Dude, that's ~17g of drag (max disparity) at common yaw angles (<10d) (from your data).

At 10d on a P5-6 (no rider) A2 could not tease out any difference between tires of completely different width, tread, and model...

The tire wasn't modeled at 95psi. I'd have to go through notes to find the exact number, but no 95psi was a surprise to find.

With respect to making a blanket statement, that wasn't the intention. I'm simply saying here's what "can" happen in a tunnel. So if your testing protocol has more variability than it should, it "could" present a problem with the results. Other rims would likely behave differently than ours, and the variance could be better or worse.

---

Chris Thornham
Co-Founder And Previous Owner Of FLO Cycling

I understand that. What I'm saying is we've seen variance "up to" 93 grams at different tire pressures. Other rims could show this type of variance, or more, or less, and any angle. Until you test every single option, you don't know. My point is simply, if you don't control your variables in your test to the best of your ability, you are opening yourself for some potentially large errors.

---

Chris Thornham
Co-Founder And Previous Owner Of FLO Cycling

Is there any reason why the Kinlin a basic deep V tested better than all the toroidal rims? Is this simply a matter of lower frontal area winning out over boundary layer control?

The XR31T is not a "basic deep V."

Actually, it pretty much is (BTW, I built one up on a PT hub last year for my TT/field test wheel):


I have a feeling that it's aero excellence is somewhat "accidental"...sorta like the Conti GP4KS ;-)

---

http://bikeblather.blogspot.com/

Fair point, but it's definitely not what I think of as a traditional "deep v."

Why? Because of the width? That's about the only thing non-traditional about it...

---

http://bikeblather.blogspot.com/

The "v-ness" of the "V"

kinlin definitely a more ovallized V. closer to a U than a V in my opinion. and i'm looking at a set right now.

Well, aside from the width, the X-section doesn't look much different from a Velocity Deep-V, which is about as traditional of a "traditional deep V" you can get ;-)



I familiar with the Kinlin, BTW. I built one onto a PT hub last year and it's my current TT/field test wheel.

---

http://bikeblather.blogspot.com/

Talking of controlling variables, what sort of size error bars do people typically get from repeat tunnel tests?

Kinlin's shape looks somewhat similar to Reynolds Assault SLG.



I wounder if the brake track protrusion in Kinlins acts as sort of a "lip" as in SLG's:

https://www.reynoldscycling.com/.../swirl-lip-generator

Can you define exactly what you mean by error bars?

---

Chris Thornham
Co-Founder And Previous Owner Of FLO Cycling

Dave (or anyone who has insight),

In light of this in your blog entry:

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.


Since the difference between the best and worst isn't huge, based on what you know and the time you have spent in the tunnel, do you think there is any real drag reduction using say a 60mm rim compared to a 30mm rim when using large tires? When I say large tires, I'm talking 35-40mm? I'm thinking if the drag values are that close with 23s, then throwing a huge tire on could 'swallow up' all the aero savings to the point where choosing the lighter rim makes sense. I'm coming at his from a gravel racing perspective...thoughts?

---

_______________________________________________

Every time something is measured there's an error associated with that measurement. I've seen plenty of graphs from wind-tunnel tests showing product X producing say 50 grams less drag force than product Y, but no indication of the repeatably of those readings. If the same test is performed in the following hour, day or week how much variability is there in that reading?

https://en.wikipedia.org/wiki/Error_bar

Now as to what type of error bar is appropriate, I've no idea!, but I can't recollect seeing wind tunnel graphs giving any indication of what the potential variability in repeat measurements might be.

Great question.

Simplest answer? You never know until you test.

Slightly longer answer? I think aero is out the windy when you're got even a cx tire on there much less a textured 40mm tire. Plus the correct read, the correct pressure, rolling resistance (those 3 of course linked but each is discrete), comfort, blah blah blah are probably more important.

Slightly snarky answer? These are races where people (people who might dust the entire field, mind you) put bottle cages on their fork blades! I don't think that's very aero.

As to error bars for wind tunnel tests, the A2 results are unbelievably repeatable. Each number give has 2 runs behind it, and anyone who's gotten a test from there could attest that the 2 runs are always near carbon copies of one another.

This test has been shared directly with an ST poster who can speak to the repeatability of the runs in this test if he chooses to.

Thanks for your opinion Dave. Based on the tests I've seen and how sensitive even having your tire AS wide, opposed to the recommended advice of ensuring your rim is 105% the width of your tire, so having a tire that is almost 2x the width of the rim can't be good!

I suppose there could be some aero benefits a higher yaw angles using very large tires on aero rims, but I guess for me, I'm thinking this is likely a situation where weight of shallower rims is better than the heavier deeper rims overall. I'm building up a gravel race bike right now, so that's why this is on my mind!

---

_______________________________________________

Does this test invalidate the 105% rule? It was said before that the rim should be 105% larger than the tire for the best aero. Is that only at higher yaw angles? 23mm GP4000s tires are ~27mm on my previous HED Jet + as well as FLO 30's. It would be similar on the XR31T and ALX wheels. Carbon wheels are typically wider(303) and can conform to the 105% rule.

I assume (or hope) that you're saying this with a little 'pink' in that text.

Can you define how a point on your graph is derived? Did you only take one measurement at one position and move on to the next. Did you take a time lapse measurement at each position? Did you do multiple runs at each YAW and take the average? ...what is meant by 'error bars' kinda depends on how the data is derived.

Typical standard deviation at the A2 Wind Tunnel is about 0.5%, which is what we observed during our tests. To give you an idea of what that means, I'll speak in terms of grams.

A2 frequently does "Air Off" testing, which means with zero load (moving air) in the tunnel, what is the drag reading for a wheel. I just spoke with Geoff from A2 on the phone, who just finished an Air Off test on a wheel this morning. His readings were predominately 1-2 grams with a couple readings as high as 5 grams.

"Air Off" testing is good, but what really matters in a tunnel is how close are the measurements when the air is on. What makes repeatable "Air On" readings so challenging to achieve is that turbulent air is not repeatable. Even with varying turbulent air flow, the majority of our readings were within 1-5 grams of each other.

I hope that helps,

---

Chris Thornham
Co-Founder And Previous Owner Of FLO Cycling

---

Chris Thornham
Co-Founder And Previous Owner Of FLO Cycling

Thanks Chris for replying. That's very helpful for getting a feel to variability.
Cheers!