I know there’s been a lot of discussion on the board about how to do rolldown tests and so on, but I haven’t been able to dig up a condensed version of best practice. (I can’t afford tunnel testing, and don’t have a powermeter. Also, what would the benefits be if I could borrow a powertap?)
What’s preferred: rolldown, pictures to calculate cda, or something else? What protocols are important to get meaningful data?
Much appreciated.
I know there’s been a lot of discussion on the board about how to do rolldown tests and so on, but I haven’t been able to dig up a condensed version of best practice. (I can’t afford tunnel testing, and don’t have a powermeter. Also, what would the benefits be if I could borrow a powertap?)
What’s preferred: rolldown, pictures to calculate cda, or something else? What protocols are important to get meaningful data?
Much appreciated.
This will depend on how small of a difference you want to detect. It’s easy to detect something like “hands on the hoods” vs. “hands on aerobars” with outdoor coastdowns – but then, you probably don’t need to do coastdowns to tell you that. It’ll be hard to detect the difference between two aero helmets, or two water bottles, or two frames.
The smaller the difference you want to detect, the better the measurements have to be. Better measurements depend on two things: better equipment, and better control of testing conditions.
This makes sense to me. So, in your judgment, would you say that with a powermeter and several runs over a flat course I’d be able to get decent resolution on fit issues (aero bar positions, steepness, etc)?
The smaller the difference you want to detect, the better the measurements have to be. Better measurements depend on two things: better equipment, and better control of testing conditions.
This makes sense to me. So, in your judgment, would you say that with a powermeter and several runs over a flat course I’d be able to get decent resolution on fit issues (aero bar positions, steepness, etc)?
Potentially, yes – though keep in mind that addresses the “better equipment” part but not the “better control of testing conditions” part of better measurements. The PM takes care of the data recording but you’ll still have to find a day and time when the wind isn’t too strong, you hold your position steady, you follow the other demands of the particular protocol you’re following, yadda yadda yadda. Like anything else, the more experience you have the better you get at following the protocols so don’t be discouraged if the first time you try field testing you get crap precision.
Coggan likes flat venues where he does controlled speed runs. You might find this link handy. Others seem to like the VE method. Depends on the terrain you have available to you.
Coggan likes flat venues where he does controlled speed runs.
Actually, I like (prefer) wind tunnel testing (unless you race on an indoor track, in which case testing on such a track may be sufficient). Sometimes you just have to make do, though, and field-testing is fun in the sense that it gives you lots of numbers with which to play around.
Better measurements depend on two things: better equipment, and better control of testing conditions.
Actually, I think the last could actually be broken down into two…
- better equipment
- better control of testing conditions (ie weather/wind “control” by selecting location and testing time/date).
and
- experimental method
The last is something that is refined over time. If you have never done field testing, the way you perform the actual tests (and how detailed your records are) CAN influence the results.
The more you do it, the better you get at doing it and the better you get at knowing what influences results and what to try and control.
g
The protocol in my experience is key, as is finding windless days and a good loop for the chung method.
One other note- you have to take INSANSE quantities of notes on what you vary in each run, and what lap your powertap thinks its on, etc etc. EVERY time i think im taking too many notes, and every time i find it relaly hard to piece together the optimal setup from my notes again.
And one note- the lap your powertap says (ie lap 3) and the lap as displayed in the computer printouts are not the same (the printouts call that lap 2 or 4, i cant remember which, each time it re-confuses me)
If you have never done field testing, the way you perform the actual tests (and how detailed your records are) CAN influence the results.
The more you do it, the better you get at doing it and the better you get at knowing what influences results and what to try and control.
Yeah. Just cuz you have a PM on your bike doesn’t mean your results won’t be lousy.
I’m actually kinda surprised that when I first began aero testing my results were as good as they were cuz my experimental technique was crap. (Must’ve been the method that saved me). I’m still not quite in Tom A territory (damn him) but I’m getting close.
I’m actually kinda surprised that when I first began aero testing my results were as good as they were cuz my experimental technique was crap. (Must’ve been the method that saved me).
Or it is easier to obtain precise (vs. accurate) results than you think.
Case-in-point: even completely ignoring variations in wind, road surface, and elevation*, back when I first started using a powermeter I was able to estimate my CdA to w/in ~0.005 m^2 from power and speed data collected during (*relatively flat) TTs.
Or it is easier to obtain precise (vs. accurate) results than you think.
Case-in-point: even completely ignoring variations in wind, road surface, and elevation*, back when I first started using a powermeter I was able to estimate my CdA to w/in ~0.005 m^2 from power and speed data collected during (*relatively flat) TTs.
field-cda-challenge.csv
Or it is easier to obtain precise (vs. accurate) results than you think.
Case-in-point: even completely ignoring variations in wind, road surface, and elevation*, back when I first started using a powermeter I was able to estimate my CdA to w/in ~0.005 m^2 from power and speed data collected during (*relatively flat) TTs.
field-cda-challenge.csv
That’s easy: based on those data, your CdA is XXX +/- 0.000 m^2, with XXX depending on the constants (mass, air density, etc.) assumed.
Or it is easier to obtain precise (vs. accurate) results than you think.
Case-in-point: even completely ignoring variations in wind, road surface, and elevation*, back when I first started using a powermeter I was able to estimate my CdA to w/in ~0.005 m^2 from power and speed data collected during (*relatively flat) TTs.
field-cda-challenge.csv
That’s easy: based on those data, your CdA is XXX +/- 0.000 m^2, with XXX depending on the constants (mass, air density, etc.) assumed.
Show your work.
Or it is easier to obtain precise (vs. accurate) results than you think.
Case-in-point: even completely ignoring variations in wind, road surface, and elevation*, back when I first started using a powermeter I was able to estimate my CdA to w/in ~0.005 m^2 from power and speed data collected during (*relatively flat) TTs.
field-cda-challenge.csv
That’s easy: based on those data, your CdA is XXX +/- 0.000 m^2, with XXX depending on the constants (mass, air density, etc.) assumed.
Show your work.
Some of my work
Thanks for the link; this gives me much more to think about. What is the VE method? Do you have a link to this?
Or it is easier to obtain precise (vs. accurate) results than you think.
Case-in-point: even completely ignoring variations in wind, road surface, and elevation*, back when I first started using a powermeter I was able to estimate my CdA to w/in ~0.005 m^2 from power and speed data collected during (*relatively flat) TTs.
field-cda-challenge.csv
That’s easy: based on those data, your CdA is XXX +/- 0.000 m^2, with XXX depending on the constants (mass, air density, etc.) assumed.
Show your work.
Some of my work
I’m not surprised.
Or it is easier to obtain precise (vs. accurate) results than you think.
Case-in-point: even completely ignoring variations in wind, road surface, and elevation*, back when I first started using a powermeter I was able to estimate my CdA to w/in ~0.005 m^2 from power and speed data collected during (*relatively flat) TTs.
field-cda-challenge.csv
That’s easy: based on those data, your CdA is XXX +/- 0.000 m^2, with XXX depending on the constants (mass, air density, etc.) assumed.
Show your work.
Some of my work
I’m not surprised.
Neither am I.
Thanks for the link; this gives me much more to think about. What is the VE method? Do you have a link to this?
Hmmm. There are a couple of spreadsheets floating around but AFAIK there’s no single place that discusses the protocol in the same level of detail as that link I gave for the regression method. Alex discussed it a bit here.
I’m not surprised.
Neither am I.
I’m hoping the things we’re not surprised by are different.
Thanks for the link; this gives me much more to think about. What is the VE method? Do you have a link to this?
Hmmm. There are a couple of spreadsheets floating around but AFAIK there’s no single place that discusses the protocol in the same level of detail as that link I gave for the regression method. Alex discussed it a bit here.
Speaking of your method…what do the results look like for the analysis shown on slide 6 here:
http://anonymous.coward.free.fr/...cda/indirect-cda.pdf
when you correct for changes in kinetic energy as you did with your method (so as to truly compare apples-to-apples, and not apples-to-oranges)?
I’m not surprised.
Neither am I.
I’m hoping the things we’re not surprised by are different.
“Hope is for sissies.” - Dr. House.