Real World Aero Testing via Chung Method - Data Thread

[rchung]

RyanH: I'll take a look at the files but I almost certainly can't do it today (real world work has my attention, dammit).

FA63: That's pretty much all it is. If it were complicated I wouldn't have been able to do it. I just solve the equations of motion under an additional constraint: that the potential energy is the same at each point for each lap. If you have an on-bike airspeed and yaw meter (almost no one does) the method can use that info but otherwise I collect all the errors and re-structure them into a diagnostic of goodness-of-fit. That's the VE profile. I guess those are the main differences from previous methods: doing laps to create a "natural" contrast that we can use as an additional constraint, and a way to evaluate goodness-of-fit. There are some minor differences but those are the big two.

[RyanH]

Day 16: Cervelo R5ca* on 404s with 23mm Veloflex Criteriums (Important, See Comments)
Tire Pressure: 90F/100R
Beginning Temp: 70*F
Ending Temp: 72*F
Wind: 0-3 mph SW
Avg HR: 143 bpm
Avg Speed: 22.2
Distance Covered: 15.639
Duration: 42:15
Estimated CdA: 0.3257
Wind Adjusted CdA: 0.3299
Quality Rating (out of 3): 1 (below average)

Ride Data (Strava)
Google Sheets Aero Data

Notes: Rapha 2015 Pro Team Jersey. Laps 3,4,5 all saw periods of 15-30s of interference with cars.

Test Setup:


Chung Method Aero Analysis


Atmospherics (beginning of session)


Comments:
Unfortunately, I think I have to invalidate all previous runs. I made the mistake of collecting the data using GPS, and I'm hoping the variation we've been seeing is due to that. I also think there were some mechanical issues with the load on the BB when swapping the Quarq around. Today's run was collected with a GSC-10, but I also did the run a lot later in the morning that usual, so there was more than usual traffic interference. I'm introducing a rating system: 0 (discard), 1 (below average), 2 (average) and 3 (above average) for the quality of the run.

Today's run should be an outlier due to the amount of traffic that I had to deal with.

* This session's power data came from an SRM vs the other bikes are using the same Quarq.

[RyanH]

I'll be off this week for a business trip, so no updates. This was a huge training block for me, 375 miles last week and 21 hours of riding; 555 miles in the past 10 days. Legs were pretty fatigued today, so this was a nice way to cap the block off:

[fa63]

Thanks for sharing this with us. As in all things science, often the first attempt ends up serving as a trial for future studies Anyways, I think the speed sensor will help. Another thought is using a wind/yaw-meter on your second round of tests, if one is available. I wouldn't mind pitching in some money to help the cause.

[rchung]

[b]Day 16
[...]
Unfortunately, I think I have to invalidate all previous runs. I made the mistake of collecting the data using GPS, and I'm hoping the variation we've been seeing is due to that. I also think there were some mechanical issues with the load on the BB when swapping the Quarq around. Today's run was collected with a GSC-10

Those two files above, those were with GPS speed?

Yeah, GSC-10 is likely better. Just for comparison, can you post this GSC-10 file, too?

[RyanH]

Hi rchung,

Yes, those two were GPS. Here's the GSC-10 run from yesterday:

https://drive.google.com/file/d/0ByYQAm ... sp=sharing

[beeatnik]

Ryan, you're a madman.

[RyanH]

GSC10 Run 2: Cervelo R5ca* on 404s with 23mm Veloflex Criteriums
Recording Device: SRM PC8
Speed Sensor: GSC-10
Tire Pressure: 90F/100R
Beginning Temp: 73*F
Ending Temp: 73*F
Wind: 0-3 mph N
Avg HR: 136 bpm
Avg Speed: 21.5
Avg Power: 226 watts
Distance Covered: 15.639
Duration: 43:37
Estimated CdA: 0.3305
Wind Adjusted CdA: 0.3305 (some wind adjustment it probably warranted but it would be a WAG)
Quality Rating (out of 3): 2 (below average)

Ride Data (Strava)
Google Sheets Aero Data

Notes: Rapha 2015 Pro Team Jersey. Laps 3,4,5 were pretty close to no wind. My legs did not want to cooperate at all today so laps 3/4 were done at Z2. Laps 1/2 had a North based wind of up to 3 mph.

Test Setup:


Chung Method Aero Analysis


Atmospherics (beginning of session)


Comments:
Considering the first two laps saw a headwind at around 3mph, which would warrant a 4% adjustment if it was constant throughout the entire session, I think this data point is in agreement with the previous run with the GSC-10. We'll see if this is a coincidence, but the distance for both runs was exactly 15.639 miles.

* This session's power data came from an SRM vs the other bikes are using the same Quarq.

[RyanH]

Thanks for sharing this with us. As in all things science, often the first attempt ends up serving as a trial for future studies Anyways, I think the speed sensor will help. Another thought is using a wind/yaw-meter on your second round of tests, if one is available. I wouldn't mind pitching in some money to help the cause.

Yeah, it's frustrating but it was a risk I knowingly took hoping the GPS data would be good enough.

Regarding the Wind/Yaw meter, I'd love to run one assuming it isn't too cumbersome to bring around (meaning, it won't require me to drive to the Rose Bowl). Any idea who sells them?

[fa63]

Something like the iBike Power Pod maybe? It is supposed to measure wind speed:

http://www.powerpodsports.com/powerpod-features/

That said, I am not sure if it allows you to download the wind speed data after a ride. Might be worth asking the manufacturer.

[wingguy]

b. The Cavendish - Kittel example is not suitable for conclusions. We know nothing about the technical-physiological data of the two rider-bike systems. Certainly it is not the S5 that gives the advantage. Even if you isolate the CdA then the decisive factor would be the lower CdA of Cav's body in comparison to Kittel's body.

The problem there is when you start talking in terms of the advantage. There's not just one advantage, there's plenty of potential advantages. What is the major factor that allows Cav to be so fast with such relatively low power figures? Of course, it's the body position. But what else has an effect on his ultimate speed on any given day - skinsuit vs traditional jerseys? Yep. Deep wheels vs 32h box? Yep. Aero frame vs non aero? Yep. Helmet choice? Yep.

When it comes down to a photo finish decided by a rim's width (or tyre, ref Kittel vs Coquard) then there really is a 'straw that broke the camel's back' situation. How many watts of equipment drag can you add back onto the winner before he doesn't make it?

[kgt]

In terms of aerodynamics the rider's body has a multiple times bigger impact than the frame. We all know that. So when comparing two different riders the frame's cda impact is really negligible. That's what I am saying.

BTW 1st, 2nd and 3rd in the Tour (the three fastest riders this is) were all on non aero frames. Unless you consider F8 an aero frame, no problem.

Anyway, I am really waiting to see this thread's outcome. So far the fastest bike is the one with the less aero frame and that's strange even for me...

[wingguy]

In terms of aerodynamics the rider's body has a multiple times bigger impact than the frame. We all know that. So when comparing two different riders the frame's cda impact is really negligible. That's what I am saying.

Yes, but how negligible is negligible when it comes down to 2.5cm in a 70kph sprint? How many watts of frame drag does it take to change that outcome?

BTW 1st, 2nd and 3rd in the Tour (the three fastest riders this is) were all on non aero frames. Unless you consider F8 an aero frame, no problem.

Tour magazine considers the F8 a very good aero frame. But A) none of these guys won in a sprint. B) none of them were riding much, if at all, off the front except Froome to Bagneres and Bardet to Saint-Gervais and C) Froome took a big chunk of his lead on a stage where he chose a full aero bike / wheel setup and none of the other GC contenders did.

[RyanH]

@fa63, rchung said above it wasn't accurate enough for aero testing. I'm waiting for him to weigh in since I believe they have developed their own.

@kgt, since we're kind of starting from scratch, it's too early to tell, but I think it's possible that the Cervelo will end up being the fastest. The drop is 1cm more but I think that puts me in a more arms bent position than the other frames, for some weird reason. This feeds into your position can matter more than the frameset, which underscores the importance of making sure the frame will offer the same position. The Scott Foil, I don't have a -17 stem option so I can't go any lower. We'll see.

P.S. It'll be awhile before we can probably draw any conclusions. I'll focus on just the R5ca with 404s and the Scott with 404s so we can amass as many data points in as short of time period as my legs permit.

[kgt]

Yes, but how negligible is negligible when it comes down to 2.5cm in a 70kph sprint? How many watts of frame drag does it take to change that outcome?

But even then you don't even consider the stiffness of the frames, the cda of the wheels, the grip and rr of the tires, the clothes, the shoes, the bars, the helmets, the total energy lost. You just attribute those 2,5cm to the cda of the S5. How do you know? In theory you may be right but in practice it does not work like that, especially in a sprint where the frame moves in multiple directions.

Tour magazine considers the F8 a very good aero frame. But A) none of these guys won in a sprint. B) none of them were riding much, if at all, off the front except Froome to Bagneres and Bardet to Saint-Gervais and C) Froome took a big chunk of his lead on a stage where he chose a full aero bike / wheel setup and none of the other GC contenders did.

The first three riders of the Tour are the three faster riders of the Tour, period. What's the point of isolating a tt here and a sprint there?

Anyway we should also wait for RyanH's test results.