A Little Tech-Talk About Wheels

I'm trying to make a decision about a racing wheelset. I've been pretty sold on classic non-aero rims based on my initial research. However, I'm beginning to think that the people I've been talking to about wheels, while having a ton of experience building wheels, don't really know very much about aerodynamics. It's definitely complicated and there are a lot of assumptions and variables to consider not to mention there are all sorts of conflicting reports out there.

I came across this spreadsheet that I kind of like:

http://www.whitemountainwheels.com/SpeedPower.html

I've made something very similar to this, the big difference being that I took the easy shortcut by not considering wind effects from varying angles which isn’t very realistic. I haven't verified his equations yet, but there is a section at the bottom called "iterations" that looks pretty thorough in accounting for wind. After a quick look at the equations it seems like it accounts for both head and cross winds, so that’s good.

The most interesting inputs to me are comparing CdA in cell B26 and weight in either C4 or C5 (CdA is the total aerodynamic drag coefficient of the rider/bike). They allow me to compare the effects of losing 1lb of wheel weight (a lot) and a 0.01 improvement in CdA. The 0.01 improvement in CdA wins by a landslide unless you have a decent climb of 5% grade or more which is comparable to the Whiteface hillcimb.

I found a graph online (which I’ve provided above, although I don’t know the source) that shows the relationship between rim depth and CdA. The CdA difference between a 20mm rim depth (about right for a Ksyrium wheelset) and a 55mm rim depth (now we're talking something like a zipp404) is about 0.01, maybe even slightly more. The default CdA for the bike/rider combo in this spreadsheet is .32 and I think this is about right. I think it was Cervelo that once bragged about getting a CdA value of .285 for a TDF TT (try this in the spreadsheet, it's a crazy difference!). According to: http://www.fitwerx.com/NewFiles/Tech%20Center/BicycleAerodynamics.html

the wheels account for 7-11% of the total aerodynamic drag. So let’s see, 0.32 * 0.11 = 0.035 and 0.32 * 0.07 = 0.0224, a difference of 0.013. Hey! That’s pretty close!

So what do I conclude from this? Well, assuming that the whitemountainwheels spreadsheet is close, wheel weight doesn’t mean much to me, right? My answer is still no! This is still overlooking two very important things that I am yet to look into myself. They are:

1) Wheel Stiffness: Being the young grasshopper of a racer that I am, I don’t really have a type of riding style that defines me yet but I see myself developing into more of a sprinter type than trying to make Floyd Landis breakaways. I think this means I need a stiff wheel! If I go to sprint like I did at Wells Ave the other week and my rear wheel flexes so much that I feel like I’m riding a mt bike, my brake pads are going to rub and Tim Mitchell is going to smoke me. But wait, what if I’m more tired from using a stiff but non-aero wheelset and I just don’t have the gas to lay down a good effort? That brings me to my next worry.

2) Aero benefit within a pack: All these numbers and calculations seem to be thorough and all, but they all make the assumption that the rider is in a time trial situation and there is no drafting. In most of the racing I do there is lots of drafting and it’s important that I consider it, but how? I might be able to to simply adjust the CdA coefficient to a much lower value to account for the drafting effects. I’d throw out a quick guess and say the CdA will be something like 0.20 with drafting (down from 0.32), so about 63% of the non-drafting CdA. Does that mean that the 0.01 improvement in switching from Kysrium Sl’s to Zipp404 is reduced to 0.0065? Maybe, and it seems like my best guess for the moment, but I really don’t know.

According to this really cool spreadsheet, if I’m sitting in a pack putting out 280W with a CdA of 0.20 on a flat rode with my Ksyrium Elites and with whatever other assumptions, I’ll be cruising at 24.96 mph. Now give me a set of zipp404’s with none of those golfball dimples thus reducing my CdA from 0.20 to .1957, I only have to put out 275W to do the same 24.96mph! Woohoo! I guess that means my heart rate will be lower, but how much lower? Gosh I don’t know.

Let’s do one other quick calculation to compare 280W vs. 275W. Let’s assume a whole bunch of typical inputs for a 10 mile TT type situation (so we’re back to the 0.32 CdA if anyone is trying to repeat my results here). The question is, how fast can I do the 10 mile TT at 280W vs. 275W? Let’s consult the magical spreadsheet! The answer is: 28:12.3 for the 280W effort and 28:25.1 for the 275W effort. Somehow that doesn’t seem like anything too crazy. I could eat my wheaties in the morning and make a way bigger difference than that. Does that mean if I go do the Charlie Baker Time Trial it’s only going to make a 13 second difference by using the Zipps?

Nope. But while we’re on a roll we’ll figure that too. Same assumptions, blah blah blah, conservative 0.01 CdA improvement at a 280W effort (gee I hope 280W is a good number to be using for a cat5 like me, I think it is…), the Mavics finish in 28:12.3 and the Zipps in 27:53.9! That’s about an 18 second or 1.04% improvement! 1% doesn’t sound like much, but if the winning non-aero rider is finishing in 24 minutes (not sure if this is true), you go from 4:12 back to 3:54 back which is over a 4% improvement.

This stuff probably only appeals to 5% (if that!) of the people reading this, but I can’t get enough of it. If anyone does any calculations of their own I’d love to check them out. Also, if anyone has had any actual experience with a similar decision involving my key factors of wheel rubbing and drafting effects I would love to hear them. Thanks for reading.