I gotta agree with the others. There is certainly sliding friction since the rollers are moving vertically and the cheese grater is moving in a circle.

And it cannot be that the 3% refers to bike + mannequin since the specialized engineer says time savings over 40k could be as much as 8 seconds. Just back of the napkin, if you combine the power and drag formulas you get P=D*V.

Let's say a rider goes 10.5m/s (23.5mph) on 220 Watts. We get D=P/V=220/10.5= 20.95 N. He covers 40k in 63.49 minutes.

So let's reduce drag by 3%: 20.95 N * .97 = 20.32 N

Now his new velocity is 220/20.32= 10.83m/s.
Now he covers 40k in 61.56 minutes. An improvement of almost 116 seconds.

Now I did not take I to account that rolling resistance and less importantly, drive train friction remain the same, so let's adjust that 116 down by 20% and we get 93 seconds.

So either the 8 second improvement is wrong or the 3% is bike only.

---

-------------
Ed O'Malley
www.VeloVetta.com
Founder of VeloVetta Cycling Shoes
Instagram • Facebook

This isn't right even for "back of the napkin." The power required to overcome aerodynamic drag is cubic with speed, so you can't just scale velocity by the ratio of CdAs. You need to scale velocity by the cubic root of the ratio of CdAs.

The cubic root of 1.03 is roughly 1.01. So the speed difference, if we only consider aerodynamic drag, is about 1%. 1% of an hour is a little over half a minute.

ok so....

D = CdA*.5*rho*V^2
P = CdA*.5*rho*V^3

P = CdA*.5*rho*V^2*V
P = D*V

So power is still scaling with the cube of V because Drag scales with the square of V. I was assuming when he said a 3% improvement in the wind tunnel, he meant drag, but I could have been wrong. It could have been a 3% reduction in CdA. In which case it would be about a 36 second difference if it was with the rider on. Still a lot different than 8 seconds!

---

-------------
Ed O'Malley
www.VeloVetta.com
Founder of VeloVetta Cycling Shoes
Instagram • Facebook

Ah right, I think I'm just confusing myself because I read the article differently. It's hard to tell because the article isn't specific... the axes in the CdA graphs aren't even labeled!

  
Yeah, I think the 3% could refer to either the way it's worded.... so who knows, but either way it's not rider-on if the 8 second improvement is correct. No labels on an axis is always a warning sign...

---

-------------
Ed O'Malley
www.VeloVetta.com
Founder of VeloVetta Cycling Shoes
Instagram • Facebook

Yeah, they could easily label the y-axis to clear this up.

Probably what happened is that the tests for rider-off and rider-on were very similar in terms of drag delta from a standard chain setup, and then the marketing guy from Ceramic Speed mistakenly stated the 3% figure for the rider-on setup. It's weird that he doubled down on that in this thread though when the numbers clearly don't add up by a factor of 4x-5x.

One other thing about the test data: it's very unusual for a setup that changes things on just the drive side of the bike to show essentially identical drag reductions across the yaw sweep. That doesn't pass an initial smell test for me.