Chest Strap Power Meter... Does it work?

[russett]

I have been out of the cycling world for a few years. Since I re-started training I have been monitoring personal statistics and have been reading about power meters as a good training tool. As you are all aware, there are several types and all are super-expensive except for the Chest-strap style from CycleOps. Does anyone have the CycleOps chest strap power meter? I think it is called the Powercal. It basically takes HR and uses an algorithm to calculate power output. Just interested in how well it works. I realize that it does nothing for individual leg output monitoring but it is cheap ($100-150 range). Appreciate any feedback from those that have tried

 

[maydog]

I got one as present for Christmas. The HRM itself appears identical to the Timex/Heartwear models. I don't have a real power meter to test this against, but in my comparisons against some decent gym exercise bikes and treadmills I have found the data to be pretty far off. For me it appears to report low during hard efforts and high during light efforts.

I am not ready to completely ignore the data yet, I do want to try it with some more consistent, controlled and longer efforts. Right now, it seems that the accuracy is no better than using a standard HRM and converting the calories per unit time to watts.

There is another thread discussing the powercal.

http://www.cyclingforums.com/t/492771/thoughts-on-the-cycleops-powercal

 

[russett]

Thank-you. I appreciate the feedback on this. I'll check out that link as well.

 

[maydog]

Your question inspired me to do another test this afternoon. I went to the gym and did a 2x20 workout on an exercise bike with power output. Other than a bad fit and uncomfortable position, the resistance is smooth and consistent between machines of the same model. It has a brushless generator and I believe that the power measurements should at least be in the ballpark.

I stayed at a consistent 290W (per the bike) for each effort. The Powercal had an approximate 200W average for the first effort and a 225W for the second effort. I assume the difference was due to the higher heart rate on the second run. I also changed my posture more on the second run, my hr is higher in an upright position and the transients are read as higher power.

Also not the spikiness in the rest periods, I was at about 200W for the warm up and then a very low output for the rest between runs.


Here is the data in TCX format.










Things are not looking good for the powercal.

 

[russett]

I have done a few tests on my own as well. I figured for the $120 it would be worthwhile. I compared to the Powertap that I borrowed from a buddy. Coclusion was that the strap was lousy at variable effort when compared to the powertap. It measured pretty reasonably on long ride at consistent effort when I set up a 30 second average. Unfortunately this doesn't do much for me with interval training. The bottom line... I might not use the strap for much other than a baseline "feel". I appreciate your input and it seems we are in agreement.

 

[maydog]

Here is another follow on. I did a few long climbs when vacationing in Hawaii. I compared the powercal results to Strava, http://www.analyticcycling.com and bikecalculator.com. The latter 3 calculators came within a few watts of each other given an average speed, slope and distance of the climb; whereas I would have needed to be towed up the hill for the powercal to be truthful. It was nearly 100 watts short in some instances.

I pretty much ignore the readings now. It's a pity too, I think that the predictions could be vastly improved if some rider specific parameters were taken into account.

 

[alienator]

Here is another follow on. I did a few long climbs when vacationing in Hawaii. I compared the powercal results to Strava, http://www.analyticcycling.com and bikecalculator.com. The latter 3 calculators came within a few watts of each other given an average speed, slope and distance of the climb; whereas I would have needed to be towed up the hill for the powercal to be truthful. It was nearly 100 watts short in some instances. I pretty much ignore the readings now. It's a pity too, I think that the predictions could be vastly improved if some rider specific parameters were taken into account.

FYI, it's not unexpected that those three sites came up with similar power numbers since they likely used very similar equations and assumptions. They all do a poor job, however, when compared with strain gauge power meters or even a well calibrated iBike (that's not being used in a group or draft of any kind).

 

[russett]

Yes. Sad but true. Thanks for all the testing and insight!

 

[maydog]

I figured that I would dig up this thread since I finally bit the bullet and bought a used powertap.

I captured data from a 10 mile section of my commute today with both the powercal and powertap and exported it to Excel to do some comparison:


The powercal power is on average 40 watts below the powertap measure. The difference in normalized power is 50 watts.

Here is the data:

 

[AyeYo]

Same chart, with Powercal offset +50%. 55-60% might look even better. I give it credit for being surprisingly close. Maybe it's a poor measure of absolute power, but could it be consistent and relative enough to train with?

 

[maydog]

I did notice that the results were correlated with level of effort. This was also a relatively easy effort after a hard ride Monday and another just hours before. My earlier tries on exercise bikes at higher workloads had worse results.

After years of riding with a HRM, I know that my heart rate and RPE varies with supplements/stimulants, riding technique (grinding vs. spinning), ambient temperature and humidity, position and how well rested I am. I suspect that these variables would affect the response, error of the powercal. I'd need to gather a lot more data under different scenarios to determine whether the offset is more or less constant.

I may end up doing just that, but I will be relying on the power tap for training data.

 

[maydog]

Here is another comparison I just ran.

This was on my ride home, a cool day at 62f and pretty windy. I tried to run two eight minute intervals with a 10 minute rest in-between. The powertap put the average power for each interval at 330 watts - the powercal much lower at just over 200 watts.


This pretty much matches my experience with the indoor exercise bikes. One hundred + watts of error is not going to be useful from a training or pacing standpoint.

 

[CAMPYBOB]

Maydog, do you use STRAVA? If so, how does their estimated average Watts match up to to what your Powertap records?

 

[AyeYo]

^I'd imagine not very well unless it's a dead calm day. Strava's biggest power measuring issue is the inability to account for wind.

 

[CAMPYBOB]

Yeah, wind...heart rate decay times...ramp slopes vs. ave. gradient...etc. I don't have a power meter of any type and I'm just curious about the approximate accuracy/correlation of the figure STRAVA throw up as calculated on an uploaded .gpx file.

I can see Garmin-like devices, now paired to smart phones, integrating local weather and wind condition data into the ride files.

 

[maydog]

I do use STRAVA. It is a bit of a PIA to do comparisons, but I have done it on a ride or two. For the most part, STRAVA underestimates my power but then overestimates when there is a tailwind. I think the last ride I compared STRAVA put the average power at 216w whereas the powertap data was 260w. The 35 mile ride had rolling hills and stiff south wind. The ride started south into the wind, then looped around.

STRAVA probably does not account for much aerodynamic change with increases with weight. Being 6'4" with a large frame makes me pretty unaero compared to my ride mates. Two of the strongest, and much smaller, riders have powertaps and dropped me a couple of times their avg power for the ride above was just about my STRAVA calculated average of 216w. We did about an equal amount of pulling.

 

[CAMPYBOB]

Thanks, Maydog. I had a feeling STRAVA was low. The odd thing is that is always yields the most vertical climbing of the four websites I upload to and and STRAVA is usually number 2 in estimating calorie data (Map My Ride is the highest on a consistent basis in that category).

I like the concept of the new Garmin pedal meters, but the execution still needs more work IMO. When Campy comes out with that new 12-speed electrically shifted group with Record Pro Fit Power pedals...I'm all in!

Kudos on your power reading, too! Keep on drilling it!

 

[AyeYo]

Originally Posted by CAMPYBOB

I like the concept of the new Garmin pedal meters, but the execution still needs more work IMO. When Campy comes out with that new 12-speed electrically shifted group with Record Pro Fit Power pedals...I'm all in!

Yours to take home today for the low low price of $25,000 and your left nut??

 

[CAMPYBOB]

Yeah, but that price includes the Cancellara Edition hidden motor with lithium polymer battery and a 235 (clipping driven at 8 Ohms) Watt RMS per channel Dynaco sound system!

 

[maydog]

A little update for those interested in STRAVA. It is hard to compare data between STRAVA and the powertap because STRAVA does not let you export the power data.

I had a somewhat easier ride today, flattish course; the first 1/3 into a 10mph headwind, 1/3 crosswind and 1/3 tailwind. Strava reported the average power as 183 watts, the powertap 210 watts.