Spoke tension meter

[Benjamin Lewis]

jim beam wrote:

> Benjamin Lewis wrote:
>> jim beam wrote:
>>
>>> all those photos show is plasic deformation. it does not even /begin/
>>> to show "undeniable photographic evidence that stress-relief can reduce
>>> local high stress in tensioned spokes". you'd need some pretty damned
>>> expensive microscopy or neutron/x-ray inferometry equipment to do that.
>> Please explain either - how a local region of the spoke can deform
>> plastically when I increase
>> the tension in that spoke unless that local region is at a higher stress
>> than the rest of the spoke, or
>
> 1. you're assuming significant inital residual stress which is not
> necessarily the case.

That is correct. If were no significant initial residual stress,
stress-relief would be unnecessary. Conversely, if increasing the tension
during stress relief causes a local region to yield, the initial stress at
that region (before stress relief, not before wheel tensioning) must have
been significant.

> 2. if you're deforming a component after too much time has elapsed from
> initial deformation, you can /add/ to residual stress, not mitigate it.
> as a general rule, deformation /creates/ residual stress, depending on
> degree of work[*], temperature, work rate, etc. all you're doing is
> advocating & photographing additional plastic deformation without
> evidence of the benefits claimed. a photo of a bent spoke, with
> respect, is not evidence of mitigation.

I'm not talking about the general case. I'm not even claiming that if you
remove the spokes from my stress relieved wheel, they will have no residual
stress. I'm claiming that if there are any regions near yield in a spoke
in a newly tensioned wheel, that temporarily increasing the tension in the
spoke will bring these regions past yield, reducing the undesirable stress
there.

The initial increase in tension via spoke squeezing causes plastic
deformation.
Subsequent increases in tension do not cause plastic deformation.

Therefore, the tension in these regions after initial squeezing *must* be
lower, in the finished tensioned wheel.

--
Benjamin Lewis

"Love is a snowmobile racing across the tundra and then suddenly it flips
over, pinning you underneath. At night, the ice weasels come."
--Matt Groening

 

[[email protected]]

On Wed, 22 Jun 2005 08:43:24 -0400, Peter Cole
<[email protected]> wrote:

>Luns Tee wrote:
>> In article <[email protected]>,
>> <[email protected]> wrote:
>>
>>>then the diagrams show what I think is roughly the
>>>disagreement for a finished spoke:
>>>
>>>Luns Jim
>>>
>>>ctNct ccttN
>>>||||| |||||
>>>||||| |||||
>>>||||| ||||| --> inner curve
>>>||||| |||||
>>>||||| |||||
>>>
>>>c = compression
>>>t = tension
>>>N = neutral
>>
>>
>> The N is a little overemphasized, and there are necessarily
>> neutral planes between the ct pairs too - it'd be cNtNcNt if you want
>> to be precise, but ctct with the implied neutral planes in between
>> would suffice too.
>>
>> You can demonstrate to yourself that the latter profile is
>> impossible for a spoke with no external forces applied.
>
>
>An interesting document which shows "ctct" pattern in automotive springs
>from actual neutron scans:
>
><http://www.ncnr.nist.gov/AnnualReport/FY1999/residual.pdf>

Dear Peter,

Thanks!

Exactly what I was hoping that someone would show me. If I'm
reading the luridly colored scan of the 14mm coil spring
right, it's showing this pattern:

outside-of-curve (cntnCnT) inside-of-curve

The compression and tension are greater (CnT) on the inside
of the curve than on the outside (cnt).

Just as interesting is the brief comment about "bulldozing,"
where the spring is squashed instead of heated to relieve
the residual stresses:

" . . . the spring has been compressed to the point where
the length of the spring is equal to the number of windings
times the wire thickness. After this the spring was allowed
to relax. A small part of this torsion strain is in the
plastic region, so this spring is slightly shorter than all
the others."

If I understand this, the idea is that a spring like this:

/ / / / / / / /

is squashed down like this until the coils touch:

//////

and re-expands to a slightly shorter length with its
residual stresses virtually eliminated (the article mentions
that the scan showed the same results for this as for heat
treatment).

So assuming that the same process would work in reverse on a
stainless steel spoke elbow, would it be correct to say that
spoke pairs would have to be squeezed together hard enough
for the tension to reach the plastic region and that there
should be a noticeable change in the curve?

That is, could Benjamin Lewis be really be seeing evidence
of stress relief in the slight change in spoke elbow angle
in his pictures, rather than just ordinary bending?

http://www.cs.sfu.ca/~bclewis/personal/bike/spoke/spoke_after.jpeg

I've always been under the impression that the claimed
stress-relief was microscopic and not measurable with
anything less than gadgets like this neutron diffraction
rig, so the possibility that results can be seen with the
naked eye is very interesting.

Thanks again,

Carl Fogel

 

[[email protected]]

On Wed, 22 Jun 2005 06:11:09 -0700, jim beam
<[email protected]> wrote:

>>Benjamin Lewis wrote:

[snip]

[Jim wrote:]

>[*] small degrees of work frequently result in much larger residual
>stress than larger ones. a small deformation can therefore be a
>significant issue.

Dear Jim and Benjamin,

Aaargh!

I've been assuming that the big ~90-degree bend of the spoke
elbow during manufacture would leave far more residual
stress than the comparatively tiny bend of correcting the
spoke line.

Have I misunderstood things all along?

Is the major original factory bend of the elbow
paradoxically trivial in terms of residual stress compared
to the tiny bend created when the wheel-builder corrects the
spoke line?

Carl Fogel

 

[dvt]

Benjamin Lewis wrote:
> ...if increasing the tension
> during stress relief causes a local region to yield, the initial stress at
> that region (before stress relief, not before wheel tensioning) must have
> been significant.

I don't understand. Can't it be possible that the "local region" was
simply stressed *more* than the rest of the spoke during the
squeezing/stress relieving? That would seem to explain your results just
as well. Maybe you go on to describe it in these sentences...

> I'm claiming that if there are any regions near yield in a spoke
> in a newly tensioned wheel, that temporarily increasing the tension in the
> spoke will bring these regions past yield, reducing the undesirable stress
> there.
>
> The initial increase in tension via spoke squeezing causes plastic
> deformation.

I'm with you up to here, based on the assumption that regions of the
spoke are near yield.

> Subsequent increases in tension do not cause plastic deformation.

I don't think we've seen any proof that "Subsequent increases in tension
do not cause plastic deformation." Even if that were true, I can't make
my brain reach your next conclusion:

> Therefore, the tension in these regions after initial squeezing *must* be
> lower, in the finished tensioned wheel.

I just don't get it. Let's follow your assumption that there is residual
stress in the spoke prior to squeezing the spokes, and that residual
stress is near yield. Now you add more stress by squeezing the spokes,
and the spoke yields. When you finish squeezing, the spoke has less
stress because it has yielded. But you would need to tighten that spoke
after yielding in order to get the tension back. How do you know that
the stress is different than it was prior to the spoke squeezing process?

I hope you don't take this wrong -- I'm not attempting to say you're
wrong, but I can't fill in the gaps of your argument. Maybe it's obvious
to Joe Mechanical Engineer, but it isn't obvious to me.

--
Dave
dvt at psu dot edu

 

[dvt]

[email protected] wrote:
> On Wed, 22 Jun 2005 08:43:24 -0400, Peter Cole
> <[email protected]> wrote:
>><http://www.ncnr.nist.gov/AnnualReport/FY1999/residual.pdf>

> Just as interesting is the brief comment about "bulldozing,"
> where the spring is squashed instead of heated to relieve
> the residual stresses:

As I read it, the spring is not squashed *instead* of heated. It is
squashed *and* heated. Three cases were tested:

-no stress relieving
-tempered stress relief applied
-tempered stress relief applied, then bulldozed

"We have looked at three cold-coiled springs. The first spring
is an as-cold coiled spring. The second one is cold-coiled followed
by a relatively low temper. The third one is identical to the second
one, but in addition to being tempered the spring has been compressed
to the point where the length of the spring is equal to the
number of windings times the wire thickness. After this the spring
was allowed to relax. A small part of this torsion strain is in the
plastic region, so this spring is slightly shorter than all the others.
In the automotive industry this process is known as “bulldozing”."

The bulldozed case reportedly had a similar profile to the non-bulldozed
case. So the bulldozing didn't affect the residual stress, but much of
that stress should have been relieved already by tempering.

--
Dave
dvt at psu dot edu

 

[jim beam]

Benjamin Lewis wrote:
> jim beam wrote:
>
>
>>Benjamin Lewis wrote:
>>
>>>jim beam wrote:
>>>
>>>
>>>>all those photos show is plasic deformation. it does not even /begin/
>>>>to show "undeniable photographic evidence that stress-relief can reduce
>>>>local high stress in tensioned spokes". you'd need some pretty damned
>>>>expensive microscopy or neutron/x-ray inferometry equipment to do that.
>>>
>>>Please explain either - how a local region of the spoke can deform
>>>plastically when I increase
>>>the tension in that spoke unless that local region is at a higher stress
>>>than the rest of the spoke, or
>>
>>1. you're assuming significant inital residual stress which is not
>>necessarily the case.
>
>
> That is correct. If were no significant initial residual stress,
> stress-relief would be unnecessary. Conversely, if increasing the tension
> during stress relief causes a local region to yield, the initial stress at
> that region (before stress relief, not before wheel tensioning) must have
> been significant.

theoratically yes, but in practice? significant stress concentrations
can exist at defects [check into fracture mechanics] under slight
elastic load which do not exist in the relaxed state, therefore local
yielding can ocurr, but it doesn't mean residual stress existed before
deformation. significant stress concentrations at a surface defect is
the reason why surface finish is such an important feature of fatigue
mitigation and why fatigue almost always initiates at such a feature.

>
>
>>2. if you're deforming a component after too much time has elapsed from
>>initial deformation, you can /add/ to residual stress, not mitigate it.
>>as a general rule, deformation /creates/ residual stress, depending on
>>degree of work[*], temperature, work rate, etc. all you're doing is
>>advocating & photographing additional plastic deformation without
>>evidence of the benefits claimed. a photo of a bent spoke, with
>>respect, is not evidence of mitigation.
>
>
> I'm not talking about the general case. I'm not even claiming that if you
> remove the spokes from my stress relieved wheel, they will have no residual
> stress. I'm claiming that if there are any regions near yield in a spoke
> in a newly tensioned wheel, that temporarily increasing the tension in the
> spoke will bring these regions past yield, reducing the undesirable stress
> there.

why? if a component is not stress relieved shortly after initial
formation, additional work can add to any existing residual stress, not
mitigate it.

>
> The initial increase in tension via spoke squeezing causes plastic
> deformation.

this is assuming residual stress is in a condition to be relieved and
can be done with a spoke squeeze of undefined magnitude...

> Subsequent increases in tension do not cause plastic deformation.

assuming there was initial residual stress and that it was relieved.

>
> Therefore, the tension in these regions after initial squeezing *must* be
> lower, in the finished tensioned wheel.
>

no. see above. even if spokes were delivered with high degrees of
residual stress, it's a gross assumption to say that this process will
produce results. there are a multitude of factors against it and only
really hope in favor of it.

but it's all an academic argument because the hard facts are:

1. the "stress relief" process /does/ seat spokes reliably in their
anchor points [simple visual inspection is possibe for the lay person to
corroborate that]

2. "stress relief" seating /is/ important for a wheel to remain true
rather than allow deformation /during/ use which makes the wheel go out
of true, [try riding a wheel that has not been "stress relieved"!!!], and

3. modern quality stainless spoke wire is inherently more fatigue
resistant [lower inclusion count, better composition].

to deny the above in favor of an arguement of anecdotes and "selected
data" is simply illogical. [captain.]

 

[jim beam]

[email protected] wrote:
> On Wed, 22 Jun 2005 06:11:09 -0700, jim beam
> <[email protected]> wrote:
>
>
>>>Benjamin Lewis wrote:
>
>
> [snip]
>
> [Jim wrote:]
>
>
>>[*] small degrees of work frequently result in much larger residual
>>stress than larger ones. a small deformation can therefore be a
>>significant issue.
>
>
> Dear Jim and Benjamin,
>
> Aaargh!
>
> I've been assuming that the big ~90-degree bend of the spoke
> elbow during manufacture would leave far more residual
> stress than the comparatively tiny bend of correcting the
> spoke line.
>
> Have I misunderstood things all along?
>
> Is the major original factory bend of the elbow
> paradoxically trivial in terms of residual stress compared
> to the tiny bend created when the wheel-builder corrects the
> spoke line?
>
> Carl Fogel

absolutely. small degrees of cold work can leave much higher levels of
residual stress than larger ones because insufficient yielding has ocurred.

 

[jim beam]

dvt wrote:
> [email protected] wrote:
>
>> On Wed, 22 Jun 2005 08:43:24 -0400, Peter Cole
>> <[email protected]> wrote:
>
> >><http://www.ncnr.nist.gov/AnnualReport/FY1999/residual.pdf>
>
>> Just as interesting is the brief comment about "bulldozing,"
>> where the spring is squashed instead of heated to relieve
>> the residual stresses:
>
>
> As I read it, the spring is not squashed *instead* of heated. It is
> squashed *and* heated. Three cases were tested:
>
> -no stress relieving
> -tempered stress relief applied
> -tempered stress relief applied, then bulldozed
>
> "We have looked at three cold-coiled springs. The first spring
> is an as-cold coiled spring. The second one is cold-coiled followed
> by a relatively low temper. The third one is identical to the second
> one, but in addition to being tempered the spring has been compressed
> to the point where the length of the spring is equal to the
> number of windings times the wire thickness. After this the spring
> was allowed to relax. A small part of this torsion strain is in the
> plastic region, so this spring is slightly shorter than all the others.
> In the automotive industry this process is known as “bulldozing”."
>
> The bulldozed case reportedly had a similar profile to the non-bulldozed
> case.

reportedly, but they don't show it! thing is, that stress profile is
perpendicular to the wire axis. "bulldozing" stresses the wire
torsionally w.r.t. the wire axis, so i'd really love to see whether the
actual result is as claimed.

> So the bulldozing didn't affect the residual stress, but much of
> that stress should have been relieved already by tempering.
>
tempering for springs like this is not just for stress relief - there
are crystal structure changes too. tempering of steel is a big topic.

 

[41]

jim beam wrote:

> if a component is not stress relieved shortly after initial
> formation, additional work can add to any existing residual stress, not
> mitigate it.

Specify deltaT. Specifiy the equation from which one calculates this
deltaT, or where it enters in any way. Specify the nature of the
"additional work", and the conditions that determine "will" and
"won't".

I don't believe you can.

 

[[email protected]]

On Wed, 22 Jun 2005 22:19:44 -0700, jim beam
<[email protected]> wrote:

>[email protected] wrote:
>> On Wed, 22 Jun 2005 06:11:09 -0700, jim beam
>> <[email protected]> wrote:
>>
>>
>>>>Benjamin Lewis wrote:
>>
>>
>> [snip]
>>
>> [Jim wrote:]
>>
>>
>>>[*] small degrees of work frequently result in much larger residual
>>>stress than larger ones. a small deformation can therefore be a
>>>significant issue.
>>
>>
>> Dear Jim and Benjamin,
>>
>> Aaargh!
>>
>> I've been assuming that the big ~90-degree bend of the spoke
>> elbow during manufacture would leave far more residual
>> stress than the comparatively tiny bend of correcting the
>> spoke line.
>>
>> Have I misunderstood things all along?
>>
>> Is the major original factory bend of the elbow
>> paradoxically trivial in terms of residual stress compared
>> to the tiny bend created when the wheel-builder corrects the
>> spoke line?
>>
>> Carl Fogel
>
>absolutely. small degrees of cold work can leave much higher levels of
>residual stress than larger ones because insufficient yielding has ocurred.

Dear Jim,

Aaaargh!

Er, I mean, thanks for explaining that.

Carl Fogel

 

[[email protected]]

On Wed, 22 Jun 2005 21:39:51 -0400, dvt <[email protected]>
wrote:

>[email protected] wrote:
>> On Wed, 22 Jun 2005 08:43:24 -0400, Peter Cole
>> <[email protected]> wrote:
> >><http://www.ncnr.nist.gov/AnnualReport/FY1999/residual.pdf>
>
>> Just as interesting is the brief comment about "bulldozing,"
>> where the spring is squashed instead of heated to relieve
>> the residual stresses:
>
>As I read it, the spring is not squashed *instead* of heated. It is
>squashed *and* heated. Three cases were tested:
>
>-no stress relieving
>-tempered stress relief applied
>-tempered stress relief applied, then bulldozed
>
>"We have looked at three cold-coiled springs. The first spring
>is an as-cold coiled spring. The second one is cold-coiled followed
>by a relatively low temper. The third one is identical to the second
>one, but in addition to being tempered the spring has been compressed
>to the point where the length of the spring is equal to the
>number of windings times the wire thickness. After this the spring
>was allowed to relax. A small part of this torsion strain is in the
>plastic region, so this spring is slightly shorter than all the others.
>In the automotive industry this process is known as “bulldozing”."
>
>The bulldozed case reportedly had a similar profile to the non-bulldozed
>case. So the bulldozing didn't affect the residual stress, but much of
>that stress should have been relieved already by tempering.

Dear Dave,

Re-reading the article, I think that you're right and that I
was wrong.

Drat.

I was over-eager to see clear evidence of some sort of
mechanical stress relief.

But it looks as if they're saying that heat treatment
relieved the residual stress (whose nice neutron diffraction
pictures were Peter Cole's point).

Squashing the already-stress-relieved coils afterwards
didn't make much difference as far as they could see.

That's just a failure to create residual stress by faint
bending.

I'm just as over-eager to see clear evidence that residual
stress is not being relieved.

But in this case the residual stress had already been cooked
off, so to speak, before the squashing, so it's a no-pitch
situation, to mix metaphors, and doesn't seem to say much
either way.

My mistake was to think that they just squashed a cold coil
and produced stress relief, but they didn't.

Drat.

Carl Fogel

 

[[email protected]]

Carl Fogel writes:

>> absolutely. small degrees of cold work can leave much higher
>> levels of residual stress than larger ones because insufficient
>> yielding has ocurred.

> Aaaargh!

> Er, I mean, thanks for explaining that.

Oops. As you know, Johnnie Walker claims there is no residual stress
so how can that clarify this position?

[email protected]

 

[Benjamin Lewis]

dvt wrote:

> Benjamin Lewis wrote:
>> ...if increasing the tension during stress relief causes a local region
>> to yield, the initial stress at that region (before stress relief, not
>> before wheel tensioning) must have been significant.
>
> I don't understand. Can't it be possible that the "local region" was
> simply stressed *more* than the rest of the spoke during the
> squeezing/stress relieving? That would seem to explain your results just
> as well. Maybe you go on to describe it in these sentences...

I think you're forgetting that if there is to be roughly equal tension
everywhere in the spoke in the finished wheel, the elbow must be the same
shape as the curve in the hub it sits next to. If the elbow is
unsupported, tension in the spoke causes it to bend either elastically
(creating an undesired region of higher tension than the rest of the
spoke), or plastically (which is what occurred in the spoke in the photo
when I stress-relieved it).

If it *does* lie flush to the hub, and there are regions close to yielding,
increasing the tension everywhere in the spoke, and there in these regions,
will clearly cause them to yield.

If it does lie flush to the hub, and there are no regions close to
yielding, then stress-relief is unnecessary.

--
Benjamin Lewis

"Love is a snowmobile racing across the tundra and then suddenly it flips
over, pinning you underneath. At night, the ice weasels come."
--Matt Groening

 

[[email protected]]

On Thu, 23 Jun 2005 16:09:07 GMT,
[email protected] wrote:

>Carl Fogel writes:
>
>>> absolutely. small degrees of cold work can leave much higher
>>> levels of residual stress than larger ones because insufficient
>>> yielding has ocurred.
>
>> Aaaargh!
>
>> Er, I mean, thanks for explaining that.
>
>Oops. As you know, Johnnie Walker claims there is no residual stress
>so how can that clarify this position?
>
>[email protected]

Dear Jobst,

Er . . .

Jim Beam and Luns Tee have been talking about where the
residual stress is located throughout this entire thread.

I've been drawing ASCII diagrams and trying to see where
they think it is. They've been quite patient with my
fumbling efforts.

Peter Cole has provided a nice link to an article with
neutron diffraction images showing bands of residual stress
in coil springs.

The question is where the residual stress is, followed by
whether squeezing pairs of spokes together with our bare
hands relieves the stress, does nothing to it, or even makes
it worse.

As a side issue, it's beginning to sound as if the major
bending of the elbow at the factory might paradoxically not
produce as much residual stress as the well-meant efforts of
a wheel-builder to bend the spokes slightly to "correct" the
spoke line.

Hope that clarifies things. The simple spoke turns out to be
a much more complicated thing than I'd been led to believe,
which is why I'm curious about it.

All sorts of people can end up apparently confused about how
spokes behave:

"In tension tests . . . all spokes tested, (swaged) butted
and straight . . . instead of breaking in the threads or at
the elbow . . . failed in their midsections."
--p. 32, "The Bicycle Wheel," 3rd edition

"Swaged spokes failed in their reduced diameter mid-sections
while the straight gauge spokes failed at the elbows."
--p.124, "The Bicycle Wheel," 3rd edition

Cheers,

Carl Fogel

 

[dvt]

Benjamin Lewis wrote:
> dvt wrote:
>
>
>>Benjamin Lewis wrote:
>>
>>>...if increasing the tension during stress relief causes a local region
>>>to yield, the initial stress at that region (before stress relief, not
>>>before wheel tensioning) must have been significant.
>>
>>I don't understand. Can't it be possible that the "local region" was
>>simply stressed *more* than the rest of the spoke during the
>>squeezing/stress relieving? That would seem to explain your results just
>>as well. Maybe you go on to describe it in these sentences...
>
>
> I think you're forgetting that if there is to be roughly equal tension
> everywhere in the spoke in the finished wheel, the elbow must be the same
> shape as the curve in the hub it sits next to. If the elbow is
> unsupported, tension in the spoke causes it to bend either elastically
> (creating an undesired region of higher tension than the rest of the
> spoke), or plastically (which is what occurred in the spoke in the photo
> when I stress-relieved it).
>
> If it *does* lie flush to the hub, and there are regions close to yielding,
> increasing the tension everywhere in the spoke, and there in these regions,
> will clearly cause them to yield.
>
> If it does lie flush to the hub, and there are no regions close to
> yielding, then stress-relief is unnecessary.
>
I understand everything you say, Benjamin. But what happens after the
spoke yields? If it yields, then the tension in that spoke must be
reduced, right? So now you bring the tension back up by turning the
nipple. How do you know that the stress in the spoke is not the same as
it was before stress relieving?

I know I'm missing some simple, fundamental point. But I don't know what
it is.

There's another thing missing in this whole discussion, IMO. What kind
of stress are we talking about? Shear, longitudinal? Maybe some
combination of these, like the Von Mises stress? It's a bit hard for me
to imagine these residual stress profiles without that bit of info.

--
Dave
dvt at psu dot edu

 

[Benjamin Lewis]

dvt wrote:

> I understand everything you say, Benjamin. But what happens after the
> spoke yields? If it yields, then the tension in that spoke must be
> reduced, right?

Not significantly, no; it is only locally yielding. The net tension in the
spoke is something like 1/3 of yield. If you pluck the spoke before and
after stress relief, you won't hear a difference in pitch.

--
Benjamin Lewis

"Love is a snowmobile racing across the tundra and then suddenly it flips
over, pinning you underneath. At night, the ice weasels come."
--Matt Groening

 

[[email protected]]

Dave vt? writes:

> I understand everything you say, Benjamin. But what happens after
> the spoke yields? If it yields, then the tension in that spoke must
> be reduced, right? So now you bring the tension back up by turning
> the nipple. How do you know that the stress in the spoke is not the
> same as it was before stress relieving?

If there were high stress locations, these would yield and when the
overload is relaxed, stress cannot return to or close to the yield
stress that existed locally from forming the spoke to it's current
line.

These elongations by yielding are in the micro-inches and have
negligible effect on spoke length, and therefore, tension. That the
spokes "bed in" is accomplished before this stage just by tension.
Therefore there is no significant change in spoke support although
some people believe that this is a major factor in stress relieving.
To those who believe this, I suggest they spoke up a wheel with new
hubs to full tension without stress relieving, and remove a spoke to
see that it is fully bedded in.

As I mentioned, stress in the elbow of a spoke supported in aluminum
is lower than in the adjacent straight portion because the elbow is
supported. It is like a winch. The cable does not snap on the
winding spool but rather on a straight run where the cable pulls
unaided.

> I know I'm missing some simple, fundamental point. But I don't know
> what it is.

It is mainly a visualization of the magnitude of these changes. They
are high stress and low elongation, steel being far stiffer than it
appears from manual bending. The Modulus of elasticity expresses that
the best.

http://www.engineersedge.com/manufacturing_spec/properties_of_metals_strength.htm

> There's another thing missing in this whole discussion, IMO. What
> kind of stress are we talking about? Shear, longitudinal? Maybe
> some combination of these, like the Von Mises stress? It's a bit
> hard for me to imagine these residual stress profiles without that
> bit of info.

This is entirely reducible to tensile and compressive stress. As
mentioned, the inside of a bend is compressed and the outside is
stretched. That is the case with spoke elbows. It is primarily the
residual tensile stresses that are interesting in this regard because
they add to the stress of spoke tension, thereby being dangerously
high for fatigue life.

Residual refers to stress that is not generated by tension in the
spoke but localized stresses caused by manufacture and lacing the
wheel, both of which plastically change (yield) the shape of the
spoke. Whenever forming a wire with other than pure longitudinal
tension, there will be residual tension and compression zones in that
wire (spoke). The reason why pure tension on a wire does not cause
residual stress is that the entire cross section is subjected to
uniform stress and when it is relaxed the entire cross section relaxes
to zero, whether it reached yield or not. Every micro-inch of yield
is "forgotten" by the spoke because it yielded and the road back is
unrestricted.

[email protected]

 

[jim beam]

41 wrote:
>
> jim beam wrote:
>
>
>> if a component is not stress relieved shortly after initial
>>formation, additional work can add to any existing residual stress, not
>>mitigate it.
>
>
> Specify deltaT. Specifiy the equation from which one calculates this
> deltaT, or where it enters in any way. Specify the nature of the
> "additional work", and the conditions that determine "will" and
> "won't".
>
> I don't believe you can.
>
eh? "T" isn't used for stress. you used it twice, so presumably it's
not a typo. you'll have to explain what are you talking about...

 

[jim beam]

Benjamin Lewis wrote:
> dvt wrote:
>
>
>>Benjamin Lewis wrote:
>>
>>>...if increasing the tension during stress relief causes a local region
>>>to yield, the initial stress at that region (before stress relief, not
>>>before wheel tensioning) must have been significant.
>>
>>I don't understand. Can't it be possible that the "local region" was
>>simply stressed *more* than the rest of the spoke during the
>>squeezing/stress relieving? That would seem to explain your results just
>>as well. Maybe you go on to describe it in these sentences...
>
>
> I think you're forgetting that if there is to be roughly equal tension
> everywhere in the spoke in the finished wheel, the elbow must be the same
> shape as the curve in the hub it sits next to. If the elbow is
> unsupported,

the elbow is usually supported, if the spoke elbow is the right length
for the hub, but you still get elastic distortion at any load interface.

> tension in the spoke causes it to bend either elastically
> (creating an undesired region of higher tension than the rest of the
> spoke), or plastically (which is what occurred in the spoke in the photo
> when I stress-relieved it).

let's assume that you bend the elbow to a more acute angle than factory,
as per your pics. that means the outside of the spoke is in tensile
stress during deformation, and springs back with some degree of
compressive residual after relaxation. that's good for fatigue, right?
[the compressive subtracts from the tensile, with a net cyclic load in
that small region /less/ than the macro load the spoke is experiencing.]
now, if that's the case, you /don't/ want to relax the residual
stress! particularly when examination of fracture surfaces shows that
most fatigue cracks initiate on the outside of the elbow.

basd on that actual fracture observation [a few do initiate on the
inside of the spoke elbow but from what i've seen, that very few] one
has to conclude that that "stress relief" is either not ocurring if
observed service lives of spokes are very long, /or/ that its effect is
[thankfully] very small in comparison to other known, researched,
corroborated fatigue initiation factors such as material composition,
quality & surface finish.

>
> If it *does* lie flush to the hub, and there are regions close to yielding,
> increasing the tension everywhere in the spoke, and there in these regions,
> will clearly cause them to yield.
>
> If it does lie flush to the hub, and there are no regions close to
> yielding, then stress-relief is unnecessary.

you /do/ have local elastic deformation with any load, but it's
incorrect to assume that applied load is evenly distributed. in this
case, the load is non-axial due to the spoke elbow offset, so you will
have a highly distributed load effect within the elbow region.

 

[[email protected]]

Johnnie Walker writes:

>>>> ...if increasing the tension during stress relief causes a local
>>>> region to yield, the initial stress at that region (before stress
>>>> relief, not before wheel tensioning) must have been significant.

>>> I don't understand. Can't it be possible that the "local region"
>>> was simply stressed *more* than the rest of the spoke during the
>>> squeezing/stress relieving? That would seem to explain your
>>> results just as well. Maybe you go on to describe it in these
>>> sentences...

>> I think you're forgetting that if there is to be roughly equal
>> tension everywhere in the spoke in the finished wheel, the elbow
>> must be the same shape as the curve in the hub it sits next to. If
>> the elbow is unsupported,

> the elbow is usually supported, if the spoke elbow is the right
> length for the hub, but you still get elastic distortion at any load
> interface.

>> tension in the spoke causes it to bend either elastically (creating
>> an undesired region of higher tension than the rest of the spoke),
>> or plastically (which is what occurred in the spoke in the photo
>> when I stress-relieved it).

> let's assume that you bend the elbow to a more acute angle than
> factory, as per your pics. that means the outside of the spoke is
> in tensile stress during deformation, and springs back with some
> degree of compressive residual after relaxation.

The point is that there is no spring-back. The spoke gets pulled into
its seated position in the hub and stays there. That is why it
remains bent in an acute angle when untensioned and removed from the
wheel as pictures showed.

> that's good for fatigue, right?

No. that leaves the spoke at yield at the outside of the bend. What
mechanism do you see for relaxing this condition if the spoke isn't
stress relieved by momentary overload?

> [the compressive subtracts from the tensile, with a net cyclic load
> in that small region /less/ than the macro load the spoke is
> experiencing.] now, if that's the case, you /don't/ want to relax
> the residual stress! particularly when examination of fracture
> surfaces shows that most fatigue cracks initiate on the outside of
> the elbow.

Did the wind change last week or what. Up to then you claimed there
were no residual stresses in spokes and now there are?

> basd on that actual fracture observation [a few do initiate on the
> inside of the spoke elbow but from what i've seen, that very few]
> one has to conclude that that "stress relief" is either not ocurring
> if observed service lives of spokes are very long, /or/ that its
> effect is [thankfully] very small in comparison to other known,
> researched, corroborated fatigue initiation factors such as material
> composition, quality & surface finish.

Your experience has insufficient statistical body to be valuable in
this discussion. People who stress relieve don't see these
occurrences and they have statistically significant samples, such as
Trek bicycles.

>> If it *does* lie flush to the hub, and there are regions close to
>> yielding, increasing the tension everywhere in the spoke, and there
>> in these regions, will clearly cause them to yield.

Sounds like stress relief to me.

>> If it does lie flush to the hub, and there are no regions close to
>> yielding, then stress-relief is unnecessary.

> you /do/ have local elastic deformation with any load, but it's
> incorrect to assume that applied load is evenly distributed. in
> this case, the load is non-axial due to the spoke elbow offset, so
> you will have a highly distributed load effect within the elbow
> region.

Why does it need to be "evenly distributed"? All it needs to do is
drive the high stress locations into yield before relaxing.

[email protected]