RowToTri wrote:
Why spend so much time debating tests that did not even happen? It still seems to me that it is much more likely that this guy did not sneak all this time in to this wind tunnel. It also seems to me that people that DO spend a lot of time in that wind tunnel, like Jean-Paul Ballard, would have heard of Hambini if he was there so much. But during the Q&A section of this Google talk, someone asks him about Hambini and he says he has no idea who that is.
http://www.youtube.com/watch?v=lmqdqcOvrlc Look, I think Hambini is a arrogant jerk too. He is also very clearly full of bias - as he has made some pretty big enemies and slags them pretty hard (partly for entertainment of his followers, but I think that is just who he is also). He also sells product/services, and has a Youtube channel that pays for views. Sure it is not his day job, but between the money and ego, it is all in his benefit to grow, build his name. Without looking, I can 100% guarantee that his Aero wheels blog post has more views than every other post combined. So there is a lot of reason for him to perpetuate that discussion. THat is why he continues to test more wheels. And the more we say his name, the more recognition we given him. And like certain politicians (and news agencies) have figured out, controversy generates interest, notoriety, and support.
Maybe he made all the data up. It would be an interesting exercise in psycology and personality disorder if he did. The coworker using his account. The flow legal letter. But the reality is that we will never really know. It all comes down to credibility, and our trust. Given there is no true peer review of his data/process, we have to go on what we see.
Even if his data was real, I would not use it to chose wheel A over B because his protocol error of 2.5% overlaps most of the wheels in each test. But even if the data was not real, or fatally flawed, the basic principals of aerodynamics that he brings to the table is worth considering. i.e. be careful of how much weight you place on data generated from a laminar flow situation, because the real outside world is often not laminar. My armchair PHD tells me that the higher the wind velocity on the day you ride, the farther off the data may be. Or is it that far off?
But at the same time, assuming his data is real, I still think it supports that our current methods of designing and aero wheels is on point. The general aerodynamic principles and trends are still there. Deeper rims = faster. Just don;t use it to say rim A is faster than Rim b, because in other situations, the pecking order might change.
IS his data more valid of real world than data from other forms of testing? The only way to know is if his data was more accurate at predicting real world performance gains. The problem is, outside of an indoor velodrome, or minimal wind days, reality has a lot of noise in the data... so the magic 8 ball answer will always be "answer unclear." But from what I understand, current protocols are pretty damn good at predicting real world performance gains. So a ground up rethink of how the industry does things may not be required. Rather, non laminar aero data may just enhance our knowledge.
What I would like to see teased from Hambinis data (or any large volume test data) would be things like spoke count, spoke type, spoke alignment, exposed nipples, rim profile, rim with to mounted tire with. Brake that up by rim depth. Show us how much rim depth vs say spokes really matter in building a fast wheel. same for tire to rim width. Maybe the poor performers just simply had poor spoke alignment... and their rim shapes were pretty good. Maybe it was that 23c tires at 120psi (I estimate could be an easy 26mm on some rims) killed the aero of some rims because they were just too damn wide. Maybe depth is more important than shape, and terroidal vs naca is really just the final 5% of gains for a rim... and the best answer depends on the environmental conditions at the time. Intuitively I think we know what will be faster to maybe 95% of the wheels potential. The last 5% is the trick, but with the rider taking 80% of drag... maybe who cares, because as a system, that 5% is now 0.3% or some similarly small number. And that 0.3% is far less than the error of any test method we have... so we can;t prove it anyway. IT is not easy to detect a pattern among the noise, and the real world has a ton of noise.
What I am taking from this, and everything I have learned of bike aerodynamics from the comfort of my couch, is that there is no one best model. And it is key to understand the limitations, and strengths/advantages of each testing method. And they all have value. But the ones with the most value are easily reproducible, and accessible. And don't take any protocol that says wheel/frame A is faster than frame B too seriously, but don't ignore it either.