I get very consistent results at paces in the low HIM to high HIM pace.

I do occasional stretches just above threshold for workout paces. I still get consistent numbers at various paces. All my aero sessions have workout session build into them.

I keep the fart around pace to warm up the tire.

What have you seen that makes you favor race pace ? Unless race is IM of course :-)

Any of you all have any issues with running the most affordable Garmin option versus the nicer models if you only plan to use this for aero testing? Like getting a used Garmin 130 instead of a 500's series?

Most of my use would be post-op anyway using the company's app anyway and GC. Given that I'm not in a rush to get one given the app can work with one.

Thanks.

I do all of 90% of my testing with a 130. The only "downsides" I have experienced so far is limited "auto lap" capabilities and it doesn't show normalized power.

For autolap, it only does autolap on 5km, you can't set 1km for example or lap by position. No big deal.

No Normalized power : I don't care

It supports Connect IQ datafields quite well.

I would have to check if it supports logging of DI2 shifts, which is useful in aero testing

It's nice and small. If you have specific needs just ask and I'll check on mine

I've done all my testing with either the (late, lamented) Little Yellow Power Tap Computer or else (the very recently deceased) Joule. No automatic laps, no Connect IQ. I'd just crunch the data after download (by wire, not wirelessly). I calculate gear changes post hoc.

But I could see the Joule was on its last legs, and it died 2 days ago. Now I have to decide on a replacement. Quickly (as in, probably today or tomorrow). Should I go Garmin or Wahoo? What Connect IQ apps would be useful? If none, I might go Bolt; if some, I might go 530.

For one, there are speed dependencies for both Crr and CdA... though we tend to treat them as constants. Another is because you'll get better CdA results if the aero drag energy is high compared to other components.

Also, ideally:
short out-backs (your runs are your aero calibration)
no cars
ride near the centerline both ways
upslope on both ends (they do *not* need to be steep enough to avoid brakes; this is for getting up to speed more easily)

And an app that will automatically trigger laps based on heading (apex of one turn), uses the odometer to map location, and cuts off the ends/turns (user defined). If you also map the elevation, then you gain the ability to determine CdA as a variable through the run, spot anomalies, and do lots of other cool stuff.

It's a shame Pierre lost interest, because it think he had the best app available, and it was practically free.

Well, that makes my decision easier.

So is your solution to ride at constant, high speed ?

Do you believe doing a test at 36km/h vs 42km/h (which will be huge effort differences) will give a very different A vs B comparison ?

What do you do if you have a headwind/tailwind ? Do you keep you GroundSpeed constant or your AirSpeed^2*GroundSpeed constant ?

You know, if you have wind, speed, gradient, and weight right, then speed dependence in estimated Crr and CdA can sometimes be a symptom of inaccuracies in power. (Typically, speed dependence is a symptom of inaccurate wind measurement; that's why I said 'if wind is right.')

Using the protocol and app I described above, it would be possible for the test section to be done at a fairly constant (race) speed. It isn't possible with current methods and software (AFAIK) without a lot of tedious work. For CdA determination you'd want the Re# to be close to what you expect to see in a race. Clothing in particular can be sensitive to Re#.

If your concerns of testing at various speeds were valid, you wouldn't be able to use constant speed in conditions with say a 4km/h headwind/tailwind. It you are riding at 36km/h your airspeed would be 40km/h in one direction and 32 in the other.

If you decide to go at constant airspeed, you wouldn't be at constant ground speed and your crr/speed concern would surface.

Long story short, it doesn't make a measurable difference at mere mortal speeds. If you are doing 55km/h on a track maybe.

There are a ton of bigger impact issues (yaw for example) that we gloss over and worry about 36km/h vs 42km/h testing

100% agree.

Wouldn't this also be true in a race? Small deviations aren't a problem, but your Re# range in testing would ideally be about the same as you expect in a race.

I'm referring to a shallow climb (little over 1% grade) with a strong tailwind where Crr dependence was observed. I can't rule out another cause, but I had a lot of experience on this course, and there wasn't another easy explanation. CdA dependence comes from Dan Bigham's testing, and pro teams who observed the same and have worked to optimize skin suits for different speeds. One advantage of being really fast is that CdA tends to drop, but Crr goes in the opposite direction.

So, for *most* of the speeds at which we ride, the Reynolds number is pretty stable so CdA is *almost* constant with airspeed. At a certain speed, the Reynolds number can pass a threshold and it can change a lot very quickly, in which case CdA will suddenly change. But it's pretty stable below that threshold. More worrisome is that your Crr also changes. The Reynolds number shouldn't affect Crr. That they both change is a symptom of inaccuracy in at least one of wind speed, or ground speed, or power.

530 works great. The Notio integrates well with it.

So couple of things in the thread :

I asked BurnTheSheep why he goes full tilt in testing and we got on the conversation of Crr/Cda dependency on speed. We can come back to that. My personal belief on "speed to test at" is to go at a pace where fatigue will not be the limiter of how much you test. I will get better data with more runs at lower intensity that being toast after a few runs. Your comment that the more "aero contribution" the better is true. But you can get very good results without having to go close to or over threshold. Also I agree on your comment of apparel possibly being more speed dependant. If I was fine tuning two TT skinsuits, maybe I would confirm A vs B differences at higher speeds. But 90% of my testing is done at HIM watts. Personal opinion there

Yes, Crr/Cda will be impacted by speed. Dan showed this in a thread not so long ago and was right. IIRC his charts were well over 50km/h. Robert explanation is in line with what I have seen with tunnel testing at various speeds. I also saw people explain other speed dependant anomalies with this theory at what I called "mremortal speeds" (40ish km/h), rather than Dan's super-human speeds. They were wrong IMO. In one case it was airspeed sensor calibration issues.

I saw another track case where they were unable to explain speed dependant anomalies so they attributed it to change in speed dependant Crr/Cda. As soon as you properly factored banking into the equation, the problem went away. Alternate sensors helped figure this out.

What I am saying is sometimes (this is not necessarily your case), we need to get to the bottom of what is causing the anomaly rather than use the Crr/Cda theory. For this, the more underlying sensor data you have the better to figure this out.

Im my case the degrading CDA number at higher speeds with tailwind is 98% of the time a windspeed calibration issue.

I can reproduce a similar thing on my favorite section of road. BTW, "favorite" to me means the most challenging to figure out. I have a section of road I love to test on because it creates all kinds of problems. There is a dip and rise in altitude, in a big arc going from West-East to North South, completely exposed, with the wind coming typically from North East and 2 expansions joints that cause a thump-thump in 2 spots. You get huge swings of head wind to high yaw to tail wind in a very short period of time. You see huge swings of instantaneous CDA. You can go from 48km/h tail wind to 34 km/h head wind in 30 seconds. And you can do it out and back to get really interesting results. Doing it in no wind and huge wind conditions allows to pick apart the different issues.

I would gladly accept that as a consequence if I was faster. Key word being "if".

I'm just now starting to get airspeeds at the lower end of the speed range some of the cool suits are made for.

I've definitely some work to do still.

marcag... was that you I heard on the Endurance Innovation podcast? Great interview and great insights.... I recommend people here have a listen very interesting items on aero sensors etc.

yes, that was me.

Garmin 130 plus perfect for an IQ data fields (like Notio data field).
Garmin 530/830/1030 if you need to run an IQ app (like Aerotune).

Here is a pic of a double setup (530 in the center for Aerotune, 130 on the side for Notio). https://hostux.pics/image/uFG[/url]

Pierre,

Have you noticed whether you get different airspeed readings if the Notio is mounted elsewhere (like, at the end of the base bar vs. under one of the aerobar extensions)?

Where did the crr correction option come from? I havent found a great source for how to adjust for temperature, so this would be a nice addition for me.

What software are you using ?

What I currently used was inspired by what Tom A published a while back. It was basically a Percentage as a function of temperature and IIRC 100% was around 20degC. So for example at 18degC CRR would be multiplied by 1.06 (making that number up) and at 28degC by 0.94. Temperature was ambient temp.

I am not sure how he got the specifics, but through experimentation I got something similar. I added an optional vibration factor as well, also found through experimentation.

He would get different values depending where he placed it. The position he chose works well for high hands. It would not work as well if his extensions were less angled.