Over next fall, I plan to re-build the rear wheel on a bike that has a 7-speed freehub, eyeletted Fir tubular rim, on a Shimano 105 hub. I plan to use 14/15 DB spokes 3x on the drive side, and 14/16 db spokes radial on the NDS. Here is my question:

I've seen semi-radial spoking done with the heads in, which is probably the strongest laterally, but looks kinda bad. Heads out looks nicer, and is probably more aerodynamic, but is probably less strong laterally. I was wondering if alternating in/out on the radial side would have any benefits, and if it would just look odd. Has anybody ever done this way of radial spoking?

I always put the heads on the inside, I don't think it really makes much difference.

Heads-out on a half-radial will improve spoke tension balance some.
But its quite uncommon to use both a thinner NDS spoke and a half-radial lace.

Yeah, I may be going a little overboard there....

If the hub was already built as a conventional 3-cross wheel, I wouldn't lace it any other way. A virgin hub? Go crazy.

I don't think radial on the rear is a good idea. Radial spokes are not good for power transfer.
And not all hub flanges can take the stress of radial spokes.

Mavic has already made over a million of them, I have a set of Aksiums with several thousand miles on them, very stiff, never had to true them.

Full radial is supposedly not a good idea. Although there is a guy here claiming to have used it w/o any issues. Never tried it myself.



Half-radial WRT power transfer to the spokes is a non-issue for almost all hubs.

If it's a DS radial, the hub spindle is usually more than strong enough to transfer the torque to the NDS flange and the (semi-) tangential spokes there.

If it's a NDS radial, well, then you have the power pick-up at the correct flange already.



Power transfer to the rim might merit some discussion.

In theory, a half-radial wheel would see more wind-up during each power stroke, as there's only half the number of spokes "perfectly" placed to carry the torque compared to a both sides (semi-) tangential wheel.

IRL and IME, I don't notice any difference.

On the wheels I have, and the spoke tensiometer I have, I see precious little increase in spoke tension as a result of standing on a pedal.

True.
But maybe not necessarily important for a NDS half-radial.
This is something I've been meaning to throw some math on for years:
- unless perfectly tangential, the force experienced by the spoke hole in the flange can be divided into two vectors, one radial, one tangential.
- nobody would think you much of a Daredevil for running a 2X rear.
- a 2X rear (depending on spoke count and flange diameter) would have a decently sized radial vector on the DS.
- NDS typically sees something like 60% of the DS tension
- so if the radial vector in a 2X DS isn't too far away from the 60% mark, then a radial NDS doesn't have to be a surefire ticket to impending self-destruction.


(Disclaimer 1: English is my 2nd language. If I haven't got the words just right, try to see the idea and not the letters.
Disclaimer 2: This is the idea in short. I don't claim to have worked it all out. Constructive additions welcome.)

And if you do anything but heads-out, kinda defeating the purpose.


The traditional reason for a NDS radial is to cure/avoid NDS spokes breaking from fatigue.
First/easiest thing people tend to do to avoid/cure spokes breaking from fatigue is to increase tension.


And heads-in will have a lower tension than heads-out, thereby contradicting the traditional purpose of a NDS half-radial.


If you use a decently flexible spoke calculator, you can check the impact on heads-in vs heads-out by changing the position of the NDS flange.


Shift the flange half the flange thickness further out, and it'll give you the tension balance of a heads-in build.
Shift the flange half a flange thickness further in and it'll give you the tension balance of a heads-out build.

I generally set spoke tension by sound. As long as the least-tensioned spokes have 'enough' tension, then I know I'm in the ballpark; once trued and centered, the entire system should be good. Never had a spoke break on a wheel I've built yet using this method.

They you're aiming to fix what ain't broke.


Which is fine, as a hobby.


Just don't expect much of a wow-factor when going from "good enough" to "presumably better".
An unbroken spoke can't get any unbrokener.


An alternating heads-in/out radial build would be very similar to a crossed build WRT lateral stiffness.
In a lab setting, it'd probably read as a tiny bit stiffer due to the shorter spokes being used.