Puzzling results

Post Reply
BillOsler
Posts: 76
Joined: Mon Apr 04, 2016 10:10 am

Puzzling results

Post by BillOsler » Mon Apr 15, 2019 6:06 am

I'm still having a hard time getting results that make sense when I do CdA testing.
Yesterday I did some comparisons between two bicycles, my road bike with clip-on aerobars and my tt bike. The road bike has shallow road rims (Velocity A23 rims, Industry 9 hubs) and the TT bike has aero wheels (Zipp 404 front, 808 back).
I used the same TT helmet and wore the a skin suit for all of the tests.
The course is an out-and-back.
The weather report at the beginning showed wind speed of about 6-7 mph and it did not feel gusty though I'm not sure the wind speed was correct; I don't have a way to measure that independently. It felt like the wind increased gradually during testing, again I don't have a way to measure.
The bikes both use C1 chainring DFPMs that appear to correlate well when compared against a 3rd DFPM (either P1 pedals or my trainer; I strongly prefer Speedplay pedals but the P1s have their uses).
Immediately prior to each of the activity files I have attached I did a full calibration ride on the same road, though the calibration ride included a longer section of the road so the final continuous riding portion could be done without braking or turning around. Immediately prior to the calibration ride I re-set the zero offset of the DFPMs.
The CdA estimates were obtained by selecting the relevant lap and using the average HR reported by Isaac.
The road bike activity files are limited to data from the actual test.
The TT bike activity file has 5 laps, laps 1,3 and 5 are the CdA tests. Lap 3 differs from all the other test laps in that I wore a jersey over my skin suit to see if I could measure the difference.
Other than the change of bikes and the use of the jersey everything under my control was exactly the same.
The results do not make sense to me:

Road with Clip-on
0.294

TT bike (lap 1, no jersey)
0.277

TT (lap 3, with jersey over skin suit)
0.264

TT (lap 5, no jersey)
0.270

Road with Clip-on
0.209

On the TT bike I did lap 5 as a measure of the stability of the estimates from Laps 1 and 3. On the road bike the second CdA measurement was used for that same purpose.

I know from extensive experience that I am significantly faster on the TT bike than on the road bike even when I use the clip-on aerobars on the road bike so the 5th test (repeat test of the road bike) makes no sense at all and the difference between the two road bike measurements is huge.
It's hard to imagine that wearing a jersey makes me more aero.

Is there an obvious problem with either the ride files or the protocol?
Attachments
TT_Bike.ibr
(475.14 KiB) Downloaded 1 time
Road_Bike_AeroBars_2.ibr
(136.46 KiB) Downloaded 1 time
Road_Bike_AeroBars_1.ibr
(138 KiB) Downloaded 1 time

Velocomp
Velocomp CEO
Posts: 5747
Joined: Fri Jan 11, 2008 9:43 am

Re: Puzzling results

Post by Velocomp » Mon Apr 15, 2019 11:03 am

It looks to me like you did your first road bike test at around 730AM. Then, you moved AP to your TT bike and did testing at 811 AM; then, you moved AP back to your road bike at 911AM for a second set of road bike tests. Each time you did a new DFPM calibration and AP calibration. Is this correct?

I assume you used the same ride position and equipment for road bike test 1 and 2. Is this correct?

I compared the profiles between road bike 1 and 2; the wind calibration is essentially identical, which is a good thing!

Assuming you used the same ride position and equipment for road bike test 1 and 2:

On Road test #1, your average speed on the return leg was 33.1 km/h, with an average headwind of 2.3 km/h, and average DFPM watts were 198.2W

On road test #2, your average speed on the return leg was 29.5 km/h, with a 4.7 km/h headwind, and average DFPM watts were 175.0

So, on the same course, assuming the same ride position and equipment, because of the reduction in bike speed in test 2, your DFPM watts decreased by 23.2W

I used http://www.aerocoach.com.au/power-from-speed/ to estimate the power differential caused by the change in bike speed and wind conditions. I used a constant CdA of 0.29 for the calculations

the website estimates 189.4W for 33.1 KM/h with 2.3 km/h headwind (test 1) (vs 198.2 DFPM measured)
Screen Shot 2019-04-15 at 11.05.19 AM.png
Screen Shot 2019-04-15 at 11.05.19 AM.png (173.24 KiB) Viewed 55 times
the website estimates 157.4W for 29.5 KM/h with a 4.6 km/h headwind (test 2) (vs 175.0 DFPM measured)
Screen Shot 2019-04-15 at 11.01.28 AM.png
Screen Shot 2019-04-15 at 11.01.28 AM.png (176.25 KiB) Viewed 55 times
So, according to the website, decreasing your bike speed in the second road bike test should reduce DFPM watts by 32.0 W

There is a 6.8W DFPM watt discrepancy between road bike test 1 and 2; that is, the website predicts your DFPM watts should have decreased by 6.8W more watts than your measurements actually report. Said differently, your DFPM watts on test 2 were 6.8W HIGHER than predicted. Higher than predicted DFPM watts would force measured CdA down.

I did an estimate, reducing your test 2 DFPM watts from 173W to 166.2W (6.8W less). The AP CdA resulting from a 166.2W DFPM reading would be about 0.294--identical to your test 1 result.

The lesson here is that EVERYTHING must remain unchanged from test to test, including DFPM calibration.

It looks to me like your DFPM calibration changed from road bike test 1 to test 2.
Attachments
Screen Shot 2019-04-15 at 10.28.20 AM.png
Screen Shot 2019-04-15 at 10.28.20 AM.png (192.95 KiB) Viewed 55 times
John Hamann

BillOsler
Posts: 76
Joined: Mon Apr 04, 2016 10:10 am

Re: Puzzling results

Post by BillOsler » Mon Apr 15, 2019 12:35 pm

Thanks for working on this.
You are correct that I moved the Aeropod twice.
That meant that I did 3 calibration rides (road bike first, TT bike second, road bike again) and I "calibrated" the DFPM (ie, re-zeroed it) immediately prior to each ride. I did not note the zero offset reported when I reset that so I do not know whether it changed.
My ride position was as close to identical for each ride on each bike as I can make it, though fatigue could have had some impact.
Clothing et cetera was all unchanged. In other words everything under my control was unchanged. Even the helmet positioning on my head was unchanged since I did not remove or adjust the helmet at any time during the testing
A quick estimate suggests that at 200 W and 80 RPM a difference of 7 W represents a difference in torque of about 1 Nm.

Are you suggesting that I should not have re-zeroed the DFPM before the second ride?

Even if there were a difference between the two road bike rides the results for laps 1,3 and 5 on the TT bike are puzzling to me.

Post Reply