Profile 4 Test Anomaly

Post Reply
WilsonT
Posts: 18
Joined: Wed Apr 10, 2019 7:20 pm

Profile 4 Test Anomaly

Post by WilsonT »

Question for John!

After a long hiatus, I finally took the time to try out the latest firmware with the updated profile 4 specific to testing. To test for consistency, I ran 4 consecutive tests using identical setups, same position, down to my head position. The day wasn't entirely ideal, as I had some amounts of crosswind, but the conditions stayed relatively steady during my test runs (the first 4 9ish minute laps I did were the test runs - an out-and-back course that was dead straight with a highway overpass.

On the first 3 runs, everything looked good and consistent, with the CdA started off relatively high as soon as I hit the lap button, and gradually dropping to what would appear to be my final CdA of ~0.222 to 0.225. However, on my 4th run, as soon as I hit the lap button, it gave me a much lower CdA value, around 0.190, and stayed low for the duration of the test run. However, as soon as I hit the lap button to complete the test run, the displayed CdA ended up showing a number that closely matched the test, at around 0.222. Could you explain what might have occurred here?

I've attached my ibr file to look over!

As an aside, I also did a couple of test runs (the last two 5-ish minute runs) wherein it was a different stretch of road - I know it's been said that the lap button must be clicked at the exact same spot, but is there any harm in doing a test run in one location (lap button at one given location), and then doing a couple test runs in a different location, with it's own lap button spot?)
Attachments
Velocomp_06_05_2021_1458_0_Miles.ibr
(300.08 KiB) Downloaded 123 times
Velocomp
Velocomp CEO
Posts: 7793
Joined: Fri Jan 11, 2008 8:43 am

Re: Profile 4 Test Anomaly

Post by Velocomp »

On lap 4 there was a big change in your tilt calibration. Why this happened I do not know, but it is the cause of your low CdA registered during lap 4. At the end of the lap the "error" was detected and corrected, and your on-screen CdA was corrected to 0.222--very close to your other laps.

Lap 1 .225
Lap 2 .221
Lap 3 .223
Lap 4 .222

These are very consistent results!

Isaac does not have access to the on-screen correction factors used in AP firmware, so when you use Isaac "CdA analysis" feature it shows a low CdA for lap 4.

In profile 4, whenever you click the lap button, AP assumes you are in exactly the same physical position as where you last clicked the lap button. It's OK to move to a different location to do testing, but you need to do a warmup lap in that new location, with lap markers, so that AP can readjust its internal settings to the new location.
John Hamann
WilsonT
Posts: 18
Joined: Wed Apr 10, 2019 7:20 pm

Re: Profile 4 Test Anomaly

Post by WilsonT »

Thanks for the reply! That's awesome to know that the device corrects itself! I'm honestly not sure what might have been the cause of the tilt shift, as the aeropod still seems to be in the same position as I had set it in (I tightened the gopro clamp bolt really nice and tight).

After the lap 4 device correction, had I continued doing laps in that spot, would the CdA display have corrected itself, or would it still show the incorrect CdA numbers until the end of the lap? (I probably should've checked this myself but I was running out of time)
WilsonT
Posts: 18
Joined: Wed Apr 10, 2019 7:20 pm

Re: Profile 4 Test Anomaly

Post by WilsonT »

Also, one additional question, where did you grab those Lap CdA values from? I'm trying to compare them to the HR data channels, but I can't seem to figure out the conversion. Wasn't it HR * 4 / 1000, or has that changed with the new firmware?
Velocomp
Velocomp CEO
Posts: 7793
Joined: Fri Jan 11, 2008 8:43 am

Re: Profile 4 Test Anomaly

Post by Velocomp »

WilsonT wrote: Mon Jun 07, 2021 7:23 pm Also, one additional question, where did you grab those Lap CdA values from? I'm trying to compare them to the HR data channels, but I can't seem to figure out the conversion. Wasn't it HR * 4 / 1000, or has that changed with the new firmware?
In profile 4 you get the most accurate CdA measurements. At the end of each lap the average CdA for the just-completed lap appears in the HR channel for the next 10 seconds of the just-completed lap (a flat line of HR).

In profile 4 only, CdA = (HR +180)/1000.

After lap 3, for the next 10 seconds after lap 3 your HR was 43. So, CdA for lap 3 = 0.223
Attachments
Screen Shot 2021-06-07 at 8.12.40 PM.png
Screen Shot 2021-06-07 at 8.12.40 PM.png (321.89 KiB) Viewed 5513 times
John Hamann
WilsonT
Posts: 18
Joined: Wed Apr 10, 2019 7:20 pm

Re: Profile 4 Test Anomaly

Post by WilsonT »

Hi John,
I did another day of testing, and I started off with another calibration ride (I tried to use the "Check Calibration" feature in isaac to export a corrected profile to profile 4, but perhaps it wasn't happy with that). In any case, after doing the calibration ride, all looked good and well, but when I started doing my baseline tests (same position as my previous file), I was getting wildly different numbers - instead of 0.222ish, it was 0.190. During a couple of my test runs, I saw numbers even drop below 0.08! Could you explain what I did wrong?

I've attached my calibration files as well as the ride file
Attachments
Velocomp_06_09_2021_1800_58_km.ibr
(579.12 KiB) Downloaded 135 times
Velocomp_06_09_2021_1754_4_km.ibr
(26.79 KiB) Downloaded 123 times
Velocomp_06_09_2021_1751_2_km.ibr
(15.22 KiB) Downloaded 126 times
Velocomp
Velocomp CEO
Posts: 7793
Joined: Fri Jan 11, 2008 8:43 am

Re: Profile 4 Test Anomaly

Post by Velocomp »

If you look at your two ride files, you'll see that the wind calibration is different from each. On your 6/9 ride you have a net headwind of 1.6mph during your tests (on closed loops it should be zero).
Screen Shot 2021-06-13 at 10.28.30 AM.png
Screen Shot 2021-06-13 at 10.28.30 AM.png (363.68 KiB) Viewed 5430 times
On your 6/5 ride you have a net headwind of 0.1mph (this is an excellent calibration)
Screen Shot 2021-06-13 at 10.29.53 AM.png
Screen Shot 2021-06-13 at 10.29.53 AM.png (421.73 KiB) Viewed 5430 times
When the calibration is "hot" (headwind) then AP watts will be high, forcing CdA down.

On your 6/9 ride, I used the Analyze/Analyze Wind... command to set average wind to 0
Screen Shot 2021-06-13 at 10.40.00 AM.png
Screen Shot 2021-06-13 at 10.40.00 AM.png (340.95 KiB) Viewed 5430 times
After doing this, to see the approximate effect on CdA, I used the Tools/CdA Analysis to check how the reduced watts affected CdA. CdA increases to 0.221, essentially the identical result from your prior test
Screen Shot 2021-06-13 at 10.41.30 AM.png
Screen Shot 2021-06-13 at 10.41.30 AM.png (150.75 KiB) Viewed 5430 times
So, the answer is this: if you have a good calibration, don't change it unless there is a good reason!

CdA measurement is very sensitive to small differences in calibration.
John Hamann
WilsonT
Posts: 18
Joined: Wed Apr 10, 2019 7:20 pm

Re: Profile 4 Test Anomaly

Post by WilsonT »

Thanks for that info!

So I've got a number of what if questions that I've come up with over the last few days of tests...

1) Am I correct to assume that the out-and-back part of the calibration ride is to set the wind sensor and tilt sensor?
2) With the wind sensor setting in mind, if I'm doing my out and back calibration, and I get passed by a whole line of cars, throwing off the calibration, what is the procedure to restart the process?
3) The instructions say that the initial setup requires that we first hold the aeropod button for 4 seconds until the green light starts flashing to pair the powermeter and speed sensor, but it says do NOT pair it to the bike computer until after the lights go off. If I started this pairing process, but my garmin was already turned on (with the Aeropod previously paired to it already from a previous ride), is this going to throw everything off? If so, do I need to restart the process?
4) For subsequent rides with an already-calibrated aeropod, is it okay to turn on the garmin before turning on the aeropod?
5) During test runs in profile 4, during an out-and-back lap, if the test is screwed up because a car passes me then drives slowly in front of me (messing up the wind), can I simply turn around and head back to the starting point, or do I *need* to complete the whole lap distance?

Being as complex a device as this is, there are equally complex questions!
Thanks in advance!
WilsonT
Posts: 18
Joined: Wed Apr 10, 2019 7:20 pm

Re: Profile 4 Test Anomaly

Post by WilsonT »

Further to this, a couple more questions regarding Isaac:

1) In the interest of accuracy, I've been weighing myself with my gear before my test rides - if my aeropod has been calibrated well and I simply want to change my weight, and/or adjust tire pressure, would it just be a matter of clicking "device > Set up device", and then going through the setup procedure, changing weight/tire pressure, but then at the end, clicking "keep accuracy of ______" ?

2) If I've got an a known Crr value for my tires, is there any way to send that directly to the device, rather than going through the setup?
Velocomp
Velocomp CEO
Posts: 7793
Joined: Fri Jan 11, 2008 8:43 am

Re: Profile 4 Test Anomaly

Post by Velocomp »

WilsonT wrote: Fri Jun 18, 2021 11:46 am Thanks for that info!

So I've got a number of what if questions that I've come up with over the last few days of tests...

1) Am I correct to assume that the out-and-back part of the calibration ride is to set the wind sensor and tilt sensor?
2) With the wind sensor setting in mind, if I'm doing my out and back calibration, and I get passed by a whole line of cars, throwing off the calibration, what is the procedure to restart the process?
3) The instructions say that the initial setup requires that we first hold the aeropod button for 4 seconds until the green light starts flashing to pair the powermeter and speed sensor, but it says do NOT pair it to the bike computer until after the lights go off. If I started this pairing process, but my garmin was already turned on (with the Aeropod previously paired to it already from a previous ride), is this going to throw everything off? If so, do I need to restart the process?
4) For subsequent rides with an already-calibrated aeropod, is it okay to turn on the garmin before turning on the aeropod?
5) During test runs in profile 4, during an out-and-back lap, if the test is screwed up because a car passes me then drives slowly in front of me (messing up the wind), can I simply turn around and head back to the starting point, or do I *need* to complete the whole lap distance?

Being as complex a device as this is, there are equally complex questions!
Thanks in advance!
1) Yes
2) You can press-hold the button for 4 seconds, which aborts the calibration. You can then restart it
3) If you've already successfully paired your AP to your Garmin, you can disregard the re-pairing to Garmin
4) I would recommend getting the AP running before starting the Garmin. There are some odd quirks in the Garmin FW that can cause it to be confused when it doesn't "see" the sensors that are paired to it
5) You can turn around and go back. Make sure to click the lap button when you return to the starting point, and click it again when you start the lap over
John Hamann
Velocomp
Velocomp CEO
Posts: 7793
Joined: Fri Jan 11, 2008 8:43 am

Re: Profile 4 Test Anomaly

Post by Velocomp »

WilsonT wrote: Fri Jun 18, 2021 1:00 pm Further to this, a couple more questions regarding Isaac:

1) In the interest of accuracy, I've been weighing myself with my gear before my test rides - if my aeropod has been calibrated well and I simply want to change my weight, and/or adjust tire pressure, would it just be a matter of clicking "device > Set up device", and then going through the setup procedure, changing weight/tire pressure, but then at the end, clicking "keep accuracy of ______" ?

2) If I've got an a known Crr value for my tires, is there any way to send that directly to the device, rather than going through the setup?
1) Yes, your procedure will work.

2) You can use setup or do the following:

A. Connect AP to Isaac
B. Go to Edit/Edit Profiles
C. Click "Get from Device". This will copy into Isaac the current profile you're using
D. Click the "Advanced" tab
E. Click the "Adjust" button to the right of the Riding Tilt % window.
Screen Shot 2021-06-18 at 1.35.47 PM.png
Screen Shot 2021-06-18 at 1.35.47 PM.png (200.39 KiB) Viewed 5315 times
F. Adjust will change to "Hold", and the Crr window will change to "Adjust".
Screen Shot 2021-06-18 at 1.35.58 PM.png
Screen Shot 2021-06-18 at 1.35.58 PM.png (202.89 KiB) Viewed 5315 times
G. Type in the Crr window the value you'd like to use.

H. Click the "Send to Device" button to send the modified profile to your AP, then click "Accept" to remember the changed profile.
John Hamann
WilsonT
Posts: 18
Joined: Wed Apr 10, 2019 7:20 pm

Re: Profile 4 Test Anomaly

Post by WilsonT »

Thanks for the responses! I feel like an "aeropod/isaac advanced users" manual should be made somewhere down the road :D

I'll keep up the testing and probably use this thread as an ongoing results/questions database! Truly appreciate your quick and detailed responses!
WilsonT
Posts: 18
Joined: Wed Apr 10, 2019 7:20 pm

Re: Profile 4 Test Anomaly

Post by WilsonT »

Hi John,
One additional question, in terms of importance for CdA testing accuracy, am I correct to say that the out and back portion (from 0-50W, then 51-70W) is the most important? Whereas 71-100 can be done "badly" (corners, braking, inconsistent position, etc), and it wouldn't have any effect on the actual CdA measurement?

I'm asking because I have a very good quiet road I can do the out and back portion on, but it's not long enough that I can avoid cornering and potentially having to brake for a light or two
Velocomp
Velocomp CEO
Posts: 7793
Joined: Fri Jan 11, 2008 8:43 am

Re: Profile 4 Test Anomaly

Post by Velocomp »

The last 30% of the ride establishes your baseline CDA. If you’re doing stops and turns then the baseline CDA will be possibly wrong, but your core calibration will still be correct. You can correct the baseline CDA afterwards in the profiles
John Hamann
Post Reply