I meant exactly as you phrased it in the words I quoted (sorry, inside joke: I've been beating the "you can't assume Cd is a constant" drum for years and years).

I did wonder about that but I sort of figured that it had something to do with cults and brainwashing which , coincidentally, is why I never wear headphones when I use this site!

Please excuse me as I step away from the computer screen. I don't want the same lightning that is going to strike you Josh for being inconsistent in your rounding methods (why oh why did you have to round the ONE result up when the other 9 were rounded down?). You KNEW that wasn't going to go over well on this forum.

Oh the humanity! When is this insane attempt to be honest and forthright going to end? [sarcasm off]

:-)

---

Craig Preston - President / Preston Presentations
Saving the world with more professional, powerful, and persuasive presentations - one audience at a time.

well aside from the Obree positions, the range is from 0.60-0.75 or pretty close to what you 1st posted.

Re the Kelmet data, that sounds like a function of how they measured FA -- maybe camera focal length and where they chose to place the reference area in the frame. IIRC, some folks have an interesting POV on whether or not it should be 'virtually attached' to the rider along the focal axis. I mean when considering the impact on FA of sliding back and forth on the saddle :-) :-)

It's been a while ...

Right, but 0.75 for a standard drop bar position? That would imply that the only benefit of aero bars is a reduction in A, which really doesn't make any sense (since you can reduce Cd in the aero position, at least at 0 deg of yaw, by increasing your aspect ratio).

okay - just read a bit. BSA were estimated using the eq. of Dubois et al and FA was estimated at 18% of that for most cases. Moser and Indurain's data were measured at 0.70 and 0.65 respectively.

So many estimates and just a few measurements in that table 2 of Padilla et al ....

say 0.7 and be done with it ;-)

it's always good to have more data- and the more 'real world' it is the better and I even own a bunch of zipp (and HED) stuff. When I see your comment- above- "So the argument about the 808-1080 .....with the 16 seconds being spot on for a front wheel" On your chart- you say 808 ft/1080 rr (saves 80 seconds) vs 1080ft/1080rr (saves 90 seconds). 90 - 80 is a savings of 10 seconds, not the 16 you mention above or on the website talking about the wheel. Am I reading something wrong here? That being said- it's nice to see the trends, and even nicer to see the real world data published. Now I'm sure Zipp probably tested other wheels- b/c you stack rank so many other wheels. Where would some of your competitors stack rank on that chart? Like a HED 3? As you mention- on the wheel area describing the 1080 wheel- which must have included a test of the Trispoke, "This all new ground-breaking torodial rim shape is designed to provide maximum aerodynamic advantage and has proven to be 29 seconds faster than a trispoke over 40 k."

So am I correct to assume that the 29 seconds- is for a wheel only test? And is that one wheel or a pair of wheels? And what were the numbers (assuming you tested the H3) in the Cancellera test?

Also- there's independent testing (not funded by zipp or HED) that said the H3 is more aero than the 808 and another company- at a "brain symposium" recently said the Hed 90 was more aero than the 808 and you made a comment on that ST thread- that the Stinger 90 is sometimes more aero, sometimes less aero... The H3, according to HED's site, is even more aero than the Stinger/JET 90- so... that should make the H3 more aero than the 808- but for this discussion- let's assume they are equal (give Zipp the benefit of the doubt). I just don't get how you guys can advertise the 1080... "as 29 seconds faster than a trispoke over 40 k." When the 1080 is only 8 to 10 seconds faster than (an equal or maybe slower other wheel- the 808 vs. trispoke) in your real world tests???

Put a "~" in front of that value and I'd be on board with it. ;-)

You are right that the chart has the 1080 at 10 sec faster than the 808 for front wheel, not 14...I'm not sure where I got that number. The thing you are not looking at, however is that this piece of data is based on a single yaw angle, 10 degrees. This sort of athlete testing is not financially feasible, nor is it physically feasible for the athlete doing yaw sweeps, so we work off of a single yaw angle...which coincidentally is the same angle used by almost every frame company and bike fitting camp to test frames and athletes, so this is a very standard test.

As for the 808 vs trispoke, I have not seen any independant data showing the 3 spoke to beat the 808 at any likely wind angle, in fact, the tests we have seen show the 404 slightly beating out the trispoke at wind angles less than 18 and the 808 handily beating it at angles less than 20. Tour magazine posted it's testing on these wheels showing the 808 and the Ritchey Carbon (non-dimpled 404 rim) beat the 3 spoke requiring 18.1 and 19.3 respective watts compared to 19.7watts at 50kph...this wattage savings was based on a weighted average of likely wind angles from 0-30 with the primary weighting placed on the 0 degree condition so assuming more than 80% of wind would fall between 0 and 10 degrees..which is the protocol specified by Tour..by this method they predict the 808 to be 8 seconds faster per 40k and the Ritchey to be 2 seconds faster...but they are assuming 50% of angles will be between 0 and 5 degrees which is where most all wheels have very similar drag, and also are conditions which our models show occur with relativley low probability...were they to move their weighting to 5 or 10 degrees the predicted time savings would nearly double using the Tour data charts.

We also use a weighted average in calculating our wheel only test based time savings (we were the first to start suggesting the use of mathmatically weighted averages based on distributions of angle and not just single angle numbers...actually John Cobb has long suggested the weighting of individual data points, and we really put complex math models to it, and also started collecting extensive field data on real world wind at bike level) based around 10-15 degrees being the most likely wind angles so our assumption (based on rider speeds and more than 1000 handheld wind speed data points from real world race situations) predicts 80% of angles occuring between 5 and 15 degrees for pro road TT and 80% being between 10 and 20 degrees for triathlon racing (lower average rider speeds yield higher effective wind angles for the same given crosswind). Since the drag curves are non-linear, you can pick and choose data if you like depending on wind angles, we could have picked 12.5 for the cancellara test, which is the angle of peak minimum drag of the 808 or 14 degrees which is the peak minimum drag of the 1080 and made the numbers look even more in our favor for this test, just as we could do the same for wheel only testing, but as it stands, we run our formulas for our published wheel time savings based on a statistically weighted average of the drag curve and not a single point. Since large data sweeps weren't possible using Fabian, we stuck with the accepted standard test protocol, and we describe that completely in this test. For the wheel time savings claims we describe that protocol as well, and we have been instrumental in creating the use of the weighted probability analysis of wind angle, which is now heavily used in the pro-cycling ranks to make equipment decisions and to predict TT times....not just for our athletes but for competitors as well, which is why we keep Cd and CdA data such a secret.

As for the 808 vs stinger 90, we have talked on this very site about us beating the 90 with 21mm and narrower tires and them slightly edging us with 23mm and larger tires...we even posted this graph last year showing all of the wheels in question showing the stinger90 to be quite fast...of course is was more than a year after the 808, has ~11% higher side force in a crosswind and is ~1lb heavier, and a few hundred$ cheaper, so how you weight that is up to you. The 808 is 4-5 grams better on average with 21mm and the stinger is 4-5 grams better average with 23mm from our numbers, though of those 4 runs the 808/21mm tire has the lowest overall drag. You can also search this site and find other wheel only testing we have published, including an analysis done by Dan I believe showing our numbers for the Hed3 matching almost identically Hed's numbers for this wheel. There is also some Tour magazine graphs posted which show them getting even better numbers for the 808 than we show here. I am not sure why Hed pushes the H3 over the stinger90, when by our testing the 90 is without a doubt the faster wheel in almost every condition so I can't speak for them, but by our thinking (as well as that of the guys at the wind tunnel and most anybody else, the 90 is the faster wheel of the 2)




Since the 1080 extends the sweet spot of the curve out to 14 degrees, this makes the 1080 a good bit faster using the weighted average of yaw as the 1080 holds it's peak minimum drag over such a wider range of yaw angles. Again, had we run the cancellara test at 14 degrees, we probably could have 'proven' the 1080 was 30+ seconds faster than the 808 and 45 seconds or more faster than the 3 spoke. Actually by chossing the industry standard 10 degree yaw for the cancellara test, we were in effect further reducing our own advantage...particularly when we focus more than 80% of our design time studying the area between 10 and 20 degrees of yaw with about 65% of that time and money spent looking at 10-15 degrees...hence our strengh in those areas. So having said all that, 29sec is a conservative number for advantage of a 1080 pair over a 3 spoke pair using a weighted wind angle average centered around 15 degrees.

In the end, anybody here can say whatever they want. I feel that we go out of our way to be honest with our testing, realistic with our numbers and we generally try to be conservative with our claims. We always divulge the protocol of the test (something I never see in any other data) and we publish more data and graphs than all other companies in this industry combined when it comes to aero. We also spend more time and money doing what I would consider basic science type testing specifically for informing the public, such as our extensive wheelset comparison test at 20mph, 25, and 30mph to debunk the popular myth that 'aero only works at high speeds like 30mph' and our publication of white papers on our website being the first to really discuss designing the rim to match the tire and choosing the tire to match the rim, as well as other studies like our direct comparison of identical wheels built with various spoke counts, aero comparison of spoke shapes, the first aero studies on tires and so on. So say what you like, but nobody contributes as much to this discussion as we do, nobody divulges as much info as we do, and nobody takes as much crap for contributing as we do...in the end, my theory is that there is no reason to inflate this data now that anybody can go to the tunnel and disprove it...certainly we can be made to look bad at a certain angle or with a certain tire or whatever, but we will not lie about this...it's just too easy to be disproven...

---

http://www.SILCA.cc
Check out my podcast, inside stories from more than 20 years of product and tech innovation from inside the Pro Peloton and Pro Triathlon worlds!
http://www.marginalgainspodcast.cc

thanks for posting that Josh. Personally I'd have been quite happy with lbf or gf drag reduction @30mph tunnel speed and stated beta but I had no issue with the data as presented.

Not meaning to go off on a total tangent but did you see this thread and linked Rolf LSWT data from ~3wks ago? Any comment on the low yaw angle performance of the low-spoke count wheels (and tire combo)?


http://forum.slowtwitch.com/...t_reply;so=ASC;mh=25;

Ditto!!!

Uhmm.. instead of worrying about 1 to 2 Watts, why not just train a little more and pedal a little harder? Seems to be the cheaper alternative to me.

-- Boris

Since when are such things mutually exclusive?

Josh: It would be really interesting to see a vector plotting of wind effect/yaw for a rider at intervals during a typical course.

Do you have much, if any, real-world data (I am sure you do) on what a typical rider actually experiences over the course of an event? Just a very simple observation, such as the vast majority of courses are closed loops, and seeing how wind and yaw change and effect would be interesting.

It would seem that 10 degrees means that wind is assumed to be a net negative factor in a ride?

Not if you're using the right equipment.

Hmmm...I may have...but I can neither confirm nor deny if it exists due to a non-disclosure agreement ;-)

However, if one were to search on "808 vs. trispoke" one may find some "independent" data appropriate to this subject. It also may cost a few ducats to find out which one tested better.

To be fair though, the testing was done with a Conti 20c tire. It's well known that the trispokes perform better with a much narrower tire. Now, what the Crr "hit" would be running that narrow of a tire is a whole 'nuther subject ;-)

---

http://bikeblather.blogspot.com/

Depends on the brand, no?

Low aero drag, low Crr, high puncture resistance: pick two of the three. :-)

Everytime I read your posts I feel like I learn something and it makes me happy that I have your wheels.

what is the Crr hit with a veloflex record 20 tire ;)

It may be a "plus" depending on whatever tire was used before...then again, a VF Record 20 clincher measures more like 21-21.5 when mounted, so it won't be as good aerodynamically on a trispoke like a true 20C or 19C tire.

You also may want to consider the negative aero effects of the "lip" on a glued-on tread type tire (like the VF) vs. the smooth surface of a "vulcanized in one piece" tire like a Conti or a Michelin. Josh may be able to add some enlightenment on that with their experiences with the Vittoria tires.

Any wonder why I try not to put many miles on my last remaining Michelin ProLight 20C tire ;-)

---

http://bikeblather.blogspot.com/

Or...go with the low aero drag and low Crr, but then fill the tube "prophylactically" with some latex sealant :-)

---

http://bikeblather.blogspot.com/

Okay- thanks for the added information- more information is always better- if it's accurate- than it's even better. Let's take a look at the above chart. I can't comment on zipp's zipp wheel data- but you can compare Zipp's HED data to HED's website. Zipps numbers re: the 60, 90 and H3 are not even remotely close to HED's numbers for their wheels (If you're paying for the tunnel time- you better get the numbers to land on the right side up!) anyway, the H3 drag on Hed's site- is 161 grams and 158 grams at 15 and 20 degrees yaw, respectively. Zipp's graph appears to be at 200 grams at 15 degrees(an increase of 24% of drag over HED's numbers for the same wheel- which makes it go from wheel number 1 to wheel number 5 at that point in the stack rank) and at 20 degrees yaw- it looks to be at 172 grams- which is the best in the stack rank- but in reality HED had it at 158 grams or 9% faster than what the chart shows... So it would be even better. And the graph is way high from 0 to 10- where HED has the wheel below 200 grams (190 grams at 0 and 192 grams at 5 and 180 grams at 10 degrees yaw)- but Zipp has the same wheel well north of 200 grams- again inflating the drag of HEDs wheels by 10-15%. Again- independent data (Cervelo) said the Stinger 90 (which is actually slower than the JET 90) is faster than the 808 (p.s. Tour didn't test the 90 or 1080- that's a shame)- but of course the data above- has the 808 faster at every single point- or at least equal. Sounds like someone (who's huge in the industry- Cervelo) called you out on the carpet on this one " but we will not lie about this...it's just too easy to be disproven..."

Also- Zipp talks about the virtues of their Tangente tire being 1-3 watts faster- which is 9 to 27 grams of drag better- than the next best comparable tire... So- why wouldn't they want 9 to 27 grams of less resistance by using their superior aerodynamic product on their own product- to give them even better numbers when they are tested? You didn't use that tire on the above charts or the recently posted Cancellara data... Why not??? Isn't it all about "The True Aero Edge"??? Ironically- their 2 ~60 mm rims (404 is actually 58 mm)- vary greatly- and the one that is much slower- in fact the slowest wheel of all the Zipp's listed above- is the one with the far 'superior tire'....

... as a personal fyi- I have owned/purchased all of the following wheels in my past- Zipp- 808, dimpled disc, non-dimpled disc, 404, 440, 303, cages, skewers, Vuka swift shift. And HED- disc, freewheel disc, JET 60, 90, clip-on bar, full aero-bar, H3... so both of the above companies have at one point or another received a lot of coin...

I think this may be some of the data to which you are referring:



When you plot it out, it appears to me that the primary difference is simply the point at which the 808 effectively "stalls". Specifically, the CSC (? I always assumed it was some other team) -generated data says that this happens around 10 deg of yaw, whereas the Zipp data indicate that it doesn't happen until around 15 deg of yaw. It seems to me that said difference could easily be explained by small differences in, e.g., the test protocol, the type of tire used (width wouldn't seem to explain it, at least based on the CSC data), etc. I suppose you could argue that Zipp has manipulated such things in their favor, but doesn't it really make more sense to think that it is the other way around, i.e., their design may have been driven in part by the test protocol that they use? After all, even the CSC data shows that both the 404 and 808 "sail" to some degree; it's just a matter of how far out such an effect remains.

(BTW, a similar difference exists in the 404 data, whereas the the Zipp-generated Specialized Ultralight vs. CSC-generated HED3 datasets are practically superimposable, as are the corresponding disk datasets.)

surely that would kill the Crr.

Not according to AFM's roller testing. No detectable difference in Crr.

---

http://bikeblather.blogspot.com/