alienator said:

I don't know what you mean by angles under which people can ride. It's not a function of angle, as much as it is a function of air speed coupled to angle. It's perfectly feasable to ride with a perfect crosswind.

Anything over a 20 degree angle of incidence is quite extreme... <10 degrees is normal.

About the "perfect crosswind"... in order to ever see that, the wind must be blowing faster than I am riding. If I'm riding at 25 mph what sort of wind would I need to get a perfect relative crosswind on myself? Draw the vector diagram, and you'll realize that you need a large tailwind to compensate for the forward speed. With wind coming at a 45 degree angle from the rear you'd need a 35+mph wind at wheel level to achieve a perfect crosswind... which is equivalent to ~70 mph wind as reported by airports! I never ride when it is that windy...

Now take an average breeze of 8 mph (at the airport) and say you are riding perpendicular to it. What is the angle of incidence if I'm riding at 25mph? At wheel level the wind speed will be about half what the airport gives, provided that you are riding in an open area (otherwise it will be even less)... so atan(4/25) = 9 degrees.

So... if it is really windy (18+mph) and it is a straight crosswind you could see a 20 degree angle... but most of the time it will be a lot less than that. And I'm pretty sure that a CX-ray would not have any strange flow separation issues at angles <20 degrees... you might even get a little lift effect.

If you'd like to get more complicated you could note that most of the spoke drag occurs on the top spokes which are traveling up to twice as fast as the bicycle... and on these the angle of incidence would be even less.

The oval shape will have a Cd of ~.6 at typical spoke Re number (10^3) compared to the round spoke with a Cd of 1. Multiply that by the width (.9mm and 1.5mm) and the CX-ray has only 36% of the drag that the Laser has.

From all the aero wheel tests I've seen too, it sure looks like the oval spokes are better... and that includes high angles of incidence. If Reynolds has some evidence that round is the way to go, they ought to publish it.