easy testing of aerodynamic drag

[Jim Parker]

I recently tested fastaerolab.com which has apparently replaced the online tool by R. Chung that used to be found at cyclingpowerlab.com for using your power meter and a quiet road loop to do Cd*A testing.

Using the old system was quite cumbersome, requiring the use of a DOS shell and converting the FIT file from your Garmin into a spreadsheet, and then manually massaging the data to be compliant with the online app. Once the those time-consuming steps were done, you needed to upload the file and fill out an online form.

With the new website, it is MUCH easier. You just register (for free), select "analyse" and then drag and drop your FIT file straight from your Garmin into the app. It will automatically parse the data based on you laps, though other options are available. You will still need to enter some variable information like weight, air temperature, air pressure, etc.

Here is an example of how I might test a new wheel (though it could be a helmet, handlebar, water bottle, etc.).:

Warm up on the bike and then get up to a comfortable speed of 18 mph while coasting. Hit the lap button and take your speed up to a sustainable but fast pace (I aim for about 23 mph average) because a faster speed is going to reveal aerodynamic differences better. On a quiet loop course, ride for 3 to 9 miles, avoiding using your brakes. Coasting is OK, so coast safely through corners and then get your speed back up.
After getting the mileage in, coast down to 18 mph (approximately) and hit the lap button as you cross the point where you started the lap. You don't have to pedal the whole time or ride at a constant speed.

Now leave your Garmin running and go put your new wheel on, and repeat the above steps. When you are done, hit the STOP button and save the FIT file. This will be the file that you will drag-and-drop at fastaerolab.com.

I did a few laps in my neighborhood on my V20 to get a feel for it. Below is the output. Lap #3 was the most aerodynamic, with a CdA of 0.187.



There are many other ways to use this great app to calculate CdA, but this should get you started.

Jim

 

[McWheels]

I like it. What did you do differently between the laps?

 

[bazzawill]

Have you ever used Autodesk Flow Design
I can most likely get an education liscence to play with. They use it in F1 in schools.

 

[LarryOz]

I feed my data in the new tool for my laps up in Chicago when I did my Century record attempt this past August: (I have added my speed and watts for each lap for comparators). This was the "prototype" Cruzbike rear end, with Kent Pol's seat at about 2% from horizontal



And then again at my "secret test facility" last weekend. It was rather windy so hard to compare each lap. Each lap was 15mins.
Lap 1-3 - regular drop handlebars, laps 5-7- My stubby ones.



You can see how much more aero I was on a nearly flat seat, but it was also a lot harder to make power on that prototype bike. Maybe if I trained with it all year I could do better.

Anyways - pretty interesting to see the CdA differences.

 

[LarryOz]

Have you ever used Autodesk Flow Design
I can most likely get an education liscence to play with. They use it in F1 in schools.

The demo was really cool. Biggest challenge would be to get a good representation of the Vendetta (with your body on it) into the program. Wow. I wonder if they have some tools for that. Maybe videos or pictures from many angles.

 

[murmur]

I think a tool like Fast Aero Lab (or the kind of coast-down testing you've done, Larry) is going to be much more valuable than CFD (Autodesk Flow Design, ANSYS, etc) for getting bike drag numbers. Not that the right CFD approach couldn't get you decent CdA, but for a bicycle I think that would be a good PhD project, because of the complexity of modeling a situation where there's lots of turbulence (i.e. the flow behind the bike, which directly affects drag because it affects the air pressure on your "backside"). A CFD study is fairly straightforward when turbulence isn't a major factor affecting the answers you're trying to get out. Otherwise, it isn't.

Source: Perennial failures-to-deliver on promises of good Cd estimates from CFD (for higher-speed vehicles, admittedly, but turbulence-prone ones) within my company. A large company with its own CFD-development staff.

 

[Jim Parker]

I like it. What did you do differently between the laps?

Removed my helmet.

 

[Jim Parker]

I feed my data in the new tool for my laps up in Chicago when I did my Century record attempt this past August: (I have added my speed and watts for each lap for comparators). This was the "prototype" Cruzbike rear end, with Kent Pol's seat at about 2% from horizontal

And then again at my "secret test facility" last weekend. It was rather windy so hard to compare each lap. Each lap was 15mins.
Lap 1-3 - regular drop handlebars, laps 5-7- My stubby ones.

You can see how much more aero I was on a nearly flat seat, but it was also a lot harder to make power on that prototype bike. Maybe if I trained with it all year I could do better.

Anyways - pretty interesting to see the CdA differences.

Those are incredible Cd*As, but I believe them. Is the second set with the drop bars vs. stubby bars on the stock V20 frame or the proto?

 

[Tigerpaw]

Power measurement from the same pedals you used at White Lake?

 

[ratz]

Where this gets equally interesting; is get two or three riders on Similar bike configuration but optimized for them; Ride around the same track at the same time, but with enough gaps to matter. Feed the data into the program and compare. Why? Because it would be interesting to have Cda Phenotypes for different size riders; from there you could try the same modifications like Add the tail box; use stubby bars, enclosed the under seat triangle; try 40 vs 60 vs 90 wheels, try a thor style seat. Determine for which type of rider does each change provides the biggest bang for the dollars. Far cheaper than a wind tunnels and perhaps more practical to execute and schedule. .....

 

[LarryOz]

Power measurement from the same pedals you used at White Lake?

Yes - but properly calibrated for correct crank length before each test.

Those are incredible Cd*As, but I believe them. Is the second set with the drop bars vs. stubby bars on the stock V20 frame or the proto?

That is from the stock frame, stock front end, curved slider. Only the Chicago Century numbers are using the Prototype - I had the stubby bars on it too.
The other "interesting" and rather "unmeasurable" item with the prototype is the great different in how hard it is to make power laying almost flat. Sure the aero is there, but if you can only make 80% of you normal watts then that cancels out the aero advantage. I believe this to be the case for even an aero upright bike. I have read of testing in wind tunnels of riders in severe aero positions could also not make and maintain power due to the extreme position.
It would be interesting to see if I could train exclusively on the flat seat for a year if eventually I could make the same power as being around 20 degrees. I do not think so at this point based on my past testing and experience between the Silvio and Vendetta. But may different people can adapt differently.

 

[RojoRacing]

Most interesting in @LarryOz's data is the variance in the CDA values which range from 140-160 in one to 160-180 in the other. That's about a 12% accuracy range in both those data runs which isn't very impressive IMO but it's free so beggers can't be choosy. It's understandable his speed, watts and lap time fluctuate from lap to lap but if he's in the same position the whole time his CDA should be constant. Now the wind and air pressure will cause some fluctuations as well as change in rider position due to fatigue, the later not really an issue on a bent with only one position available.

Will this work with my Berrego Springs data from last year? We could pull up the atmosphere data from weather history for that day.

 

[Jim Parker]

Most interesting in @LarryOz's data is the variance in the CDA values which range from 140-160 in one to 160-180 in the other. That's about a 12% accuracy range in both those data runs which isn't very impressive IMO but it's free so beggers can't be choosy. It's understandable his speed, watts and lap time fluctuate from lap to lap but if he's in the same position the whole time his CDA should be constant. Now the wind and air pressure will cause some fluctuations as well as change in rider position due to fatigue, the later not really an issue on a bent with only one position available.

Will this work with my Berrego Springs data from last year? We could pull up the atmosphere data from weather history for that day.

That's probably worth a try.
Braking or stopping during a test lap will yield a higher Cd*A for that lap unless there is some type of smart adjustment based on the FIT data, which includes elevation. I don't know if that's built into the app, but I suspect it is not. Last year at Borrego the crossing guard made me come to a full-stop almost every lap on a nice fast straightaway for NO REASON other than a stop sign with miles of open desert for visibility. At most events the guard will wave you through if it's obviously clear. If you want to use this app during Borrego this year, you might want to hit the lap button before you have to come to a stop. Then after you get going again, hit the lap button again. The app appears to discard data from very short laps, so that is good.

 

[ratz]

Yep I can confirm that one; the math requires that all decreasing is speed come from Cdr and Cda; if you hit the brakes at all in a test segment the data is invalid the math just falls apart. Chung suggests placing a MAT on the group and hitting LAP each time you cross it. If you have a course the requires braking. Then you put two matts down 1 before and 1 after that zone. and Hit the lap button as you cross each. But separate them enough that your re-acceleration zone is in between the matts in the section of data you want to discard. Worst can you can delete those short laps with RWGPS Pro account and have a clean file to feed to the site for Analysis. I use this with golden cheetah aerlab two years ago and got good repeatable results should map forward to the new better tool

 

[LarryOz]

I think wind conditions will affect the result greatly as evidenced by average speed and the wattage it takes to maintain that speed around a track with zero wind compared to a windy day. You always loose speed with windy conditions.

 

[ratz]

I think wind conditions will affect the result greatly as evidenced by average speed and the wattage it takes to maintain that speed around a track with zero wind compared to a windy day. You always loose speed with windy conditions.

Chung has always said to test on days with stable winds <= 2mph...... Pretty sure that happens here about once every 10 years or so.

 

[Jim Parker]

I just discovered that you only get 5 files processed after you sign up with www.fastaerolab.com. After that, you need to buy credits, which are very expensive.
So I will keep looking for free alternatives. Golden Cheetah has an aerolab function, but it will not run on my Windows 10 computer. Apparently others have the same problem with it.

 

[LarryOz]

I just discovered that you only get 5 files processed after you sign up with www.fastaerolab.com. After that, you need to buy credits, which are very expensive.

I saw that too when I did it. You have lots of different emails you can register under, so you might be able to do 20-30 times!
Does that mean the "old" way that you described earlier in the post is no longer available?