enough, no-name and serious
surface quality issues. all he needed was a rebuild with a quality
brand spoke and his breakage issues have been eliminated.

I have had a few DT spokes fail (at elbows) before I began stress
relieving, none since.

were they from the famous "long shank" era? they broke rapidly because
of geometry issues.

Quoted message said:

Quoted message said:

Quoted message said:

I had trouble following this "experiment". I assume that after
tensioning the spokes had very different "starting angles" before
heat was applied? It's very difficult to know exactly what was going
on, specifically if the various changes in angle with heating have
anything to do with residual stresses at all.

absolutely they do! that's why frames need to be re-set after heat
treatment for example.

Carl didn't "heat treat".

no. but he demonstrated the effect of residual stress.

Quoted message said:

Thermal warping has many causes.

not many at those temperatures and that quickly. #1 candidate is
thermal stress relief. nothing else is close.

Quoted message said:

If even the
direction (never mind the magnitude) revealed residual stresses, why did
carbon steel spokes bend the other way?

it's undergoing a phase change at the same time. austenitic stainless
is not.

Quoted message said:

Quoted message said:

Quoted message said:

This method of relieving residual stress is well known and used in
various industrial processes.

indeed. but with a quality material without residual stress, it's
pointless and can initiate fatigue, not eliminate it.

You assume that spokes are delivered w/o residual stresses and that none
are introduced in wheel building. The only way the first could be true
is if the manufacturer annealed the spokes -- I have seen no claims of
this -- you?

with respect, the years of argument on this topic have all been
concerned with the efficacy of mechanical stress relief. why then
eliminate it and say that only thermal is possible?

Quoted message said:

If, after unlacing a tensioned wheel, spoke elbow angles
are no longer equal, there must be residual stresses introduced.

fogel has demonstrated that mechanical relief occurs are low levels.
mere bending is not evidence of residual stress.

Quoted message said:

Quoted message said:

when's the last
time you saw a bridge being "stress relieved"?

Perhaps not bridges, but I posted a link to an overload stress relief
procedure required by the Coast Guard for large pressure tanks.

and like i told you, i do not accept that explanation. every fracture
mechanics student will tell you that that procedure is done to reveal
crack flaws. there is no evidence that it performs mechanical relief,
and that's not the true objective. i can only assume that piece was
written by a well intentioned but mistaken intern.

Quoted message said:

At least
some companies market equipment for vibrational stress relief of large
structures.

indeed. and they precipitate yielding. but weren[t you just arguing
that stress relief has to be thermal?

Quoted message said:

The principle is the same. I don't know what parts of a
bridge you think might need stress relief, but for known problems like
welds, mitigation is done.

yes, it's thermal. elevated temperature allows hydrogen to diffuse out
before welds crack.

jim beam said:

Peter Cole said:

jim beam said:

Peter Cole wrote:

Quoted message said:

> I find your "pencil test" a bit unconvincing.
>
> I have a new, never spoked, LX rear hub (Shimano SH-FHM570-32). I
> found that a 8d finish nail is almost an interference fit in the
> spoke holes. I can state with absolute certainty that the holes are
> drilled parallel to the axle, not perpendicular to the flange.

ok, i accept that. maybe it's a high flange vs. low flange thing.
maybe it's a quality level thing - mine was xt. but my hub /definitely/
has holes perpendicular to the canted flange.

but all this is dodging the fundamental questions:

If you say so, but I'd still like to see evidence of a hub drilled
perpendicular to the flange.

i gave you a pic peter! the pencil may not have been a wise choice of
indicator for a hardened disbeliever like you, but i assure you, it's
jammed in so it's perpendicular to the hole - the wood deforms to a
cylinder that perfectly fits and is square to the pencil axis on removal.

I tried a pencil first, too, and I could believe that the hole was
canted. I tried various other things that weren't a tight fit and found
them equally ambiguous. It wasn't until I found something with an
interference fit that I became convinced that the holes were not canted.

Quoted message said:

Quoted message said:

Quoted message said:

1. does or does not hub hole conformation lessen the required spoke
angle?
2. does or does not "correcting the spoke line" /prior/ to hub hole
conformation exaggerate the "correction", i.e. make it excessive?
3. how does elastic stress from an undeformed spoke bending toward the
rim compare to that where the spoke bends around its interleaving
partner? from what i can see, the angle is very similar so skin stress
has to be of similar magnitude.

First, I'm not aware of anyone (other than Sheldon, & I don't agree
with him) correcting the spoke line before the wheel is tensioned, so
I'm not sure who you're arguing with.

jobst brandt!!! its specifically stated in "the book"!

Not from my reading (which I quoted earlier). The step is described
after the section on tensioning. Furthermore, he expressed it in terms
of the spoke "may" need correction.

If a wheel is brought up to tension, and the spokes still have an overly
obtuse angle (as evidenced by bowing out from the flange), I think that
needs correction -- the spoke path is not in line with the load path.

It's Mavic who seem to propose stress relieving a partially tensioned
wheel -- for reasons I don't understand (nor do they, I'm willing to bet).

If you claim that all failed elbows you've examined failed from the
outside (as BenC says), then I don't understand why you're concerned
about an over acute elbow angle. That would put extra tension on the
inside of the bend. Perhaps all your spokes fail on the outside because
you haven't corrected the spoke line an have an over obtuse angle.

Quoted message said:

Quoted message said:

In all of the above cases, the only important thing is that the spoke
take as straight a line as possible. If the spoke has a bend in it,
the spoke will flex under change of tension, accelerating fatigue.

no, the spoke elbow is fundamentally going to flex regardless. the
elbow's offset means load is non-axial, hence the elbow /will/ bend. and
it is /that/ which causes fatigue, not the ethereal dragon that is
alleged to be "stress relief".

If the spoke line matches the load line the spoke won't bend with
changes in load.

Quoted message said:

Quoted message said:

If, after the wheel is tensioned, the spokes have a bow in them where
they exit the flange, it is necessary to correct the line to prevent
flex in operation.

but if that's true, it should also include the necessity to "correct"
the bow at interleaving crossover!

Since spokes have finite thickness, the spoke line has to deviate
slightly from the load line, there's no way to correct for that.

Quoted message said:

Quoted message said:

If they don't, it isn't. The elbow is a critical fatigue location, the
crossover is not.

agreed. but we may have different reasoning for that. to me, the
crossover is subcritical because the straightening of the loading cycle
runs /with/ it's natural shape and is /not/ fighting an offset like the
elbow. additionally, it's not subject to the surface defects that
inevitably accompany a bending operation like elbow forming. "orange
peel effect".

I think it's because the bend angle is much smaller and the spokes
support each other and prevent much bending.

Peter Cole said:

jim beam said:

Peter Cole said:

jim beam wrote:
> Peter Cole wrote:

>> I find your "pencil test" a bit unconvincing.
>>
>> I have a new, never spoked, LX rear hub (Shimano SH-FHM570-32). I
>> found that a 8d finish nail is almost an interference fit in the
>> spoke holes. I can state with absolute certainty that the holes are
>> drilled parallel to the axle, not perpendicular to the flange.
>
> ok, i accept that. maybe it's a high flange vs. low flange thing.
> maybe it's a quality level thing - mine was xt. but my hub
> /definitely/
> has holes perpendicular to the canted flange.
>
> but all this is dodging the fundamental questions:

If you say so, but I'd still like to see evidence of a hub drilled
perpendicular to the flange.

i gave you a pic peter! the pencil may not have been a wise choice of
indicator for a hardened disbeliever like you, but i assure you, it's
jammed in so it's perpendicular to the hole - the wood deforms to a
cylinder that perfectly fits and is square to the pencil axis on removal.

I tried a pencil first, too, and I could believe that the hole was
canted. I tried various other things that weren't a tight fit and found
them equally ambiguous. It wasn't until I found something with an
interference fit that I became convinced that the holes were not canted.

well, i accept that it may be different for different hubs, but i am
certain of the situation regarding the hub i tested.

Quoted message said:

Quoted message said:

Quoted message said:

> 1. does or does not hub hole conformation lessen the required spoke
> angle?
> 2. does or does not "correcting the spoke line" /prior/ to hub hole
> conformation exaggerate the "correction", i.e. make it excessive?
> 3. how does elastic stress from an undeformed spoke bending toward the
> rim compare to that where the spoke bends around its interleaving
> partner? from what i can see, the angle is very similar so skin stress
> has to be of similar magnitude.

First, I'm not aware of anyone (other than Sheldon, & I don't agree
with him) correcting the spoke line before the wheel is tensioned, so
I'm not sure who you're arguing with.

jobst brandt!!! its specifically stated in "the book"!

Not from my reading (which I quoted earlier). The step is described
after the section on tensioning. Furthermore, he expressed it in terms
of the spoke "may" need correction.

If a wheel is brought up to tension, and the spokes still have an overly
obtuse angle (as evidenced by bowing out from the flange), I think that
needs correction -- the spoke path is not in line with the load path.

but again, "correcting" prior to hub hole deformation has the spoke out
of the load line too! do you dispute the geometry i illustrated?

Quoted message said:

It's Mavic who seem to propose stress relieving a partially tensioned
wheel -- for reasons I don't understand (nor do they, I'm willing to bet).

presumably because, just like they were smart enough to figure out that
j-bend spokes are fundamentally fatigue challenged, they also realized
that seating was essential /before/ anything else that could influence
final spoke line.

Quoted message said:

If you claim that all failed elbows you've examined failed from the
outside (as BenC says), then I don't understand why you're concerned
about an over acute elbow angle.

i don't claim that. it's roughly 4:1 outside to inside. and again,
"outside" the elbow is different from the skin which is where 100% of
fatigue initiates.

Quoted message said:

That would put extra tension on the
inside of the bend. Perhaps all your spokes fail on the outside because
you haven't corrected the spoke line an have an over obtuse angle.

see above.

Quoted message said:

Quoted message said:

Quoted message said:

In all of the above cases, the only important thing is that the spoke
take as straight a line as possible. If the spoke has a bend in it,
the spoke will flex under change of tension, accelerating fatigue.

no, the spoke elbow is fundamentally going to flex regardless. the
elbow's offset means load is non-axial, hence the elbow /will/ bend.
and it is /that/ which causes fatigue, not the ethereal dragon that is
alleged to be "stress relief".

If the spoke line matches the load line the spoke won't bend with
changes in load.

but it fundamentally /cannot/ match the load line for a j-bend spoke,
only straight pull.

Quoted message said:

Quoted message said:

Quoted message said:

If, after the wheel is tensioned, the spokes have a bow in them where
they exit the flange, it is necessary to correct the line to prevent
flex in operation.

but if that's true, it should also include the necessity to "correct"
the bow at interleaving crossover!

Since spokes have finite thickness, the spoke line has to deviate
slightly from the load line, there's no way to correct for that.

straight pull spokes address it as effectively as you could wish.
j-bend fundamentally cannot.

Quoted message said:

Quoted message said:

Quoted message said:

If they don't, it isn't. The elbow is a critical fatigue location,
the crossover is not.

agreed. but we may have different reasoning for that. to me, the
crossover is subcritical because the straightening of the loading
cycle runs /with/ it's natural shape and is /not/ fighting an offset
like the elbow. additionally, it's not subject to the surface defects
that inevitably accompany a bending operation like elbow forming.
"orange peel effect".

I think it's because the bend angle is much smaller and the spokes
support each other and prevent much bending.

do the math for the stress in the outer skin at the bend [3-point] -
you'll be surprised.

jim beam said:

Peter Cole wrote:

Quoted message said:

Quoted message said:

I have had a few DT spokes fail (at elbows) before I began stress
relieving, none since.

were they from the famous "long shank" era? they broke rapidly because
of geometry issues.

I don't think so.

Quoted message said:

Quoted message said:

Thermal warping has many causes.

not many at those temperatures and that quickly. #1 candidate is
thermal stress relief. nothing else is close.

I don't think thermal "opening up" of bends proportional to the degree
of bend is a slam-dunk demonstration of residual stresses.

Quoted message said:

Quoted message said:

If even the direction (never mind the magnitude) revealed residual
stresses, why did carbon steel spokes bend the other way?

it's undergoing a phase change at the same time. austenitic stainless
is not.

You're guessing.

Quoted message said:

Quoted message said:

You assume that spokes are delivered w/o residual stresses and that
none are introduced in wheel building. The only way the first could be
true is if the manufacturer annealed the spokes -- I have seen no
claims of this -- you?

with respect, the years of argument on this topic have all been
concerned with the efficacy of mechanical stress relief. why then
eliminate it and say that only thermal is possible?

With equal respect, why should DT, or any other, mess about with
mechanical stress relief in formed parts when annealing would be so much
easier? On the other hand, annealing a built wheel presents some thorny
issues. I thought those arguments were obvious.

Quoted message said:

Quoted message said:

If, after unlacing a tensioned wheel, spoke elbow angles are no longer
equal, there must be residual stresses introduced.

fogel has demonstrated that mechanical relief occurs are low levels.
mere bending is not evidence of residual stress.

I think he showed thermal deformation, the cause of which is unclear.

Quoted message said:

Quoted message said:

Quoted message said:

when's the last
time you saw a bridge being "stress relieved"?

Perhaps not bridges, but I posted a link to an overload stress relief
procedure required by the Coast Guard for large pressure tanks.

and like i told you, i do not accept that explanation. every fracture
mechanics student will tell you that that procedure is done to reveal
crack flaws. there is no evidence that it performs mechanical relief,
and that's not the true objective. i can only assume that piece was
written by a well intentioned but mistaken intern.

Sure, if you say so.

Quoted message said:

Quoted message said:

At least some companies market equipment for vibrational stress relief
of large structures.

indeed. and they precipitate yielding. but weren[t you just arguing
that stress relief has to be thermal?

No, just that it's the easier way sometimes.

Quoted message said:

Quoted message said:

The principle is the same. I don't know what parts of a bridge you
think might need stress relief, but for known problems like welds,
mitigation is done.

yes, it's thermal. elevated temperature allows hydrogen to diffuse out
before welds crack.

So bridges are stress relieved?

Peter Cole said:

jim beam said:

Peter Cole wrote:

Quoted message said:

Quoted message said:

I have had a few DT spokes fail (at elbows) before I began stress
relieving, none since.

were they from the famous "long shank" era? they broke rapidly
because of geometry issues.

I don't think so.

Quoted message said:

Quoted message said:

Thermal warping has many causes.

not many at those temperatures and that quickly. #1 candidate is
thermal stress relief. nothing else is close.

I don't think thermal "opening up" of bends proportional to the degree
of bend is a slam-dunk demonstration of residual stresses.

i'm not saying it quantifies because it doesn't, but it reliably
demonstrates.

Quoted message said:

Quoted message said:

Quoted message said:

If even the direction (never mind the magnitude) revealed residual
stresses, why did carbon steel spokes bend the other way?

it's undergoing a phase change at the same time. austenitic stainless
is not.

You're guessing.

it's definitely undergoing phase change. the "guess" is whether it
explains the difference, but it's an educated guess based on facts i
know, not stuff i've made up. can you explain better?

Quoted message said:

Quoted message said:

Quoted message said:

You assume that spokes are delivered w/o residual stresses and that
none are introduced in wheel building. The only way the first could
be true is if the manufacturer annealed the spokes -- I have seen no
claims of this -- you?

with respect, the years of argument on this topic have all been
concerned with the efficacy of mechanical stress relief. why then
eliminate it and say that only thermal is possible?

With equal respect, why should DT, or any other, mess about with
mechanical stress relief in formed parts when annealing would be so much
easier?

because it will soften the steel! spoke steel is worked to ~1000N/mm^2
yield. annealed, i doubt it would be 400N/mm^2.

Quoted message said:

On the other hand, annealing a built wheel presents some thorny
issues. I thought those arguments were obvious.

apparently not.

Quoted message said:

Quoted message said:

Quoted message said:

If, after unlacing a tensioned wheel, spoke elbow angles are no
longer equal, there must be residual stresses introduced.

fogel has demonstrated that mechanical relief occurs are low levels.
mere bending is not evidence of residual stress.

I think he showed thermal deformation, the cause of which is unclear.

there is no impetus to deform other than that caused by [internal]
residual stress. unless he fabricated results of course. but i doubt
that since his results are consistent with materials theory.

Quoted message said:

Quoted message said:

Quoted message said:

> when's the last
> time you saw a bridge being "stress relieved"?

Perhaps not bridges, but I posted a link to an overload stress relief
procedure required by the Coast Guard for large pressure tanks.

and like i told you, i do not accept that explanation. every fracture
mechanics student will tell you that that procedure is done to reveal
crack flaws. there is no evidence that it performs mechanical relief,
and that's not the true objective. i can only assume that piece was
written by a well intentioned but mistaken intern.

Sure, if you say so.

Quoted message said:

Quoted message said:

At least some companies market equipment for vibrational stress
relief of large structures.

indeed. and they precipitate yielding. but weren[t you just arguing
that stress relief has to be thermal?

No, just that it's the easier way sometimes.

depends on application. you /don't/ want to thermally stress relieve
cold drawn wire unless you want to soften it too.

Quoted message said:

Quoted message said:

Quoted message said:

The principle is the same. I don't know what parts of a bridge you
think might need stress relief, but for known problems like welds,
mitigation is done.

yes, it's thermal. elevated temperature allows hydrogen to diffuse
out before welds crack.

So bridges are stress relieved?

no - temperature is insufficient for that. usually they're pre-heated
to a few hundred degrees to minimize stress gradient on weld pool
solidification and retro-heated to allow hydrogen to diffuse out. you
can argue that hydrogen presence causes internal stress, but it's not
the result of weld pool phase change or contraction.

On 2006-12-21, Peter Cole <[email hidden]> wrote:
[snip]

Quoted message said:

If you claim that all failed elbows you've examined failed from the
outside (as BenC says)

Misunderstanding: by "outside" I meant the outside _surface_, i.e. the
exterior of the wire, not the outside of the bend.

IIRC jim beam had examined failed spokes with a microscope and found
that fatigue had started (nucleated) on the exterior, never on the
interior, where there are no surface cracks or imperfections since
there's no surface. The point is that the residual stress would be in
the interior, although I'm still reading your post about skin stresses
and neutral planes and things.

The ratio of failures that start on the exterior outside the bend to
ones that start on the exterior inside the bend may also be interesting,
but noone's claimed it's 1:0.

jim beam said:

Peter Cole wrote:

Quoted message said:

Quoted message said:

If a wheel is brought up to tension, and the spokes still have an
overly obtuse angle (as evidenced by bowing out from the flange), I
think that needs correction -- the spoke path is not in line with the
load path.

but again, "correcting" prior to hub hole deformation has the spoke out
of the load line too! do you dispute the geometry i illustrated?

I saw nothing to dispute, just simple trig.

A pin in a plate with clearance will cant. If you put a load on the pin,
the hole may wallow and the pin cant further. Most of the wallowing must
occur during initial spoke tension (compared to any additional during
stress relieving and/or riding).

If, after tensioning (and associated wallowing) the spoke angle is still
too obtuse, it must be corrected, or the spoke will flex during load
cycles. Furthermore, if the load brings the spoke near yield, it will
operate at that point, drastically reducing fatigue life. That's why
most of your spokes failed at the outside of the bend.

Quoted message said:

Quoted message said:

It's Mavic who seem to propose stress relieving a partially tensioned
wheel -- for reasons I don't understand (nor do they, I'm willing to
bet).

presumably because, just like they were smart enough to figure out that
j-bend spokes are fundamentally fatigue challenged, they also realized
that seating was essential /before/ anything else that could influence
final spoke line.

Most seating is accomplished by initial tensioning.

By stress relieving a partially tensioned wheel, they run at least some
risk of over correcting the spoke line. Their method (pressing down on
the rim while supporting the axle) could yield both the holes and the
spoke bends. I think it's safest to correct a fully tensioned wheel as
Jobst indicates, and then only if it needs it, as he also says.

Quoted message said:

Quoted message said:

If you claim that all failed elbows you've examined failed from the
outside (as BenC says), then I don't understand why you're concerned
about an over acute elbow angle.

i don't claim that. it's roughly 4:1 outside to inside. and again,
"outside" the elbow is different from the skin which is where 100% of
fatigue initiates.

Quoted message said:

That would put extra tension on the inside of the bend. Perhaps all
your spokes fail on the outside because you haven't corrected the
spoke line an have an over obtuse angle.

see above.

Ditto.

Quoted message said:

Quoted message said:

If the spoke line matches the load line the spoke won't bend with
changes in load.

but it fundamentally /cannot/ match the load line for a j-bend spoke,
only straight pull.

Why not?

Quoted message said:

straight pull spokes address it as effectively as you could wish. j-bend
fundamentally cannot.

Why not?

Quoted message said:

Quoted message said:

Quoted message said:

> If they don't, it isn't. The elbow is a critical fatigue location,
> the crossover is not.

agreed. but we may have different reasoning for that. to me, the
crossover is subcritical because the straightening of the loading
cycle runs /with/ it's natural shape and is /not/ fighting an offset
like the elbow. additionally, it's not subject to the surface
defects that inevitably accompany a bending operation like elbow
forming. "orange peel effect".

I think it's because the bend angle is much smaller and the spokes
support each other and prevent much bending.

do the math for the stress in the outer skin at the bend [3-point] -
you'll be surprised.

Since spokes don't take a permanent bend at the crossing, yet they do at
hub exit, the former aren't at/near yield, the latter are (as Jobst
points out).

Peter Cole said:

jim beam said:

Peter Cole wrote:

Quoted message said:

Quoted message said:

If a wheel is brought up to tension, and the spokes still have an
overly obtuse angle (as evidenced by bowing out from the flange), I
think that needs correction -- the spoke path is not in line with the
load path.

but again, "correcting" prior to hub hole deformation has the spoke
out of the load line too! do you dispute the geometry i illustrated?

I saw nothing to dispute, just simple trig.

right, and that trig supports the spoke angle at which spokes are
manufactured, not arbitrarily bent.

Quoted message said:

A pin in a plate with clearance will cant. If you put a load on the pin,
the hole may wallow and the pin cant further. Most of the wallowing must
occur during initial spoke tension (compared to any additional during
stress relieving and/or riding).

but by definition, if subsequent loading makes it "wallow" deeper, then
something is occurring during riding to make it occurred. and that is
tension increase, not decrease.

if you think about it, that's also what's happening when you "stress
relieve" a wheel in the truing stand. it's /definitely/ happing when
you use the mavic method.

Quoted message said:

If, after tensioning (and associated wallowing) the spoke angle is still
too obtuse, it must be corrected, or the spoke will flex during load
cycles.

no, it's going to flex anyway. it has no choice because it's offset
from the load axis.

Quoted message said:

Furthermore, if the load brings the spoke near yield, it will
operate at that point, drastically reducing fatigue life. That's why
most of your spokes failed at the outside of the bend.

but "residual stress theory" has /compressive/ residual stress on the
outside of the bend...

Quoted message said:

Quoted message said:

Quoted message said:

It's Mavic who seem to propose stress relieving a partially tensioned
wheel -- for reasons I don't understand (nor do they, I'm willing to
bet).

presumably because, just like they were smart enough to figure out
that j-bend spokes are fundamentally fatigue challenged, they also
realized that seating was essential /before/ anything else that could
influence final spoke line.

Most seating is accomplished by initial tensioning.

not if you employ the mavic method. spokes can be seated before any
significant tension is applied.

Quoted message said:

By stress relieving a partially tensioned wheel, they run at least some
risk of over correcting the spoke line.

thank you!

Quoted message said:

Their method (pressing down on
the rim while supporting the axle) could yield both the holes and the
spoke bends. I think it's safest to correct a fully tensioned wheel as
Jobst indicates, and then only if it needs it, as he also says.

i agree with what you say, but jobst says to bend before the spokes are
tensioned, and that's premature.

Quoted message said:

Quoted message said:

Quoted message said:

If you claim that all failed elbows you've examined failed from the
outside (as BenC says), then I don't understand why you're concerned
about an over acute elbow angle.

i don't claim that. it's roughly 4:1 outside to inside. and again,
"outside" the elbow is different from the skin which is where 100% of
fatigue initiates.

Quoted message said:

That would put extra tension on the inside of the bend. Perhaps all
your spokes fail on the outside because you haven't corrected the
spoke line an have an over obtuse angle.

see above.

Ditto.

Quoted message said:

Quoted message said:

If the spoke line matches the load line the spoke won't bend with
changes in load.

but it fundamentally /cannot/ match the load line for a j-bend spoke,
only straight pull.

Why not?

because the load is offset from the axis of the spoke?

Quoted message said:

Quoted message said:

straight pull spokes address it as effectively as you could wish.
j-bend fundamentally cannot.

Why not?

because the load is offset from the axis of the spoke? shall i do a
diagram?

Quoted message said:

Quoted message said:

Quoted message said:

>
>> If they don't, it isn't. The elbow is a critical fatigue location,
>> the crossover is not.
>
> agreed. but we may have different reasoning for that. to me, the
> crossover is subcritical because the straightening of the loading
> cycle runs /with/ it's natural shape and is /not/ fighting an offset
> like the elbow. additionally, it's not subject to the surface
> defects that inevitably accompany a bending operation like elbow
> forming. "orange peel effect".

I think it's because the bend angle is much smaller and the spokes
support each other and prevent much bending.

do the math for the stress in the outer skin at the bend [3-point] -
you'll be surprised.

Since spokes don't take a permanent bend at the crossing, yet they do at
hub exit, the former aren't at/near yield, the latter are (as Jobst
points out).

i have some used spokes that show no difference between inner and outer
bends, hence they were not yielded by being laced & tensioned. and
that's supported by "simple trig", so i don't buy elbow bending as being
inevitable - more likely the result of over-enthusiasm.