Aero data from 19 wheels by Hambini

[hambini]

To answer a few questions.

For riding position, the overall power number for each wheel will change but the order will not.

This is more Aerospace focussed but nonetheless, I will discuss them here. There are two methods for dealing with drag that most Aero Engineers work to:

1. Active drag management (separation reduction)
With this scenario, you prevent separation from occurring. Deeper wheels will inherently have better active drag management. Another avoidance are sharp "cliff edges"

2. Reactive drag management (post separation control)
With this scenario, the separation has occurred and there needs to be geometry to control the separation or reattach the flow.

My opinions:

You really want Active drag management to be as good as you can get it because it makes the reactive easier (less to deal with). In the case of the FLO and the Hunt wheels, the active drag management is very poor which is why their performance is bad.

The bontrager and Reynolds wheels have very good Active drag management.

The 105% rule is a crude form of Active drag management. Looking at the charts for tyre widths, it's obvious the Shimano wheel cannot handle the turbulence left behind by the ballooned 25mm tyre - so it gets hammered for both it's active and reactive drag management.

[Ed72]

But now that I ride an M5 recumbent my normal speeds are 40-60 km/h on rolling terrain and I feel effects that I never felt on an upright. Trees, building, passing cars, and even passing other riders can be felt in my hands and there is just more movement in the bike. Maybe Hambini is onto something here.

Recumbent? That would completely change the equation and how much the wheels contribute to the overall drag. You're right, you won't see a 30W difference in your case.

I doubt I'll see it on my Cervello S3. We'll see. The difference could actually be larger on the recumbent. What is clear is any reference to A2 type drag data is useless in this discussion.

[Ed72]

It is almost impossible for me to relate to the reported data because I only need ca. 310 watts for 50 km/h and around 90 watts for 30 km/h (FLO 60 wheels)

You must be a bit of an exception because 30km/h and 90watts is IMO quite low.

Is it possible to transform the wheels chart into time saved over a 40km loop?

It would depend on the speed you were going at or is there a generic system that people use?

Hambini

Yes. As I had stated, my overall CdA is very, very low. Incredibly low. 29.1 mph on 247 watts is also very low. 612 watts is also an exception.
....400-420w would be more typical on an aero TT bike with good skinsuit (for me anyway).....others would need much less power because they have worked harder on position and kit than I have with my upright TT bike (CdA about 0.265.....not great).

Got the new M5 2-Spoke front wheel today from Holland, unfortunately terrible weather from the hurricane. It is probably the fastest wheel in existence. The front wheel leads on both my upright and recumbent. They see the same or similar flow. The reattachment 5-10 inches afterwards might be different but the profile into the wind is identical, so, why wouldn't I expect similar savings. Same wheel size and the front wheel is the leading edge. My aerodynamic fork is probably more helpful than the Enve or Cervelo forks on my uprights.

I just want to thank you Hambini. I had actually been thinking quite a bit about the performance of my wheels. Your video and explanations have got me to think about a solution to my problem and hopefully I'm on the right track.

[RyanH]

Yes. As I had stated, my overall CdA is very, very low. Incredibly low. 29.1 mph on 247 watts is also very low. 612 watts is also an exception.
....400-420w would be more typical on an aero TT bike with good skinsuit (for me anyway).....others would need much less power because they have worked harder on position and kit than I have with my upright TT bike (CdA about 0.265.....not great).

622 watts calcs out to an implied CdA of 0.34ish, which is appropriate for the stated position they conducted the testing at: relaxed position on the hoods. I wrote out the different watts required at various CdA's above but I'll repost them here:

0.34@50kph: 622w
0.30@50kph: 5558w
0.28@50kph: 526w
0.26@50kph: 494w

Also, those extra 3.2 kph require almost 84 extra watts over 29.1mph at a CdA of 0.26.

[WMW]

Yeah, that's why I was deliberate about saying "active" and not "joined". My point was you're not totally unbiased here is all - and the timing speaks to it. Don't mean to throw too much shade, just suggesting there's more behind the energy from all sides - and it's an unfortunate distraction.

I'm pretty certain that Tom and a lot of other engineers are active in this thread for the same reason I am. This is some real science! Or it could/might be,,,,

The reason why the tone seemed to get "hostile" from the perspective of the peanut gallerie is because questions weren't getting answered, and there was no supporting data. And then Hambini resorted to ad hominem and it got more absurd. I figure when someone does that they have something to hide, and could even be a total fraud. I no longer believe that, but at the time I'd never heard of him before, so anything was possible.

As I've stated a few times in this thread, I have always had a very positive attitude regarding the concept of testing in unsteady flow. IMO the tunnel environment is not very "real" and though there is some supporting evidence that bridges tunnel data and field testing data, it isn't very convincing on a detailed/precise level. So I love the idea of what Hambini seems to be doing.

But, if someone is presenting data derived from a "revolutionary new test protocol" then everyone with a scientific bent is going to want to know the details. Exactly what it is, and why it's better. Bold claims need to be supported with evidence and explanation. There is nothing weird about that. People who can't grasp the technical aspects though, may not understand this.

I hear ya, but is there a "context" that would be satisfactory - other than Hambini (and co) running a more "standard" test with a few benchmark wheels to serve as a reference point?

No need to run a "standard" test, just explain the one he did. I get that it could take time to put that together, and there aren't many engineers who like to do write-ups.

[WMW]

Deeper wheels will inherently have better active drag management. Another avoidance are sharp "cliff edges".... The bontrager and Reynolds wheels have very good Active drag management.

In the Trek/Bontrager D3 whitepaper they mentioned that flow separation occured much sooner (lower yaw) from the leading edge of the rim (tire front) than the trailing (tire rear). The reason they gave is that the round leading edge is aerodynamically poor. That makes sense (at least at sub-critical Re#). It seems like a decent solution to that would be to make the rim wider than the tire at the brake track and shape it properly in that region. I know the Reynolds rims are designed this way. Is there something else you see that lends to the better performance of the Bontrager and Reynolds rims?

[aeroisnteverything]

In the end a round shape and fulfilling the 105% rule seems to help after all: the Bontrager are 27mm wide outside, 19.5mm inside, with a 23mm tire that's quite a good match.

And Hambini added the new Light Bicycle rim with 30mm outside, 23mm inside - this is also quite good, especially at high speed:

Does it though? Bora 50s perform virtually identical to the 56mm deep LightBicycle and also same as the Bontys. Mavics at 45mm are only marginally worse, and still within the margin of error. How do we explain this? Is this is a fluke?

@hambini: would be good to hear more on the why we are seeing what we are seeing. You say that FLOs have bad active separation maangement and Bontrager wheels have "good" active mgmt. But why? What's the difference? Is it the shape, the width, tyre/rim interface? And how do we explain the performance of Campy wheels despite their being 'only' 24.2mm wide - narrower than the tyres, since Continental 23mm tyres measure out 25mm when on the rim? Same question re Mavics. And why do Enve 3.4s perform so badly (even taking depth into account), despite their deeper brethren being among the lead pack.

Secondly- and I asked this before, but no answer - what is the statistical significance of these results? How many tests do you run on each wheel? What is the standard deviation of results? Otherwise, when looking at how tighly clustered most of the wheels are, I fear folks might mistake random variation for an actual difference in performance.

Lastly: Roval CLX 50s, CLX 64s, and Mavics Cosmic Carbon SL-Cs (UST or otherwise) remain fairly prominent omissions on this test. Would be great to get these tested.

[Calnago]

Speaking of omissions, I’d like to get a plausible explanation why, in the 50mm profile range, Lightweights are without doubt the worst behaving wheel in gusty conditions I have ever ridden. Is it the increased surface area due to the spokes (each spoke is about 3x wider than say, the spoke in a Bora wheel for example). Is it the high thin flanges on either side of the hub shell? Is it the flat sides and narrow width of the rims. They don’t seem “that” much different than the old Boras in that regard but the difference in behavior is striking. Is it simply that the wheels are so incredibly stiff that any wind gust affects them immediately and harshly compared to less stiff wheels? I’m not sure, but I work on bikes with Lightweights (about 8 different sets in all) and every test ride end up with the same bad impression. Coffee shop wheels as far as I’m concerned. At least compared to the other alternatives.

[morganb]

Thank you for adding the light bikes rims! I've been thinking about a pair for a while but haven't seen any aero data on them before. This makes my decision a lot easier.

[WMW]

Does it though? Bora 50s perform virtually identical to the 56mm deep LightBicycle and also same as the Bontys. Mavics at 45mm are only marginally worse, and still within the margin of error. How do we explain this? Is this is a fluke?

I think it would be good to remember that he is estimating +-2.5% error. A wheel in the middle of the range will have overlapping error bars with nearly every other wheel in the test. For this test the wheels would be expected to only contribute ~5-10% of the drag of the system, and since they are the only variable being tested, that +-2.5% error is more like a +- 25 to 50% error on the drag of the wheels themselves.

With a rider on board this is to be expected, as they represent a large unwanted variable. Hambini has already said he is looking into using a dummy for future testing so that the experimental error can be reduced.

[bilwit]

as cringey and teeth-pulling as it was to have all this mudslinging, I have to say that looking at the updated blog post now, it was for the better!

also like how he snuck in another dig directly at the biggest online perpetrators

The FLO cycling and Hunt wheels performed badly, they appear to have been designed by individuals with a limited understanding of aerodynamics of rotating objects. As such they generated unnecessary separation and could not deal with the separated airflow

[Hexsense]

Does it though? Bora 50s perform virtually identical to the 56mm deep LightBicycle and also same as the Bontys. Mavics at 45mm are only marginally worse, and still within the margin of error. How do we explain this? Is this is a fluke?

Tested with fixed 23c tire, All wheels in the test while some of them do not pass 105% theory, does not have a ton of tire overhang. Actually, 23mm internal make 23c bolloon out a lot (~28mm?) on LB 23mm int/30mm ext, while 17mm internal width of Bora inflate 23c to only 24-25mm.
Story change when fitting 25c on, as shown in a few pages ago, Shimano wheels which is relatively narrow got more hit than wider wheels.

105% is more like a guide line ensuring to always have perfect aerodynamic of that rim. Observed from a Zipp U-shape wheel where brake track is narrower than middle of the rim.
From observation, it seems like brake track width is more important than width of center of the rim. Therefore V-like shape like Bora can afford a tire wider than 105% guide line without getting much impact.

[mpulsiv]

Delta is clearly visible where 25mm Continental GP4000S (balloon to 27mm) is ~4.6mm wider than Shimano C60 (22.4 wide) wheel. In contrast, Enve 7.8 wheel (29mm wide) is ~2mm wider than 25mm Continental GP4000S (balloon to 27mm) tire. To yield every watt, invest into wide wheels. Unfortunately, most wide wheels (e.g. wider than 27mm) are disc.

[silvalis]

Delta is clearly visible where 25mm Continental GP4000S (balloon to 27mm) is ~4.6mm wider than Shimano C60 (22.4 wide) wheel. In contrast, Enve 7.8 wheel (29mm wide) is ~2mm wider than 25mm Continental GP4000S (balloon to 27mm) tire.

You'd expect the tyre to balloon wider on the enve rim than the C60, not both end up at 27mm.

[mpulsiv]

Delta is clearly visible where 25mm Continental GP4000S (balloon to 27mm) is ~4.6mm wider than Shimano C60 (22.4 wide) wheel. In contrast, Enve 7.8 wheel (29mm wide) is ~2mm wider than 25mm Continental GP4000S (balloon to 27mm) tire.

You'd expect the tyre to balloon wider on the enve rim than the C60, not both end up at 27mm.

Internal width:
Shimano C60 = 17mm
Enve 7.8 = 19mm
When it comes to progressive tire width measured with a caliper, we are talking about ~0.4mm difference. Now, if we step up to 21mm or crazy wide 25mm on Enve SES 4.5 AR, the tire will balloon to max achievable width. We may see up to 28mm.
To put this into perspective, 25mm Continental GP4000S will measure ~27mm on Shimano C60 and ~27.4mm or less on Enve 7.8 wheel.
My 25mm Continental GP4000S measure ~26mm on an ancient 10-speed wheel, which is 14mm internal and 17mm external width.

By end of the day, my vote goes to Reynold ATR 700 but it's disc only You could achieve maximum aero gain with 28mm tire or wider https://reynoldscycling.com/collections ... ts/atr-700