Temperature drift with the PowerTap

Triathlon Forum
1

Hi all

I was hoping for some advice surrounding a recent FTP measurement.

I have a PowerTap and a Fluid 2 trainer, and I use TrainerRoad (just finished Sweet Spot Base I). Yesterday evening I performed another FTP test (20 minute one) to help reassess my fitness for my next TrainerRoad program.

I was very happy with my execution during the test. I knew I had made some good gains in the last couple of months (I am reasonably untrained), but where I was expecting a new FTP of perhaps 240 (best case scenario), I instead pulled out an NP of 275 for the 20 minute interval, with TrainerRoad suggesting a new FTP of 260.

As happy as I was with that, I felt it was a little on the high side. Now I have a feeling that this may be due to the temperature in the room. I calibrated the PowerTap at the start of the workout (where the room will have been around 15 degrees), but god knows how hot the room got by the time I finished the 20 minute interval (its only a small study-sized room). I have a big fan (albeit blowing more on me than down near the hub) and I had the window cracked open about 6 inches (it was about 3 degrees outside), but the room was definitely quite hot by the time I finished.

Here is a shot of the test:

http://i.imgur.com/0LpjPzO.png

Now there is obvious upward drift in my power during the interval. I feel this can be explained to a large extent by my execution, as I started a bit too conservatively and slowly kept pushing harder and harder according to what I thought I could manage for the whole interval. I just have a sinking suspicion that a certain degree of that slope can be explained by a rising room temperature, and consequently an inaccurate power measurement from the PowerTap. You will notice my cadence was very stable over the interval, yet power was rising. This is a known issue with the Fluid 2 trainer, as the resistance increases as its temperature rises. This is fine and does not affect my PowerTap measurements in absolute terms.

I have read this study…

http://i.imgur.com/fe824lc.png

When calibrating a PowerTap at an ambient temperature of 8 degrees, the wattage error was discovered to be around 8.4% over an hour. However, I can only assume this is due to the internal temperature of the PowerTap increasing as the wheel is spun. What I need to estimate, is the effect that the ambient room temperature has on the PowerTap readings, which is not reported in their study.

I have no idea how to quantify this, and I don’t want to waste my FTP test.

I think I have two options:
I rest up a little and redo the test, this time re-zeroing the PowerTap right before the 20 minute interval (although this will not control for a temperature rise during the interval, of which there will be). I will open the window wider as well :)I use the above test, but with help from the hivemind apply an adjustment downwards to try and control for this phenomenon
I’m not sure what to do about it. My objective is just to get a reasonably reliable FTP measurement to help guide my training over the next few weeks, and also to be able to remove these biases for future tests. I am very sure 260 is too high, as my previous FTP was 215 and I haven’t exactly been finding the workouts easy.

Thanks a lot for reading!

Benno

2

In my experience if you stop and zero the powertap about your warm up period, there isn’t any large shift due to temperature drift after that. I certainly don’t think those upward slants during your intervals are due to temperature drift, too large of an increase.

I have thought about getting a 2nd fan to blow on the powertap but there hasn’t been any weird data so I haven’t actually done it. You might crack the window open more so the heat has somewhere to go. Or maybe crack the door and the window so the air can flow.

3

An inaccuracy in power of 8% over temperatures of 8c and 21c is terrible. That is only normal temperature range of approx 47 Fahrenheit and 70 Fahrenheit which are hardly extremes of temperature. You could start a ride in the morning and find by mid day the temperature has increased that much.

Makes comparing tests done on different days where the temperature is different rather difficult - which is one of the most important things you want from a power meter.

Are there power meters on the market which don’t suffer from this sort of problem with temperature?

4

Was the PowerTap on auto-zero? And did you coast while on the trainer?

I’ve had trouble with this before. I’m guessing that the deceleration resistance of the trainer is much higher than what the wheel would usually do, and this throws off the PT zero. I concluded that it’s better to turn auto-zero off while on a trainer, and simply use manual zero.

5

In my experience if you stop and zero the powertap about your warm up period, there isn’t any large shift due to temperature drift after that.

Keeping in mind, of course, that the PowerTap reports the zero offset with significantly less precision than, say, an SRM or a Quarq, and that a wheel turns far more frequently than a crank. IOW, significant temperature-induced drift can occur, yet be overlooked and/assumed to be insignificant.

6

Benno, this is normal. As a long time Powertap user, here is what you need to do:

  1. Calibrate it just before the 30 minute mark.
  2. Coast on the trainer. You can’t do this during the FTP test though.
  3. Attempt to keep the room the same temperature. When the temp changes outside, coasting auto-zeros a powertap. Inside on a trainer, coasting isn’t normal. You can use a fan pointed at the trainer/rear-wheel if it’s super-sensitive.

If I notice the room has significantly warmed up, I’ll unclip and manually calibrate using my Garmin. Typically this is at least once every workout.

Good luck!

7

No suggestions about the drift but for the test, why not use that FTP for a week or so and if it’s impossible to hit intervals it’s probably too high. If not then stick with it. I’ve never had issues with the power tests and drift being too high and I close my bedroom door when I ride to keep hoise from bothering my girlfriend.

8

Trev, the power tap will rezero itself when you coast for a few seconds. So in outdoor rides it is almost never a problem.

On the trainer, sometimes you never coast, so you need to rezero manually after a warmup.

Quarq and Power2Max both offer electronic temperature compensation, which seems to work pretty well, though I haven’t any experience with it personally, yet.

An inaccuracy in power of 8% over temperatures of 8c and 21c is terrible. That is only normal temperature range of approx 47 Fahrenheit and 70 Fahrenheit which are hardly extremes of temperature. You could start a ride in the morning and find by mid day the temperature has increased that much.

Makes comparing tests done on different days where the temperature is different rather difficult - which is one of the most important things you want from a power meter.

Are there power meters on the market which don’t suffer from this sort of problem with temperature?

9

Trev, the power tap will rezero itself when you coast for a few seconds. So in outdoor rides it is almost never a problem.

On the trainer, sometimes you never coast, so you need to rezero manually after a warmup.

Quarq and Power2Max both offer electronic temperature compensation, which seems to work pretty well, though I haven’t any experience with it personally, yet.

An inaccuracy in power of 8% over temperatures of 8c and 21c is terrible. That is only normal temperature range of approx 47 Fahrenheit and 70 Fahrenheit which are hardly extremes of temperature. You could start a ride in the morning and find by mid day the temperature has increased that much.

Makes comparing tests done on different days where the temperature is different rather difficult - which is one of the most important things you want from a power meter.

Are there power meters on the market which don’t suffer from this sort of problem with temperature?

Thanks for that - just so I understand you here. You are saying that provided you free wheel occasionally to allow it to zero it will not read wrong due to temperature?

So in the study they didn’t zero before testing at 8c and 21c?

I just assumed they must have calibrated / set to zero for each test.

10

IIRC In the study, they took a PowerTap that was in an 8 degree environment, calibrated it, then did an hour workout (without mid workout calibration). The resultant error at 1 hour was around 8.4% due to the temperature changing in the PowerTap, and it not being controlled for. They did not measure the internal temperature of the PowerTap.

The 21 degree environment was a separate test (and had nowhere near the same level of error).

11

Thanks jackmott

Yeah I had my door closed, which would have accentuated the issue. Next time I will completely open my window, and also my door (sorry girlfriend) to get a decent draft in. I do not have a second fan to focus on the PowerTap unfortunately.

12

Hi Andy

I calibrated the PowerTap at the start of the trainer ride, and as you can see from the plot above, at no point did I coast (so it could not auto-zero itself). Is the auto-zero the same as the calibration on my Garmin unit? I get an initial offset of about 506. Does the PowerTap recompute this every time I coast?

13

Hi Andy

I calibrated the PowerTap at the start of the trainer ride, and as you can see from the plot above, at no point did I coast (so it could not auto-zero itself). Is the auto-zero the same as the calibration on my Garmin unit? I get an initial offset of about 506. Does the PowerTap recompute this every time I coast?

Yes, it’s (incorrectly) called “Calibrate”. Yes, the PowerTap will recompute this number when it thinks you’re coasting. If you brake ever so slightly during this process, you will fool the PT into recomputing a faulty zero.

14

Andrew, just what sort of degree of imprecision are we talking about here? I know its hard to quantify, but what would your gut suggest?

I would consider myself quite an average/typical indoor PowerTap user: my room is not particularly special, I have a window which remains open, and a large fan to circulate the air. My point is that I just wonder how many other people are receiving inaccurate feedback from phenomena such as this without even knowing.

15

Benno, this is normal. As a long time Powertap user, here is what you need to do:

  1. Calibrate it just before the 30 minute mark.
  2. Coast on the trainer. You can’t do this during the FTP test though.
  3. Attempt to keep the room the same temperature. When the temp changes outside, coasting auto-zeros a powertap. Inside on a trainer, coasting isn’t normal. You can use a fan pointed at the trainer/rear-wheel if it’s super-sensitive.

If I notice the room has significantly warmed up, I’ll unclip and manually calibrate using my Garmin. Typically this is at least once every workout.

Good luck!

Thanks Nick

I am going to do this next time. Along with opening the window as wide as possible and keeping up the air circulation. I will recalibrate at the 28 minute point.

What do you mean about manually calibrating the Garmin unit? Will the Garmin not take into account the PowerTap’s own recalibration?

16

No suggestions about the drift but for the test, why not use that FTP for a week or so and if it’s impossible to hit intervals it’s probably too high. If not then stick with it. I’ve never had issues with the power tests and drift being too high and I close my bedroom door when I ride to keep hoise from bothering my girlfriend.

I’m tempted to try, but I am extremely confident that its too high (not just by 5-10 watts, but closer to 20). I might drop my FTP to 240 manually (which is still a great improvement from last time) and see how I do in training. I will be fair to myself and up the wattage if I think I have more in the tank. After Sweet Spot II is out the way, I will retest with the advice I have received here which will hopefully control for these potential temperature issues.

17

The fact that your power output was higher during your last 5 minutes of the 20’ test than the 5’ blow out at the beginning tells me you didn’t do the test properly or yes there was some drift.

18

By calibrate, I mean setting the Powertap’s zero offset. These are 2 totally different things I often use interchangeably (as Garmin does!). To zero out my PowerTap, I know of 2 options. One is to coast for 10 seconds, which is sometimes difficult on the trainer. Another option is to go into the Power menus on my Garmin and do the “calibrate” option listed there. (I’m not sure if TrainerRoad has this, but they might.) The Garmin will return a number of 512 +/- 2 as reported from the hub.

I’ll typically do the Garmin “calibrate” operation at the beginning, and then again once the temperature has warmed up. Seems to keep things consistent when riding indoors.

19

I haven’t read the study but I would assume they didn’t zero when investigating the effect of temperature. I would have noticed swings that big in our data. Though that may be a gen 1 powertap they were studying.

Anyway, yes, when you freewheel for around 5 seconds, powertaps will re-zero if certain conditions are met. In practice I find that it works well enough that you can ignore manual zeros in normal training rides.

It is a good idea to manually zero before a time trial or triathlon where you might not ever coast though.

So in the study they didn’t zero before testing at 8c and 21c?

I just assumed they must have calibrated / set to zero for each test.

20

It is a good idea to manually zero before a time trial or triathlon where you might not ever coast though.

But I read pedaling over 30mph is silly in a race!