I’m pretty sure the two tires are dimensionally identical.
They aren’t…just because of the construction. There’s a “lip” on the TC where the tread is glued to the casing that’s not present on the Turbo
Ah, gotcha. I wonder how it would have fared with the 24mm Force. That seems to be like a really good bet on the new generation of ultra-wide wheels.
Yeah…but they’re not likely to show data of their wheels with other brands of tires 
I’m pretty sure the two tires are dimensionally identical.
IIRC, if you look at Tom A’s wind tunnel data you will see that the Turbo tire is much more aero than the Turbo Cotton tire.
It was a much more aero tire on that wheel. It may or may not be the same on a different wheel.
True. But judging by how all “open tubulars” seem to do in aero testing (i.e., early stall), I wouldn’t bet too much money on the Turbo Cotton being an aerodynamically good choice on any wheel.
You’re probably right. But, as Tom says, “low Crr can make up for a lot of aero sins” ;).
They tested with a 24mm Turbo Cotton
Nope, they were tested with their Turbo tire:
“Win Tunnel testing were all done around a Specialized S-Works Turbo 24c tire (measuring ~26mm wide)”
It’s nice that they included rotational drag, but then they go and mess up the axis (and I would like to hear how they measured the rotational drag, by the way). Why include units on the axis if you are not even going to put numbers on it?
Actually, it says they designed around the S-Works Turbo, but the testing was with a Turbo Cotton.
I’m dismayed by the lack of units on the charts…at least tell us what each division represents. Otherwise, there’s no way of knowing if we’re looking at small or large differences :-/
I see that now. Confusing. Smells a bit like the Turbo Cotton gave better results compared to Zipp (like mentioned above). Why else not use the tire you designed around? (Maybe they should have designed around the Turbo Cotton in the first place, but that’s another thing).
Yes, at least give us numbers for the divisions. I can see Mr. Yu cringe as he saw what the marketing department ended up with for a final release of this paper 
That’s Dr. Yu now
Here’s the thing about Turbo vs Turbo Cotton…as we’ve seen, the much lower Crr of the TC tends to make that a faster choice overall, especially for racing. I have no problem with them using that for the aero testing, since that’s the most likely choice for a racer to use…AND it more closely represents their fastest tubular tire offering as well.
In other words, don’t get caught up in chasing high yaw angle performance of tires that aren’t as low of a Crr…that’s not “keeping your eye on the ball”, so to speak.
I’m pretty sure the two tires are dimensionally identical.
IIRC, if you look at Tom A’s wind tunnel data you will see that the Turbo tire is much more aero than the Turbo Cotton tire.
It was a much more aero tire on that wheel. It may or may not be the same on a different wheel.
True. But judging by how all “open tubulars” seem to do in aero testing (i.e., early stall), I wouldn’t bet too much money on the Turbo Cotton being an aerodynamically good choice on any wheel.
You’re probably right. But, as Tom says, “low Crr can make up for a lot of aero sins” ;).
Yep, it’s probably still the faster of those two tires on most/all wheels when taking Crr into account 
They tested with a 24mm Turbo Cotton
Nope, they were tested with their Turbo tire:
“Win Tunnel testing were all done around a Specialized S-Works Turbo 24c tire (measuring ~26mm wide)”
It’s nice that they included rotational drag, but then they go and mess up the axis (and I would like to hear how they measured the rotational drag, by the way). Why include units on the axis if you are not even going to put numbers on it?
Actually, it says they designed around the S-Works Turbo, but the testing was with a Turbo Cotton.
I’m dismayed by the lack of units on the charts…at least tell us what each division represents. Otherwise, there’s no way of knowing if we’re looking at small or large differences :-/
I see that now. Confusing. Smells a bit like the Turbo Cotton gave better results compared to Zipp (like mentioned above). Why else not use the tire you designed around? (Maybe they should have designed around the Turbo Cotton in the first place, but that’s another thing).
Yes, at least give us numbers for the divisions. I can see Mr. Yu cringe as he saw what the marketing department ended up with for a final release of this paper
That’s Dr. Yu now
Here’s the thing about Turbo vs Turbo Cotton…as we’ve seen, the much lower Crr of the TC tends to make that a faster choice overall, especially for racing. I have no problem with them using that for the aero testing, since that’s the most likely choice for a racer to use…AND it more closely represents their fastest tubular tire offering as well.
In other words, don’t get caught up in chasing high yaw angle performance of tires that aren’t as low of a Crr…that’s not “keeping your eye on the ball”, so to speak.
Him being a Dr. just make me think he cringes even more
I know, the TC is most likely the faster tire in most circumstances. Which is why they perhaps should have designed around that. But then the graphs don’t come out as nice! So you got to figure out if you want to go fast or look fast (lucky for me a lot seem to go for the latter! ).
Hey gang, I’m running between tests and meetings this morning but rest assured answers are coming. All great questions.
Hey all,
Thank you for the comments - we’re working to be as thorough as possible with our data. We’ll work through these questions as quickly as possible.
To answer the Y-Axis unit scale question - the divisions on the graph are 0.001m^2. The range for each graph was 0.020m^2. Ultimately, we decided to not include specific absolute values as testing set up (facility, fixturing, build, even the actual tire used) can cause differences in the absolute values between tests. All of the data collected in these graphs was collected in one test session, (though deltas were confirmed on multiple test sessions), same exact tire for all tests, and inflated to the same pressure. We will answer the specific rotational drag questions shortly, but the rotation drag was always an addition to the translational drag. The graph’s range just shifted slightly higher.
Chris can discuss further on the tire choice and development of rim shapes, but the Turbo Cotton tire was selected for drag testing to show the fastest overall package (translational, rotational, Crr).
Cam
why didn’t you include aero results for the enve 4.5? weight comparison given where you come out very well but then no aero - did you lose?
roval wheels have always done well on weight, often at the expense of other qualities…
Hey all,
Thank you for the comments - we’re working to be as thorough as possible with our data. We’ll work through these questions as quickly as possible.
To answer the Y-Axis unit scale question - the divisions on the graph are 0.001m^2. The range for each graph was 0.020m^2. Ultimately, we decided to not include specific absolute values as testing set up (facility, fixturing, build, even the actual tire used) can cause differences in the absolute values between tests. All of the data collected in these graphs was collected in one test session, (though deltas were confirmed on multiple test sessions), same exact tire for all tests, and inflated to the same pressure. We will answer the specific rotational drag questions shortly, but the rotation drag was always an addition to the translational drag. The graph’s range just shifted slightly higher.
Chris can discuss further on the tire choice and development of rim shapes, but the Turbo Cotton tire was selected for drag testing to show the fastest overall package (translational, rotational, Crr).
Cam
Thanks Cam. I realized after I made that post that I could just scale the units off of the plot you did last year with all of the S-Works Turbos and Turbo Cottons on the CLX64 wheel, which included the TC 24…which is posted right on my blog
Hey gang, I’m running between tests and meetings this morning but rest assured answers are coming. All great questions.
I am particularly interested in how you set up the Win Tunnel to measure rotational drag. Some other questions:
P. S. If you need a set of Specialized Tri-spokes to test the last 2, I have a set you can borrow :-). I would love to know if my Tri-spoke front, with a 20mm SS, still rules on my Nose-cone Shiv at low yaw angles.
https://www.bikerumor.com/wp-content/uploads/2017/01/Roval-CLX-50-Whitepaper-Summary.docx
I make so many writing mistakes that it is unreal, but how do they mess up “Aeolus” repeatedly in this.
Ok, to answer some of the questions in this thread:
Turbo 24 vs. Turbo Cotton 24: Simple explanation - when we started doing the shape design of the current family of CLX wheels (32,50,64) the Turbo Cotton didn’t quite exist yet. All our historical CFD and benchmark data was based off the Turbo 24. As you can imagine, it takes a while to fully quantify and correlate a non-exact shape like an inflated tire casing. However, the comparison testing was done much more recently (e.g. includes the recently launched Zipp 454) and in our testing, and as Tom A. pointed out, the fastest combination of translational drag, rotational drag, and Crr was with the Turbo Cotton tire. This was true for the Zipps as well. As a result, we felt it would make the most sense to publish comparison data using the fastest available total combination.
Rotational drag (“power-to-spin”): As many of you know, it is extraordinarily hard to be able to reliably/consistently measure this with adequate resolution. Then there are complications like accounting for, or isolating tire Crr or bearing drag. In fact, Tom A. probably remembers us telling him we were working on various different methods (most of which we ditched) to do this in the Win Tunnel when he visited over a year ago. Our engineers recently developed and validated a protocol and equipment to be able to isolate out the rotational aerodynamic component at a resolution of ~50 mW. Since we devoted a lot of time and resource into developing the equipment and protocol to achieve this, I think you’d understand if we’d like to keep the specifics proprietary. Hint: involves a very careful and modified application of the Chung method in the tunnel while understanding what error terms can be neglected vs. not.
Anyways, a lot of geeking out. The CLX50s are FAST and ride very responsively.
Chris
Ok, to answer some of the questions in this thread:
Turbo 24 vs. Turbo Cotton 24: Simple explanation - when we started doing the shape design of the current family of CLX wheels (32,50,64) the Turbo Cotton didn’t quite exist yet. All our historical CFD and benchmark data was based off the Turbo 24. As you can imagine, it takes a while to fully quantify and correlate a non-exact shape like an inflated tire casing. However, the comparison testing was done much more recently (e.g. includes the recently launched Zipp 454) and in our testing, and as Tom A. pointed out, the fastest combination of translational drag, rotational drag, and Crr was with the Turbo Cotton tire. This was true for the Zipps as well. As a result, we felt it would make the most sense to publish comparison data using the fastest available total combination.
Rotational drag (“power-to-spin”): As many of you know, it is extraordinarily hard to be able to reliably/consistently measure this with adequate resolution. Then there are complications like accounting for, or isolating tire Crr or bearing drag. In fact, Tom A. probably remembers us telling him we were working on various different methods (most of which we ditched) to do this in the Win Tunnel when he visited over a year ago. Our engineers recently developed and validated a protocol and equipment to be able to isolate out the rotational aerodynamic component at a resolution of ~50 mW. Since we devoted a lot of time and resource into developing the equipment and protocol to achieve this, I think you’d understand if we’d like to keep the specifics proprietary. Hint: involves a very careful and modified application of the Chung method in the tunnel while understanding what error terms can be neglected vs. not.
Anyways, a lot of geeking out. The CLX50s are FAST and ride very responsively.
Chris
Sounds like you found another “nail” for RChung’s “hammer”…cool
(Now I’m going to be racking my brain trying to figure out how you do it…does tire Crr come into play at all?)
So, on the plots that show the combined translational drag, they show the units as m^2 of CdA as well…does that mean the “power to rotate” varies in proportion to V^3, like the translational drag? Do you think the differences between the CLX50 and 64 are just the different spoke lengths? Or, is something else at play?
Thanks for coming by and talking about these in some detail.
Hubs are a major drag component, and are hardly ever optimized for aero in the way that rims or frames are. Could you tell us more about what you’ve done with the ‘Aero Flange’ hubs?
And what explains the large low-yaw rotational drag advantage of these over the 808 NSWs?
Ok, to answer some of the questions in this thread:
Turbo 24 vs. Turbo Cotton 24: Simple explanation - when we started doing the shape design of the current family of CLX wheels (32,50,64) the Turbo Cotton didn’t quite exist yet. All our historical CFD and benchmark data was based off the Turbo 24. As you can imagine, it takes a while to fully quantify and correlate a non-exact shape like an inflated tire casing. However, the comparison testing was done much more recently (e.g. includes the recently launched Zipp 454) and in our testing, and as Tom A. pointed out, the fastest combination of translational drag, rotational drag, and Crr was with the Turbo Cotton tire. This was true for the Zipps as well. As a result, we felt it would make the most sense to publish comparison data using the fastest available total combination.
Rotational drag (“power-to-spin”): As many of you know, it is extraordinarily hard to be able to reliably/consistently measure this with adequate resolution. Then there are complications like accounting for, or isolating tire Crr or bearing drag. In fact, Tom A. probably remembers us telling him we were working on various different methods (most of which we ditched) to do this in the Win Tunnel when he visited over a year ago. Our engineers recently developed and validated a protocol and equipment to be able to isolate out the rotational aerodynamic component at a resolution of ~50 mW. Since we devoted a lot of time and resource into developing the equipment and protocol to achieve this, I think you’d understand if we’d like to keep the specifics proprietary. Hint: involves a very careful and modified application of the Chung method in the tunnel while understanding what error terms can be neglected vs. not.
Anyways, a lot of geeking out. The CLX50s are FAST and ride very responsively.
Chris
Thanks for the reply. It is always cool to get insight from the engineers that really know the how and why of what make a bike fast.
Your Chung clue is intriguing, but VE doesn’t make much sense to me when there is no elevation change.
I always wondered why windtunnel a don’t measure current draw for the motor that turns the wheel at the wind speed of the tunnel. Don’t watts = amperagexvoltage and you “know” tire Crr from roller tests. You could then measure bearing drag on a seperate fixture and sutract from the wattage consumed by the motor. This is a kinda piecemeal way to calculate rotational drag, but is there something I am totally overlooking?
Sounds like you found another “nail” for RChung’s “hammer”…cool
(Now I’m going to be racking my brain trying to figure out how you do it…does tire Crr come into play at all?)
So, on the plots that show the combined translational drag, they show the units as m^2 of CdA as well…does that mean the “power to rotate” varies in proportion to V^3, like the translational drag? Do you think the differences between the CLX50 and 64 are just the different spoke lengths? Or, is something else at play?
I’m going to have one of our super engineers (Ingmar Jungnickel) who helped develop this method jump in to explain a bit about how CdA plays out in this situation. As for your question re: differences: spoke length certainly plays a big part. But were also able to measure differences in how different rim shapes, hub flanges, and tires rotate through the air.
Thanks for coming by and talking about these in some detail.
Hubs are a major drag component, and are hardly ever optimized for aero in the way that rims or frames are. Could you tell us more about what you’ve done with the ‘Aero Flange’ hubs?
And what explains the large low-yaw rotational drag advantage of these over the 808 NSWs?
You answered your own question
I always wondered why windtunnel a don’t measure current draw for the motor that turns the wheel at the wind speed of the tunnel. Don’t watts = amperagexvoltage and you “know” tire Crr from roller tests. You could then measure bearing drag on a seperate fixture and sutract from the wattage consumed by the motor. This is a kinda piecemeal way to calculate rotational drag, but is there something I am totally overlooking?
In theory, that’s sound. In practice there is a huge error stack on a (relatively) tiny signal.
Well, they’re very nice bearings. You’ve put together a tidy system which justifies the investment.
Even small changes in that can have relatively large lateral stiffness effects. Damon Rinard found that out a long time ago when he looked at the lateral stiffness of a radially laced wheel, where the only change was a “spoke heads out” vs “spoke heads in” configuration in the lacing
I could be mistaken, but I recall Jobst writing about that well before Damon came on the scene. (RIP Jobst)