Carbonfiberboy 

Totally disagree. Spokes break, not through lack of "strength," but through fatigue. Fatigue happens in a spoke when there are cycles of loading and unloading. The whole idea of using DB spokes to increase wheel longevity is to reduce fatigue. DB spokes do this because they stretch more than SG spokes. Since they are thinner, they stretch more at the same tension than does a SG spoke.

I've been building my own wheels for many years and I ride a lot. I've never broken a spoke from fatigue. I once broke a spoke in a paceline crash, but that's it. That may be nothing unusual, except that for the last 10 years I've done almost all my road riding on a tandem with a team weight of 285 lbs. That's serious Clyde weight! Plus when we go loaded touring, our all-up weight is ~370 lbs., depending on exact configuration. For the last few years, I've been building both our tandem and single bike wheels with CX-Ray spokes. IIRC they are 18g for most of their length. Before I used the CX-Rays, I used DT 14-15 spokes. Because we ride in the rain so much up here, I have to replace a rim about once/year. I always reuse the spokes and nipples.

But back to the OP's question: For almost any wheel build, double butted 14-15 (2.0-1.8-2.0) spokes will produce a very durable wheel. It's not the "method of tensioning" which produces the most durable wheel, it's the amount of tension that does it. All rims have a maximum tension specification. Some manufacturers publish it, some don't, but AFAIK it's universally ~130 Kgf (kilogram-force). They call it force because a kilogram is technically a measure of weight. That doesn't mean that you tension everything to 130 Kgf. Drive side and non-drive side spokes will usually have very different tensions. The DS will be tensioned to 120 or so and the NDS to whatever produces the correct dish. Front wheels are usually tensioned to 100-105 because they bear less weight on a single bike. Though on our tandem, I tension both wheels the same because they bear about the same weight.

Be all that as it may, and people will quibble about the exact numbers I quote above, the idea is to build a wheel with very nearly the same tension in every spoke that's on one side of a wheel. How do you do that? Use a tension gauge! Park Tool makes the venerable and inexpensive TM-1 tension meter, which makes getting the tension right a piece of cake. So that's how to build a durable wheel: DB spokes and a TM-1, plus a fair bit of time to get it just right.

I've read in many places, including this thread, that one of the keys to building a durable wheel is stress relieving it several times during the build. I used to do that, but then I wondered, "What does it do?" and quit bothering with it. Not doing it made no difference in build quality. The only thing I can think of is that stress-relieving allows the nipples or spokes to unwind if they have been wound up during tensioning. But since I don't wind up my spokes anyway, there's nothing to unwind. I suppose it's possible that older rims and nipples could have alignment problems between spoke and nipple and stress relieving allowed the nipple to align itself a bit. However I've only been building wheels for ~20 years, so I don't have experience with older rims.

Bike Gremlin

Great post, just a few notes:

Proper stress relieving doesn't unwind twisted spokes.


During hard braking, front wheel carries all the bike's weight.

Front wheels usually have no dish and don't carry pedalling torque, making all the spokes on both sides work more equally, sharing the load. The main reason to tighten front spokes less than rear drive side ones is that tightening all the front spokes to the same tension as rear DS ones would be too much for the rim to handle. At the rear wheel, NDS spokes are a lot less tight, so total rim tension is not exceeded.

Carbonfiberboy 

Then what does stress relieving do? What stress is being relieved?

Our tandem is 145 O.L.D. in the rear. Thus there is less then 2mm difference between DS and NDS flange spacing. So I've been tensioning both sides to ~115 and the same on the front. Never had a rim problem doing that. My understanding is that max spoke tension specs are purely based on the strength of the spoke bed. Some paired spoke wheel designs do have trouble on a tandem with the nipples pulling through the bed because so much force in concentrated on such a small rim bed area. Never had a problem with conventional spoke spacing.

We tension to prevent fatigue. Since hard front braking happens so seldom, I guess it's not a fatigue issue, so we can use lower tension on single bikes. On modern road bikes, NDS spokes break more frequently than DS due to the lower spoke tension. Just my opinion.

Bike Gremlin

Tension in the spoke elbows during production. Quote from "Bicycle Wheel", by Jobst Brandt:

"
STRESS RELIEVING
Spokes are cold formed. After cold forming steel always springs back a certain
amount. The spring-back is incomplete because part of the material went
beyond its elastic limit and part did not. These disparate parts are fighting each
other, and when spokes are tensioned, one or the other of these elements will
be stressed additionally. This stress can be, and often is, at the yield stress and
must be relieved when the wheel is completed.
After correcting the spoke line, and when the wheel is true and tensioned, its
spokes may appear to be in perfect alignment. However, some of the spokes have
a good line at the elbow and rim only because they are tensioned. Besides, spokes
have residual stresses at their elbows, heads, and threads from their forming
process. As the wheel was laced the spokes may have been bent to make them
conform to the hub and nipples. Since they were brought to their yield stress to
bend them into place, the addition of tension guarantees that they remain at the
yield point. When stressed to near their yield point, spokes have a short fatigue
life. These stresses must be relieved to make the wheel durable.
HOW STRESS RELIEVING WORKS
Stress relieving can be regarded as correcting the spoke line at a microscopic
level. The process momentarily increases spoke tension (and stress) beyond the
yield point, but only in the parts of the spoke that are near yield. At the high stress
points the spoke will deform plastically and take a permanent set. When the
stress relief force is removed these areas cannot spring back, having, in effect,
lost their memory, and relax to a lower stress. The wheel may lose tension during
stress relieving, but not because the spokes have stretched. Any length change
occurring at the high stress points is microscopic. Loss of spoke tension comes
from the spoke elbows seating into the flange.
Stress relieving also provides an accurate method for determining the maximum
safe spoke tension for a wheel. A 50% increase in tension of four spokes
realistically represents a momentary overload. If the wheel is too tight, it will
warp into a saddle shape during stress relieving. A properly tensioned wheel
should withstand a firm squeezing of two spoke pairs with only slight loss of
alignment if any. To avoid rim damage, testing for wheel's tension limit should
be undertaken carefully. The grasp should be increased gradually to full force
while watching for sudden loss of alignment.
HOW TO RELIEVE STRESS
Spokes are best stress relieved by grasping each of the most nearly parallel spoke
pairs at midspan on the left and right sides of the wheel and forcefully squeezing
them together successively around the wheel. This process is sometimes
accompanied by the sound of spokes untwisting in the nipples. No sound will
occur if the residual twist in the spokes was removed, as it should have been while
truing the wheel. Stress relieving is not intended to free twist in the spokes
because twist should be eliminated during proper spoke tightening."


The weakest link is the rim itself - if all is made properly.

I agree.

Carbonfiberboy 

^Thanks.

FatInBike

Thanks, of course, but I'd like to resume discussion on this epic thread (no sarcasm neither irony). Epic, and important for amateur wheelbuilders as I am.
The topic: most of us read every basic text/articles on the subject, expecially by "our" venerated Brandt & Brown, included the explanation of the process, according which spokes on the NDS should have a lower gauge than the DS has.
However, by the "official" handbooks I would have preferred a different scientific approach initially based on statistical data, any speech sounding as follows: «In my life of wheelbuilders I observed that the wheels built with a DB with lower gauge than gauge used in the DS lasted much more than the wheels assembled with the same spokes used on the NDS and the DS, being the wheels assembled with same tension on the DS, equal to the an x percentage of the elastic field...and so on. My explanation is stress, yield, elasticity field and bla-bla-bla and suggest to overcome possible problems by using differential gauges».
So tailed, it'd be truly convincing.

The Science, every science, has ever-ever first observed phenomena, then has tried to explain them. Until now, all the defenders of the differential gauges between DS and NDS have cyclicly repeated the Guru's mantra (by the same Guru' I pursue for my few in house assembled wheelsets) but did not reported their (if any) experience in order to confirm the theory.
Who is in position to report ONLY a personal statistics about it, to confirm the theory or to contradict it?

(My point of view is a little peculiar, as I am a mechanical engineer).