The AOA Curves of the Rail and RFSCs

Last time I wrote about the performance of the Zipp 404 FC wheel we sent to the windtunnel as a benchmark across a range of Angles of Attack (AOA), and tried to make the point that the wheel that has the lowest trough when plotting drag against AOAs isn't necessarily the fastest wheel. And by fastest I'm just talking about aerodynamically fastest here - a wheel that crushes all comers in the tunnel is not the best choice for all races and all conditions. Its aero slipperiness has to be weighed against its handling, road feel, stiffness and weight. As you can imagine, a race with a lot of climbing would favor a wheel with low weight over pure aerodynamics, while a crit with a hundred hard jumps out of corners would be better suited to a wheel with a good balance of aerodynamics, low weight, stiffness and handling. So while we're going into a lot of detail about our findings at the windtunnel, we don't want you to lose sight of the big picture - that aerodynamics alone do not a fast wheel make. That's the heart of the design philosophy we brought to the Rail.

Recalling what I discussed last time, when reading the AOA curves we need to remember that the wider AOAs (10 degrees and up) are more common at slow speeds, while the narrower AOAs (7.5 degrees and down) are more prevalent at higher speeds. A mnemonic device to use when looking at AOA charts is to think fo them as the bike path you reluctantly admit you ride on once in a while - ride slowly on the right, pass quickly on the left.

Here then are the curves for our full range of RFSC wheels - 38mm, 50mm, 58mm and 85mm depths:

A couple of immediate observations:

- The wheels show a remarkably similar performance at very narrow and very wide AOAs, suggesting that adding depth isn't necessarily adding unqualified speed
- There is almost no measurable difference at all between the 50s and 58s (as evidenced by the 40K TT graphic here, showing that if you ride balls to the wall for a full hour the time difference between the wheels is 0 seconds)
- All of the wheels show an increase in drag in the middle AOAs experienced at all speeds. The wheels are faster than the FSWs, but these are still far from ideal curve shapes. These wheels are all at their best when there is zero wind, or when you're riding so fast that the AOA becomes very very narrow.

Our RFSCs were likely designed to be aerodynamic, but it is pretty clear they were never optimized for any particular application, and maybe never even saw the inside of a windtunnel until we sent them. Our philosophy is that you don't need an array of supercomputers running CFD in parallel to design an aerodynamically sound rim that performs exceptionally well, but we learned from this trip to the tunnel that you can't just pluck a shape out of thin air and call it fast either. Dave and I chewed through a lot of fingernails waiting for the tests of the Rail to come back, knowing that we were fully prepared to scrap the design and start again if it didn't produce the results we expected from the design we chose. Fortunately it did exactly what we were after, which is why we were able to greenlight it for production.

So how did the Rail perform across a range of AOAs? Like this, plotted alongside the curve for the Zipp 404 FC as a reference:

The curve shapes, you'll notice, are pretty similar. They start and end in almost exactly the same place and both have troughs in the mid-low AOAs. The gap between the two widens at about 7.5 degrees and continues all the way through 17.5 degrees, making it appear as though the 404 enjoys an enormous aerodynamic advantage over the Rail. The Rail performs better at narrow AOAs, but the difference does not appear as prounounced from the curves. The reality though is because of the distribution of AOAs at different speeds, the 404 is only 2 seconds faster than the Rail over a 40K TT at 30mph; the small advantage the Rail has at narrow AOAs goes a very long way towards mitigating the 404's edge in the wider AOAs, which occur far less frequently at racing speeds. (Or in other words, right slowly on the right, pass quickly on the left.)

We never expected to be faster than the 404 FC, and even when we get to the production version with 4 fewer spokes than our prototype, the Rail will still have 4 more spokes in each wheel than the Zipp and a 6mm shallower depth. But at 18mm of inside width and the added stiffness of the extra spokes, we think it's going to be about the fastest around for the whole race course. And yes, we are looking at some ways to quantify that position as well as the windtunnel measures aerodynamics.

Here's the whole collection of wheels so you can see how they all plot against each other:

I realize that the engineers and aero geeks in our audience are cringing at what they see as a rather hamfisted interpretation of exceedingly complex aerodynamic data. There is a ton of nuance that influences an analysis like this, most of which I've glossed over in order to explain what we found at the windtunnel and the conclusions we draw. What we do on the blog here is no white paper, but it isn't a marketing slick either. We're just sharing what we observe and learn as candidly as we know how. Even if the Rail isn't for you, we hope what we're doing here helps you better evaluate what other wheel makers are saying (and not saying) about their products.

We still owe you data we promised on our tests of different spoke counts, which I'll get to within the week.

Back to blog


Since you seem to be aiming your products at amateur racers / those of us that have to buy our own gear (and therefore, probably have jobs, responsibilities, limited ride time, etc) do you happen to have more speed-appropriate data available? I personally don't ride a 40 K TT at 30 MPH and don't know a whole lot of guys that do. It would be interesting to see data for something like 25 MPH up to 30 MPH at 1 MPH increments. Just a thought.


Hi Asad-I think the answer to your first question is also the answer to the question in the 2nd paragraph. We used the same 23mm Vittoria Corsa Evo CX tire on all the wheels we tested. We've since seen other tests of multiple tires which suggest the Vittoria may be more aerodynamically challenged than some other tires. But we didn't modify the data in any way.What the dip you refer to shows is a lower aerodynamic drag at some AOAs, or rather, improved performance compared to a relatively flat line. If we had tested a range of tires at the outset, we may have found one with a more pronounced dip. Naturally when brands do this, they are inclined to publish the curves that show the best performance. (I think the curves that Zipp publishes for the 404 FC were generated using a 21mm Zipp Tangente tire created expressly to optimize aerodynamics on Zipp's wheels.) Our objective was not to find the best set of data to show off our wheel; it was to test our wheel against relevant benchmarks using a tire that we know is very popular among our customers. But the tests that have since been published that measure the performance of different tires show that tire selection may play a much larger role in aerodynamics than was previously known. This isn't a surprise – if a big brand is spending a ton of money on a proprietary design and R&D that nets a gram or two of drag improvement over a rival, it doesn't want to call too much attention to the fact that a change of tires could have netted the same gain. Now that that horse is out of the barn, people are paying attention to tire choice to get the most gain out of the wheels they're using. So we're sending the Rail to the tunnel again with a range of tires to help them decide.

Mike May

Is there a reason your wind tunnel curves for the Firecrest don't show the pronounced dip in the middle of the yaw range that is seen in other Firecrest wind tunnel data (not only Zipp's, but also in Trek's Aeolus white paper where they compare the Aeolus to Zipp and others)? That dip seems to be a general trend not just for Firecrests but also other wide, blunt-nosed designs like HED Jet/Stinger, Bontrager Aeolus, Flo, etc. Were the tares removed from your data before plotting?And what tires were used? The Flo guys got drastically different results when tested with a GP4000S vs. with a Michelin Pro3 Race. In fact, the Pro3 Race wind tunnel curves have a much less pronounced dip in the mid-yaw range than the GP4000S curves.


So what you're saying is I should have read a little more before shooting my mouth off :) I've just recently stumbled across your site and am slowly working my way through articles in reverse order and apparently hadn't gotten that far yet. Thanks for all the data you're providing.


Hey Kevin-We agree that the 30mph 40K TT is an unfortunate standard. We give some data on the aero performance of the Rail and other tested wheels at lower speeds here:

Mike May

Leave a comment

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