Wheelbuilding myths debunked

[tallsam]

2. Spokes do not fail because they are too tight.  They fail because they are too loose (movement leads to fatigue).  The ultimate tensile stress of a spoke is way above anything the rim can take, and riding the wheel makes them go slacker, not tighter.

Not strictly true. The spokes nearest the ground will go slack when riding which will in turn tighten those opposite the ground a bit. Its all relative. They'll go slack and tight as you ride but its all equal (also if its built right it'll be so minute it won't matter).

I assume you mean "less tight" rather than "go slack"

Just using the OP's terminology.

[drossall]

Maybe it's a matter of trusting the builder. Round about 1979, I was a member of Leeds St Christopher's CC while a student. We had Keith Lambert as a speaker one evening. He had just collected his (one and only) team bike for the new season, and was off to take the wheels to bits and rebuild them.

[PeeJay]

I inferred that he means that deliberately building wheels with lower tension in the spokes (as has been suggested to me by someone who should know better),..

That person didn't also advocate tied and soldered spokes perchance ...   

Nothing wrong with tied and soldered on the rear wheel of a track bike used by a strong sprinter.

I lent a pair of track sprints, built for a tandem, to the England pair for the Commonwealth Games (they won gold - Geoff Cooke and Tony Brocklehurst). They were returned to me re-built and tied and soldered, the front wheel in a two cross/one radial pattern. They were done by the German team mechanic. Boy, those wheels are solid, and totally bomb-proof, and have survived several crashes since.

I've known several very good wheel builders, and I'm convinced that there are good, very good, and then, not very good, builders. One of the greatest is Steve Snowling who used to work the sixes for Tony Doyle. He tied and soldered (note - not just tied - you have to scape the chrome off chrome spokes, or use rustless, to properly solder the wires to the spoke) some track wheels. Again, they never, ever, went out of true.

Just because some wheel builders tie and solder doesn't mean that it makes the wheel stronger. As said by others up post there is no advantage in doing so other than to stop broken spokes flailing around.  The wheels you mention were obviously very well built wheels, buit not because of the tie and soldering.  If anything the practice just adds weight and makes the wheel harder to re-true.

[PeeJay]

Here's some proof

http://sheldonbrown.com/brandt/tied-soldered.html

[GraemeMcC]

I lent a pair of track sprints, built for a tandem, to the England pair for the Commonwealth Games (they won gold - Geoff Cooke and Tony Brocklehurst). They were returned to me re-built and tied and soldered, the front wheel in a two cross/one radial pattern. They were done by the German team mechanic. Boy, those wheels are solid, and totally bomb-proof, and have survived several crashes since.

I'd always been lead to believe that track wheels (from pre 1980s) were routinely tied and soldered because of the lateral and inline stresses arising from sprinting round banked tracks. Especially for tandem loading!

It sounds like these wheels are in the Crow's Foot pattern.

Re: earlier posts, laced spoking:
Another factor for rear wheels is to dish the wheel and give spoke clearance to the rear jockey-wheels cage for innermost sprockets. Again, in 5 & 6 speed freewheel days, big sprocket sizes may have been as low as 19 or 20T (for straight-up TT use). Not like the more generous 23 - 32 T dinnerplates commonly in use today. And 5 speed chains and cages were wider than modern 9/10/11sp cages. Any un-laced spoke is more likely to clash (unless it is fitted from the cassette side into the flange).

[Biggsy]

Jobst Brandt's view on tying and soldering from http://www.sheldonbrown.com/brandt/tied-soldered.html:

While writing The Bicycle Wheel, to conclusively determine what effect tying and soldering of spoke crossings in a wheel had, I asked Wheelsmith to lend me an untied pair of standard 36 spoke rear wheels, on Campagnolo low and high flange hubs. I had an inner body of a freewheel machined with flats so that a wheel could be clamped into the vise of a Bridgeport milling machine while the left end of its axle was held in the quill.

With the hub rigidly secured, with its axle vertical, dial gauges were mounted at four equally spaced locations on the machine bed to measure rim deflections as a 35lb weight was sequentially hung on the wheel at these positions. The deflections were recorded for each location and averaged for each wheel before and after tying and soldering spokes.

The wheels were also measured for torsional rigidity in the same fixture, by a wire anchored in the valve hole and wrapped around the rim so that a 35 lb force could be applied tangential to the rim. Dial gauges located at two places 90 degrees apart in the quadrant away from the applied load were used to measure relative rotation between the wheel and hub.

Upon repeating the measurements after tying and soldering the spokes, no perceptible change, other than random measurement noise of a few thousandths of an inch, was detected. The spokes were tied and soldered by Wheelsmith who did this as a regular service. The data was collected by an engineer who did not know what I expected to find. I set up the experiment and delivered the wheels.

[PH]

Jobst Brandt's view on tying and soldering from http://www.sheldonbrown.com/brandt/tied-soldered.html:

While writing The Bicycle Wheel, to conclusively determine what effect tying and soldering of spoke crossings in a wheel had, I asked Wheelsmith to lend me an untied pair of standard 36 spoke rear wheels, on Campagnolo low and high flange hubs. I had an inner body of a freewheel machined with flats so that a wheel could be clamped into the vise of a Bridgeport milling machine while the left end of its axle was held in the quill.

With the hub rigidly secured, with its axle vertical, dial gauges were mounted at four equally spaced locations on the machine bed to measure rim deflections as a 35lb weight was sequentially hung on the wheel at these positions. The deflections were recorded for each location and averaged for each wheel before and after tying and soldering spokes.

The wheels were also measured for torsional rigidity in the same fixture, by a wire anchored in the valve hole and wrapped around the rim so that a 35 lb force could be applied tangential to the rim. Dial gauges located at two places 90 degrees apart in the quadrant away from the applied load were used to measure relative rotation between the wheel and hub.

Upon repeating the measurements after tying and soldering the spokes, no perceptible change, other than random measurement noise of a few thousandths of an inch, was detected. The spokes were tied and soldered by Wheelsmith who did this as a regular service. The data was collected by an engineer who did not know what I expected to find. I set up the experiment and delivered the wheels.

And what?  The problem I have understanding most things written by Brandt (Usually quoted by you) is that although he's prepared to go to great lengths to explain the method he doesn't bother to give any of the reasoning.  Here we have a meticulously executed experiment in hanging 35lb weights of various parts of a wheel in various ways, all independently verified.  Just how does that demonstrate the point? In what way does that replicate me riding the bike?  If that's not explained, it's meaningless.

[Ian H]

Although I'm sceptical of a lot of Brandt's theories, the tests he made  seem quite simple: vertical and tangental stresses, the kinds of thing you'd want to know about a wheel.

[cuddy duck]

yebbut, nipples lubed or nipples dry?

[Ian H]

yebbut, nipples lubed or nipples dry?

I wear loose clothing when building wheels.

[PH]

Although I'm sceptical of a lot of Brandt's theories, the tests he made  seem quite simple: vertical and tangental stresses, the kinds of thing you'd want to know about a wheel.

So now we know about the stresses caused by a 35lb weight.  Call me thick, but  I still don't know the relevance of that.

[LittleWheelsandBig]

Changing the wheel stiffness, perhaps by tying and soldering spokes, should result in a difference in deflection for a fixed load. No difference in deflection means no difference in stiffness, means tying and soldering doesn't change wheel stiffness.

[Rhys W]

You're trying to replicate the forces experienced by a wheel when riding, but in a controlled environment where the mechanical effects on the wheel can be measured. I'll take that any day over somebody's subjective opinion as to how it "feels" when riding - that's when you get Cycling Plus type "stiff yet responsive" nonsense.

[tom_e]

And what?  The problem I have understanding most things written by Brandt (Usually quoted by you) is that although he's prepared to go to great lengths to explain the method he doesn't bother to give any of the reasoning.  Here we have a meticulously executed experiment in hanging 35lb weights of various parts of a wheel in various ways, all independently verified.  Just how does that demonstrate the point? In what way does that replicate me riding the bike?  If that's not explained, it's meaningless.

Surprisingly enough, the fact that he didn't explain something to you doesn't make it meaningless.  He's writing as an engineer, for other engineers, discussing an engineering modification to a bike. 

So now we know about the stresses caused by a 35lb weight.  Call me thick, but  I still don't know the relevance of that.

No, he measured deflections, not stresses.  You're not thick, you're just not an engineer so it's not your area.  The productive response to that is to ask for explanations, not get antsy and imply it's his fault.  You didn't pay him to explain it to you.

The point is that while a metal structure like this is having forces applied to it which are in its normal expected use, the deflections, or flex in it, will be directly proportional to applied force.  This means that double the force, double the flex, half the force, half the flex.  But the ratio of force to flex is called the stiffness of the structure (in this particular direction).  Measuring the movement for a 35lb force tells you that.

Furthermore, measuring it in one direction and then another at 90 degrees tells you how much it flexes in each of those directions.  Again with forces within normal expected use, the combination also tells you how much it will move for forces inbetween those angles.

So the experiment is measuring exactly how stiff or flexible the wheel is.  Objectively, with numbers.  The sort of thing engineers do to figure out if a change (such as tying and soldering) is worthwhile.

And no, I don't agree with Brandt on everything either, but this one does appear to be pretty clear.

[mrcharly-YHT]

Has anyone managed to measure the lateral torque between axle and wheel rim, when sprinting?

I'm wondering why he chose 35lb.

[Tewdric]

Not strictly true. The spokes nearest the ground will go slack when riding which will in turn tighten those opposite the ground a bit.

It's easy to demonstrate that the upper spokes don't measurably change in tension when you place a load on the wheel, whereas the tension of spokes below the hub reduce in tension.  In a static situation in any case! 

The lowest spokes go slacker because the rim deforms slightly due to the ground pushing against it. Because the lower spokes aren't pulling down so hard the weight on the hub is supported by the unchanged tension in the upper spokes.

That's a bit misleading.  Provided the spokes all remain in tension then the weight bearing down on the hub is supported by a reduction of tension of the lower spokes within the pre- stressed structure of the wheel.  There is no "need" for the rim to deform, although, of course, in a dynamic real world situation it will, a tiny bit, both as weight is borne by the wheel and as lateral and driving or braking forces dynamically affect spoke tension. 

I think of it as the rim deforming because the spoke tension is changing, not the other way round.

[Tewdric]

I forgot to comment on tensiometers. Biggsy has mentioned getting a feel for the correct tension for your particular wheel. That's all you need. For beginners, bringing the wheel into correct tension, systematically and carefully will ultimately lead to getting the wheel right.
Tensiometers are not like, say, multimeters. They all apparently give different results. You would need to know which instrument, and what rims and spokes were used to arrive at a recommended tension. Then, if you have a tensiometer, you need a well built wheel of known tension to calibrate it.
You can check for evenness of tension by ear.

How many have you tested?  I've compared mine with two others (an FSA and a DT Swiss) and they all gave the same result to within 5 percent.

How is a beginner supposed to know what the correct tension is without being able to measure it? 

Most people are happy to talk numbers when it comes to things like tyre pressures; I don't understand the widespread cynicism about tensiometers!

[Biggsy]

For the record, I'm not against tensiometers, I just think that you can build good wheels without one.  Beginners can compare by feel with another similar wheel, or learn through trial and error.  Admittedly, the error of cracking a rim through over-tension is not a cheap one.

[GraemeMcC]

I'm too too sure where we're all going with this (the tied/soldered/laced/stiffness debate) ?

Tewdric is correct (#65 above): spokes should always be in tension however the level of tension must fluctuate as the loaded bike wheel is moved along. And that variation in tension results in fatigue loading which ultimately affects spoke and hub flange life. And that variation in tensile stress/fatigue will be present regardless of whether tied & soldered spokes were used, because that is only a function of the axial tensile loads in the spokes. The fact that a spoke is tied cannot detract from its load wanting to take taking the shortest/stiffest load path available which is to carry on along the spoke. Engineering fact: stiffness = strength.

So, the tying and soldering can only affect lateral stiffness. By tying and soldering the cross-overs, a node point is created at each tied cross-over which forms a series of stiffer triangulated sub-frames in the rim/spoke/hub structure. So, the spokes cannot freely displace relative to each other as untied spoke could, under their fluctuating axial tension. So, the lateral spring stiffness of a tied wheel must be increased. So, it will be less likely to deform before exceeding/reaching the elastic yield point (of spokes) or permanent plastic deformation, or flange shear failure.

So, tied and soldered laced spokes must create a wheel that is laterally stiffer and has a greater capacity to withstand lateral forces (such as those acting on wheels on a steeply banked velodrome where gravitational and centripetal force effects are present) than an untied wheel. But tied and soldered spokes have no benefit for an upright  wheel rolling normally along a straight road.

End of...

[tom_e]

Engineering fact: stiffness = strength.

Er, wtf?  Is that meant to be a joke?

So, the tying and soldering can only affect lateral stiffness. By tying and soldering the cross-overs, a node point is created at each tied cross-over which forms a series of stiffer triangulated sub-frames in the rim/spoke/hub structure.

Stiffer how exactly?  Sure, if you consider each sub-frame it is stiffer, but when you put two in series as there are then you add deflections for the same force and it becomes pretty similar again.  Not exactly the same in all directions, which is why you'd want to do a proper analysis or test it.

So, the spokes cannot freely displace relative to each other as untied spoke could, under their fluctuating axial tension. So, the lateral spring stiffness of a tied wheel must be increased.

So why doesn't it show up in Brandt's test?  The one where the axle is vertical and the rim loaded with weights is a direct test of lateral stiffness.

...lateral forces (such as those acting on wheels on a steeply banked velodrome where gravitational and centripetal force effects are present) ...

Those forces are balanced by the lean angle.  There may be plenty of lateral forces on the wheel from sprinting and generally throwing the bike about.  And the bends will generate increased vertical loading (relative to the wheel/bike).  But the corners and banking do not directly affect lateral forces on the wheel.

End of...

Er, end of what exactly?  Does that mean you don't want anybody else to talk anymore?

[Somnolent]

End of...

Er, end of what exactly?  Does that mean you don't want anybody else to talk anymore?

Quite.... and tbh I find the attempt to re-open the tying and soldering debate as pointless as a h.....t debate.   Nobody is saying GraemeMcC (or anyone else) should cease to tie and solder if they so choose.   Nor, I hope, is anyone trying to impose tying and soldering on us mere mortal wheelbuilders who can build a wheel that is satisfactory for our own purposes without doing so.

[rogerzilla]

Tying and soldering gives you a trick-looking wheel which is a PITA to fix at a later date, should anything go wrong - so you'd better get it spot on.

Apart from the looks, it stops a broken spoke from flailing about and impaling something but (a) injury is a pretty remote possibility - only Nutty or Phil could probably manage it - and (b) if you do break a spoke on your tied and soldered wheel, you're in for a very tough time trying to swap it out.

[mark]

I'm told that tying and soldering was fairly crucial on ordinary/penny-farthing bicycles. The spokes on the front wheel of an ordinary were long enough to present a hazard when they broke, and the primitive state of metallurgy back then meant that they broke much more frequently than is the case today.

[frankly frankie]

End of...

Er, end of what exactly?  Does that mean you don't want anybody else to talk anymore?

 
ISTR the last time we had a "does a wheel stand on its lower spokes or hang from its upper ones" debate it ran to about 50 pages ...

[rogerzilla]

It does neither.  It "hangs" by ALL the spokes except the bottom two or three.  Brandt was being provocative in the way he described it, although he understood it perfectly well.