November in the wind tunnel: "semi-aero"


Tunnel testing alloys was actually more exciting for us than it should have been.  While a lot of people choose to buy builds from us with the rims on test, this essentially amounts to a "public good" test, simply because these rims are so widely available.  The excitement mostly came from the fact that the internet was seemingly ready to have an aneurism there was so much demand for us to do this (which is a funny concept when you think about it - no one else spends the money to do this for people, not even the rim makers), and because so many "eyeball wind tunnel" experts had definitively pronounced the results long ago. 

We have to once again stress that directly comparing this test on this day with that test on that day is a fool's errand.  The way that we have designed our testing protocol, we will always be able to make significant back-references, but direct comparisons between different tests are worthless. Relative results of one wheel vesus another from test to test are valuable, which is why we have tied ourselves to the benchmark of the 404 wheel we used in our original test.  We will always keep that wheel as the reference standard. 

We tested these with a 23mm Continental 4000s II tire - one tire used in all tests, as it must be. That is perhaps the most popular tire used with the rims we tested, and is a standard for the wind tunnel.  Significantly, you must take note of the tire width as it is on each wheel when looking at these results.  The same tire will inflate differently on different rims: the wider the interior dimension between the brake tracks, the wider the tire will set up.  Among the rims in this part of the test, the differences are fairly small.  As we get into other rims, this difference will become more significant. 

 The biggest lesson out of this test is how much closer the Kinlin and Pacenti rims come to the performance of carbons than what the A23 did.  The overall delta between worst and first in this test was much smaller than it was the last time.  Tony Martin's not going to read this and think "you know, maybe I ought to use some nice mid-depth alloys in my next time trial," but the gap is much smaller.  Again, we are only able to make this statement because the A23 was tested against our reference wheel using the same tire (although we used a different type of tire last time), and the same reference wheel was used to measure the gap this time.  There is the noise of "maybe that's just a particularly bad tire for the A23" in there, but that's at most a very small part of the story.  Using our back-check method, the Rail 34 is faster than an RFSC38, while the alloys in this test are slightly slower.  

The Kinlin was the faster alloy.  As we will explore later, the tire's increased width on the Pacenti may account for all of the difference between the two.  It is impossible to normalize one tire to the same width on two different rims, and you wouldn't want to anyway.  What you could do is test a narrower tire on the wider rim and interpolate from there, but with these rims we were more interested in seeing if there were any gross differences between the two, and how they stacked up compared to the carbons on test.  We will explore the effect of inflated tire width in more detail in subsequent posts. 


Older Post Newer Post

  • pyf on

    Dave,Having been in the tunnel myself I'm completely aware of the one day of testing = one set of results that you can't replicate 100% the next day. Also aware of one tire is not the same as the very next tire produced.Having said that… the differences we are talking about are not in the order of making a good wheel one day a bad wheel the next day, etc. or all of this would be pointless ! Imagine if the rail 34 was good one day and you come back again the next day in the tunnel with just a new tire on a new day and all of a sudden it's much better or absolutely worse. I guess we can agree that in the order of +/-5% for grams of drag (it's probably less) or a +/-1° stalling point results are going to follow the same trend.Please have a look at this folder I created here just for the sake of talking with you ;-) : at the Zipp 303 firecrest carbon clincher tested by Zipp and tested by Boyd, both at A2 wind tunnel and both with GP4000S 700×23 (flashy green curve on Zipp graph), you can see that even though they were tested at a different date and with a GP4000S from a different production batch that numbers are "close" enough.So when I say it would be great if you could provide Y axis in grams of drag, I think as long as you use the same balance as Boyd, Zipp and Flo do (Hed also tests at A2 but uses a different balance I think :-( ) then your results won't be apple to apple since on a different day and with a GP4000S generation 2 tire but as long as we don't look at decimal of grams of drag then general trends can still be interesting…Hope you understand a bit better what I was asking, again I don't expect to look at decimals but general trends are probably possible to look at…Ps : I've just added 202 firecrest with GP4000S 700×23 and 18 spokes 404 firecrest VS 18 spokes 404 firestrike with GP4000S 2 700×23 in the folder… not completely related to our talk but maybe you want to have a look (again for general trends) at Rail 34 VS 202 firecrest and also to 404 firestrike results which is somehow interesting —> Zipp managed to decrease drag at lower yaw with firestrike even though internal width is 1mm wider thus probably making the same tire wider than on 404 firecrest !

  • Dustin @ Southern Wheelworks on

    Great info guys! A question that's always bothered me about these type test…how is one wheels slower than another if they're both going 30mph? I know the answer is that one requires less power to do so, aka at a given power one wheel is faster than the other. But when you say "Wheel A is 2secs faster than Wheel B in a 40km TT at 30mph" my immediate thought is that 30mph is 30mph is 30mph. Doesn't matter if you're in a Ferrari, on a bicycle, or riding backwards on a unicycle – the speeds is the same. I know it's a convient and easy way to compare the data since no one can really wrap their head around what saving X watts would actually mean.Also, people need to keep in mind just how significant these differences are. So the Kinlin is 2.2secs faster than the Pacenti in a 40k TT at 30mph. Well a 40k TT at 30mph takes 49mins and 36seconds. Saving 2.2 seconds is a whooping 0.074% reduction in time – less than a tenth of a percent. I'd argue that difference isn't significant, and other factors are more important when comparing the two rims. Such as cost, tubeless capability (which you may or may not care about), brake track smoothness, overall build quality, finish, etc.Looking forward to more info!

  • Dave on

    Velocity A23. Slowest wheel in our last test.

  • JP on

    What is the A23?

  • pyf on

    Dave,Here is some very nice information about A2 tunnel repeatability ! this is helpful and give you some confidence in some of the results you get from one tunnel trip to another !

Leave a comment

Please note, comments must be approved before they are published