spoke fatigue troll

bjw@mambo.ucolick.org

You know what's great about this thread? It finally
names the true cause of spoke failure. It's right
in the subject header. It's the spoke fatigue troll.
He hides under cattle guards and reaches up
and grabs your spokes when you aren't looking.
That's why your hub flanges get deformed - it's
that damned spoke fatigue troll yanking on them.

Ben
You have to admit it's a better explanation than
some on RBT.

Ben C

On 2008-04-28, Peter Cole <peter_cole@verizon.net> wrote:
> Ben C wrote:
>
>> Recently I said I thought stabilization mostly worked by bending the
>> spokes but in some cases may also work by deforming the hub. daveornee's
>> recent pictures, which are very similar to those posted long ago by jim
>> beam, tip things a bit towards hub deformation in my mind. But I think
>> it still may be a bit of both and that it depends what hub and spokes
>> you're using.
>
> Holes typically deform as shown in the pictures. Just from eyeballing
> those, it looks like about 0.5mm. Taken over an elbow length of 6mm,
> that works out to a (rough) angle change of 5 deg. The depth of deform
> should be less than linear with spoke tension, so the nominal 50% over
> load of stress relief should represent less than a third of the depth.
> Even that amount will reduce the spoke tension significantly (full spoke
> tension = 0.75mm). Stress relief may reduce problems with tension loss
> caused by spokes bedding in, but it's unlikely that has any impact on
> spoke fatigue failure. There isn't even a plausible hypothesis for that.

Well my hypothesis (which perhaps you don't regard as plausible is
that this bedding in reduces the bending moment on the spoke.

It doesn't matter whether the spoke bends or the hub deforms. The result
is a spoke closer to the flange so less moment.

> Bottom line is that the deformation, both the majority that comes from
> initial tension as well as whatever smaller contribution that may come
> from over loading, doesn't predict the direction or degree of bending
> stress in the tensioned spoke.

I don't understand what you're saying here. Surely for all spokes the
effect of hub deformation is going to be to reduce moment on the elbow?

[...]
> As I said before, these threads have gone through an evolution that
> began with a denial of the presence of residual stress and have morphed
> into a grudging admission with the dismissal for irrelevance. Residual
> stresses are present, they may or may not be a factor in any given spoke
> failure, but that's impossible to know without knowing all the operating
> stresses. No postmortem is going to tell you whether the failure was
> caused by a raiser, bending stress, residuals or any combination
> thereof.

You would expect to see more failures starting in the regions of tensile
residual stress (i.e. inside of inbound spokes) if residual stress from
manufacturing is a factor.

There are other factors of course, but there ought to be a statistical
bias.

[...]
> These (endless) arguments follow the same pattern. Something is taken
> out of context and a straw man is created. Like spoke tension and wheel
> strength. The FEA leaves no doubt about what happens when a wheel is
> overloaded, it doesn't matter that an unspoked rim will support body
> weight without collapsing, the loads we're talking about are the several
> g's that happen when the wheel hits a pothole, that's when strength
> becomes important. Normally tensioned wheels loose spoke tension at
> about the same magnitude of deflection as rim damage starts, looser
> tension will mean impact damage will happen earlier. As the loaded area
> of the rim becomes slack, the rim looses lateral stiffness while still
> under compression and therefore is susceptible to buckling.

I follow your account of that. But I reserve judgment on whether a
tightly spoked wheel's rim may yield before its spokes go slack.

> Instead of focusing on those important issues the threads degenerate
> into a critique of spoke bed cracking, despite the fact that this is a
> problem for only a few rims and the straw man is raised that Jobst
> advised builders to exceed manufacturer's specs.

Fogel was probably right when he said there were no manufacturer's specs
in those days, and that the method of going to the taco point and back a
bit was good advice for the rims of the day.

It is worth pointing out though that that is not a suitable method for a
modern deeper-section Mavic rim. I'm not claiming Jobst ever said it
was, but the details get lost and people over-tension their rims.

I probably would have myself if it hadn't been for jim beam's
explanations of why it causes fatigue.

> When several sources are found that confirm the anodization fatigue
> connection, all that is dismissed with talk of anisotropy and
> extrusion flaws -- factors (obviously) still there whether anodization
> is or is not.

That is a much harder one to call. Some of these debates can be well
understood with basic mechanics and understanding of stress/strain and
S-N curves, which are simplified macroscopic views of how materials
behave.

But the effects of anisotropy and anodization on fatigue life is getting
much deeper into the structure of metals. Yes we know in the most
general terms they both can be factors but that's a long way from
understanding it enough to know how to apply it to bicycle rims.

andresmuro@aol.com

On Apr 29, 2:49*am, "b...@mambo.ucolick.org" <b...@mambo.ucolick.org>
wrote:
> You know what's great about this thread? *It finally
> names the true cause of spoke failure. *It's right
> in the subject header. *It's the spoke fatigue troll.
> He hides under cattle guards and reaches up
> and grabs your spokes when you aren't looking.
> That's why your hub flanges get deformed - it's
> that damned spoke fatigue troll yanking on them.
>
> Ben
> You have to admit it's a better explanation than
> some on RBT.

This is the reason you need to learn to bunny-hop cattle guards. You
don't let the troll grab your spokes.

Peter Cole

Ben C wrote:
> On 2008-04-28, Peter Cole <peter_cole@verizon.net> wrote:

> Well my hypothesis (which perhaps you don't regard as plausible is
> that this bedding in reduces the bending moment on the spoke.
>
> It doesn't matter whether the spoke bends or the hub deforms. The result
> is a spoke closer to the flange so less moment.
>
>> Bottom line is that the deformation, both the majority that comes from
>> initial tension as well as whatever smaller contribution that may come
>> from over loading, doesn't predict the direction or degree of bending
>> stress in the tensioned spoke.
>
> I don't understand what you're saying here. Surely for all spokes the
> effect of hub deformation is going to be to reduce moment on the elbow?

No. Hole deformation will change the elbow support angle slightly. It
will always enlarge the angle, if the angle was too big to begin with,
it will increase the moment. If the angle was initially too small, the
angle may be improved, but there still may be a bending stress.


>> As I said before, these threads have gone through an evolution that
>> began with a denial of the presence of residual stress and have morphed
>> into a grudging admission with the dismissal for irrelevance. Residual
>> stresses are present, they may or may not be a factor in any given spoke
>> failure, but that's impossible to know without knowing all the operating
>> stresses. No postmortem is going to tell you whether the failure was
>> caused by a raiser, bending stress, residuals or any combination
>> thereof.
>
> You would expect to see more failures starting in the regions of tensile
> residual stress (i.e. inside of inbound spokes) if residual stress from
> manufacturing is a factor.
>
> There are other factors of course, but there ought to be a statistical
> bias.

Who says there is or isn't? If spokes are bowed at the flange and the
spoke line is not corrected, the bending forces introduced as the spoke
is tensioned will swamp any residual stresses. Those spokes will
generally fail at the outside of the bend. There are those who refuse to
correct a bad spoke line then cite these failures as evidence of the
irrelevance of residual stress. It's comical, really.


> [...]
>> These (endless) arguments follow the same pattern. Something is taken
>> out of context and a straw man is created. Like spoke tension and wheel
>> strength. The FEA leaves no doubt about what happens when a wheel is
>> overloaded, it doesn't matter that an unspoked rim will support body
>> weight without collapsing, the loads we're talking about are the several
>> g's that happen when the wheel hits a pothole, that's when strength
>> becomes important. Normally tensioned wheels loose spoke tension at
>> about the same magnitude of deflection as rim damage starts, looser
>> tension will mean impact damage will happen earlier. As the loaded area
>> of the rim becomes slack, the rim looses lateral stiffness while still
>> under compression and therefore is susceptible to buckling.
>
> I follow your account of that. But I reserve judgment on whether a
> tightly spoked wheel's rim may yield before its spokes go slack.


Jobst did an FEA. It's in the book. A similar FEA is here:
http://www.astounding.org.uk/ian/wheel/3c_rim.html

A loaded spoked wheel doesn't deform into a uniform oval, it develops a
flat spot. As the length of the flat spot grows (with increasing load)
the bending stress on the rim increases. When that stress reaches a
critical value the bend is permanent. The rate at which the flat spot
grows with load is related to the initial spoke tension. That's it. That
you can ride around on a loosely spoked wheel doesn't say anything about
what happens when it meets a pothole.

>> Instead of focusing on those important issues the threads degenerate
>> into a critique of spoke bed cracking, despite the fact that this is a
>> problem for only a few rims and the straw man is raised that Jobst
>> advised builders to exceed manufacturer's specs.
>
> Fogel was probably right when he said there were no manufacturer's specs
> in those days, and that the method of going to the taco point and back a
> bit was good advice for the rims of the day.
>
> It is worth pointing out though that that is not a suitable method for a
> modern deeper-section Mavic rim. I'm not claiming Jobst ever said it
> was, but the details get lost and people over-tension their rims.
>
> I probably would have myself if it hadn't been for jim beam's
> explanations of why it causes fatigue.

Again, if you haven't read the book then you don't have a leg to stand
on. You're just repeating a misquote. Jobst (correctly) refers people to
his book. That's the "advice" you should take. The "details" don't get
"lost" if you buy the book.

>> When several sources are found that confirm the anodization fatigue
>> connection, all that is dismissed with talk of anisotropy and
>> extrusion flaws -- factors (obviously) still there whether anodization
>> is or is not.
>
> That is a much harder one to call. Some of these debates can be well
> understood with basic mechanics and understanding of stress/strain and
> S-N curves, which are simplified macroscopic views of how materials
> behave.
>
> But the effects of anisotropy and anodization on fatigue life is getting
> much deeper into the structure of metals. Yes we know in the most
> general terms they both can be factors but that's a long way from
> understanding it enough to know how to apply it to bicycle rims.

Nonsense. The "defenses" of anodizing boil down to an assertion that the
rim extrusions are so crappy that it doesn't matter if they take an
additional reliability hit from anodizing. If you have to derate spoke
tension below what all the other wheel components can tolerate just to
prevent socket cracking then you've just got weak sockets. Of course if
you don't understand the benefit of high spoke tension you don't think
you've given anything up.

jobst.brandt@stanfordalumni.org

Peter Cole <peter_cole@verizon.net> wrote:

> Nonsense. The "defenses" of anodizing boil down to an assertion
> that the rim extrusions are so crappy that it doesn't matter if they
> take an additional reliability hit from anodizing. If you have to
> derate spoke tension below what all the other wheel components can
> tolerate just to prevent socket cracking then you've just got weak
> sockets. Of course if you don't understand the benefit of high
> spoke tension you don't think you've given anything up.

I have a collection of things I don't throw away in my basement where
they are not in the way. Among these I have more than 40 rims from
the days of tubulars and subsequent MA-2 clinchers that were discarded
in wheelbuilding sessions held at my place when riders took care of
their own equipment. None of these rims has cracks, none were
anodized and all had 36 spoke sockets and eyelets. These include
Fiamme, Super Champion, Nisi, Weinmann, and Mavic rims.

The Nisi rims (tubulars) used steel flat washers and had no eyelets.
To work with them was a pain because loose washers at times got lost
in the rim. For this a glob of rim glue was put inside the hollow
section of the rim and the washer rattled around until it got stuck.

Since then Solomon bought Mavic and changed their focus on how rims
should be made. Today we read about the results in this newsgroup
often.

Those were the "good old days" but then I still have enough MA-2 rims
to keep me rolling for a long time.

Jobst Brandt

Ben C

On 2008-04-29, Peter Cole <peter_cole@verizon.net> wrote:
> Ben C wrote:
>> On 2008-04-28, Peter Cole <peter_cole@verizon.net> wrote:
>
>> Well my hypothesis (which perhaps you don't regard as plausible is
>> that this bedding in reduces the bending moment on the spoke.
>>
>> It doesn't matter whether the spoke bends or the hub deforms. The result
>> is a spoke closer to the flange so less moment.
>>
>>> Bottom line is that the deformation, both the majority that comes from
>>> initial tension as well as whatever smaller contribution that may come
>>> from over loading, doesn't predict the direction or degree of bending
>>> stress in the tensioned spoke.
>>
>> I don't understand what you're saying here. Surely for all spokes the
>> effect of hub deformation is going to be to reduce moment on the elbow?
>
> No. Hole deformation will change the elbow support angle slightly. It
> will always enlarge the angle, if the angle was too big to begin with,
> it will increase the moment.

I'm not sure the elbow support angle is the important thing to measure
here, but the perpendicular length of unsupported spoke.

It seems to me that will always get shorter as you are crushing the
fulcrum the spoke is bearing against.

> If the angle was initially too small, the
> angle may be improved, but there still may be a bending stress.

I think there will _always_ be a bending stress unless you've got a
straight pull spoke. So long as it's small enough you're OK though.

>>> As I said before, these threads have gone through an evolution that
>>> began with a denial of the presence of residual stress and have morphed
>>> into a grudging admission with the dismissal for irrelevance. Residual
>>> stresses are present, they may or may not be a factor in any given spoke
>>> failure, but that's impossible to know without knowing all the operating
>>> stresses. No postmortem is going to tell you whether the failure was
>>> caused by a raiser, bending stress, residuals or any combination
>>> thereof.
>>
>> You would expect to see more failures starting in the regions of tensile
>> residual stress (i.e. inside of inbound spokes) if residual stress from
>> manufacturing is a factor.
>>
>> There are other factors of course, but there ought to be a statistical
>> bias.
>
> Who says there is or isn't? If spokes are bowed at the flange and the
> spoke line is not corrected, the bending forces introduced as the spoke
> is tensioned will swamp any residual stresses.

Exactly.

> Those spokes will generally fail at the outside of the bend. There are
> those who refuse to correct a bad spoke line then cite these failures
> as evidence of the irrelevance of residual stress. It's comical,
> really.

Well I don't know who you're talking about there.

>> [...]
>>> These (endless) arguments follow the same pattern. Something is taken
>>> out of context and a straw man is created. Like spoke tension and wheel
>>> strength. The FEA leaves no doubt about what happens when a wheel is
>>> overloaded, it doesn't matter that an unspoked rim will support body
>>> weight without collapsing, the loads we're talking about are the several
>>> g's that happen when the wheel hits a pothole, that's when strength
>>> becomes important. Normally tensioned wheels loose spoke tension at
>>> about the same magnitude of deflection as rim damage starts, looser
>>> tension will mean impact damage will happen earlier. As the loaded area
>>> of the rim becomes slack, the rim looses lateral stiffness while still
>>> under compression and therefore is susceptible to buckling.
>>
>> I follow your account of that. But I reserve judgment on whether a
>> tightly spoked wheel's rim may yield before its spokes go slack.
>
>
> Jobst did an FEA. It's in the book. A similar FEA is here:
> http://www.astounding.org.uk/ian/wheel/3c_rim.html
>
> A loaded spoked wheel doesn't deform into a uniform oval, it develops a
> flat spot. As the length of the flat spot grows (with increasing load)
> the bending stress on the rim increases. When that stress reaches a
> critical value the bend is permanent. The rate at which the flat spot
> grows with load is related to the initial spoke tension.

Yes, I know, and I am familiar with Ian's FEA.

> That's it. That you can ride around on a loosely spoked wheel doesn't
> say anything about what happens when it meets a pothole.

I don't think it's so easy to predict what happens when you meet a
pot-hole.

[...]
>>> When several sources are found that confirm the anodization fatigue
>>> connection, all that is dismissed with talk of anisotropy and
>>> extrusion flaws -- factors (obviously) still there whether anodization
>>> is or is not.
>>
>> That is a much harder one to call. Some of these debates can be well
>> understood with basic mechanics and understanding of stress/strain and
>> S-N curves, which are simplified macroscopic views of how materials
>> behave.
>>
>> But the effects of anisotropy and anodization on fatigue life is getting
>> much deeper into the structure of metals. Yes we know in the most
>> general terms they both can be factors but that's a long way from
>> understanding it enough to know how to apply it to bicycle rims.
>
> Nonsense. The "defenses" of anodizing boil down to an assertion that the
> rim extrusions are so crappy that it doesn't matter if they take an
> additional reliability hit from anodizing.

Well, you can put it like that, but that could be misleading.

Designing something like a rim is an optimization problem with
constraints on things like strength, fatigue life, stiffness, weight,
cost and parameters like choice of alloy, shape of extrusion, whether
you anodize.

It won't necessarily be possible to get on the limits of all the
constraints: there may be more constraints than parameters.

It's rather like the recent discussion of forks: for aluminium forks the
fatigue constraint probably dominates, for steel forks, strength, or
resistance to crumpling or getting dented.

So, it may be that the best way to make a rim to satisfy all the other
requirements leaves it just on the limit of fatigue due to anisotropy.
If that were the case then it wouldn't do any harm to anodize it.

Understanding how to design something like that and making educated
guesses which are likely to be the dominant factors _does_ involve
understanding the materials properly.

There is also the evidence that the cracks tend to be oriented on the
path along which the rim was extruded. Now some people have suggested
anodizing could cause a crack to start and then hoop stress or
anisotrophy cause it to propagate on that path. Unforgiven98, on the
other hand, who seems to know what he is talking about, has said that
anodizing and anistropy do not work together ("collude" I think was the
term).

Finally just because anodizing causes fatigue in some applications it
doesn't follow that it does in all. I'm sure there's more too it--
different kinds of anodizing, what sort of aluminium you're using, etc.

> If you have to derate spoke tension below what all the other wheel
> components can tolerate just to prevent socket cracking then you've
> just got weak sockets. Of course if you don't understand the benefit
> of high spoke tension you don't think you've given anything up.

I think there's a similar sort of balance there too.

frkrygow@gmail.com

On Apr 29, 2:52 pm, Peter Cole <peter_c...@verizon.net> wrote:
>
>
> Again, if you haven't read the book then you don't have a leg to stand
> on. You're just repeating a misquote. Jobst (correctly) refers people to
> his book. That's the "advice" you should take. The "details" don't get
> "lost" if you buy the book.

Strictly speaking, you don't have to buy the book. If, that is,
you're afraid of contributing money to Jobst's nefarious plan to
dominate the world, starting with its bicycle wheels.

Just go to your library. If they don't have _The Bicycle Wheel_, they
can get it through interlibrary loan.

But Peter's right. It's a bit silly to join arguments about what the
book contains if you haven't bothered to read it.

- Frank Krygowski

Ben C

On 2008-04-30, frkrygow@gmail.com <frkrygow@gmail.com> wrote:
> On Apr 29, 2:52 pm, Peter Cole <peter_c...@verizon.net> wrote:
>>
>>
>> Again, if you haven't read the book then you don't have a leg to stand
>> on. You're just repeating a misquote. Jobst (correctly) refers people to
>> his book. That's the "advice" you should take. The "details" don't get
>> "lost" if you buy the book.
>
> Strictly speaking, you don't have to buy the book. If, that is,
> you're afraid of contributing money to Jobst's nefarious plan to
> dominate the world, starting with its bicycle wheels.
>
> Just go to your library. If they don't have _The Bicycle Wheel_, they
> can get it through interlibrary loan.
>
> But Peter's right. It's a bit silly to join arguments about what the
> book contains if you haven't bothered to read it.

I agree, that would be silly.

Peter Cole

Ben C wrote:
> On 2008-04-29, Peter Cole <peter_cole@verizon.net> wrote:

>> No. Hole deformation will change the elbow support angle slightly. It
>> will always enlarge the angle, if the angle was too big to begin with,
>> it will increase the moment.
>
> I'm not sure the elbow support angle is the important thing to measure
> here, but the perpendicular length of unsupported spoke.
>
> It seems to me that will always get shorter as you are crushing the
> fulcrum the spoke is bearing against.
>
>> If the angle was initially too small, the
>> angle may be improved, but there still may be a bending stress.

Picture a 4" nail driven half way into a wooden wall. Hang a weight from
it. There are 2 forces, shear & moment. If you hang the weight closer to
the wall, the moment goes down (eventually to zero) while the shear
remains constant.

If the weight is very close to the wall, and very heavy, the wood will
crush from shear and the nail will sag. If the nail is flexible enough
it will bend also, the bend being inside the wall surface.

If you started with the weight further from the wall, the nail would
bend outside of the wall surface, too.

If you replace the straight nail with one bent to a (downward) 90 degree
angle, flush to the wall and hang a weight, the shear crushes the wood,
the nail droops and the weight tries to open up the bent nail. Moving
the nail out from the wall does the same, but also tries to bend the
nail on the (nearly) horizontal section.

What we're seeing in this model is a new radius of curvature trying to
form inboard of the existing one. The forces are trying to create a new
bend and straighten the old one. The bigger this distance from where the
bend is to where "it wants to be", the larger the forces become. The
crushing of the material only moves the radius of curvature inward,
increasing that distance and those forces.

An angular mismatch (elbow angle to spoke line angle) will cause bending
stress at the elbow, the direction can be either way, depending on the
direction of the mismatch. Generally, the worse mismatch is for outbound
spokes, where the spoke line angle is smaller than the elbow angle. This
is evidenced by the bowing of the spoke away from the flange. Too long
elbows cause a bending stress inboard of the elbow bend. Hub hole
deformation will help or hurt in the first case, but will only hurt in
the second.

> I think there will _always_ be a bending stress unless you've got a
> straight pull spoke. So long as it's small enough you're OK though.

The bending stress for hooked spokes can be made small enough for
reliability if the wheel is built properly. That is clear from my (and
countless others) experience. Straight pull spokes don't so much solve
the problem as move it around. Heads become the problem spot rather than
elbows.


>> Those spokes will generally fail at the outside of the bend. There are
>> those who refuse to correct a bad spoke line then cite these failures
>> as evidence of the irrelevance of residual stress. It's comical,
>> really.
>
> Well I don't know who you're talking about there.

Check the archives.

>> A loaded spoked wheel doesn't deform into a uniform oval, it develops a
>> flat spot. As the length of the flat spot grows (with increasing load)
>> the bending stress on the rim increases. When that stress reaches a
>> critical value the bend is permanent. The rate at which the flat spot
>> grows with load is related to the initial spoke tension.
>
> Yes, I know, and I am familiar with Ian's FEA.
>
>> That's it. That you can ride around on a loosely spoked wheel doesn't
>> say anything about what happens when it meets a pothole.
>
> I don't think it's so easy to predict what happens when you meet a
> pot-hole.

Since a wheel stands on its spokes, strong spokes make a strong wheel. A
slack spoke is the same as a missing spoke. If the FEA's don't predict
what happens in an overload, what does?

>> Nonsense. The "defenses" of anodizing boil down to an assertion that the
>> rim extrusions are so crappy that it doesn't matter if they take an
>> additional reliability hit from anodizing.
>
> Well, you can put it like that, but that could be misleading.
>
> Designing something like a rim is an optimization problem with
> constraints on things like strength, fatigue life, stiffness, weight,
> cost and parameters like choice of alloy, shape of extrusion, whether
> you anodize.

Sure, engineering 101, but if a device has early failure, in normal
operation, in a singular mode, it's been badly optimized.

One of my cars eats head gaskets, the other one doesn't. One cracks
suspension springs, the other one doesn't. Both have similar engines and
curb weights. Predictably, the manufacturers attempt to blame the victim
(poor maintenance, excessive loads, etc.), but this doesn't hold up when
you consider that similar vehicles exposed to the same conditions don't
have the same failure rates. The thing that prevents recognition of the
true causes is that usually it's only the manufacturers who know what
the actual rates are. Manufacturing defects are understood, so are the
necessary margins needed to reduce them to tolerable limits. When car
motors or suspensions fail catastrophically in high percentages, it
means somebody has screwed up. Anybody who has worked in engineering can
tell stories about this for hours. Usually, the screw up was either bad
original engineering or good engineering turned bad by
economics/marketing driven compromises. Those of us who take
professional pride in our capabilities take a very dim view of both of
those problems and are not inclined to look favorably on the apologists
for them, particularly if they blame the victim, often (literally)
adding insult to injury. The obvious irony here (rbt) is that it always
seems to be the engineers who blame the engineering, while the
non-engineers blame the users.

Ben C

On 2008-04-30, Peter Cole <peter_cole@verizon.net> wrote:
> Ben C wrote:
>> On 2008-04-29, Peter Cole <peter_cole@verizon.net> wrote:
>
>>> No. Hole deformation will change the elbow support angle slightly. It
>>> will always enlarge the angle, if the angle was too big to begin with,
>>> it will increase the moment.
>>
>> I'm not sure the elbow support angle is the important thing to measure
>> here, but the perpendicular length of unsupported spoke.
>>
>> It seems to me that will always get shorter as you are crushing the
>> fulcrum the spoke is bearing against.
>>
>>> If the angle was initially too small, the
>>> angle may be improved, but there still may be a bending stress.
>
> Picture a 4" nail driven half way into a wooden wall. Hang a weight from
> it. There are 2 forces, shear & moment. If you hang the weight closer to
> the wall, the moment goes down (eventually to zero) while the shear
> remains constant.
>
> If the weight is very close to the wall, and very heavy, the wood will
> crush from shear and the nail will sag. If the nail is flexible enough
> it will bend also, the bend being inside the wall surface.
>
> If you started with the weight further from the wall, the nail would
> bend outside of the wall surface, too.
>
> If you replace the straight nail with one bent to a (downward) 90 degree
> angle, flush to the wall and hang a weight, the shear crushes the wood,
> the nail droops and the weight tries to open up the bent nail. Moving
> the nail out from the wall does the same, but also tries to bend the
> nail on the (nearly) horizontal section.
>
> What we're seeing in this model is a new radius of curvature trying to
> form inboard of the existing one. The forces are trying to create a new
> bend and straighten the old one.

OK, I follow you so far.

> The bigger this distance from where the bend is to where "it wants to
> be", the larger the forces become.

The shear force is always constant. But yes, the bigger the distance the
bigger the moment. I would measure the moment as the perpendicular
distance from the line of force (i.e. the straight part of the spoke) to
the nearest point to that line on the hub where it's supported.

> The crushing of the material only moves the radius of curvature
> inward, increasing that distance and those forces.
>
> An angular mismatch (elbow angle to spoke line angle) will cause
> bending stress at the elbow, the direction can be either way,
> depending on the direction of the mismatch. Generally, the worse
> mismatch is for outbound spokes, where the spoke line angle is smaller
> than the elbow angle. This is evidenced by the bowing of the spoke
> away from the flange. Too long elbows cause a bending stress inboard
> of the elbow bend. Hub hole deformation will help or hurt in the first
> case, but will only hurt in the second.

Thinking of the nail in the case where you pull it at an angle greater
than that of its original bend, the original bend will open up as the
new bend forms. The more the wall crushes while you're doing this the
larger the final angle on both ends. If the wall crushes it also means
the old bend moves towards the wall more quickly as you pull.

But I can't see that any of that matters.

The part of the spoke that's embedded inside the hub doesn't matter.
It's supported. It doesn't matter what the angle on its elbow is.

>> I think there will _always_ be a bending stress unless you've got a
>> straight pull spoke. So long as it's small enough you're OK though.
>
> The bending stress for hooked spokes can be made small enough for
> reliability if the wheel is built properly. That is clear from my (and
> countless others) experience.

Agreed.

> Straight pull spokes don't so much solve the problem as move it
> around. Heads become the problem spot rather than elbows.

Interesting-- are you saying you get fatigue at the heads because
they're stress risers or something?

[...]
>>> Nonsense. The "defenses" of anodizing boil down to an assertion that the
>>> rim extrusions are so crappy that it doesn't matter if they take an
>>> additional reliability hit from anodizing.
>>
>> Well, you can put it like that, but that could be misleading.
>>
>> Designing something like a rim is an optimization problem with
>> constraints on things like strength, fatigue life, stiffness, weight,
>> cost and parameters like choice of alloy, shape of extrusion, whether
>> you anodize.
>
> Sure, engineering 101, but if a device has early failure, in normal
> operation, in a singular mode, it's been badly optimized.
>
> One of my cars eats head gaskets, the other one doesn't. One cracks
> suspension springs, the other one doesn't. Both have similar engines and
> curb weights. Predictably, the manufacturers attempt to blame the victim
> (poor maintenance, excessive loads, etc.), but this doesn't hold up when
> you consider that similar vehicles exposed to the same conditions don't
> have the same failure rates. The thing that prevents recognition of the
> true causes is that usually it's only the manufacturers who know what
> the actual rates are. Manufacturing defects are understood, so are the
> necessary margins needed to reduce them to tolerable limits. When car
> motors or suspensions fail catastrophically in high percentages, it
> means somebody has screwed up. Anybody who has worked in engineering can
> tell stories about this for hours. Usually, the screw up was either bad
> original engineering or good engineering turned bad by
> economics/marketing driven compromises. Those of us who take
> professional pride in our capabilities take a very dim view of both of
> those problems and are not inclined to look favorably on the apologists
> for them, particularly if they blame the victim, often (literally)
> adding insult to injury.

I share the dim view. Next time get a Japanese car

> The obvious irony here (rbt) is that it always seems to be the
> engineers who blame the engineering, while the non-engineers blame the
> users.

A wheel-builder is not exactly a user. It would be reasonable for Mavic
to say to someone it's your fault your rim cracked because you used too
much tension.

Peter Cole

Ben C wrote:
> On 2008-04-30, Peter Cole <peter_cole@verizon.net> wrote:

>> Picture a 4" nail driven half way into a wooden wall. Hang a weight from
>> it. There are 2 forces, shear & moment. If you hang the weight closer to
>> the wall, the moment goes down (eventually to zero) while the shear
>> remains constant.
>
> Thinking of the nail in the case where you pull it at an angle greater
> than that of its original bend, the original bend will open up as the
> new bend forms. The more the wall crushes while you're doing this the
> larger the final angle on both ends. If the wall crushes it also means
> the old bend moves towards the wall more quickly as you pull.
>
> But I can't see that any of that matters.
>
> The part of the spoke that's embedded inside the hub doesn't matter.
> It's supported. It doesn't matter what the angle on its elbow is.

The point it that it's not really supported if it crushes. That allows
the spoke to bend there. If the hole didn't deform, the peak stress for
the cantilever load would be right at the hole's edge. As the hole
deforms, the peak stress must move inward (some amount) effectively
increasing the moment (overhang). Hole deformation, by itself would seem
to increase fatigue, not reduce it.


>>> I think there will _always_ be a bending stress unless you've got a
>>> straight pull spoke. So long as it's small enough you're OK though.
>> The bending stress for hooked spokes can be made small enough for
>> reliability if the wheel is built properly. That is clear from my (and
>> countless others) experience.
>
> Agreed.
>
>> Straight pull spokes don't so much solve the problem as move it
>> around. Heads become the problem spot rather than elbows.
>
> Interesting-- are you saying you get fatigue at the heads because
> they're stress risers or something?

They may be, Peter White seems to think so, preferring the smoother head
transition of Wheelsmith spokes to DT's. But, in any case, heads are
cold formed, so there will be residual stresses there too. It's
interesting to think about how much of the spoke tension actually gets
to the head. The relative bend radius and spoke radius would seem to
restrict much of the spoke tension from getting there. A flexible cable
bending over a large radius would transmit all its tension to the anchor
point, but a spoke is far from that. With a straight pull spoke all the
tension would go to the head.

>> One of my cars eats head gaskets, the other one doesn't.

> I share the dim view. Next time get a Japanese car

I've had those, and German, and Swedish, and Italian. US doesn't have a
monopoly on bad engineering, although they may have the lead.

>> The obvious irony here (rbt) is that it always seems to be the
>> engineers who blame the engineering, while the non-engineers blame the
>> users.
>
> A wheel-builder is not exactly a user. It would be reasonable for Mavic
> to say to someone it's your fault your rim cracked because you used too
> much tension.

Perhaps, but I'd have more sympathy if they didn't hide their max
tension specs behind password-protected web sites. But this presumes the
guilt of the builder. Who's to say that the spoke bed failures were the
result of over-tension -- OK, rhetorical question, we know who. The real
question is whether the spoke bed failures that have been reported were
a result of out-of-spec tension -- I doubt we'll ever know.

What Keith Bontrager said (over 9 years ago on rbt):
http://tinyurl.com/6sx5nz

"With Open 4 and Open Pro sections this might no longer be a good idea.
In those sections Mavic has taken the extrusion process a little
farther. Many of the walls are thinner (though the actual spoke bed is
thick) and the section is "larger" for a given rim weight. Even though
they are socketed (double eyelets), many of the failures I observe in
shop recycling bins are due to spoke bed failures, usually drive side
rear spokes. "

"But that's the design compromise that makes the newer rims a little
different than the older style, thicker wall rim. It raises the short
term performance characteristics at the cost of fatigue strength (and
brake wall wear replacement intervals). "

"If you make the walls of the rim continuously thinner and the section
larger it's clear that, at some point,you can't ignore spoke bed loads
and subsequent fatigue failures, even with reinforcement schemes.
That's the point at which these rims are."

Re MTB rims (but applicable):

"They are not that many riders who really NEED them though. A heavier
rim with thicker walls is easier for everyone to deal with in almost
every respect, is stronger on the trail, and lasts longer too. The
difference in performance to a recreational rider between a 450g + rim
and a 400 gram rim is small but the difference in durability on the
trail is big."

Perhaps the final word:

"Of course, the light section rims also make wheels that are more
attractive and easier to sell, so most popular rims are that way. These
are the only ones that get attention from the press as well.."

Comment on sockets:

"The sockets are fit into the rim and require (due to mounting
requirements and tolerance accumulation) some deflection of the rim
inner walls to bear much of the load. The cups are also not shaped very
well to support the load."

The basic point he makes is that these rims aren't designed to be
durable and may be actually intended/known to fatigue before they wear
out. All of this to shave a few grams (and perhaps make them pretty).
He's a guy who sells rims, so he ought to know.

Ben C

On 2008-04-30, Peter Cole <peter_cole@verizon.net> wrote:
> Ben C wrote:
>> On 2008-04-30, Peter Cole <peter_cole@verizon.net> wrote:
>
>>> Picture a 4" nail driven half way into a wooden wall. Hang a weight from
>>> it. There are 2 forces, shear & moment. If you hang the weight closer to
>>> the wall, the moment goes down (eventually to zero) while the shear
>>> remains constant.
>>
>> Thinking of the nail in the case where you pull it at an angle greater
>> than that of its original bend, the original bend will open up as the
>> new bend forms. The more the wall crushes while you're doing this the
>> larger the final angle on both ends. If the wall crushes it also means
>> the old bend moves towards the wall more quickly as you pull.
>>
>> But I can't see that any of that matters.
>>
>> The part of the spoke that's embedded inside the hub doesn't matter.
>> It's supported. It doesn't matter what the angle on its elbow is.
>
> The point it that it's not really supported if it crushes. That allows
> the spoke to bend there. If the hole didn't deform, the peak stress for
> the cantilever load would be right at the hole's edge. As the hole
> deforms, the peak stress must move inward (some amount) effectively
> increasing the moment (overhang). Hole deformation, by itself would seem
> to increase fatigue, not reduce it.

OK, well I think I understand what you're saying now. But as the hole
deforms, the spoke is still touching the sides of the hole all the way
to the edge, so it's not clear that it's bending about any point except
where it leaves the hole.

Unless it's starting to behave a bit like a cable wrapped around a
cylinder, where the point it's bending around is effectively moving
continuously as it bends. It's probably good if it is doing that since
the result will be that the load is distributed over a larger area.

>>>> I think there will _always_ be a bending stress unless you've got a
>>>> straight pull spoke. So long as it's small enough you're OK though.
>>> The bending stress for hooked spokes can be made small enough for
>>> reliability if the wheel is built properly. That is clear from my (and
>>> countless others) experience.
>>
>> Agreed.
>>
>>> Straight pull spokes don't so much solve the problem as move it
>>> around. Heads become the problem spot rather than elbows.
>>
>> Interesting-- are you saying you get fatigue at the heads because
>> they're stress risers or something?
>
> They may be, Peter White seems to think so, preferring the smoother head
> transition of Wheelsmith spokes to DT's.

Yes I remember reading something about that-- the heads were popping off
some spokes I think.

> But, in any case, heads are cold formed, so there will be residual
> stresses there too. It's interesting to think about how much of the
> spoke tension actually gets to the head. The relative bend radius and
> spoke radius would seem to restrict much of the spoke tension from
> getting there. A flexible cable bending over a large radius would
> transmit all its tension to the anchor point, but a spoke is far from
> that. With a straight pull spoke all the tension would go to the head.

A herring I raised in a thread long ago was about how well supported the
spoke was in the hub hole. If the hub hole doesn't deform much at all
(steel hub for example) then the edge of it is probably digging into the
spoke somewhere and that's where all the force is concentrated. If
they're deformed together nicely then the situation is a little bit more
like a cable wrapped around a cylinder. You might get some force going
all the way to the head, but in any case, the force is spread over a
greater area, so less stress.

I was speculating that this might be a benefit of putting oil on the
elbows as Gene does.

[...]
>> A wheel-builder is not exactly a user. It would be reasonable for Mavic
>> to say to someone it's your fault your rim cracked because you used too
>> much tension.
>
> Perhaps, but I'd have more sympathy if they didn't hide their max
> tension specs behind password-protected web sites.

Absolutely. Every other manufacturer seems to be capable of supplying a
fold-out instruction sheet larded with safety warnings in 32 languages.
That would be a good place to state the recommended tension. It wouldn't
be a bad idea also to put it on the sticker on the rim itself along with
the max tyre pressure.

Even once you've got the password you have to look quite hard in all
those pdfs for recommended tensions. I don't think I ever found any for
the normal rims, only for the "boutique" stuff.

> But this presumes the guilt of the builder. Who's to say that the
> spoke bed failures were the result of over-tension -- OK, rhetorical
> question, we know who. The real question is whether the spoke bed
> failures that have been reported were a result of out-of-spec tension
> -- I doubt we'll ever know.
>
> What Keith Bontrager said (over 9 years ago on rbt):
> http://tinyurl.com/6sx5nz
[...]

Great link, thanks!

Tom Sherman

Peter Cole wrote:
> [...] Those of us who take
> professional pride in our capabilities take a very dim view of both of
> those problems and are not inclined to look favorably on the apologists
> for them, particularly if they blame the victim, often (literally)
> adding insult to injury. The obvious irony here (rbt) is that it always
> seems to be the engineers who blame the engineering, while the
> non-engineers blame the users.

Then there are the ex-metallurgists who are resentful of engineers.

--
Tom Sherman - Holstein-Friesland Bovinia
The weather is here, wish you were beautiful

Tom Sherman

Peter Cole wrote:
> Ben C wrote:
>> On 2008-04-30, Peter Cole <peter_cole@verizon.net> wrote:
> [...]
>>> One of my cars eats head gaskets, the other one doesn't.
>
>> I share the dim view. Next time get a Japanese car
>
> I've had those, and German, and Swedish, and Italian. US doesn't have a
> monopoly on bad engineering, although they may have the lead.
> [...]

Nonsense. Buy a vintage British car, and you will gain understanding.

--
Tom Sherman - Holstein-Friesland Bovinia
The weather is here, wish you were beautiful