You ever get a chance to test 19 different wheels in that wind tunnel at yaw?

Ha ha, no that wasn’t me! I read that discussion with interest, but I don’t know Sachin Hambini or those guys.

There was suggestion that they work for Airbus, but that’s not true (certainly not in my Aerodynamics department anyway).

Alex-

I was trying to play with your "speed from power" and "power from speed" tools this weekend. I noted that both the Crr and CdA sliders were stuck at "0", and would snap back to zero when I move them. I tried Android Chrome, Windows 10 Chrome, Win10 Edge, and Win7 IE-11. All did exactly the same thing.

They both worked for me last week.

Thanks for the heads up. I'm not sure what's happened there.

Calculoid have been making some changes of late. I'll see what they say.

---

http://www.cyclecoach.com
http://www.aerocoach.com.au

It's working again.

Calculoid said there was some problem at their end which they fixed. I'm still yet to have the default value setting fixed (they should sit mid range rather than at zero) but at least the sliders should be working again.

---

http://www.cyclecoach.com
http://www.aerocoach.com.au

Cool. Thanks. Works for me, also.

I did some position tweak testing this weekend and wanted to confirm the estimated speed improvment at race power....Beyond the usual rule of thumb.

I did my first attempt at field testing this weekend.Didn't have too much time but was able to get a 5 loop uninterrupted run on a small loop in a nearby park, closed to traffic, no one around (before a roller blade guy showed up). It's completely flat and about 525m long as measured by my speed sensor.



Wanted to run through my protocol and see if I'm on the right track here or if there's anything extra I need to do.


1. Measure weight (bike 10.2kg, suit+helmet+shoes 1.8kg, me: 73kg). Total weight 85kg
2. Found a wunderground weather station less than a mile away from the park
3. Using Garmin 520: GPS turned off, using the Garmin hub based speed sensor, disabled auto calibration and set wheel circumference to 2096 (700x23c).
4. Pumped tires to 80psi (HED Jet 6+ with GP4000s 23c and butyl tubes). Assuming a crr of 0.00411 based on this test
5. Zero PM (using crank P2M).
6. Rode a few loops to get familiar with the course and gauge speed in corners etc. Averaged around 190-200W. Kept it in the same gear, no braking, no coasting.
7. There's an actual start line (they do crits on this course). Hit Lap button on Garmin at the start line and rode 5 loops all recorded on the same lap / interval on my Garmin.
8. Download data in Golden Cheetah
9. Calculate air density based on weather data (69F, 100% humidity, 0 wind, 30.01 pressure) -> 1.21821
10. Fire up Aerolab
- set crr, mass, Rho
- Elevation line looks almost flat - as expected - can I just mark the "constant altitude box"?
- move CdA slider to make blue line as horizontal as possible? There are some micro dips in the line but every hard to align.

Question: how do I make my numbers not suck? just kidding... but if you could give me some feedback on my protocol that would be awesome. I plan on going back the following weekends and experiment some more and see if I can get repeatable numbers before I start doing some A/B testing.








EDIT1:
Okay... Read a bunch of the previous threads and I'm adding these findings:
- Fudged air density calculation: correct should be 69F, dew point 69F, pressure 30.1 ==> 1.19816
- I need to loop my Garmin every loop. It helps visualize the VE for every loop of the run and weed out loops with anomalies. I did not do this and ended up with one interval with no way of knowing where each loop starts and ends unless I split by distance.
- Crr of 0.0041 as per the roller test is probably not accurate on the road. This loop while it's not super crappy asphalt is not exactly super smooth either. Problem is that the difference between a Crr setting of 0.004 and 0.006 is the difference between solving for a CdA of 0.277 and 0.240 which is pretty significant. How do I get a better Crr estimate?
- I'm at sea level and since this is a flat loop with a 100m radius, actual elevation changes are probably not important so removing the Garmin elevation data and then trying to flatten the VE graph as best as possible should be the way to go.
- Set Eta to .98 for crank PM

- Will try to mark the course turns to make sure I take similar lines for as much as possible.

---

What's your CdA?

no feedback to your protocol, but I wanted to mention that I've been looking for somewhere to do some testing and the thought of using the piccolo road course had crossed my mind!! i'm very interested in how this goes for you

Yep, happy to share any info :)

---

What's your CdA?

If you know that the loop is 525 m long you don't really need to hit the lap button every lap.

Uncheck the "constant altitude" box.

Thanks!

Any suggestions on how to better estimate or calculate actual Crr? It seems like it's pretty crucial to get an accurate CdA #.

It might not matter that much for an A/B test scenario where I can assume some constant (.004 or .005 whatever it may be) and look at the differences in CdA between A and B. I guess this is ultimately what matters but would be nice to have a semi accurate CdA as well :)

---

What's your CdA?

You are correct in that an accurate crr is necessary for an accurate cda but not necessary for the purposes of A/B testing and getting faster. The simplest way I've found to coax out crr on the road is to climb a very, very steep hill at a low speed. There's another method but it eludes me at the moment.

If you vary the speed across the multiple laps, then only a unique set of CdA AND Crr will both "level" the plot AND result in each lap elevation variation being the same. Alternatively, if you KNOW the actual elevation variation within a lap, that helps too.

See "Shen Method"...http://nyvelocity.com/...ess/the-shen-method/

---

http://bikeblather.blogspot.com/

Yup. Andy Shen's Method works pretty well in this.

The underlying reason is that aero drag varies with the cube of speed while rolling drag varies linearly with speed, so in order to "pry apart" the components due to CdA and Crr, you need to vary speed and power. The wider the range of speed and power, the easier it is to pry them apart. That's why I often include a little ramp up and coast down even on a flat surface (like the one trailerhouse is using), and I'll do that it at least a couple of different places.

Great stuff! I'll look into all of this.

One more question since you mention coasting. Does zero cadence - zero power data points influence negatively the VE calculation. I want to know like how paranoid should I be about this stuff :)

---

What's your CdA?

Actually, it's my opinion that zero power data points help the accuracy and precision of the VE calculations, in that most power meters report zero power fairly accurately AND precisely ;-)

Because of that, I will "soft pedal" (i.e. turning cranks but not engaging freewheel) on the downhill portions of my "half-pipe" style course, with the added benefit from using a PT hub set to "autozero on coasting" of having the PM re-zeroed ("calibrated" in Garmin parlance) continuously throughout the test run. With a crank-based PM, the soft-pedaling should result in the power reading being zero as well (or very close)...if not, then the PM may need to be re-zeroed, or you might want to look at cleaning the drivetrain ;-)

For flat courses, I like to follow the "Shen Method" protocol of continuously increasing speed and power across a few laps, with the final lap having as much coasting as possible.

---

http://bikeblather.blogspot.com/

Awesome, my Power2Max reports zeros and it's supposed to auto zero as well when you coast.

---

What's your CdA?

It seems to me that you should get on the bike and roll it marking the distance for one revolution of a wheel to get an accurate measurement of wheel circumference (i.e., circumference is dependent on more variables than just wheel diameter and tire size, e.g., rim width, tire pressure, etc.)

Good point. Thanks!

---

What's your CdA?

Yes, though some PMs lag when starting up again so you can see the speed increasing even though the power will be zero for a second or two. That's annoying.



Trailerhouse: Your should listen to Tom A.'s suggestions. He gets better precision from testing than I do, damn him.

I just want to note that testing requires a slightly different mindset with regards to power application. It actually takes a bit of practice to intentionally soft pedal and not engage the freehub. Personally I try to avoid stationary legs as a slight change in where one stops pedaling (level vs one leg down), heel drop, etc can really throw one’s results off.

Another tip I would add is to start with doing repeats in the exact same position and setup. I would do at least 3 and preferrably 5-6. Until you've done that you don't really know the precision/repeatability of your testing. This will test evertything from powermeter precision to your ability to hold the same position. Once you know your repeatability you know which kind of changes in drag you should be able to test for - no idea trying to test for differences of ~1W if your precision is 10 times that.

Thank you. I want to make sure my protocol is effective so I won't be making any kind of changes in my position and equipment until I can get some consistent results across different runs and different sessions.

What's a good CV to aim for? I'm thinking <1.5% between runs.

When I was out this weekend I also used the CdaCrr Android app
It looks like the numbers correlate well with Aerolab and that's another control I'll use. What's nice about this app is that it gives you a CdA for each lap when you hit the lap button. In my first run I got a 1.7% CV and as I got used to riding the course I got 1.0% on the second run.

---

What's your CdA?

I found it somewhat useful to do a series of ABBA runs with gross differences like sitting up in the hoods vs. TT position when tuning my protocol. To me this was better for protocol tuning than just doing AAAA runs. YMMV.