Your sensor is severely interfering with what you're trying to measure. Any way you can mount it near a brake? Keep taping further below and then add a "behind the arms" bottle like Sam. As I've said, when you close up the front you get rid of interference drag at the cost of pressure drag. Getting rid of interference drag does not necessarily reduce total net drag on its own. Now pressure drag is simple to deal with, just fill the low pressure zone in the back, with anything. Sam had a bottle there, but you can add stuff to make it better; gels, a taped up bag with your tools/flat kit, anything goes. Bento box is good too and while you're a it stuff a bottle down your jersey placed longitudinally with the flow.

As for placing it on the brake, placing these devices offset to one side or another leads to poor measurement in yaw conditions.

There are pros and cons of several positions as people that have tried placing it on the basebar and a pole two feet out have discovered.

...

What "factor" are you using? Or is that not a Notio? That CdA seems pretty high. I know we're after "what is better than what", but still. Upper 250's?

It's not a Notio.

Road helmet, road clothes, IM type position. It matches what has been measured at A2.

We could have a whole other conversation on realistic CDAs. Underestimating CDA is the new overestimating FTPs :-)

I can't thank you enough for playing along. Yes, I'm two hours east of you. We've been very fortunate with the season, but we're indeed done now.

I think we can all agree that we can't do things in isolation and expect results to be the same for everyone as we are all intricate aerodynamic systems, each with their own specific print of vortices, separation bubbles and resulting pressure maps.

But I am also convinced that we need to follow logical pathways to thoroughly investigate and test ideas in an informed way. Testing without understanding undeniably leads to missed opportunities. Sometimes, you need to do C as well for B to be faster than A. It's physics, there are rules, we just don't understand them all, but when we do, we can make predictions.

Thanks again

So…the results. A few caveats at the end….

A super large data sample of 10 tests 😊 10x5km

All testing was done along the Ottawa River. It’s a 2.5km out and back. Good bike lane, medium-rough roads. We do measure road vibration but in this case did not correct for it. Some traffic, nothing crazy. I used a road helmet.


I am not a small guy at 80kg.

In the tunnel, I typically tested .003 from run to run with no changes.

I typically get very consistent results from day to day after I compensate CRR for temperature as the day progresses. The two days were within 2deg C of each other. I chose two jerseys I know test almost identically. The ability to get consistent numbers day to day is something I find important.


Day 1 was a per my first set of pics, Day 2 was to get closer to your suggestions. I did not have a Bento. I thought I had a water bottle but only had the WD40 and a 4x500ml Sapporo cans. The beer cans were flat at both ends so I went for the WD40.

I normally do something like a AA-BBB-A, often throw out the first test. So for the first day I show A-BBB-A. For the second day, I wanted to leave B setup, so I did AAA-BBB, and throw out the first A.

Day 1



Day 2




The first thing you notice is A and B from day 1 are pretty consistent with day 2. Within my .003 anyways. They are also very consistent with my results from hundreds of tests. I was happy to confirm the placement of the sensor did not impact differently than day 2.

I hesitated to put the sensor on the side because I know it is more impacted by yaw and it’s not great when you need to brake. Wind on day 2 was slightly less and lower yaw so it was ‘ok’. You can see in the underlying data that it is more asymmetrical, but still I have good confidence in the number. We can talk about this a little more later.

While the difference between A and B is small, the pattern are pretty consistent.

The slowest runs were B and they were quite consistent.

A is not taped, B is taped

I suspect that if I did a better job of taping, I would probably get a tie (for me...more later). It also increases my confidence that if I got the fancy pants armrests that cover more my arm I would not lose from an aero perspective and would gain from a comfort perspective.

I’ll bring your attention to the red and green line. Green is an elevation using a high precision barometer. You get atmospheric pressure drift, a bit of noise from passing cars. It’s useable but brings in noise and error. In red is a corrected elevation. Correction is done with fusion of multiple signals from a IMU. The reason you want to keep the device nice and stable is to get good IMU values that will result in better altitude. Perch the device way out in front and as BurnTheSheep has said, you get it bouncing around and introducing error. You will get +2points benefit in airspeed signal quality and -8 in elevation quality. If you are on 0 elevation, silky smooth velodrome, you don’t get the -8, so by all means take the +2. The ultimate sultion is out in clean air and a very stable sensor. But one point you have to look at cost/benefit and value of getting good data by just strapping something on and going for a ride.

To your comments on testing at race pace, I personally test at a pace I can do for a long time. You introduce more error with fatique than the aero difference between 36 and 40km/h. Testing at 40 vs 36 does little for yaw challenges. We measure it, report it and take it into consideration. I confirmed yaw in the position we tried was “ok” for this ride. I was actually lucky, yaw was lower than typical.

First of all let me say these numbers could probably be quite different on someone else and just emphasize the need for each person to test. For example, I am one if the few people where wider pads are a benefit and NOT because of head position. What I have experimented many times is some air flow under the arms around the chest seems to help. We also use IMUs on the body to look at other changes that may have occurred with a change. If Sam tested he probably got very different numbers than me.

I "think" you are in the same neck of the woods as me. We got a surprise two days of riding in Novembers. Snow Flurries forecasted for Sunday, so changes are I cannot test the Sapporo cans as a source of hydration. I will have to ship devices elsewhere for more testing. But it was fun to do and needed to test a few other things.

I finally got around to doing a similar test (no WD40 can, but did have a Bento). I used electrical tape, and wrapped (accept this may not be optimised). Overall my results seem similar to Marc's - taped seems to be consistently slower. This was on a outdoor velodrome, I did runs of AABBAABB, where A was taped.

Delta between A and B was about 0.008 (probably 5 or 6 watts) in favour of B.

I suspect it only works if you have a moto up ahead :-)

But seriously, thanks for the test. I am sure it was cleaner than my setup.

Wouldn't something with that much area like the tape probably be pretty yaw/speed dependent. I'm guessing Sam was riding a lot faster than we likely test at. Which would reduce the yaw angle.

Also, what tilt was this tested at?