Work Flow for using Aeropod

Post Reply
BillOsler
Posts: 116
Joined: Mon Apr 04, 2016 10:10 am

Work Flow for using Aeropod

Post by BillOsler »

There are so many documents with instructions and information that I've long since lost track of what is actually current and what a "best practice" might be for using the Aeropod.
What I'm trying to sort out is the best sequence for certain steps in the process of calibrating and using Aeropod. It would be helpful if all of this were available in a single location.
There are a couple of points that seem unclear to me despite having reviewed what I believe to be the most recent versions of the various documents.
When I am not doing CdA testing: The instructions state that the first 5 minutes of the ride should be done at a fairly constant pace with no sharp turns. That 5 minutes is used to calibrate or zero-offset the accelerometers. Does it matter whether or not the DFPM has been "calibrated" prior to starting the initial 5 minute process for the Aeropod? Or, put another way, does the Aeropod use the DFPM data as part of its calibration check during that first 5 minutes? Since DFPMs can be somewhat temperature sensitive I usually ride for a while prior to doing the zero-offset procedure and I don't know whether I should turn the Aeropod on when it is convenient to do so or whether I should wait until after the PM zero offset is done.
When Aeropod was released it did continuous assessment of the accelerometer offsets so that if the Aeropod's tilt changed during a ride the calibration would be reset within several minutes after the change. The the firmware was changed to fix the accelerometer calibration check during the 1st five minutes and not do continuous reassessment of the calibration. Is it still the case that the accelerometer calibration settings are fixed after the first 5 minutes of each ride and not adjusted to compensate for any subsequent change in the Aeropod tilt?
Is there any way to test/adjust the wind speed sensor calibration/zero offset?
Velocomp
Velocomp CEO
Posts: 7793
Joined: Fri Jan 11, 2008 8:43 am

Re: Work Flow for using Aeropod

Post by Velocomp »

During the first 5 minutes of any new ride, Tilt calibration is being checked. In profile 3 (CdA Trends), the most accurate tilt calibration measurement will happen when speed is fairly constant and there are no sharp turns. Of course, sometimes this is not possible; if so, in profile 3 tilt calibration will continue to be adjusted for the remainder of the ride, just at a slower rate. In this case it can take up to 12 minutes total for the tilt calibration to stabilize. But it will eventually settle-in.

During the 5 minute AeroPod calibration period, DFPM readings are not important.

In profile 3, accelerometer calibration is now continued to be updated slowly after the first 5 minutes. If there was a "problem" during the first 5 minutes, or if for some reason there is a tilt drift after the first 5 minutes, this will be corrected.

In profile 4, accelerometer calibration is locked-in at the end of the first 5 minute period; however, each time a lap is completed afterwards, the calibration is updated. This allows highest accuracy CdA readings. Note that if the lap button is touched in a place other than the location of the true lap-mark, it can make CdA measurements much worse.

You can make wind adjustments in Isaac, but the only way to do it on the road is through a new cal ride. For what it's worth, I have 3 AeroPods on my bike, and I have not touched their wind settings in months.
John Hamann
BillOsler
Posts: 116
Joined: Mon Apr 04, 2016 10:10 am

Re: Work Flow for using Aeropod

Post by BillOsler »

Thank you for the information.
Would it be correct to conclude that in Profile 4 if one is doing an out/back calibration ride instead of laps the lap button should be pressed only at the beginning and end of the full out/back and should not be used to mark the turn-around point?
If the lap button is pressed at times other than a full lap during a Profile 4 ride is there a way to correct the resulting erroneous tilt calibration results during post-processing in Isaac?
Velocomp
Velocomp CEO
Posts: 7793
Joined: Fri Jan 11, 2008 8:43 am

Re: Work Flow for using Aeropod

Post by Velocomp »

For the cal ride, in any profile, there is no reason to click the button at the turnaround point. I don't think clicking it at the turnaround point OF THE CAL RIDE ONLY would have any effect, but I have never tried this.

After the cal ride, in profile 4, the lap button should NOT be clicked at the turnaround point of an out-and-back; it should only be clicked when you get back to the starting point of the lap.

If you are in profile 4 and you click the lap button at places other than the correct lap location, bad things will likely happen to your on-the-road CdA readings, and also to your ride file. You can fix the mistakes after the ride in Isaac with the "Analyze Route" command, but your CdA/HR data will be bogus.

Profile 4 is designed for CdA Testing on closed loops (like a velodrome) or out-and-back routes. If you're not doing CdA Testing according to the instructions of profile 4, then I recommend that you ride in profile 3.
John Hamann
BillOsler
Posts: 116
Joined: Mon Apr 04, 2016 10:10 am

Re: Work Flow for using Aeropod

Post by BillOsler »

Thank you for the information.
In my case I don't click the Aeropod button at all ... but for reasons unrelated to CdA data collection I do use the lap button on my Garmin head unit to help me keep track of events during the ride. That does insert a lap marker into the Aeropod file. If that causes a re-calibration of the accelerometers then there are some things I'll need to re-think in terms of how the testing is done.
One example that is relevant: In order to merge data from my head unit with data from my Aeropod without data loss I use lap markers to help with syncing the data. Isaac does a reasonably good job of syncing my Garmin data with my Aeropod data, but it does not preserve all of the data from the Garmin (eg: pedaling dynamics, HRV data) and tools like Golden Cheetah do not reliably sync the data if start/stop times do not match so having one or more markers to aid in synchronizing the data streams is useful.
Bogus real-time CdA data is not a concern: I have not found the real-time CdA useful since it is so heavily influenced by environmental conditions, especially head/tail wind. Those things average out during out/back or closed loop circuits (obviously) but they make the "real time" CdA less than useful since it is hard to be certain whether a change in the displayed CdA represents a meaningful change.
Velocomp
Velocomp CEO
Posts: 7793
Joined: Fri Jan 11, 2008 8:43 am

Re: Work Flow for using Aeropod

Post by Velocomp »

If you're using Garmin lap markers in profile 3, there is no problem. They have zero influence on accelerometer calibration.

Clicking the AP button does not add a lap marker.

I agree that real-time CdA is more of a time-averaged quantity. You can get useful trend data, even when not on closed loops, but the data needs to be averaged over a period of a few minutes.
John Hamann
BillOsler
Posts: 116
Joined: Mon Apr 04, 2016 10:10 am

Re: Work Flow for using Aeropod

Post by BillOsler »

I have been puzzling over how to integrate all of this information into a workflow.
When I want to do CdA testing I need to be able to change equipment or make an adjustment to the bike between tests.
All of the available locations that are suitable for making those changes are unsuitable as locations at which to begin a CdA measurement because of atypical road surface conditions, need to negotiate turns in the roadway or whatever else.
That means if (for example) I have a test ride with wheel A and then want to test wheel B there is some riding involved between the test location and the location of my stashed equipment. If I'm riding from home it's about 1.5 miles each way, somewhat less if I park my car in the lot at a nearby school. If I go to a nearby track my equipment is still a couple of hundred yards away from the track.
I've been using lap markers on my Garmin to help to identify the actual test intervals.
The process is problematic since Garmin laps do not necessarily appear in the Aeropod ride file at exactly the right time point but it at least gives some information.
If riding in Profile 4 (or in Profile 3 for that matter) and using lap markers in this way weirds the calibration then it becomes challenging to identify the beginning and end of the actual tests.
How then should I go about obtaining meaningful data without weirding the calibration?
Post Reply