So I'm trying to build my first rear derailer wheel that's
actually up to the maximum safe level of tension the rim can
handle. In the past I have built 1 front wheel, 1 rear
derailer wheel, and 2 dishless rear wheels. These came out
okay but not as high-tension and windup-free as possible.
I'm using a new 36-hole 700C Sun CR-18 (575 grams, 23mm outside/18mm-
ish inside width, eyelets not sockets, brushed), straight
2mm spokes (Sapim on one side and DT on the other due to an
ordering error, and I didn't have the money for butted or I
would have gotten them), brass Sapim nipples, and a new LX
rear hub. This is going to be a 9-speed wheel for commuting,
distance riding and loaded touring. I know all about how
much 9-speed sucks strength-wise and am partially just
building a 9-speed wheel and setting my bike up 9-speed to
get a feel for how well such highly dished wheels can hold
up. I have read The Bicycle Wheel and Sheldon Brown's
wheelbuilding page several times. I'm using a tensiometer to
check for inordinately loose/tight spokes, balanced tension,
and to establish a standard for how much tension my future
wheels with CR-18s can handle.
I laced the wheel 3-cross using Tri-Flow to lube the spoke
threads and white grease on the nipple seats. I
accidentally dunked the non-drive side spoke threads in
lube before I remembered not to (as advised on Sheldon's
page), so I wiped a bunch of lube off with a rag and
decided to see how the wheel would fare longevity-wise with
lightly lubed non-drive spokes.
The first mistake I made was overestimating how much tension
this rim could take before beginning to overload and go potato-chip-
shaped (this is also a reason I didn't just use degreaser on
the lubed non-drive-side spoke threads - you wouldn't really
have to worry about that with a strong enough rim even if it
was highly dished, right?). In fact, I was a bit under the
impression that with this rim I wouldn't be able to get the
tension that high up at all before nipples started getting
impossible to turn. This was a big mistake, I know now, but
there is not exactly much material out there about learning
to predict exactly what rims will be over-tensionable and
which won't be, even though there are hints that such a
clear distinction exists. Anyway, during what I then thought
was not quite the end of the building process, I checked the
wheel's centering and discovered the rim needed to be pulled
over to the drive side by quite a bit. The drive side spokes
were at I think (can't remember exactly) about 100-110 KgF
average. At this point, the rim was true laterally and was
mostly true radially and tension balanced with a few
exceptions (see below). I figured it would be safe at this
point to make the centering adjustement by just tightening
all the drive-side spokes a half-turn, rather than loosening
the non-drive spokes and tightening the drive ones. So I did
that and when I spun the wheel immediately afterwards to
check for typical small necessary lateral truing
corrections, instead I saw that the wheel had become fairly
potato-chip shaped. Up to this point I had not been
following The Bicycle Wheel's procedure of adding layers of
tension and then stress relieving to check for how close one
is to approaching the rim's maximum safe tension - again, I
didn't think it was really necessary with this rim to do so
yet. After this occurred, I backed off all the drive-side
spoke tensions by the half-turn I had just added. The potato
chip shape remained. Then I backed off the tension by I
think a quarter turn on both sides and it still remained.
Afraid that I had permanently warped the rim, because the
wheel didn't go back to being true, I then removed all
tension from the spokes and reset them to the initial stage
with the spoke threads just barely covered by the nipple,
added a small amount of tension, did some very minor truing
that one always needs to do at the beginning
tensioning/truing phases, and saw that the rim was just as
true as it was at this point the first time through, leading
me to believe that my over-tensioning did not permanently
warp the rim. The small bits of truing necessary did not
match the shape of a slightly collapsed rim.
My questions about this episode are: If the rim was not
permanently warped, what exactly is the reason that it did
not go back to its previous true shape after I removed the
tension I just added, and then some? I assume now that this
is the same reason why a rim that's just gotten stress-
relieved enough to overload it, as per The Book, will remain
in the slight potato-chip shape even after you stop
squeezing it, right? In later attempts at building this
wheel, I added a quarter-turn layer of tension, stress-
relieved it all around, saw a very slight potato-chip shape
of maybe 2mm away from the centerline on either side
happening as I spun the wheel in the stand, backed it off a
half turn on both sides, all as per The Bicycle Wheel, and
the result after the de-tensioning was that it was about as
true as it was prior to the last additional tension layer.
On the initial build, if after my overtensioning I had just
backed off the tension all around a bit more instead of
starting over, would the wheel have gone back to trueness,
presuming that all my nipple-turning had been accurate
enough? Does it depend on exactly how overloaded and
deformed the rim became? In The Book, it says that after you
overload a rim during stress relieving, some truing will be
necessary after you back it off half a turn all around. But
none of the potato-chip shape from overloading is supposed
to be remaining at that point, right? If it was, and you
trued the rim in reaction to that, wouldn't everything just
get really screwed up? So is the truing you'll be needing to
do at that point just in reaction to inaccuracies in your
nipple-turning? How exactly do all the rules about all this
apply to rims of different types and weights? When you
stress relieve a rim to test it for overload, are you
supposed to be watching the rim for a potato-chip shape
happening as you squeeze each group of 4 spokes and then
stop if you see one, or should you just go through and
stress relieve all the spokes and then check to see if the
deformation occured somewhere in the process?
There was another time when I stress relieved all around and
then saw that the rim was in a shape that resembled the
slight potato chip/saddle shape of an overloaded rim, but
not quite. Whereas the usual shape is a series of curves
where there's an apex veering to the left, followed by an
apex veering to the right 90 degrees later, and an apex to
the left another 90 degrees from that, etc. this series of
curves went more like apex to the left followed by one to
the right 45 degrees later, followed by the next one to the
left 135 degrees later, followed by one to the right 45
degrees later, etc, such that both curves going to one side
were still 180 degrees from each other, but the overall
shape was weird. Again, this was after I tensioned and
stress relieved the wheel. I figured that this was just a
sign of overload but am still not really sure if I'm missing
something. Does rim deformation just happen this way
sometimes? Are there other variations on the typical imploded-
rim shape?
When stress relieving spokes, all internal stresses are
relieved after you make one complete round of squeezing
adjacent spokes beyond yield, right? If so, does that mean
that if you were building up a rim that you absolutely knew
was strong enough that difficulty in turning nipples was
going to be the tension bottleneck for the wheel, you would
only really have to stress relieve once, when you hit the
point where no more tension can be added?
I'm also wondering about how exactly techniques to eliminate
spoke windup work. When you overshoot a quarter turn and
then back up to eliminate windup, is the idea that somewhere
in that extra quarter turn, the spoke's increasing torsional
load will become enough to overcome the amount of friction
between its threads and those of the nipple? What exactly
keeps the spoke from winding up in the other direction when
you back up the nipple? When you go to back up the nipple,
isn't there just going to be more friction than you started
with because now the spoke is tighter by a quarter turn plus
whatever adjustment you wanted to make? Is it the best idea
to keep one hand on the spoke you're adjusting as the other
turns the spoke wrench, so that you can feel when a spoke is
winding up and when it unwinds? Is it possible for a spoke
to only unwind partially?
I'm also wondering why hardly anything I've read about
wheelbuilding mentions the possibility of tightening drive
vs. non-drive-side spokes according to a ratio based on how
much they pull the rim due to their differing angles, and
how much tension each side will have in total when the wheel
is done. The ratio is something like 8:5 for most 9-speed
rears, isn't it? So why not just do your tensioning layers
and truing adjustments by turning the drive side something
like twice as much all the way through? If you just act like
both sides pull the same amount and therefore you make even
increments on both sides when you're tensioning, dishing, or
truing, aren't you bound to create lateral/dish errors that
must be dealt with using the same flawed process? I was
experimenting with this and it seems like there may be
something to it, but this time around I was confused about
enough things that throwing this in the mix was more than I
really wanted to deal with.
In The Bicycle Wheel, it's written that wheels with
unbalanced spoke tensions will equalize themselves in use.
Is this just for the obvious reason that the spokes with low
tensions will get looser and out of true in use, which
causes all sorts of havoc, or is there something subtle I'm
missing here that causes problems when some spokes are also
inordinately high in tension, other than increasing the
likelihood of eyelet cracking? Does this imply, for example,
that a wheel with generally closely balanced, high tensions
but a few spokes at inordinately high tensions for some
reason would all equalize in use?
A final group of questions has to do with interactions
between radial trueness, tension balancing, and rim
imperfections. I got my wheel to a point several times where
the total tension difference between the slackest and
tightest drive-side spokes was about 30 KgF, with a few at
about 95, a few at about 125-130, and most at about 105-120,
and a similar bunch of disparities on the non-drive-side.
This was at about the max safe tension for the rim using the
Jobst method unless I'm very confused. There was still quite
a bit of radial truing error, perhaps 1mm between high and
low points, but it was arranged in the classic annoying
pattern where the bumps are tighter spots and the dips are
looser spots. In other words, if I just went through and
made the tension on each spoke exactly the same without
regard to how true it would make the wheel, then the wheel
would be a total mess. I was left with the definite feeling
that I was just encountering imperfections in the rim, since
I worked on it for a long and it seemed like there was
little further I could do without compromising either
tension balance or reasonable trueness. On the other hand,
I'm fairly new at this and I don't want to put undue blame
on the rim. My question is just how bad are the tolerances
on Sun rims, or CR-18s in particular for those who have lots
of experience with them, and what kind of tension
disparities do you usually end up with?
Thanks for reading and replying, Nate Knutson