Zipp had issues with my Revolution Wheelworks review, and the subsequent Zipp article has now upset Steve Hed. I got an email from Hed stating that while he and Zipp have a cordial relationship and would like to continue to do so, he wanted to rebut some of the claims in the article.
Regarding Poertner's claim that Hed optimizes his wheels for a yaw angle of 26 degrees, Hed wrote:
We do not base our wheel designs on a 26 degree wind angle. It is true that our website points out that a slower rider (19mph) can see an angle of 26 in a 10mph sidewind. The preceding sentence on our website states that a 27mph rider in the same wind sees a 17 degree apparent wind, and it illustrates that at different speeds riders encounter different wind conditions. This is not a statement that we base our rim designs on 26 degrees.
As for this statement, "before Zipp introduced the 808 in 2004, neither company produced rims of this shape, and that before the 808 ventured to be 28mm wide, neither company was making super wide toroidal rims that are now the staple of both companies", Hed responded:
Our patent was granted in 1990, the same year we produced the toriodal shaped CX wheel. All our wire spoked wheels have used this toroidal shape continuously since then. The original CX had a 26mm wide rim section - certainly wide for the time, and still wide by current standards. The Jet 90 and CX Deep entered into production in 1996, nine years before the 808 was available.
And regarding this, "He also points out that over the last few years the Hed rim shapes and wind tunnel data have started to look an awful lot like Zipp rim shapes and Zipp data."
In the early 90s Zipp's wheels were deep V shapes with flat sides and a rounded nose section. Our CX had a 21mm width at the rim bed, and bulged out to 26mm partway down the rim before narrowing to a rounded nose at the spoke bed. We were using wind tunnels in the late '80s, and nothing has changed about the way air flows over a wheel since then. The fact is that since Zipp acquired our patent (sometime in the late 90s) their wheels have changed shape to more closely mirror the wheels we started selling 18 years ago. We have continually improved them since then, but the underlying aero shapes are still similar. As Zipp's wheel shapes have changed to more closely resemble ours, it only follows that their wind tunnel data is more like ours too.
I asked Hed what angle or range of angles does he optimize his wheels for?
The answer is that there is not a likely yaw angle, there is a percentage where one wheel is faster or slower, and a percentage where they are roughly equal. What we tried to convey (and what was misconstrued) on the website is that this percentage changes for different speed riders. Figuring the range of wind angles from straight headwind (zero degrees) to tailwind (170 degrees) the numbers for H3 vs Jet9 work out like this:
A 20mph rider in a 7 mph wind is faster on the H3 in 38% of possible wind angles
Faster on a Jet 9 in 26% of possible wind angles
Roughly the same in 36% of wind angles.
With this wind, in 3/4 of wind angles the H3 on a 20mph guy is better off or not hurt on the H3. In 1/4 of wind angles, the Jet 90 will be the faster wheel. Same conditions with a 27mph guy? 90 will be faster 53%, H3 will be faster 26%, and about the same is 21%. The argument that courses have turns and direction changes is quite valid. It makes it very likely that riders will "see" several wind angles, and it also makes it likely that they will see close to the full 180 degrees of possible wind. Without modeling a specific course we think that the "faster, slower, equal" method makes a lot of sense.
I asked Hed if his way of figuring gives equal weight to all yaw angles regardless of probability. If it is a bell curve distribution, shouldn't those percentages be weighted to not only what's possible, but what's more probable as well?
Yes. That is why if one has a "drag" chart of our wheels, and uses the yaw calculator- they can model a course for a certain wind condition, and direction. If you know that in your area, you are just doing a local TT on Saturday morning and wind speeds are never over 4 mph and you a 27 mph fast guy- you just need to look at under 5 degree data. This will help you narrow down your wheel choices.
The percentage idea, is just a way of giving a customer something to think about, as a way of explaining the difference between different wheels. All of our aero wheels decline in drag with yaw, but at some point they "stall" and the drag climbs. A three spoke maybe doesn't drop as far as a Stinger 90, but it has lower drag at higher yaw angles, where the Stinger 9 has higher drag.
The drag chart is here on Hed's site. I had a feeling Zipp would object to Hed's data, so I tried to pre-empt another response from Zipp by asking Hed why the Zipp Sub 9 never goes negative in drag in Hed's chart as Zipp claims, even though it was pretty easy to make it move forward in a crosswind:
There is of course no real way to reconcile why the Sub 9 goes negative in Zipp testing and not in Hed testing. Neither company would ever want to fully disclose the test procedure because it would then allow other companies to duplicate the processes for their own benefit.
My guess as to why the Sub 9 is not negative is that Zipp tests without a cassette installed. We always test with a Dura Ace cassette installed - it is either an 11-21 or 11-23 I have forgotten as that cassette permanently lives at LSWT in San Diego. They may have tested with a skinnier tire than our test which was a 21mm Vittoria CX. The skinnier tire would give the Sub 9 (the same is true of the Stinger disc) a slightly later stall point and thus lower drag. I would guess that if we used a 19mm tire and no cassette that the Sub 9 would go negative. Besides Hed and Team Columbia I have never heard of anyone else testing a Sub 9 in a wind tunnel so no one has ever independently confirmed negative drag for that wheel.
As you can see, it's very hard to wade through each side's claims and hope to reach an objective conclusion. Here's a 3rd party study that weights the various angles based on probability and pops out one number for each wheel. The only Hed wheel they tested was the Hed3, which absorbed 3 more watts than the Zipp 808 at 50kph, the equivalent of 22 grams of drag. Here's those two wheels on Hed's drag calculator for you to assess what that 22 gram difference looks like over a range of angles. To muddy the waters even more, the Roues Artisanales distribution curve peaks at zero, while Zipp would argue that it should peak somewhere between 10 and 15 degrees.