SolarEnergy said:
I have no intention of being negative or to argue just for the fun of 'winning' an argument. But for the sake of those who are in the process of buying a powermeter, this post here
http://www.cyclingforums.com/showpost.php?p=2868500&postcount=30
inspired me to do a pubmed research over polar s710. And so this one here was the result.
http://www.ncbi.nlm.nih.gov/entrez/..._uids=12740731&query_hl=6&itool=pubmed_docsum
Note that their system might have improve since this study was conducted, but it may partially explain why Polar has a reputation of being not that accurate...
Actually...I'm glad you pointed that out since, if anything, it highlights the dangers of making conclusions based on abstracts. I'm assuming you haven't read the entire paper, correct? If not, I can send you a copy.
Anyway...here's my "critique" of the study and it's methods. First, the measurements were taken in 2 parts. One part was a lab test where the bikes were attached to a wind trainer and the riders did three 5-minute runs at 150W with the cadence for each bout set at 60, 90, and 110 rpm. Obviously, this testing was done prior to the reports and discovery of the Polar's "flakey" performance on a stationary trainer. Additionally, the investigators found that the Polar read higher than the SRM with increasing cadence at this relatively low power on the trainer. This is now known to be a symptom of "relatively low power" and high cadence with the Polar. In those conditions, the vibration signal gets weaker as the chain tension decreases, which then makes it more likely that the signal coming from the chain pins passing the sensor can get coupled into the frequency calculation. This will result in higher power readings.
Basically, the "lab test" just went to show that the Polar doesn't do well on stationary trainers. Well...we all know that now. Even so, the investigators DID say this about the Polar's lab results: "...compared to laboratory ergometers, a bicycle with an S710 may be more suitable for power output measurement during cycling, especially for constant pedalling cadence protocols. The S710s accuracy at low cadence seems better than other ergometers that have been studied."
The other part was an uphill "field test". In short, for this the Polar read higher than the SRM by a consistant percentage. That tells me one of 2 things, either the "slope" setting of the SRM was off, or one of the values (chain weight, chain length, or chainstay length) entered into the Polar was incorrect. No info was given in the study about how they mounted the Polar and what values were put into the head. It's very possible that they didn't abide by the later recommendation that the chain be no further from the sensor that 30 mm. Additionally, although the authors state that "Prior to each trial, the SRM crankset was calibrated according to the manufacturer's recommended procedures", I've got a feeling they're just talking about resetting the zero offset. I highly doubt they hung weights and did a multipoint slope check. It's well known that very few SRMs come from the factory with the correct slope setting. I know of many people (coaches) that do a full calibration on any SRM they buy new.
So...what are we left with? First...the Polar doesn't do so well on trainers. OK, we know that. Even so, it actually performs better than lab ergometers in low cadence conditions. How about that?
Then, in the field, the Polar is shown to be highly reliable in it's power measuring. In the test, there was basically a fixed % difference to an SRM of which we don't know if the slope was set properly. Sounds to me like a simple calibration of the SRM, or checking of the Polar's input values may take care of that difference.
Here are 2 quotes directly out of the paper:
"In addition, the S710 appears to have a better accuracy than Powertap, reported to read 8% higher than SRM."
"The present results show that the S710 is a highly reliable powermeter. The repeatability of S710 appears better than some other devices (e.g. Kingcycle, Politechnika, Monark) available for recording power output, but similar to the SRM."
Anyway...that's the way I see it. Even though a reading of the abstract tends to shed a bad light on the Polar's performance, a careful reading of the entire paper and a reasoned look at the methods and results actually show evidence that the Polar can be a pretty good performer (IMHO, that is).
I just wanted to put that out there as well for those in the process of buying a powermeter.
Actually, IMHO the BEST and most reasoned comparison of all the powermeter options is in a chapter of Hunter Allen and Andy Coggan's book "Racing and Training with a Powermeter". If you're planning on diving into buying a powermeter, I HIGHLY recommend getting that book just for that chapter alone.