Aero Position Testing

Afterwards, test threshold power in each position indoors on a trainer. You are looking for the most aerodynamic position at a given wattage. But, finding the position which gives maximum watts while paying a lesser aerodynamic penalty would be the optimum, no?

M. DeKelver said:

Afterwards, test threshold power in each position indoors on a trainer. You are looking for the most aerodynamic position at a given wattage. But, finding the position which gives maximum watts while paying a lesser aerodynamic penalty would be the optimum, no?

I'm operating under the assumption that I will be able to generate the same power in the most aero position after practicing in the position sufficiently. IOW, if I do tests to determine my max power in different positions, there will no doubt be positions that will prevail in the first test. But, if I find the position that results in the least drag even though I can't generate as much power initially, it's only fair to practice that position a few times and then see if I can generate a comparable power. I think that with practice I will indeed be able to generate a comparable power. I won't be riding a TT of any realistic duration at my max power for shorter durations. That's why I intend to conduct my tests at a power that is realistic for the most common TT durations (20K and 40K). The main reason I intend to manage speed and not power on my practice runs is that I find that I can control my speed more precisely than my power. I would have a very difficult time riding at precisely 300w, but I find it easier to ride at precisely, say, 28.0 mph.

RapDaddyo said:

What I'm interested in is developing a practical test protocol to evaluate different positions and bike setups. While I am not now focused on wheels, I will use whatever test protocol I settle on to evaluate wheels at a later date... I'd greatly appreciate any comments on this planned test protocol or any better protocol.

Paul: To suggest a completely different approach. Most professionals optimize their aero position in a windtunnel. The cheapest equivalent is coasting downhill for a standard distance and timing the run. To state the obvious, the least drag will produce the fastest time; power is not an issue, and hence plays no part in the test. The biggest problem is standardizing the variable of wind. Granted this does not deal with the issue of power production, but you mentioned in an earlier post that you wanted to first find the optimal position, and then determine whether it hindered power production.

RapDaddyo said:

I have decided to enter the NV Senior Games cycling events, which include two TTs (5K & 10K -- yeah, I know, they're short). My primary objective is to get a benchmark on my speed in the range of power levels that I anticipate using in longer TTs next spring. This raises the issue of testing different subtleties of aero positions and my bike setup (e.g., bar height and seat position). I'm well aware of the primary elements of aero positions. What I'm interested in is developing a practical test protocol to evaluate different positions and bike setups. While I am not now focused on wheels, I will use whatever test protocol I settle on to evaluate wheels at a later date.

My plan is to find a road segment that is as flat and smooth as possible for a distance of at least 3/4 mile. I want a segment of ~2 mins for timing. I have a roadway segment very near my residence that was re-paved in June that will probably work out pretty well. Second, I will do tests on days when the wind is <10mph and is primarily a crosswind. The road segment I have in mind runs E-W and the prevailing winds here are out of the south. I plan to record the wind direction and speed for each test day with a portable digital anemometer. I plan to ride the test segment in various aero positions at a constant speed that is about as fast as I can smoothly maintain for the duration of the run. I find that the further up the power curve I go the more uneven my stroke becomes. So, I don't want to be at my 2 min MP, but rather ~90% of that power. That should put my speed in the 25-30mph range. I will attempt to ride each test run at precisely the same speed, within 0.1mph. I will ride each test twice so that I can throw out a test in which my speed varied outside of an acceptable range (e.g., +/- 0.1mph). The variable of interest is average power. The lowest average power means that is the most aero position or bike setup. The benchmark I will use to gauge all positions is a standard training ride position with hands on the hoods. All tests will be done with a TT bike arrangement (e.g., 1 bottle, no pump or wedge). I'd greatly appreciate any comments on this planned test protocol or any better protocol. BTW, I don't have convenient access to a track, indoor or outdoor.

Wouldn't riding a longer distance at a relatively low power give you greater confidence that you were accurately getting information about aerodynamics? Or even the same distance at a fairly low power? Is there a reason to ride at 90% 2MP? Absolute time/distance differences due to aerodynamics will be greater at lower power outputs. Why not ride at 200W (or even less) and record time over a given distance or distance in a given time?

It is probably a good idea to repeat the tests multiple times both on the same day and on other days. I might take the average of 3 tests over a single test. I also would try to standardise the tests somehow - eg, if you have a road bike with a PM as well as a TT bike, take both bikes out and do a test on your road bike in a standard position before each session so you can scale the data against that number, or see if the data is too variable to be worth anything. Be careful to wear the same clothing each test. Remember that wearing an aero helmet will change what is the best head position.

I think wind and air pressure will mean environmental factors still come into play too much. You may want to just set up as Ric and Andy describe elsewhere and leave it at that until you have access to an indoor track and wind tunnel. Otherwise keep records of humidity and barometric pressure as well as wind direction and speed. You will be fast as hell on days with high humidity, low barometric pressure and no wind.

Different yaw angles will affect which position and equipment is best. For example, Zipp makes a disc, an 808 and a 404, in descending order of aerodynamics (I know they make other wheels too - this is just an illustration). They are almost indistinguishable with a yaw angle 0-5 degrees but there is a significant difference at yaw angle 15+ degrees. Consequently on a hilly course with no wind or a wind from directly behind or in front, two 404s might be the best choice to avoid weight penalties, whereas on a flat course, especially with a sidewind, the disc rear/808 front is probably best (assuming the wind isn't gusty, rendering the disc difficult to control - in this case a tri spoke rear like a HED might be best). Maybe try to work out what the likely yaw angles and std deviation in wind speed at your target event is and think about that.

palewin said:

Paul: To suggest a completely different approach. Most professionals optimize their aero position in a windtunnel. The cheapest equivalent is coasting downhill for a standard distance and timing the run. To state the obvious, the least drag will produce the fastest time; power is not an issue, and hence plays no part in the test. The biggest problem is standardizing the variable of wind. Granted this does not deal with the issue of power production, but you mentioned in an earlier post that you wanted to first find the optimal position, and then determine whether it hindered power production.

Thanks for the suggestion, Pete. First, there's no way I am going to shell out several thousand dollars for wind tunnel time. I am aware of the coasting methodology, but it seems a little artificial in that the positions I can get into coasting are not necessarily the positions I can get into while putting out 300w of power. I am not only interested in my back angle and hand/arm positions, but also in my pedaling style (knees in toward the top tube, for example). I will feel more comfortable with results obtained while turning the cranks over with race-like effort. I may not know that I can hold the position for an hour, but at least I'll know I can hold the position for 2 minutes. What's 58 minutes among friends?

Roadie_scum said:

Wouldn't riding a longer distance at a relatively low power give you greater confidence that you were accurately getting information about aerodynamics? Or even the same distance at a fairly low power? Is there a reason to ride at 90% 2MP? Absolute time/distance differences due to aerodynamics will be greater at lower power outputs. Why not ride at 200W (or even less) and record time over a given distance or distance in a given time?

I'm already concerned about finding an absolutely flat stretch of road 3/4 mile in length, much less longer. And, I purposely want to ride at a relatively high speed (even faster than I may be able to maintain in a 1 hr TT) because I want the air resistance to be high in order to accentuate differences in aerodynamics. I'd do it at 50 mph if I could make my bike go that fast for 2 mins (wishful thinking).

Roadie_scum said:

It is probably a good idea to repeat the tests multiple times both on the same day and on other days. I might take the average of 3 tests over a single test. I also would try to standardise the tests somehow - eg, if you have a road bike with a PM as well as a TT bike, take both bikes out and do a test on your road bike in a standard position before each session so you can scale the data against that number, or see if the data is too variable to be worth anything. Be careful to wear the same clothing each test. Remember that wearing an aero helmet will change what is the best head position.

I agree about multiple tests, especially for the positions that appear to be the best. I think I will be able to rule out certain options with one test. And, I agree about keeping clothing and equipment constant. At this point, I don't want to introduce equipment into the equation (bikes, helmets, shoe covers, skinsuits).

Roadie_scum said:

I think wind and air pressure will mean environmental factors still come into play too much. You may want to just set up as Ric and Andy describe elsewhere and leave it at that until you have access to an indoor track and wind tunnel. Otherwise keep records of humidity and barometric pressure as well as wind direction and speed. You will be fast as hell on days with high humidity, low barometric pressure and no wind.

I do plan to record wind direction and velocity, baro pressure and humidity for each test day/time. As I narrow down what appear to be my optimal positions, I will re-test them on different days. Positions that get soundly trounced will get tossed, never to be heard from again.

Roadie_scum said:

Different yaw angles will affect which position and equipment is best. For example, Zipp makes a disc, an 808 and a 404, in descending order of aerodynamics (I know they make other wheels too - this is just an illustration). They are almost indistinguishable with a yaw angle 0-5 degrees but there is a significant difference at yaw angle 15+ degrees. Consequently on a hilly course with no wind or a wind from directly behind or in front, two 404s might be the best choice to avoid weight penalties, whereas on a flat course, especially with a sidewind, the disc rear/808 front is probably best (assuming the wind isn't gusty, rendering the disc difficult to control - in this case a tri spoke rear like a HED might be best). Maybe try to work out what the likely yaw angles and std deviation in wind speed at your target event is and think about that.

All my position testing will be done without deep dish or disc wheels. I know I will be faster with such wheels and that choice of wheels will depend on the specifics of the course and environmental variables, but at this point my focus is on position and bike setup (e.g., bar drop).

RapDaddyo said:

I'm already concerned about finding an absolutely flat stretch of road 3/4 mile in length, much less longer. And, I purposely want to ride at a relatively high speed (even faster than I may be able to maintain in a 1 hr TT) because I want the air resistance to be high in order to accentuate differences in aerodynamics. I'd do it at 50 mph if I could make my bike go that fast for 2 mins (wishful thinking).

Hmm. I guess by going very fast you increase the amount of power required to overcome air resistance. By going at a low constant power you increase the amount of time saved/lost due to aerodynamics. So it all depends on maximising the variation in the variable you want look at. I wasn't really thinking. I guess your method makes as much sense as mine.

I was thinking time for a known course, given power.

You are thinking power for a known speed, given speed.

Obviously minimising either of these achieves what you want.

Roadie_scum said:

I was thinking time for a known course, given power. You are thinking power for a known speed, given speed. Obviously minimising either of these achieves what you want.

Exactly. But, my experience to date suggests that it is really hard (for me) to ride at an absolutely constant power. I have tried riding at an absolutely constant power, speed and cadence (independently of course). I do best at managing cadence, but managing speed to a tolerance of 0.1mph gives me an error rate of ~1/280th at 28mph, whereas cadence gives me an error rate of ~1/100th at 100rpm, assuming I can maintain speed at +/- 0.1mph and cadence at +/- 1rpm respectively.

RapDaddyo said:

Exactly. But, my experience to date suggests that it is really hard (for me) to ride at an absolutely constant power. I have tried riding at an absolutely constant power, speed and cadence (independently of course). I do best at managing cadence, but managing speed to a tolerance of 0.1mph gives me an error rate of ~1/280th at 28mph, whereas cadence gives me an error rate of ~1/100th at 100rpm, assuming I can maintain speed at +/- 0.1mph and cadence at +/- 1rpm respectively.

But cadence and speed should belinearly related shouldn't they? Is part of this because of differing sampling rates for each measurement? That is, your wheel magnet will go past the speed sensor about 20 times per second, whereas your cadence magnet will only go past about 1.5-2 times per second.

I've noticed that most bike computers I've had don't respond instaneously or even a close approximation thereof to speed changes, so you'd also need to take into account the calculation rate of your computer (or powertap as I guess it is in this case).

Roadie_scum said:

But cadence and speed should belinearly related shouldn't they? Is part of this because of differing sampling rates for each measurement? That is, your wheel magnet will go past the speed sensor about 20 times per second, whereas your cadence magnet will only go past about 1.5-2 times per second.

I've noticed that most bike computers I've had don't respond instaneously or even a close approximation thereof to speed changes, so you'd also need to take into account the calculation rate of your computer (or powertap as I guess it is in this case).

Yes, they are linearly related. The problem lies in the precision of the display. If cadence were displayed to 0.1rpm, then it would allow for tighter error management (e.g., 1/1000th). But, as it is, cadence is displayed in even rpms, whereas my PT speedo is displayed to a precision of 0.1mph. In both cases, you have to set the averaging interval to the shortest available to make the instrument as precise as possible. The good news is that I can download the file and examine it at a 1 sec data recording interval. If this degree of precision and attention to detail doesn't allow me to differentiate between two positions, then I don't suppose it's going to matter much in my elapsed time in a race.

We should talk about this in Colorado Springs next month...over the past few years I've done a fair bit of field testing to determine aerodynamic drag characteristics on both a velodrome and a flat straight road, and may be able to give you some pointers that would help you get cleaner data, or clean data sooner.

acoggan said:

We should talk about this in Colorado Springs next month...over the past few years I've done a fair bit of field testing to determine aerodynamic drag characteristics on both a velodrome and a flat straight road, and may be able to give you some pointers that would help you get cleaner data, or clean data sooner.

Wonderful. I would welcome your insights and suggestions. I look forward to seeing you in CO Springs. Cheers.

I was flipping through a Triathlete magazine and there is a place in San Diego that offers wind tunnel position/power testing for about $700. Sure, it's expensive but if you are truly serious about getting good information without the variables that you introduce trying to guesstimate things outside, there is no better way.

park said:

I was flipping through a Triathlete magazine and there is a place in San Diego that offers wind tunnel position/power testing for about $700. Sure, it's expensive but if you are truly serious about getting good information without the variables that you introduce trying to guesstimate things outside, there is no better way.

I am familiar with Allied Aerospace's Low Speed Wind Tunnel in San Diego, but I am not aware of a firm in SD that offers wind tunnel testing for $700. Do you happen to know the name or url for this firm? Chances are that I would apply the $700 (plus travel and time) to a secret weapon I am working on rather than wind tunnel testing, but I would still like to know where I could go for more precise testing once I have ruled out clearly inferior positions with my own, clearly less precise but still functional, approach. Thanks for the suggestion. And, yes, I am serious.

RapDaddyo said:

I am familiar with Allied Aerospace's Low Speed Wind Tunnel in San Diego, but I am not aware of a firm in SD that offers wind tunnel testing for $700. Do you happen to know the name or url for this firm? Chances are that I would apply the $700 (plus travel and time) to a secret weapon I am working on rather than wind tunnel testing, but I would still like to know where I could go for more precise testing once I have ruled out clearly inferior positions with my own, clearly less precise but still functional, approach. Thanks for the suggestion. And, yes, I am serious.

I just threw the latest issue away. I'll see if I can find specifics for you.