CdA readings from tests

Post Reply
aeropod.user
Posts: 20
Joined: Sat Jul 20, 2019 2:08 am

CdA readings from tests

Post by aeropod.user »

Hello,

last weekend I conducted calibration + 4 out/back tests with Profile 4. In general I wanted to compare two helmets, both tested in the same TT position and bike setup.

I send results of this session attached to this post. Every test was approx. 9,5km in distance. I did calibration ride in helmet 1, and then:
Test 1: helmet 2
Test 2: helmet 1
Test 3: helmet 2
Test 4: helmet 1

I have 2 questions to results, which I cannot answer myself. Maybe you can help me:
1) In all tests I see the same pattern: lower CdA when "Wind wrt Road" is positive, and higher CdA when "Wind wrt Road" is negative. The difference is on the level of 0.03. Why is it so? Shouldn't my CdA be more independent from wind direction?
2) Result were very consistent for rides with helmet 1, each ride (calibration + test 2 + test 4) gave me CdA on the level of 0.23. For helmet 2 results were not so consistent although I did my best to keep the same position and similar power. Test 1 gave me CdA of 0.213, test 3 resulted with 0.228. Do you see something in the file, what could affect this difference?

Thank you in advance for your hints and answers.

Bartosz
Attachments
calibration_and_tests.zip
(193.03 KiB) Downloaded 125 times
Velocomp
Velocomp CEO
Posts: 7793
Joined: Fri Jan 11, 2008 8:43 am

Re: CdA readings from tests

Post by Velocomp »

This will be a long post but hopefully also an informative one!

When using AeroPod in profile 4 mode (lap testing), one of the things that happens at the end of each closed lap is that the tilt correction is fine-tuned automatically by AeroPod firmware. The fine-tune correction is applied to the measurements of the just-completed lap. The tilt-corrected CdA is recorded in the ride file HR window, and is displayed during the first 10 seconds of the next lap. Additionally, the tilt correction is applied to the next lap. This means that, starting with second closed-loop lap, tilt corrections should be less and CdA readings should be more consistent.

In your ride file Your first full closed-loop test lap I starts at the 2 mile mark of your ride. I call this “Lap 1”. There are four closed-loop laps in your ride file.

The tilt-corrected CdA at the end of each closed-loop lap test shown in the HR channel of the ride file, for the first 10 seconds following the end of the lap. CdA = .001 * (180+ HR reading for first 10 seconds after test lap ends).

Here are the tilt-corrected HR/CdA values of CdA for your tests

Lap 1 CdA 0.229 Helmet 1
Lap 2 CdA 0.219 Helmet 2
Lap 3 CdA 0.227 Helmet 1
Lap 4 CdA 0.225 Helmet 2

These CdA results are not as consistent as they should be. Why?

At the end of each closed lap, the average wind speed should be the exactly same as average bike speed. However, with current versions of firmware, AeroPod does not automatically apply any wind correction at the end of each lap. So, this is a potential source of error in CdA measurements.

If you look at your ride file, you’ll see that average bike speed and wind speed are not the same for each lap (as they should be):

Lap 1 Bike Speed 25.9 Wind speed 25.5
Lap 2. Bike Speed 26.0 Wind speed 25.3
Lap 3. Bike Speed 26.2 Wind speed 25.1
Lap 4. Bike Speed. 26.8 Wind speed. 25.6

What happens if average wind speed is not the same as average bike speed? If reported wind speed is less than bike speed, then opposing watts will be low, forcing CdA up. If wind speed is greater than bike speed, then opposing watts will be high, forcing CdA down.

So, for the most accurate and consistent CdA measurements, it is important that wind speed corrections be applied at the end of each lap.

The next paragraphs show how to MANUALLY correct CdA measurements for wind speed differences. This is a complicated procedure; that said, these correction procedures will soon be made automatically in AeroPod firmware.

Here is what I did with your ride file:

1) From your ride file I got the tilt-corrected CdAs for each lap. Those values are listed above
2) Using the “File/Combine-split” command, I created four separate ride files, one for each lap.
3) For each lap, I used the Tools/CdA Analysis command to determine the Isaac-calculated CdA for the lap. For example, for lap 2, the Isaac-measured CdA is 0.233
Screen Shot 2021-08-16 at 12.56.54 PM.png
Screen Shot 2021-08-16 at 12.56.54 PM.png (155.46 KiB) Viewed 2713 times
4) I then used the “Analyze/Analyze Wind” command to adjust wind speed. When I open this window for Lap 2, it shows that average ground wind speed for the lap is -0.7 mph less than average bike speed (25.3 vs 26.0, highlighted in black).
Screen Shot 2021-08-16 at 1.04.20 PM.png
Screen Shot 2021-08-16 at 1.04.20 PM.png (289.3 KiB) Viewed 2713 times
5) In the Avg Ground Wind window at the bottom of the page, I adjust the offset from -0.8 to -0.1 (increase wind speed by 0.7 mph). Note that Average wind speed is now 26.0, and watts have increased from 351 to 366
Screen Shot 2021-08-16 at 3.24.03 PM.png
Screen Shot 2021-08-16 at 3.24.03 PM.png (348.19 KiB) Viewed 2713 times
6) I click Accept, then go back to “Tools/CdA Analysis”. CdA drops from 0.233 to 0.221. Why? The average opposing wind is higher, forcing opposing watts to be higher, forcing CdA to drop accordingly.
Screen Shot 2021-08-16 at 3.25.51 PM.png
Screen Shot 2021-08-16 at 3.25.51 PM.png (96.53 KiB) Viewed 2713 times
7) So, the CdA effect of the wind speed correction is to reduce the CdA by 0.233 to 0.221, or by 0.012

8) Finally I apply the wind-CdA correction of -0.012 to the Lap 2 value of 0.219. The corrected wind-corrected CdA is 0.219 – 0.012 = 0.207

I know this is a complicated process. The good news is that we are in the final stages of automating it.

I applied the above steps to all four of your laps. Here are the results:
Screen Shot 2021-08-16 at 3.31.45 PM.png
Screen Shot 2021-08-16 at 3.31.45 PM.png (45.65 KiB) Viewed 2709 times

As you can see, the only outlier CdA is lap 1. Why is this? When doing CdA testing, the first out-and-back lap may have more tilt correction than the subsequent laps. In other words, for the first test of any day, it’s best to do two laps, before changing things. So, laps 2-4 are the ones to investigate.

Laps 2-4 show a very consistent CdA of 0.207. Maybe lap 3 is 0.001 higher (different helmet) but more data would be needed to zero-in on the difference.

The wind correction feature will soon be implemented in a firmware update, getting rid of all of this manual manipulation.

But in the mean time, you can see the kind of consistency that is possible!
John Hamann
aeropod.user
Posts: 20
Joined: Sat Jul 20, 2019 2:08 am

Re: CdA readings from tests

Post by aeropod.user »

Hello John,

as always - I'm very impressed with the level of details and clarity you provide in your answers. Thank you for that!

I'm really happy that you are going to release related firmware update soon, as - from what I understand - it will bring Aeropod to higher level of usability in field testing. If you needed some more real-life testers of this feature before firmware release, I would be more than happy to help!

Regards,
Bartosz
jules.b
Posts: 16
Joined: Thu Feb 25, 2021 3:06 pm

Re: CdA readings from tests

Post by jules.b »

Likewise!
Velocomp
Velocomp CEO
Posts: 7793
Joined: Fri Jan 11, 2008 8:43 am

Re: CdA readings from tests

Post by Velocomp »

jules.b wrote: Wed Sep 29, 2021 5:02 pmLikewise!
We have been working on this firmware. The issue is, surprisingly enough, Isaac compatibility. We want Isaac to provide the same answers AFTER the ride that AeroPod provides DURING the ride.

This has turned out to be a surprisingly difficult thing to do, but we think we have it sorted.

Stay tuned...
John Hamann
Post Reply