Well THAT didn't work!

[DrivenUnder]

Hi all, first post. New PowerPod user, went for a calibration ride on Weds, then a 40 mile club ride, and all seemed well with the PowerPod.

Just went for an AeT ride, a VERY flat 3 mile loop, five laps. Kept my heart rate at 137 very nicely. Since the HR was so even, I would have expected the power to be quite even also, when smoothed a bit.

The problem was the 20 mph side wind.

PowerPod totally botched the job. Side wind absolutely threw it for a loop. I've attached the file, and you can see how badly the wind skews the power numbers. Despite speed reduction on account of the stiff quartering headwind on the back section of the loop, my HR stayed flat, and power should have done also. Instead, it goes through the roof on those sections, practically doubling which is just silly.

I asked Isaac to analyze and correct it, but it failed, and only gave me a marginal correction. So now I don't know what to do. I'm still hopeful that the device can be given better calibration numbers, but I'm not clear on how to do that.

Any ideas out there on how to correct this?

[Velocomp]

It is really tough to evaluate this file:

1) You have huge head and tail winds

2) Though you are making a loop, it is not a true out-and-back ride (the headwind part of the course is different from the tail wind part). So, I don't know if the winds should be symmetrical (they are not)

3) Your tail wind portion isn't as "negative" as the headwind portion is "positive". So, while it may be that your wind scaling is incorrect, I can't be sure

4) It appears that, on the tail wind portion of the course, the tail winds were so strong that, at the bike speed you held (16.9 mph) you still had a net tail wind pushing on you from behind. PP will read low if it is not seeing a net head wind

5) Constant HR is a great idea, but it appears that, at the HR you chose, you were getting a "push from behind" on the tail wind portion of the course.

PP assumes that, in tail wind situations, the rider is moving fast enough such that there is a net head wind. This does not appear to be the case in your ride file; if so, that is why there is power variation.

Said differently: for your constant HR test to work, you'd have to ride faster, at a higher HR, such that your PP always saw a net head wind, even when the wind was behind your back.

[DrivenUnder]

Agreed, the wind numbers are large, but really, anyone who has ever lived or ridden in the midwest would call these middlin' winds. A fairly steady 20 mph quartering wind.

1) Huge is relative. 20 mph is strong but not unusual by any standard.

2) I have already done the out and back ride, and calibrated the PP on a not-so-breezy day.

3) Correct, it was a quartering wind. When behind me, it was still pushing on the side, and the downwind section was more protected that the upwind section.

4) This is what bothers me. I was still keeping a steady zone 2 HR (137 bpm), so the power output should not have been dramatically different. The increased ground speed accounts for the additional benefit of the wind.

5) Not sure what this means. Tailwinds ALWAYS push from behind, and PP needs to calculate that properly. If you look at the laps (starting at exactly mile 4, each is 3.2 miles long) you'll see that the PP shows a: strong tailwind, some slight wind, and strong headwind, as the course is roughly triangular. The wind readings look fine, but the power is totally jacked.

I'm hoping that there is some correction that can be applied. The collected data streams look fine, but the resulting power estimate is way off. I'm not seeing accelerometer data, but that is sampled at a huge rate and would not be meaningful for me to review, I'm sure.

[DrivenUnder]

Sorry, John, missed the last part of your reply.

Unless I'm misinterpreting your statement, it seems like you are saying the PP is somewhat useless on windy days.

What I mean is, there is a very standard method of determining aerobic fitness, and it is to go on a ride with a steady HR in zone 2. You then compare the ratio of power to HR over time, and if that ratio falls during the ride, then aerobic fitness is lacking. It is really one of the most crucial performance metrics for an enthusiast cyclist. Now you seem to be saying that it can't be measured by the PP on anything but bluebird days?

I don't think it should be like that. Even with a tailwind, a cyclist maintaining 17-19 mph in the flats should be generating some power. And with a headwind, I don't think the power should go up so much when the HR is keeping steady and speed is reduced. Is this a scaling factor issue?

Edit:

Played around with the CdA Analysis On Selection tool. If I select a lap and change any of the graphs to Wind wrt Road, it looks like the wind speed data looks better than the ride as a whole. The issue is the PP power estimate is overemphasizing the wind as a whole. Tailwinds are eating up too much actual wattage, and headwinds are adding too much wattage.

[Velocomp]

I guess windy is a state of mind...Riding in steady winds of 10 mph, with gusts of 15 mph or more, is not usual. And turning around and riding at a speed where bike speed is comparable tail wind speed is equally unusual. Yours is, in fact, the most unusual ride file for wind I've seen in 13 years.

Also, yours is the first ride file I've seen where HR was held constant, as opposed to power. I get what you're doing but, for the situation you set (HR = 139) you happened to hit on a situation where tail wind was comparable to bike speed. PowerPod won't do well in that situation; it will underreport power.

What I still cannot tell is if your PP is properly calibrated. Your ride file issue may be one of overall calibration, but I won't be able to determine that until I see a ride file under more normal circumstances.

I don't think PP is over-reporting watts in the head wind; if anything, it is under reporting in tail winds. But that is a consequence of the very strong tail wind.

[Velocomp]

I played around this morning with your ride file a little bit.

The original ride file suggests that wind speed might be over-reported. This would boost watts in the headwind section.

I changed the wind scaling factor in the profile to 0.9, a value that gives approximately equal values of headwind and tail wind in your loops.

With a lower wind scaling the headwind watts drop as expected and out-and-back sections look more "equal".

You had a lot of gusty wind conditions on this ride, so comparing head and tail wind sections is not easy. You have places where tail winds really gust a lot, resulting in near zero net opposing wind. Watts drop in situations of significant tail wind--as they should--because you're riding on the flats where the only opposing force is rolling resistance.

I note that, where there are strong tail winds, there also are HR drops (about 6 bpm) around these low-wind sections. HR is NOT a sensitive indicator of power; it doesn't change much when body load changes, and there is a time lag too.

The loop where wind conditions seem to be least gusty is the loop from mile 15 to 18. On the head wind section you had 106W, with HR of 137 (135 to 140). On the tail wind you had 81W, with HR varying between 130 and 140. Interestingly, at the bottom of this loop, where you go from headwind to tail wind, you had a short power spike of 279W--and your HR jumped from 131 to 144. So, a small variation in HR seems to imply a large variation in power.

In your 3 earlier loops (mile 6 to 15) the wind appears to be much more gusty (or perhaps you're riding around other cyclists); there are places where tail wind speed equals bike speed. PP will not do a good job of measuring power when net opposing wind speed is near zero.

Most certainly I am not saying that PP won't do a good job on windy days. We ride on windy days ALL THE TIME here in Florida...it's how we simulate hills

You might benefit from doing an FTP test. Your average power for this ride is about 91W. If your FTP is significantly above 100W, you can boost your average ride power, go faster in the tail wind sections, (which will improve PP accuracy) and get a better workout.

Modified ride file and profile attached.

[DrivenUnder]

Most unusual in 13 years? There should be some kind of award! Thanks for looking into this, BTW, John. Much appreciated!

A zone 2 ride with steady HR is the AeT fitness testing method Joel Friel has been using since pre-2007, and therefore, it is the one Training Peaks uses also, so that is not so unusual. FTP is helpful, but I'm in base mode right now, so FTP comes later.

I've attached the previous club ride data, and this one seems fine, with the exception of the slow paceline back into town on the flats. We rode into a stiff quartering headwind on that, so the paceline draft was not very effective, but the PP shows negative pressure the whole time. Despite that, the power numbers do seem fair.

Is the issue simply that PP is not dealing properly side winds? When quartering into a side wind, it can't distinguish it from a front wind (how could it?) and when quartering away, the venturi effect makes it see negative pressure. Is the answer to have a side wind profile with correction factors for that scenario?