Wednesday, July 04, 2007

Poor man's wind tunnel

I wanted to know what effect my TT bike, Aero helmet, and skinsuit had on my drag coefficients and TT finish times, but didn't want to pay $2500 for a couple of hours in a wind tunnel. So I found the longest fairly low-traffic hill near my house, a 0.6-mile climb with a 3.5% average grade. Starting at a constant 20mph at the top to eliminate momentum factors, I coasted down the hill lots of times with different aero setups to determine my drag coefficients. I know that my power was 0 watts, so I was able to eliminate power measurement from the list of uncertainties. My typical average speed was in the 23mph range, so I know my numbers will be valid for my average TT speeds.

The only factor pushing me down the hill was gravity, so weight was important. I carefully weighed myself in my kit/shoes/helmet and the bikes prior to testing.

I ran at least 3 'no-car' runs for each setup and averaged them. (I found that each car that passed me during the tests cut about 1 second from my finish time, so I threw out all the runs where cars passed me during the test). There was little or no breeze, it was 88 degrees, and I alternated test runs with various equipment in case the environmental conditions changed during the testing.

The spreadsheet I used required entry of a drag coefficient and/or a bike and rider frontal area and factored those with an air density number to get an 'A2' factor that it used in the calculation. I got everything to work well in the spreadsheet, but I still don't know exactly what my frontal area and drag coefficient are, but I'll work on that later. I used a rolling resistance of 0.004 (which I realize now might be too low for my tubular tires - it's probably nearer .0055).

Here are my averaged results for each setup:
Orbea Orca with standard team kit and helmet, riding in drops - 1:34.5
Cervelo P2K with standard team kit and standard helmet, in aerobars - 1:30.0
Cervelo P2K with standard team kit and TT helmet, in aerobars - 1:29.3
Cervelo P2K with skinsuit and TT helmet, in aerobars - 1:25.8

I was pleased to find that even my low-tech testing was able to accurately measure the effect of even changing helmets on a minute and a half section. The repeatability of the runs was very good, usually within 1 second.

The biggest surprise was the skinsuit. Changing from the team kit to the skinsuit has as much effect at going from the road bike to the tri bike! And the numbers are huge. For a 40K TT, the skinsuit would save 1 minute, 46 seconds. That's a lot of time. Based on my time in the recent half iron man, I saved 4 minutes and 18 seconds purely by wearing the skinsuit and TT helmet. For the short 10-12 mile TTs usually included in the Georgia Cup races, I'll save almost a minute by wearing the skinsuit -- that's amazing to me.

And these numbers are not wild guesses, they are real world measurements that can be easily and accurately extrapolated to racing.

Don't stop training hard -- the bike won't power itself. But for God's sake, buy a skin suit if you're going to ride TTs.

Next I plan to play around with my bar height, bar angle, etc. to see if I can get any measurable effect. And rolling resistance is also on the list of things to investigate.

Since originally posting this earlier in the year, I've fine tuned my analysis spreadsheet. You can view or download it at Badger 3 if you are interested.

1 Comment:

Douglas Ellmore, Sr. said...

Could you share your spreadsheet? How many runs did you perform with each configuration? Have you ever done a Design of Experiment approach?

Finally, one of the factors that I think is missing is the ability to produce watts with certain configurations.

I think you are on to something though. If you can get a basic time dialed on down a hill, then you can tweak from their for power effeciency.

The only thing I find, could my configuration be different for different length races, Sprint Tris, Olympic/40K TT, 1/2 Iron, etc.

Doug Ellmore
www.ellmore.net