Index Home About Blog
From: (Jobst Brandt)
Subject: Re: Broken Spokes
Date: 27 May 1998 20:55:02 GMT

Chris Hall writes:

> I've been suffering from a rash of broken spokes, almost always on
> the rear on the cog side. I've tried using a new wheel but the
> problem remains. The breakages tend to happen not when I'm climbing
> or pushing it but almost always when I'm taking it easy.

You don't say what sort of wheel this is or who built it.  On the
surface of it, the most common cause for spoke breakage, assuming the
materials are reasonable quality, is that the spokes were not stress
relieved after building the wheel.  You can still do that and improve
the durability of the wheel but you may get another couple of failures
after that.  Stress relieving is described in the FAQ:


Subject: 8.46  Stress Relieving Spokes
From: Jobst Brandt <>

> I wonder if "stress-relieving" is entirely correct?  I see it as a
> yielding/hardening process, in which the yield load is increased by
> embedding the spoke elbow in the hub, bending the elbow to a
> different angle, etc.  When unloaded from a high load, this area of
> the spoke should be more or less elastic.

> So I think the term should be "overloading" or "hardening" -- any
> thoughts??

Yes.  I am certain that the concept of stress relieving is obscure to
many if not most people because after seeming to understand it,
comments like this one surface.

A spoke is cold formed from wire that is (at least DT) as hard and
work hardened as it will get.  The process after tensioning a wheel
does not further harden the spokes.  The wire is straightened by
running it through staggered rollers in X and Y directions.  The
rollers have, like a degausser, ever diminishing excursions.  This
gets rid of the natural curl left from being shipped in a coil.  If
the wire was not curled before winding it would be a dangerous weapon
on the spool because if the end got loose, all hell would break loose,
making a huge birds nest.

Anyway, the straightening process leaves the spoke with internal
stresses that are well balanced and relatively low.  I haven't given
this a lot of thought but it seems that if there were a large number
of rollers, the stress might approach zero.  After this process, the
spoke gets its head forged on is cut to length, threaded and, and
lastly its head is crudely but accurately knocked to one side to
produce the elbow.

The threads, head, and elbow, contain metal that went beyond yield as
well as metal that did not.  The metal in these zones is stressed one
part against another, one wanting to return to the condition before or
during forming, and the other to the formed shape.  On lacing the
spokes into a wheel, the elbow is additionally bent (brought to yield)
and upon tensioning this stress remains at or reaches the yield point
it if it wasn't already there.  The threads, that have locked in
stresses (all stresses are ultimately tension and compression) is
selectively stressed at the contact points with the nipple thread and
in tension in the core that already was in tension because thread
rolling stretches this portion of the spoke slightly.

The result is that a freshly built wheel has spokes locations where
stress is guaranteed to be at the yield point.  If used this way, the
cyclic load with each wheel revolution will cause spokes to fail in
fatigue at these high stress points.  The load on the wheel only
unloads spokes but because the spoke is operating up to the yield
point, it cannot withstand many stress cycles.  The greater the load
(unloading) the sooner it will fail because when operating close to
the yield stress a metal cannot survive.  Only the lightest riders who
ride smooth roads might not experience failures.

The purpose of stress relieving is to relax these high stress points
in the spokes.  The purpose is not to bed the spokes into the hub.
Bedding in has usually already occurred sufficiently for practical
purposes during tensioning.  By stretching each spoke with a strong
grasp, its tension can be temporarily increased by 50 to 100%.
Because a spoke operates at about 1/3 its yield stress, this operation
has little to no effect on the spoke as a whole.  Stress relieving
affects only the microscopic zones of the spoke that are at high
stress (near or at the yield stress).  By stretching these zones and
relaxing the load afterward, the margin to yield is as much as the
overload or more.

A whimpy grasp of the spokes during stress relieving is close to
worthless and dropping the wheel, bending it in a partially opened
drawer, pressing on the rim with the hub on the floor and the like is
as close to useless as you can get.  The only method that I have seen,
but do not recommend, is walking on the wheel while wearing tennis
shoes and carefully stepping on each pair of crossed spokes.  The
problem with this is that it bends the rim and it is difficult to be
sure each spoke gets a good stretch.

IT IS STRESS RELIEVING!  At least that's what I am referring to by the

Jobst Brandt      <>


From: (Jobst Brandt)
Subject: Re: adjusting spokes with the tool?
Date: 7 Oct 1998 16:42:51 GMT

Anonymous writes:

> Anyone can make a wheel true.  Building a true, round and balanced
> wheel is an art.  Built properly a wheel can last many years and
> thousand of miles without a broken spoke of loss of alignment (with
> minor maintenance).  Wheel building is a philosophy not a chore.

This perception was nearly universal when I first wanted to build
wheels.  In fact it was claimed to be so difficult that unless I was
as gifted as the old masters, I shouldn't try.  I don't believe in
shrouding a technically simple process in mystique.  I believe most of
the bad advice I got was not through ill will but more from technical

It was this aspect that sharpened my interest and led me to write "the
Bicycle Wheel" so that others would be spared the scavenger hunt I
went through to discover the facts.  I found much more than I
expected, because the bicycle wheel does not work the way it was
commonly thought.  It's stress is not higher in the top spokes from
which the hub was believed to hang.

That stress relieving was the missing ingredient in most wheels that
came from some of the "best" wheel builders, who prided themselves in
super straight wheels, was probably the most important part of the
process that I discovered.  Once exposed, the mystery can be reduced
to a scientific and repeatable process.  I think it has reached that
point for those willing to read.  I say that because there are bicycle
shops that sell the book but whose mechanics have chosen not to read
it and continue to work in the myth and lore expressed anonymously

Jobst Brandt      <>

From: Mike Prime <>
Subject: Re: Best way to bed spokes ...
Date: Fri, 05 Feb 1999 13:30:29 -0700

Rafael Raban wrote:

> I really do not understand your arguement with regard to "residual
> stresses."  Residual stresses that are generated during the elbow
> bending process are *GOOD*.  When you bend an elbow, you create a
> compressive residual stress on the inside surface of the elbow, and a
> tensile residual stress on the outside of the elbow.

The residual stresses caused by bending plastically and then unloading
elastically have the opposite sign to this. The side that was originally
in tension has compressive residual stress and vice versa.

At the risk of having this come through poorly via ASCII, let me try to
sketch this (try viewing with courier or other fixed-width font). The
following is the through-thickness variation of stress in a bent beam.

        Sy  -1.5Sy     -0.5Sy
        |      \           \
        |       \           \
        |        \           \
        |         \           \
   -----           \      -----
  |                 \     \
  |                  \     \
  |                   \     \
  |                    \     \
 -Sy                 1.5Sy   +0.5Sy

Sy = yield stress
Assumption: elastic-perfectly plastic, elastic unload
The reason the peak stresses are 1.5 Sy for unloading is that the
unloading distribution must have the same moment as the load.

Conclusion: the side originally in tension has compressive residual

A similar figure appears in the introductory textbook from which I
learned solid mechanics:
"Introduction to mechanics of solids," Igor Popov. I trust that newer
textbooks contain a similar treatment.

Michael B. Prime, Ph.D., P.E.

From: (Jobst Brandt)
Subject: Re: Spoke stress relieving method
Date: 6 Apr 1999 00:35:05 GMT

Ray Green writes:

> I need an alternate method for stress relieving spokes when I'm
> building a wheel.

> So I thought about laying the wheel down on the floor and using the
> heel of my foot to apply some pressure at the crossing point of two
> spokes, then going around the wheel at each cross, then turning over
> and repeating.

Doing this gives the wheel a one sided load and must be done with
special care.  The first time I saw a wheel builder do this it was a
5ft or so little guy who walked on the spokes carefully with tennis
shoes.  It was this and the experience I had with trying to break
spokes on an unreliable wheel that had formerly broken many spokes
that made me realize that stress relieving was the key to durable
wheels... besides good spokes.

> Is there a danger of applying too much pressure and damaging the rim,
> hub or spokes?

Yes because you could put your entire weight on one spoke.

> Can this technique still be used for also determining the maximum
> spoke tension the rim can take (at the onset of buckling).

No, because what you can grasp with the hands raises four spokes an
amount that coincidentally is about the safety margin a wheel should

> Are there any alternate stress relieving methods.

People have devised various ways of not using the hands.  From what I
have heard, wheel building machines press inward from two sides on two
pairs of spokes.

Jobst Brandt      <>

From: (Jobst Brandt)
Newsgroups: rec.bicycles.misc
Subject: Re: Thanks to Sheldon Harris!
Date: 30 Aug 1999 17:21:41 GMT

Andy Katz writes:

>> There's more to it than tension.  The causes and remedies for spoke
>> failure are described in great length and detail in "the Bicycle
>> Wheel" that you can find at most bike shops or at

> Now I've just finished reading the sections pertaining to spoke
> failure. It's very interesting that you point out that failure is
> common in climbing, since that's where I've had the problems. You also
> say weight is involved. I weigh about 240 (which you may interpret to
> mean that I really weigh 255-260;), and have put about 3,000 miles on
> these wheels, with no truing problems or spoke breakage until very
> recently. Someone suggested that the wheel should be rebuilt with new
> spokes, otherwise the problem will continue.

Climbing is not the cause however.

> I'm able to reduce or increase spoke tension, retrue, and stress
> relieve the wheel myself. But I don't feel confident in rebuilding a
> rear wheel myself, particularly as I'll be riding in the AIDS Ride
> soon, which leaves me less latitude to learn by my mistakes (I bought
> your book to study building a front wheel).

> I wonder if you agree that the wheel out to be rebuilt for maximum
> confidence, or are intermediate measures adequate?

I don't know what sort of spokes you have and how often they have been
breaking.  If they are swaged DT spokes, and you approach it as though
it were a freshly built wheel, you should be able to make it into a
durable wheel by improving the spoke line of the outbound spokes,
balancing tension and stress relieving the whole wheel.

My first experience with this was years ago when I had frequent spoke
failures and wanted to be done with it by attempting to break any
spoke that was about to break.  That is where I stumbled onto stress
relieving.  I broke two more spokes by forcefully stress relieving the
wheel and then there were none for a long time.

I don't think you should shy away from building a rear wheel.  It is
no more difficult than a front wheel except that you can't forget
where you are in the process, left or right side.  I tested the
building instructions on my grade school sons when I first wrote it.
Both were able to build a set of wheels without any assistance from
the pages from a copier.

Jobst Brandt      <>

From: (Jobst Brandt)
Newsgroups: uk.rec.cycling,
Subject: Re: Spoke breakages
Date: 29 Nov 1999 23:33:09 GMT

Nick Maclaren writes:

> I am neither a metallurgist nor engineer, and cannot be certain
> whether mechanical "annealing" is a standard practice nor whether
> the term "stress relieving" has ever been used for it.

On the one hand, I am glad to see that you read the book apparently
thoroughly, on the other hand, I am disappointed that the text and its
examples were not able to conveigh the effect that is taking place.
Stress relieving has nothing to do with work hardening nor seating or
setting the spoke into the hub or spoke nipple.  It is the reduction
of stress in the spoke.

> But I know quite enough about the strength of materials to know that
> there are alternative explanations for why stressing spokes as part
> of building might improve their reliability.  And one alternative
> explanations is (a) simpler, (b) traditional practice and (c) much
> easier to explain in terms of standard material strength theory.

> What I want is a PROPER scientific reference to any evidence that
> (a) the technique really does reduce the residual stresses in the
> spokes (e.g. electron microscopy) and (b) that it really does increase
> their resistance to fatigue (i.e. a proper statistical test, or the
> mathematical calculations.)

Unless you are a Maxwell deamon, you cannot enter into the material to
take roll call of molecules for residual stress.  We can only see the
side effects.  In the example below, a bent piece of hardened wire
placed in a tensile tester will become straight when taken beyond
yield and be perfectly straight when taken out of the tester.  It has
no residual stress because all parts were taken to yield and all parts
were relaxed whn the load was removed.  For there to be any residue,
there must be a gradient in the elastic modulus.

From: Mike Prime <>
Newsgroups: uk.rec.cycling,
Subject: Re: Spoke breakages
Date: Tue, 30 Nov 1999 08:43:44 -0700

Nick Maclaren wrote:
> In article <>, Tho X. Bui <> wrote:
> >
> >I think the problem here is that you are viewing the "stress relieving"
> >of bicycle spokes as something that occurs within the material.
> >
> No, that is not what I meant.  The question is whether the activity
> described in The Bicycle Wheel as "Stress Relieving" works by making
> sure that the spokes are properly bedded in and straightened as
> necessary or by changing the internal stresses of the material.
> Jobst Brandt says the latter - but provides no evidence.

The stress relieving of spokes as described by Jobst does indeed "change
the internal stresses of the material." This process is used for
aluminum much more commonly than for steel because steel can be stress
relieved by heating whereas aluminum loses too much strength at the
temperatures required for stress relief.

Common temper designations for aluminum reflect the stress relief
process. For example in 7075-T7451, the final 51 refers to "stretch"
stress relief. 52 refers to compressive stress relief--it works in
either tension or compression. For more info see the ASM Handbook Volume
2 (nonferrous alloys) or any good reference on aluminum.

For some results of some actual residual stress measurements I did on
7050 aluminum plate before and after stretching see:
The residual stress was reduced by about a factor of 10 by the stress
relief process.

Somewhere, I have a figure from a presentation that illustrates how
stretch stress relief works. If there is sufficient interest, I can dig
it out and out it on my web page.

Michael B. Prime, Ph.D., P.E.

From: (Jobst Brandt)
Newsgroups: uk.rec.cycling,
Subject: Re: Spoke breakages
Date: 13 Dec 1999 22:07:40 GMT

Nick Maclaren writes:

> The "tight rivet" effect is when one rivet is much tighter than
> its neighbours.  Because the others can shift marginally, the tight
> rivet takes a higher stress than its neighbours, and therefore
> usually fails  first.  I.e. it is not the loose rivets that fail,
> but the tight ones.  Look in works on 19th century shipbuilding
> for the references.

So what does that have to do with bicycle wheels?  Spoke failure is
what is being discussed here and in the range of spoke tensions of a
reasonable wheel, there are no tighter and looser spokes within a set
or the wheel would be crooked.  What's more. spokes on the right and
left side of rear wheels often have a tension ratio of 2:1 and it is
the loose left side spokes that often fail.

> You are definitely wrong that spokes do not bed in with steel
> flanges.  I inspected a new spoke, tightened it appropriately,
> removed it and reinspected it.  I could clearly see where the
> zinc had formed to match the shape of where it fitted into the
> flange.  What do you call that, if not "bedding in"?

That they deform zinc is not unusual but chromed steel hubs, as they
most often are, do not appreciably "bed in".  By that I mean that they
don't begin to approach the contact to which aluminum flanges deform.
No doubt, nicks are visible to the extent that one can reconstruct what
the spoke pattern might have been on a bare hub but it does not
develop that which makes people believe that this is an important
aspect of stress relieving or for that matter the essence of it.

>>> The point is that your "stress relieving" procedure would complete
>>> the bedding in and straightening, if this were the case, and thus
>>> improve the life of the wheel.  And over-stressing structures in
>>> order to bed things in and straighten wires is a VERY old practice.
>>> Probably Roman or earlier :-)

>> Well?  Which way do you want it?  Did the ancients stress relieve or
>> not?  You say the Romans did it?  The point is that it was not common
>> practice before I wrote about it and being contentious, people who
>> felt upstaged, have since spent much time constructing specious
>> arguments to the contrary.  In spite of that, the procedure has become
>> common practice for wheel building.

> Well, I have been following your procedure since the late 1960s,
> though I then regarded it as a way of bedding in and straightening
> the spokes.

Why do you make that assumption and what do you believe causes spoke
failures if it isn't residual stress overlayed on stress from
tensioning?  Bear in mind that spokes are generally tensioned no more
than 1/3 their yield stress and that their failure is from tensile

> Overtightening fixtures as a way of bedding them in is most
> definitely ancient, and I first saw the technique in a motor vehicle
> manual.  It said something like "tighten up to 50 pounds-feet
> without the special washer in place to bed the bolt in, remove the
> bolt, replace the washer, and retighten to 25 pounds-feet."

Please explain that.  What sort of bolt and where, and what was
supposedly being accomplished?  Why would a bolt be better off for such
a procedure?

> Straightening wire by pulling it sideways is also ancient, and has
> been standard practice when making wire fences by hand probably
> since they were invented.  Every farm worker knows THAT one, and
> it is a standard trick when tightening ropes.

So what's your point and why do these people who were doing this from
antiquity do it?  Somehow it was not known in the bicycle industry
judging from the ubiquitous failures in the past.  No reasonable
bicyclist went on a ride without spare spokes and a freewheel remover.

>>> Is there any hard scientific evidence that this is not the reason
>>> that the way that stress relieving procedure works, and that your
>>> explanation is the correct one?

>> How hard do you want it?  Tensile testing, modeling, and practical
>> experiments demonstrate the concept independently from the positive
>> effects it has on wheels whose spokes have been subjected to this
>> process.  You sound like a believer in the flat earth concept.  They
>> also were impossible to convince.

> As I have posted all along, your procedure works - that is not the
> matter being debated.  You are confusing association with causality.

I have clearly explained how residual stress from manufacture and
installation of spokes can be observed and that it exists beyond a
doubt.  I point out how these stresses can be reduced.  Fatigue
failures occur in steel when cyclic loading oscillates near the yield
point.  Spokes not stress relieved, break readily.  To what are you
attributing spoke failure at the elbow and threads if it isn't
residual stress?

> The issue is whether it works by relieving internal stresses or by
> bedding in and straightening the spokes.  You may be right that it
> is the former, but have presented no evidence that I can see.  You
> say that all those experiments prove your point, but seem reluctant
> to provide any references.

Since this procedure works for steel hubs with no functional bedding
in, how do you explain that it works there as well?  How does this
affect spoke failure in the threads?

>>> The other problem that I have is that I tried modelling the internal
>>> stress model using a simplistic structure and what I understand of
>>> what you are saying, and the effect of the overstressing in use was
>>> to cause the overstressed part to flow, and the load distributed
>>> more uniformly.  This would mean that wheels would "stress relieve"
>>> themselves in use, which is not what is observed in practice.

>> I think you didn't read the book after all.  The highest stress spokes
>> see in a built up wheel are in the static condition with no load on
>> the wheel.  Riding the wheel unloads spokes.  It does not additionally
>> tension spokes.  This is a prestressed structure and such structures
>> generally work in that mode.  If riding a wheel were to cause spokes
>> to approach yield, then spokes would break continually.

> Are you SERIOUSLY saying that spokes will break on wheels that are
> not being used?

Where do you see that stated or implied?

> If no part of a spoke every approached yield point, they would never
> break.  The point is that, as you say in your book and can be
> observed on a broken spoke, part of a spoke approaches yield point
> while the structure as a whole is still well below it.

I didn't say that.  That is your interpretation.  Let's take one point
at a time because you mix and switch so readily that it defies

Spoke work in a wheel by relaxing their tension, not by increasing it.
With a bump in the road, spokes above the hub do not increase in
tension.  Therefore, spokes will not bed in from use because their
highest tension occurs when sitting idle.

That spokes are at the yield point and don't break immediately also
occurs because they only slacken from operating loads and then return
to their highest stress.  This high stress is not from tension but
rather from tension overlayed with residual stress from manufacture
and installation into the wheel.

> My analysis was what the effect of that occurring would be.  You
> may be right that it was unrealistic, which is why I was asking for
> a reference to a better one.

>>> That is why I asked for references to hard scientific papers, a
>>> mathematical explanation or other such research.

>> The book gives both an explanation and experimental methods by which
>> you can convince yourself of these effects.  Had someone written about
>> it previously, I would not have written "the Bicycle Wheel".  Much of
>> what the book contains could previously not be found in any
>> literature.  The work of Karl Wiedemer is cited.

> Your book contained no mathematics on this topic, no results of
> proper controlled experiments on the cause of the effects and no
> references that I could look up.  In particular, I was extremely
> surprised that it didn't seem to contain any references to where
> you had published your analyses and experiments in the scientific
> literature.

Well!  I guess that means it is all wrong as you state.

> Yes, I have tried doing a literature search, but found nothing.

As you see, I found more than you.  With help I located Karl
Wiedemer's publication on the subject.  Prof. Wiedemer, now retired,
did his analysis at the same time I did and he also had no references
because it was new work in a field that had progressed without
analysis for a long time.  Prof. Pippard in England wrote extensively
on the subject but never discovered the mode of wheel loading and
deflections that I and Wiedemer presented.  As I said, I made the
analysis by measurement and was rejected by professors of engineering.
When I presented the finite element analysis, these same people chose
to change the subject and get back to "serious" work.

I think the burden of proof lies with you.  The book has had peer
review for nearly 20 years and found to stand on its own.  Your
challenge comes a bit late and is based entirely on personal beliefs
with no explanation other than, on the one hand to say this is nothing
new and on the other, that it must be wrong because I don't cite prior
work.  There wasn't any as you see.

Jobst Brandt      <>

From: (Jobst Brandt)
Subject: Re: Question for Jobst
Date: 28 Jul 2000 00:23:29 GMT

cLaWs ??? (who dat?) writes:

>> To stress relieve, grasp most parallel pairs of spokes and squeeze
>> them together as though you were trying to break them by over
>> tensioning.  This relaxes the high stress points by yielding them.

> How hard should I squeeze?  Hard enough to make the rim move
> laterally in the stand?  Only hard enough to make the spokes flex?
> Hard enough for my knuckles to turn white and my fingertips to
> nearly bleed?

Use some common sense.  I do it with bare hands, one hand on each side
of the wheel, making a quick grasp.  Unless you are unusually strong
and have more calloused hands than I, you won't damage the wheel.  The
exception would be rims under 300g like Scheeren Weltmeisters that at
280g take a bit of care.  The wheel should withstand the stretching
grip of an athletically inclined person.  I can do one wheel, once
around before I feel my hands need a rest.  I wouldn't do a whole lot
of wheels in succession without leather gloves to protect my hands.

> I have been doing the first one, watching the rim deflect about 2-3
> mm, and it seems to work; I am afraid I may be over doing it and
> shortening the rim or spoke life.

Not to worry, you won't shorten spoke life.  You might bend the rim,
but only if the spokes are already too tight.

The way I discovered this was that I had a pair of wheels that forever
broke spokes, wheels that I subsequently seldom used.  I had them
shipped to me on a bike tour after I had crashed both wheels on
Monitor Pass CA.  After riding from Reno (NV) to Fort Bragg (CA) on
the Pacific Coast. I had two spoke failures.  With a 5-speed Regina FW
but a large flange hub, I could get at the spokes by disassembling the
FW at a gas station, I could take off the rotor to expose the spokes
for removal.  (There were no bicycle shops around in those days.)

I had plenty of spokes, as one did then, and began replacing them,
each time trying to break others that might already be cracked by
"stress relieving" although I was unaware of the effect at that time.
In all I replaced five spokes, as I recall, and reassembled my FW.
Afterward, I rode that wheel for a long time without further failures.

A couple of years later, in Florence (I) I saw a team mechanic finish
a wheel and lay it on the floor to walk on the spokes with his tennis
shoes.  I asked what this awful looking routine did to which he
replied with a sly grin that it made wheels last a long time.  It was
then that I put it together.  You see, I had not noticed that my bad
wheels stopped breaking spokes after I had stretched them, they just
worked the way wheels should.  Seeing Mara walk on wheels that he had
built for Gastone Nencini caught my attention.

I am dismayed that I cannot persuade Holland Mechanics to call their
"Stabilizer" what it really is, a stress reliever on their web site: (under PRODUCTS, STABILIZER).  This
machine presses spokes in pairs from opposite sides of the wheel
toward the central plane, thereby stretching and yielding their high
stress points to relax residual stress concentrations.

They remind me of my experience explaining to the former Wheelsmith
bicycle shop how important this process is and how it achieves its
effect, to which then they responded by calling it "seating the
spokes" or "pre-stressing".  This became more obvious when Wheelsmith
reviewed my book for Bicycling Magazine, panning it as light weight,
lacking depth.

Jobst Brandt      <>

From: (Jobst Brandt)
Subject: Re: Life of frame after crash
Date: 17 Aug 2001 22:04:29 GMT

Tim McNamara writes:

>> The wrinkle has typical residual stresses of any element that has been
>> deformed, part of the metal having yielded and part not, or at least
>> not as much.  There is always 'spring back" or the desire to do so and
>> this is the evidence of a residual couple.  If the whole piece yielded
>> equally, there would be no spring back as we typically see when bending
>> a coat hanger wire, for instance.

> So, when I bend a coat hanger to, say 120 degrees, it springs back to
> 90 (to pull numbers out of thin air).  Is this what you're talking
> about?

> Is there residual stress still present in the bend of the coat
> hanger, or was that stress relieved by the visible "spring back?"

The spring-back would be complete if there weren't a residual stress
holding it at 90 degrees, part of the material wanting to straighten
the other wanting to be at 120 degrees.

> And is this situation analogous to a bent tube in a bike frame or is
> the situation different because the ends of the tube are not free
> being brazed to other tubes?

The wrinkle is a bend and for bending, regardless of how far you bend
it, there will be layers inside the metal that did not reach yield and
will have full 'memory' of where they came from.  That there is
spring-back indicates that parts of the metal have memory and these
are, the inside layers for a flat strap, visualizing the piece to be
made of many laminations.  These layers are not highly stressed in a
bend and as a piece of paper that has not been creased, would return
to their original shape if free to do so.  In fact for many large
radius bends only the outer skin yields and spring-back is nearly
total.  However, the surfaces are subsequently in tension and

In contrast, tensile loads that cause yield, leave no residual stress
because all 'fibers' of the cross section are stressed equally.  If
one yields with any significant strain, all yield and have no memory
or spring-back.  That is the essence of spoke stress relieving.  If
properly installed, the entire spoke is loaded only in tension,
assuming the spokes are supported in the elbow.  That is why improving
the spoke line after lacing the wheel is important before stress

Jobst Brandt    <>

From: (Jobst Brandt)
Subject: Re: Life of frame after crash
Date: 18 Aug 2001 22:59:10 GMT

Chris Reeder writes:

>>> Is there residual stress still present in the bend of the coat
>>> hanger, or was that stress relieved by the visible "spring back?"

>> The spring-back would be complete if there weren't a residual
>> stress holding it at 90 degrees, part of the material wanting to
>> straighten the other wanting to be at 120 degrees.

> None of the material wants to remain at 120 degrees.  Even the
> outermost parts of the coat hanger that were at yield in the 120
> degree position will still want to spring back elastically to say,
> the 110 degree mark.

I agree, that is correct but for practical purposes the spring-back
from the yielded material is small, its return path on the stress
strain curve being offset from the path to yield, but outer fibers
having a greater strain/stress ratio, do not cause a significant
return while elements that did not yield attempt a return that is
larger.  I haven't done an analysis of this other than by observation.
It is residual stress that causes spokes to break in fatigue and
wrinkled frames to crack.

> So there's the inner part that wants to straighten out to 0 degrees,
> and there's also the outer radius of the bend that wants to
> straighten but only to 110 degrees.  In between the inner part and
> the outer bend radius the "desired/remembered" position varies
> continuously from 0 degrees to 110 degrees.  When you release the
> wire, the stresses within it fight it out to determine the
> equilibrium position of the wire, which we're saying is 90 degrees.
> Of course the parts of the cross section at the inner and outer
> radius of the bend get the most say in the matter, because they are
> the furthest from the centroid of the cross section and therefore
> have the most leverage.

Jobst Brandt    <>

Subject: Re: Snaping Spokes
Date: Fri, 21 Sep 2001 05:29:56 GMT

Chris Reeder writes:

>>> What advantage is it that they have been used, other than the nice
>>> stress relief job?

>> None other than expense.  A stress relieved set of spokes is not a
>> trivial advantage.  You don't start over.  The point is that most
>> people believe that a crashed wheel ruined its spokes.  That is not
>> true.

> I understand that crashing won't hurt the spokes, and that time and
> money can be saved by not having to fish out the old spokes and put
> in new ones.  But what is so non-trivial about the effort required
> to stress relieve a set of spokes after building the wheel?  Sounds
> like just a couple minutes time to me.

Ah, that is the problem.  Stress relieving isn't all that binary and
forming the spokes at the hub not a perfect art.  Even when a wheel is
carefully built, it may break a spoke or two in the first 10000 miles.
If it doesn't happen there it probably won't.  I occasionally break a
spoke and it invariably has marks from a stick in the spokes or a kink
from a chain drop-in or the like.  It's always reassuring to find that
there was a dent.

Jobst Brandt    <>

Subject: Re: Snaping Spokes
Date: Sat, 22 Sep 2001 18:39:34 GMT

Jonathan Levy writes:

> The method outlined by someone about an Italian builder who placed
> the wheel on the floor and then walked on the spokes themselves
> could be approached in a similar manner to the hand method.  Only
> big is the foot, what is the weight, and how many spokes are
> trod on simultaneously ?  and presto again !

It was I who witnessed this in Mara's bicycle shop in Florence.  He
was a small man, little more than five feet tall, whose shop was not
open to everyone.  He worked only for a select group of professionals
so there was no sign over the door.  It should be noted that almost no
one but professional and semi professional (dilletante) racers rode a
racing bicycles in those days (1950's) anyway.  A friend (Rick
Bronson) who lived in the area introduced me to him because he thought
I would be interested in his shop.

When Mara finished a wheel, he placed it on the floor and gently
walked on the spokes with his tennis shoes, one side at a time.  He
explained that this would make the wheel last a long time.  He had on
idea why but it reminded me of my own experience.  Rick and I rode
over the Sierra a few years before where I crashed on Monitor Pass
folding both wheels.  I Phoned home to have my second set of wheels
sent to Reno by Greyhound express.  These were large flange Campagnolo
wheels that regularly broke spokes.

By the time we had ridden to the coast at Fort Brag, via the Feather
river, I had two broken spokes, but as one did in those days, I had
plenty of spokes in my saddlebag.  I removed the freewheel and after
replacing the spokes, I used the two hand "stress relief" grip to find
whether there were others about to fail, breaking two more spokes.  I
rode those wheels for a long time with no more failures and after
seeing Mara a few years later do his spoke walk it all came together
for me.  He said the magic words for me, "It makes the wheels last a
long time."

> The method outlined by Marquis spreads load over a number of spokes,
> and just cannot achieve similar forces - its not worth the hassle of
> the more complex calculation needed.  Jobst may wish to comment.

Spoke forces for lateral displacement are shown in the book and they
are not sufficient to stress relieve although spokes readily go slack
or at least lose tension rapidly with even small deflections.  What is
apparent from the computed graphs is that cutting the spokes on the
left side of the rear wheel will not cause the right side spokes to
become slack.  Although this is evident from the graph, it is still
interesting to observe when cutting the spokes out of a rear wheel.

Jobst Brandt    <>

Subject: Re: Snaping Spokes
Date: Sun, 23 Sep 2001 01:14:31 GMT

Mark McMaster writes:

>>> The method outlined by Marquis spreads load over a number of
>>> spokes, and just cannot achieve similar forces - its not worth the
>>> hassle of the more complex calculation needed.  Jobst may wish to
>>> comment.

>> Spoke forces for lateral displacement are shown in the book and
>> they are not sufficient to stress relieve although spokes readily
>> go slack or at least lose tension rapidly with even small
>> deflections.

> What is your opinion of the stress relieving table used by some
> wheel manufacturers and some shops?  The table has a round hole in
> its center just smaller than the diameter of a rim.  The wheel is
> laid on the table so the rim is supported around the edges of the
> hole.  A ram then comes down from above and applies a force on the
> end of the axle normal to the plane of the rim.

> It seems that the characteristics of this apparatus are that it
> supports the rim laterally around its entire circumference, and that
> at least for non dished front wheels the increase in spoke tension
> on one side of the wheel will be similar to the decreases in tension
> on the other side.


I suppose it should work because all spokes of one side get tighter
and this can be taken to twice normal tension for each side of the
front wheel without compromising the rim.  Rear wheels, on the other
hand cannot be taken to that stress on the left side, the right side
not losing much tension when axial load is applied to the right side.

Just the same, higher individual spoke loads can be achieved by
squeezing pairs of spokes, even on rear wheels.

> I guess my questions are: In the case non dished front wheels, will
> the tension changes be equal and opposite on opposite sides of the
> wheel (so that overall circumferential compression on the rim will
> be constant)?  Or will radial rim flex prevent there from being
> significant tension increases?  Rolf wheels are reported to use this
> type of fixture - would the pairing of spokes (so that there is no
> rim flex between right and left spokes) make any difference?  In
> dished rear wheels, would the right spokes experience smaller
> changes in tension on this fixture, therefore making this fixture
> even less able to stress relieve rear wheels?

I think Rolf wheels need this type of fixture because the nipples
cannot be turned unless spoke load is decreased.  With 16-spoke
wheels, tension is to high to achieve without unloading the spoke.

Are there any other reasons for not using low spoke count wheels?

Jobst Brandt    <>

Subject: Re: speaking of jobst: wheelbuilding query
Message-ID: <SveGa.2193$>
Date: Fri, 13 Jun 2003 06:51:30 GMT

Jasper Janssen writes:

>> Spokes don't stretch.  Old spokes are better than new since they've
>> proven that they've been properly stress relieved.  Unless of
>> course your spokes are all breaking.

> But, uh, well.. don't you have to do the stressrelieving again when
> you make a new wheel? It's all in the elastic deformation phase, as
> I understood it.

What stresses do you believe need relieving.  The purpose of
transferring spokes, one at a time, to a new rim is to assure that the
spoke shapes remain identical after transfer to their shape in the
previous rim.  Therefore, no stress relieving is necessary, although
it never hurts.

Jobst Brandt
Palo Alto CA

Subject: Re: Making spokes..
Message-ID: <F8%Bd.16156$>
Date: Sun, 02 Jan 2005 23:07:17 GMT

Carl Fogel writes:

> I'm beginning to wonder about something that should have occurred to
> me earlier. Could both sides be right, but looking at different
> eras?

> That is, in 1981 (I just looked at the copyright dates in my second
> edition, and that's the earliest) spoke-squeezing was applied to
> spokes so inferior to modern spokes that it worked in a blatantly
> obvious manner.

> By 1993 (third edition), spokes had improved so much that Jobst
> could say that the failure rate had dropped so much that the effect
> would be hard to see.

> And now in 2005, the spokes are so good that some people are
> wondering what the fuss was about.

Let me recall for you that I am riding wheels with spoke from 30 years
ago, ones that have served well for all this time and are still
working after close to 300,000 miles and many tours over the Alps,
trails, dirt roads and what have you.

> If this is even partly true, people who began squeezing spokes in
> 1981 would understandably believe that the process is valuable--it
> really worked back then, so why would they stop doing it?

As i pointed out, we regularly have reports here about people with
good quality components (spokes) breaking in short order and requiring
rebuilds.  It is not as though the problem is gone.  It has gotten

> And new wheel-builders reading about the process would use it, find
> that their spokes lasted impressively, and assume that the squeezing
> was responsible, even when most of the improvement is due to better
> spoke manufacture.

They would do well to consider these procedures as warranted,
considering that stress concentrations must occur in spokes while
wheel building.  Otherwise, used spokes would not have a different
elbow shape than new ones.

> There are plenty of holes in this scenario, but it might explain
> some of the confusion.

I think blowing smoke all over the issue, as you are, does not lessen
confusion.  Your "But what if..." writings are not helping.

Jobst Brandt

Index Home About Blog