Aerodynamics thoughts

This week, the Al33 and a number of other rims will be tested in the A2 wind tunnel in North Carolina. We aren't paying for it, it isn't our test, but we agitated for it to happen, facilitated all of the arrangements, and designed the test. Without those inputs from us, the test wouldn't have happened. Call our stake sweat equity. We won't be there, but the US distributor will be, and we will be there in virtual.

I'm excited. I love testing, I love the wind tunnel, and I've learned a lot every time we've done it.  

The test will use a 2017 model Zipp Firecrest 303 as a baseline. Tested wheels will be a HED Belgium+, a Kinlin XR31T, the Al33, and a Flo30. All wheels are 20h rims built with CX Rays and standardized hubs, except for the 303 which is of course an 18h wheel which uses Zipp's own hub and CX Sprint spokes. 

The test will be done as a wheel only test. This has plusses and minuses, but it's proven to be an accurate way to test wheels and it allows wheels to be tested in a time efficient manner, which means we can include more wheels.

Test tire will be a Continental GP4000sII in 23c size. There will also be a 25c GP4000 there, how much testing gets done with it depends on time available.

All the quantitative data on each rim/wheel will be presented - depth, weight, inner and outer width, and retail price. 

We'll be doing the standard 20* sweep in 2.5* increments, on one side. Since these are all symmetrical front wheels, doing both sides would take time that would reduce the number of wheels we could test. We will also include steering axis force data as provided by A2. 

Not entirely certain how the data will be presented. My inclination is to show the standard graph like you've all seen 100 times, and then overlay some of the more defensible angle of attack distributions over top of those. It makes the most sense to then use those distributions to create a one number score for each wheel. 

I'd encourage anyone who wants to get the most from this information to become as informed as possible about the benefits and limitations of aerodynamics testing. Tour Magazin is an amazing resource, and you can go to the App Store, download their app, and buy issues for about $3 each. Issue 8 from 2016 is particularly good. Become familiar with the other methods like Chung and Alphamantis. 

A few bullet point thoughts:

1. Any quantitative test will have some strengths and weaknesses, but no wheel can make aerodynamics claims without credible quanification that allows at least some comparison to relevant standards. It shocks me how many brands still try to skirt past with a "trust us, we're fast!" line of bull. If the whole sales proposition for any wheel is that it's fast, yet it shows no data, I think you know what I'd say to that. This principle is why we insisted that the Al33 absolutely needed to be tested. 

2. Depth and speed are not interlocked. We first showed this four years ago when doing the original Rail 52 test, where the 52 proved faster than even the 85mm wheel than we'd been using, and was faster than the deeper Zipp 404 at angles from 0 through 5 degrees. 

3. We're still using the GP4000 in 23c size because that's been the standard, and it's still a VERY widely used tire in situations where aerodynamics are important. Our previous tests showed a reliable pattern that wider tires had a linear and predictable negative effect on outright aerodynamics. 

4. The Zipp 303 gets used "as is" because it's a wheel system, and its value as a baseline is in using it as it's been used in other tests. That allows you to make worthwhile comparisons to the greater universe of what's out there.

5. Have reasonable expectations. In the Tour Magazin test I referenced earlier (seriously, download it), the difference between best and worst was 13 watts. That's 40ish seconds in a 40k TT at 30mph between a Mavic Ksyrium and a 404 and DT Swiss 65, which were the fastest wheels in the test. That's about .4mph, worst to first. Anyone telling you you're going to go 2 or even 1mph faster by just switching to more aero wheels is selling you a load of crap. 

Okay, that's it for now. Looking forward to Thursday. 


Back to blog


So the post script to my answer to Chris K from yesterday evening: the difference would be more or less bupkus when you're really ripping. Why? Because when you're really ripping, angle of attack (aka yaw) is really low. Like 0*. And really low attack angles is where differences between wheels are at their least. So if we look at our old test from July 2014, the difference between a Kinlin XC279 based wheel and a Zipp 404 is 1w at 0* at 30mph. Say the Ksyrium is another watt down on that, so 2 watts down. Difference in speed? More or less nada. And if the angle of attack isn't more or less 0 and you're going 35, you're going to go faster with the more easily controlled wheel. So there's that. Sometimes riding makes me dumb, sometimes smart. This little jewel was courtesy of today's lunch ride.

Dave Kirkpatrick

Rail 34s were available with any hub we sold at the time. Effective price delta between Rail 34 with T11s and RFSW3 with T11s is $550. We sold very few Rail 34s with Nimbus Ti hubs actually. And the price gap between Nimbus Ti and T11 was nowhere close to $200 – as I recall it was $60. Nimbus Ti was an impossibly good value, though. The weight gap is closer to 40g per set. This is a big part of why alloy makes sense at this rim depth – the weight differences are very small. Carbon wheels need mass at the brake track for a heat sink. An Enve 3.4 front that we have built in the shop weighs 12g less than the Al33 front that's going to A2 – same hub and spokes. And if the weight is the same as carbon, and the strength and stiffness allows the same spoke count as carbon, and the aerodynamics are at par (which we'll see – no claims made yet at all), then what is carbon buying you? Entirely disregarding the price gap, I'd rather have the alloy wheels in this situation. The manufacturing process is less susceptible to error, there are fewer variables in the rim's lifespan, braking is just plain better, and there are no use case restrictions. That's not me hating on carbon, that's me believing in sound logic. At the time of the Rail 34, it seemed that it was the right product, and it was what some significant part of the market wanted (or at least wanted to want – there was always more interest in than sales of Rail 34). We just plain learned that some of our, and really the world's, premises were wrong. Trust me, it's not easy to swim upstream like this. The elephant in the room is that you make more money selling carbon – it creates its own demand, differentiates itself from the rest of the market more easily, and it has better raw margins by quite a bit. Wheel builders don't want you to know that, but it's true. It's hard f-ing work making it go without carbons, and I'd bet most brands would fold somewhere just outside of instantly without it. When you're into wheels fast enough to make a substantive difference in time trials or whatever "hairs that get split are paris that count" venue, then carbon can do things that alloy can't. But that's not how most people ride. From a personal perspective, it sucks to be the outcast and people call us names and we're freaking heretics or worse to a lot of people, but we think we're right.


I'm sure Dave will correct me if I'm wrong but I believe the Rails always came with the November branded White Industries hubs, which probably cuts a hundred or two off the price delta. That being said for equivalent aerodynamic performance I think 70 grams is a small price to pay for aluminum braking, even if the price had been equal!

Ben Gordon

I'm looking forward to seeing these test results. But until then, I'm guessing the alloy Al33 will perform very closely to the carbon Rail 34 in the wind tunnel, since they share pretty similar dimensions and shapes. So if that's true, then it may be fair to compare the RFSW3s built with Al33 rims to the discontinued Rail 34 wheels. And if you consider the slight weight penalty (the Al33 rims weigh 470-475g, vs Rail 34 rims weighing about 440-450g) in exchange for the substantial difference in price (RFSW3s are $715 currently, vs $1285 for Rail 34 wheels about 3 years ago, both built with "standard" Novatec hubs?), then the RFSW3 offers a crazy amount of bang for the buck. And I'll gladly accept that extra ~40-70g per wheelset to save ~$570, especially if the aerodynamics of the Al33 are as close to the Rail 34's aerodynamics as I suspect. Hmmm… I think I just talked myself into buying a pair of RFSW3s (albeit the disc version)!

another dave

Hi Chris – I gotta get Mike to get out the old scale-o-matic (there's a function to scale speed that he has from A2), but you're more in the range of 36, possibly 36.5. Go into the Sagan Supertuck and you're going 40+.Dave


Leave a comment

Please note, comments need to be approved before they are published.