Just listened to your excellent podcast - really interesting and I totally agree with Marc about the most important areas of development for aero devices.
I'm Barney Garrood, founder of Aerosensor (
www.aerosensor.tech). As John Buckley (of Streamlines Aero) mentioned earlier in this thread, he and I started development some years ago before amicably diverging last year.
Our initial focus was on an accurate wind speed sensor, able to cover high wind yaw angles. We had both independently seen from our F1 days that traditional multi-hole pitots don't tend to perform well in low speed, high yaw conditions, such as hairpins. We have our theories why. We ended up patenting our probe shape which is now back in F1 running on two teams' cars this year, and as one of their aerodynamicists put it "for the first time ever we can trust low speed track aero".
Our second target was altitude measurement, and we spent an enormous amount of time looking at different ways to do this, tested many different barometers etc. We now have a system that works well, and resolves down to around 10cm. Of course in windier conditions this is more challenging, but in general outdoors over a 1km out and back we can achieve repeats within 1-1.5% on CdA and this is very much down to consistent altitude measurement.
Our third challenge was body position measurement. Many times we have seen rider position drift (typical lift) as they fatigue, which could easily wipe out small changes at the Velodrome. Our "Aerobody" device uses an optical device on the stem to measure head and chest height. Of course this is far from the whole picture but we have found it to be enough to improve consistency by an order of magnitude. Now I wouldn't want to test without it as rider position is potentially the biggest source of error.
We have developed "Aerodrome", which uses a wireless lap trigger at the velodrome to very accurately synchronise lap position. This allows us to reliably account for lean angle (knowing track geometry), and deliver CdA and laptime to the Garmin bike computer at the end of each lap. On an indoor velodrome, over a single lap we can achieve CdA repeatbility better than 1% with this system.
As Marc says, current systems are too hard to use to be an everyday training tool. Our objective is to make an aerodynamic device which can be used like a power meter. Aero testing should be part of every ride. Our devices are supported by a Garmin Connect IQ app, allowing all settings to be set with no need for extra phone apps or software. No post-processing is "necessary" but for those that would like to it is possible in GC, WKO and even excel. No dedicated calibration ride is necessary as the calibration factor is provided with each run - averaging on an out and back or over a lap of a velodrome or road circuit gives good results. This can then be used to post-correct data if absolute accuracy is important, but if all you want to know is which config is faster, then even this isn't necessary. For accurate aero testing I believe out and back testing will always give better results, but the rolling-average CdA value shown on the Garmin during general rides is plenty to see how your consistent your CdA is. We can also tell you other metrics, such as percentage of power consumed by aero, useful to know when it is most important to focus on your aero position, or where it will cost you less to take a breather and hydrate.
We are talking to other head unit manufacturers about getting support on their devices. Since the ANT+ aero profile, that John and I had spent an enormous amount of time on, this side of things has been frustrating.
On our current situation, we are doing an
Indiegogo campaign as a means of pre-selling the first batch of products (with a 25% discount), which is now in its final week. The amount we are seeking is relatively low (£45k) as we have taken substantial investment in the last 6 months enabling us to proceed with tooling for production regardless. Indiegogo is not critical to our success but has been extremely useful to dip our toe in the water and get a better understanding of the market. We will start shipping products in February 2023.
By way of example, a couple of weeks ago I did a test with Matt Bottrill at an outdoor Velodrome (Halesowen), where we tested different aero helmets. Unfortunately it was pretty windy - 10-13mph so the wind was very variable around the track. However it was fairly consistent from run to run. Given the variation in yaw I wouldn't have a lot of confidence in absolute numbers, but relative difference between configs should be reasonable.
For each config we ran 8 x 400m laps. The repeatability was actually excellent (<0.5%) and is as much a testament to Matt's ability to hold his position as as anything else! Here are the results, just showing direct averages of the 8 laps for each "run". Other than this averaging, and application of a constant cp correction factor the data has had no other post-processing.
We did a direct repeat of baseline (runs 1-2), and also repeated the best helmet (runs 3-6).
Below are histograms of wind yaw angle (in degrees) and relative wind speed (positive headwind, negative tailwind) for each run.