Hambini,

Thanks for sharing your data and thoughts. Any thoughts on the Zipp 454 "hyper foil" design with regards to both aerodynamics and stability? Specialized tested the 454 and seemed to indicate it performed similarly to 50mm "toroidal" rim profile. Do you think there's any validity to Zipps' claims with regards to stability?

Sadly, some people here might take you seriously. Myself included. Perhaps that's why I'm single...

Very interesting !!

What do you think about the assumption that a U-shaped rim, having a more symetrical shape between the leading and trailing edges, creates a less important differential of "lift" (than a V-shaped rim) between the front and the rear of the wheel, thus more stability with less rotation around the axle of the fork ... don't know if my question is clear ?

Regards, E

For what it's worth, if you go back and look at the white paper Trek did on their D3 rim design (prior generation) that was exactly their theory. They seem to have since reversed course with their new rim design. Knight supposedly tries to balance lift between the rear of the leading portion of the rim (ahead of the fork) and the rear of the trailing portion of the rim (behind the fork) to minimize steering torque in crosswinds.

Understood, I guess what I was trying to get is your thoughts on steering torque caused by an imbalance of the leading and trailing edges of the wheel.
1) Does Cl vs alpha look similar when the profile is travelling backwards?
2) Does Flow separation occur similarly when the profile is travelling backwards?

The other question would be:
Given a blank sheet, and a speed of ~30-40 kmh, what sort of wheel would you design (or look for feature-wise)? Or is it too hard to just guess without doing the testing?

Cheers!

This post is getting slightly derailed by tire pressure discussion rather than aero differences. Whilst it is true tire pressure varies with ambient temp these differences are fairly easy to calculate and have been done in various spreadsheets such as this one (https://goo.gl/99PG6Z). For example if you had a 25mm and inflated inside at 20degC you would inflate at 87psi to get 90 psi outside at 30degC. However ambient temp causes pressure differences more than volume change because volume is constrained by the casing. OK now I am getting distracted! The point is tire pressure has a very small effect on tire size in MTB, Cx and road. 25psi might give a 1-2mm change in tire width or height (see http://forums.mtbr.com/...im-width-756818.html). In turn this also a tiny effect on aerodynamics measured as few seconds over 40km or 90km. Of course tire pressure will have a big effect on CRR but thats not the topic. For any wheel manufacturer to say the results are invalidated by choice of pressure in the tire is really over reaching (unless they are invalidating their own results, which is rare indeed!). This will be the least of all possible variables. There is probably more difference in the aerodynamics of two identical tire out of the box due to manufacturing & installation differences than the same tyre in two tests at 85psi and 90psi.

Now on the topic of these results, what is interesting to me is the slight re-ordering between 30kph and 50kph which implies some wheels/tire combos cope better with deep vs shallow yaws. What I like about this test is that they appear to have tried to make the wind tunnel protocol simulate outdoor wind variation.

@MattyK

It's not the amount of torque that makes the wheel unstable. It's the hap hazardness.

If you had constant cross wind then you would tilt your body to compensate. What generally happens is the wind pulses and you find yourself with either too much or too little body tilt and the instability arises.

However to answer your questions

1. Cl to alpha would cut off at a smaller angle backwards and the Cl term would be lower. Basically the line would be a shallower slope. This is summarised by lifting line theory, the maths is a bit heavy but it's here https://en.wikipedia.org/.../Lifting-line_theory
2. Flow separation will still occur around 12 degrees for reasonable geometry.

Hope that helps
hambini

---

HELLO HAMBINI FANS!!!

Good stuff in the D3 whitepaper: https://www.google.com/...GxtN4hrYidZmZ0jAg2oS

"It’s easy to design a rim cross section to have good rim-leading performance. In fact, the rim-leading cross
section mimics the shape of the KVF tube shape found on the Trek Speed Concept [1], where the tire acts
as the truncated airfoil tail. Like the KVF airfoils, the rim-leading side can be designed to mitigate flow
separation—the key challenge in bicycle aerodynamics. As a result, the rim-leading side of the wheel can
achieve very low or even negative drag until finally stalling at a relatively high yaw angle, as shown in the
previous figure.
It’s a different story on the tire-leading side of the wheel, where the tire leads the rim into the wind—an
obvious but critical observation. The tire’s circular cross section is a very poor start of an airfoil. As we know
from the Speed Concept white paper, circles have early flow separation and, as a result, very high drag. This
flow separation off the tire is the largest source of aerodynamic drag for the entire wheel. As such, it is the
largest opportunity for potential drag savings, and therefore the primary focus of the Aeolus D3 design."

Leads me to believe that small tires are best....

Also, Hambini has added the Reynolds wheels... and they are the best. Yes, this rim shape is the new winner:

I just notice the link in the OP has been updated with a lot of additional information. I'll update the title.

Interestingly, the Reynolds, Enve 7.8, and the new Bontrager rims all share a very similar rim shape. I'll add that Venn Cycling also came up with a very similar rim shape about a year ago http://www.venn-cycling.com/.../ctl-rims#venn507tcc

Sadly, they're not offered in the fancy-pants filament wound construction :( yet.

It's funny how we went from pointy to oval and back to pointy. Only now the pointy rims are much wider. Wider than the tire. I guess back in the day you could have achieved the same thing with a very narrow tire, which the smart people did. With less drag than modern rims also I'd wager if you are matching Cd but making the A smaller.

It looks like Reynolds has actually made their rims less extreme (pointiness) than they were a couple years ago. They used to be concave at the very end of the tail.

I found the paper on the D3 very interesting regarding the performance of the trailing vs leading edge of the rim. I figured the tire leading part would perform better (subject to good tire choice), but they say the opposite. I wouldn't expect round to be bad for the leading edge, you just need the proper shape and width of rim behind it.

Trek has been at the bleeding edge of bike aero for a while IMO. I recall that one of Carl's departing remarks was that their in-house CFD had gotten so refined that they almost didn't have to go to the tunnel to test as 99% of the time their CFD predictions matched what they saw in the tunnel. The only area where they still "struggled" was with helmet design.

A personal suspicion of mine is that the Mavic CXR 60 with a 23mm Corsa Speed and Mavic's "tire" strip fitted is a very fast combination.

I find it interesting to compare the Enve 7.8 to the XXX 6 and the Venn 507. All very similar shapes. It would appear that the longer the aerofoil the less of a need there is to truncate it into a "toroid" shape.

Thank you for stating (and backing up) what I have been intuitively believing all along. thanks

Have they? Both rruff and I have been attempting on the weight weenies thread to get more detail about the results of their data gathering step, and in particular the apparent wind probability distribution, and how that squares (or doesn't) with the wind tunnel test protocol. No luck so far...

I've asked for some sort of data or calculations that show this particular protocol is actually better at predicting "real world" performance than simple CdA vs. yaw angle data taken in steady-state wind tunnel conditions...which actually work quite well (e.g. modeling like BestBikeSplit). Again, no luck...

---

http://bikeblather.blogspot.com/

Doc...care to comment on the variability of the wind conditions during the data runs for this?:


I recall you saying conditions were quite gusty, no?

---

http://bikeblather.blogspot.com/

It does kill me when people are suspicious when a company does well in their own testing. Because they should do well in their own tests. Since they should be making a test that evaluates what makes the best wheel, they should be designing their wheel to do well in that test. It is not automatically dishonest, it just the outcome of a well designed wheel. The thing that all these companies should do is to provide detail of their protocol and why they chose that protocol, so that consumers can evaluate if the company optimized for the right conditions.

Flo obviously does a very good job in detailing their protocol and why they do it. Bontrager is similar.

I don't know if an industry standard protocol is the best idea, because what is the best protocol? Will it constantly be revised when new tires come to market?

Quite gusty indeed.

Despite this, the wind tunnel (steady flow) data were strongly predictive of 'real world' power requirements.

That is presumably because 1) parts of the bike+rider system exhibiting significant yaw sensitivity we're only a small part of the whole, and/or 2) the time scale over which yaw angle may have flucuated was too small to matter (such that knowing just the overall average as we did was sufficient).

Thanks for the trip down memory lane... that study is >20 y old now!

Well our protocol was tire independent. The protocol was designed to ensure you had good measurements but gave you the freedom to use any combination of wheels, tires, and pressure that you wanted.

---

Chris Thornham
Co-Founder And Previous Owner Of FLO Cycling

Yeah...I tried to show "hambini" that as an example of how steady-state data appears to predict power requirements under somewhat variable conditions (i.e. wind velocity with SD ~ 0.4 m/s according to the paper...so, +/- 3-sigma of ~4.3km/h). I asked for any evidence that his new protocol works any better at predicting "real world" performance than steady-state data...with no answer to that question :-/

---

http://bikeblather.blogspot.com/

Up to 5 years ago I had read every SAE paper there is on aerodynamics when I was involved in road vehicle efficiency. So I figured the auto industry would have looked into transient flow analysis. I don't have access to everything that is published anymore, but a quick google search landed the following:

https://www.google.ca/...8LCLxOo1mnJTZ4_B_ctv

I'm sure more can be found on the subject and I feel there is something non negligible at play here that applies to the field of cycling as well.

I don't believe transient flow evaluation is related to yaw angle distribution specifically, but really to turbulent flow, which is something wind tunnels typically try to avoid by smoothing out the flow as much as they can.

If we talk about yaw angles distribution I really think they should generally be low, but again I feel that is an unrelated concept. Wind gradient diminishes with height following a power law and factoring an hellmann exponent for the environment you're cycling through, so the closer to the ground you get the lower the wind speed is with varying degree, which have a direct impact on effective yaw angles measured at the wheel. Technically there are differences between the top and bottom of the wheel, exacerbated by the rotation of the wheel itself, but I do not know whether these are significant or not.

Our wind tunnel vs. field test data indicate otherwise.

Yeah...as I mentioned above, I've tried to point that out to Hambini on the WeightWeenies thread on this...and when either I or rruff try to ask questions about the data gathering and protocol, we basically get shouted down and accused of being shills.

https://weightweenies.starbike.com/...=153138&start=60

---

http://bikeblather.blogspot.com/

What makes the most sense, laminar or turbulent flow? Good enough doesn't mean something else isn't better. You just don't know how to explain divergences from the model that might very well be explained by something you haven't tested for. I don't pretend I know, I just don't dismiss it.

It certainly can't be ruled out that variability in the field test data that is not explained by the wind tunnel data is due, at least in part, to turbulent vs. highly laminar flow.

At a maximum, though, this means that the effect amounts to only 3%.

Having built, validated, and performed numerous experiments using my own wind tunnel, I'm betting that the true effect is actually much less than this.

Well that was a bizarre read. Entertaining in a train wrecky sorta way. The discussion seems rational enough until it suddenly took a left turn at Albuquerque. WJW.

Do you think a 40 second blast in your study 20 years ago is a good validator against a wind tunnel test?

---

HELLO HAMBINI FANS!!!