About the rear wheel with the ZIPP 404 rim that I built the other day
↑This one
As someone pointed out in the comments (← awesome)
the spoke lacing was wrong in just one spot.
↑At the hub flange peaks on the freewheel side, out of 6 locations,
5 of them look like this,
but one spot is reversed.
This hub can be laced either way,
but judging from the subtle chamfering on the flange holes,
it's actually suggesting the lacing shown in the upper image,
which means the wrong ones aren't just this one spot
but rather the other 5 spots.
However, I had my reasons
and when I changed the lacing here during the pre-assembly,
there was one spot I forgot to change back.
This rear wheel uses a final crossing lace pattern,
but if it didn't,
I might have reversed this lacing.
Also, because the lacing is different in just one spot,
the spokes are contacting each other at the first crossing.
With the 5-spot lacing method, it doesn't look that way,
but at the first crossing the spokes just barely avoid contacting.
So I relaced it.
And it's built.
This hub has a very slight difference
in flange diameter between the two holes
coming out in different directions from the same flange peak.
With my lacing method,
the spokes in the "porcupine direction"
repeat through the large flange 6 times.
I don't really worry about the spoke tension differences
between the two spokes that make up the final crossing
at this level of variance, but as a more extreme example of this phenomenon:
there's a Shimano 7700 series 16-hole wheel.
To be precise, the image above is the front wheel of a WH-7701,
and the 7700 uses this lacing on both front and rear,
while the 7701 rear freewheel side is actually radial laced.
The Dura-Ace rear hub has radial lacing on one side,
but this one uses tangential lacing on both sides.
The lacing on the far flange relative to my view isn't a mirror image
of the near flange—rather,
it has a JIS-style structure where flipping the wheel left and right
doesn't change its appearance.
So when you look at the resultant force of the two spokes that make up the final crossing,
on the right side you get 0-cross lacing with spokes twisted clockwise,
and on the left side you get 0-cross lacing with spokes twisted counter-clockwise.
↑This is a ZIPP rear hub from a certain era,
and the spokes in the porcupine direction
alternately come from the large flange and the small flange.
So when you look at the first crossing,
you see crossings between spokes from the large flange
and crossings between spokes from the small flange
alternating with each other.
This rear hub is 24-hole, but
this lacing method isn't possible with 20-hole.
The number of large-to-large and small-to-small crossings
needs to be even (20-hole is 5 crossings).
So to use this lacing method,
you need (4×n)-hole with n being even,
but 8-hole with n=2 doesn't have enough spokes
to make a functional wheel.
Next, 16-hole with n=4,
but if you use this lacing method,
↑it ends up looking like this, with spoke tension variation
that alternates back and forth in a way that's too noticeable to ignore—
it feels wrong to me.
The earlier 7700-series 16-hole wheels are also paired-spoke wheels,
so this lacing method should be avoided,
and indeed they don't use it.
Beyond that you have 24-hole with n=6, 32-hole with n=8,
40-hole with n=10, and so on,
but realistically the options are 24-hole or 32-hole,
and for 32-hole straight-spoke complete wheels,
off the top of my head I can only think of
Mavic's Ksyrium SL and
its successor the Ksyrium SLR,
so the lacing method of alternating large/small flange
for the porcupine-direction spokes is essentially
24-hole exclusive.
As for the Dura-Ace hub,
in the diagram above I've exaggerated the high-low of the flange peaks a bit,
but first you'd just thread the porcupine-direction spokes
through the large flange side
and repeat that 6 times to thread all the spokes, but instead
first I threaded the spokes through every other flange peak
for 3 iterations,
and the remaining 3 spots would be the reverse of that—
this is what I attempted during pre-assembly,
but I didn't properly revert it,
and that was my mistake this time.
That article also had
another comment.
"With this rim and hub, wouldn't you get a better wheel
by halving the non-freewheel side
and doing 18-hole 2:1 lacing with aero spokes
that are nearly 100% spoke weight ratio?
Or is that not the case?"
It's a seriously fantastic comment,
as if the wild Fulcrum has appeared from some bike blog.
100% spoke weight ratio flat spokes
correspond to Sapim CX.
The nominal weight for 260mm length 64 pieces is 423g,
so 423÷64÷260÷0.0257
equals 0.989131..., which is
about 99%, but
from my accumulated measurements it's actually 100.3%.
Either way it's in the range of about 100%.
100% class aero spokes won't fit through
a 2.3mm round hole (CX is 2.8mm wide),
so it's impossible with this Dura-Ace,
and you'd need either a flange with a slot hole
or a hooking-type rather than threading-type design.
Setting aside those logistical considerations,
if you build an 18-hole rear wheel with 2:1 lacing
using a 24-hole rim and 24-hole hub,
you'll have an idle phase at the non-freewheel rim holes
and spare rim holes.
When an idle phase occurs with 18-hole,
truing becomes extremely difficult,
so you'd need a rim with stiffness designed
with that in mind.
For example, the Bora Ultra 80,
which is an 80mm deep rim brake Bora,
uses Campagnolo G3 spoke bundles that are
6 pairs / 18-hole rather than 7 pairs / 21-hole,
and since it's G3 lacing it has a long idle phase,
and the rim stiffness is designed with that lacing in mind.
Aside from the hub holes, I'm not particularly fond of
having spare rim holes,
but I have done repairs on 21-hole rims
with 28-hole rims when unavoidable (→here).
So if you did it with an 18-hole rim
that had no hole offset,
it might make for an interesting wheel.
Thanks for the comments!
↑This one
As someone pointed out in the comments (← awesome)
the spoke lacing was wrong in just one spot.
↑At the hub flange peaks on the freewheel side, out of 6 locations,
5 of them look like this,
but one spot is reversed.
This hub can be laced either way,
but judging from the subtle chamfering on the flange holes,
it's actually suggesting the lacing shown in the upper image,
which means the wrong ones aren't just this one spot
but rather the other 5 spots.
However, I had my reasons
and when I changed the lacing here during the pre-assembly,
there was one spot I forgot to change back.
This rear wheel uses a final crossing lace pattern,
but if it didn't,
I might have reversed this lacing.
Also, because the lacing is different in just one spot,
the spokes are contacting each other at the first crossing.
With the 5-spot lacing method, it doesn't look that way,
but at the first crossing the spokes just barely avoid contacting.
So I relaced it.
And it's built.
This hub has a very slight difference
in flange diameter between the two holes
coming out in different directions from the same flange peak.
With my lacing method,
the spokes in the "porcupine direction"
repeat through the large flange 6 times.
I don't really worry about the spoke tension differences
between the two spokes that make up the final crossing
at this level of variance, but as a more extreme example of this phenomenon:
there's a Shimano 7700 series 16-hole wheel.
To be precise, the image above is the front wheel of a WH-7701,
and the 7700 uses this lacing on both front and rear,
while the 7701 rear freewheel side is actually radial laced.
The Dura-Ace rear hub has radial lacing on one side,
but this one uses tangential lacing on both sides.
The lacing on the far flange relative to my view isn't a mirror image
of the near flange—rather,
it has a JIS-style structure where flipping the wheel left and right
doesn't change its appearance.
So when you look at the resultant force of the two spokes that make up the final crossing,
on the right side you get 0-cross lacing with spokes twisted clockwise,
and on the left side you get 0-cross lacing with spokes twisted counter-clockwise.
↑This is a ZIPP rear hub from a certain era,
and the spokes in the porcupine direction
alternately come from the large flange and the small flange.
So when you look at the first crossing,
you see crossings between spokes from the large flange
and crossings between spokes from the small flange
alternating with each other.
This rear hub is 24-hole, but
this lacing method isn't possible with 20-hole.
The number of large-to-large and small-to-small crossings
needs to be even (20-hole is 5 crossings).
So to use this lacing method,
you need (4×n)-hole with n being even,
but 8-hole with n=2 doesn't have enough spokes
to make a functional wheel.
Next, 16-hole with n=4,
but if you use this lacing method,
↑it ends up looking like this, with spoke tension variation
that alternates back and forth in a way that's too noticeable to ignore—
it feels wrong to me.
The earlier 7700-series 16-hole wheels are also paired-spoke wheels,
so this lacing method should be avoided,
and indeed they don't use it.
Beyond that you have 24-hole with n=6, 32-hole with n=8,
40-hole with n=10, and so on,
but realistically the options are 24-hole or 32-hole,
and for 32-hole straight-spoke complete wheels,
off the top of my head I can only think of
Mavic's Ksyrium SL and
its successor the Ksyrium SLR,
so the lacing method of alternating large/small flange
for the porcupine-direction spokes is essentially
24-hole exclusive.
As for the Dura-Ace hub,
in the diagram above I've exaggerated the high-low of the flange peaks a bit,
but first you'd just thread the porcupine-direction spokes
through the large flange side
and repeat that 6 times to thread all the spokes, but instead
first I threaded the spokes through every other flange peak
for 3 iterations,
and the remaining 3 spots would be the reverse of that—
this is what I attempted during pre-assembly,
but I didn't properly revert it,
and that was my mistake this time.
That article also had
another comment.
"With this rim and hub, wouldn't you get a better wheel
by halving the non-freewheel side
and doing 18-hole 2:1 lacing with aero spokes
that are nearly 100% spoke weight ratio?
Or is that not the case?"
It's a seriously fantastic comment,
as if the wild Fulcrum has appeared from some bike blog.
100% spoke weight ratio flat spokes
correspond to Sapim CX.
The nominal weight for 260mm length 64 pieces is 423g,
so 423÷64÷260÷0.0257
equals 0.989131..., which is
about 99%, but
from my accumulated measurements it's actually 100.3%.
Either way it's in the range of about 100%.
100% class aero spokes won't fit through
a 2.3mm round hole (CX is 2.8mm wide),
so it's impossible with this Dura-Ace,
and you'd need either a flange with a slot hole
or a hooking-type rather than threading-type design.
Setting aside those logistical considerations,
if you build an 18-hole rear wheel with 2:1 lacing
using a 24-hole rim and 24-hole hub,
you'll have an idle phase at the non-freewheel rim holes
and spare rim holes.
When an idle phase occurs with 18-hole,
truing becomes extremely difficult,
so you'd need a rim with stiffness designed
with that in mind.
For example, the Bora Ultra 80,
which is an 80mm deep rim brake Bora,
uses Campagnolo G3 spoke bundles that are
6 pairs / 18-hole rather than 7 pairs / 21-hole,
and since it's G3 lacing it has a long idle phase,
and the rim stiffness is designed with that lacing in mind.
Aside from the hub holes, I'm not particularly fond of
having spare rim holes,
but I have done repairs on 21-hole rims
with 28-hole rims when unavoidable (→here).
So if you did it with an 18-hole rim
that had no hole offset,
it might make for an interesting wheel.
Thanks for the comments!