Road Cycling - Does your polar upload data correctly?

Okay, so I purchased the Polar 520 and am learning to use it. Upload of data to computer is no problem, except the data transferred to the computer is not the same as on the 520.

For example, the 520 display shows my last ride was 58.6 miles, 20.1 ave speed, 148 ave heart rate and so on.

The computer upload of this ride shows a distance of 54.1 miles, 19.3 ave speed, 152 ave heart rate.

I have yet to get a reply from Polar on this question.. Anyone else experience this?

What software are you using? PC Coach? Something else? Regardless, make sure you have the latest software patches.

Just guessing: Might one take into account a warmup and cooldown phase (would account for lower average, but unlikely a shorter distance and higher heart rate)?

FWIW, I have the Polar 510 and I've noticed things like that too. I haven't looked lately at the display vs. the computer for differences. I'll do that on my next long(ish) ride. I suspect the differences would be greater the longer the ride.

I have no problems like that, but then again I have an S720i, so it uses IR communication rather than Soniclink.

I noticed mine was off today also. Never checked before. I am going to put my old cyclometer on the bike and compare on my next ride.

I have no problems like that, but then again I have an S720i, so it uses IR communication rather than Soniclink.

Same here. The 720i is a very good product. I've been very happy with mine since day 2. Day 1 was frustrating, getting it set up. Since that time, it's been quite good. I have the power option, and it's pretty accurate, considering it's sampling rate.

Same here. The 720i is a very good product. I've been very happy with mine since day 2. Day 1 was frustrating, getting it set up. Since that time, it's been quite good. I have the power option, and it's pretty accurate, considering it's sampling rate.

When I first set mine up I had the speed sensor right at the maximum recommended distance from the computer. Every couple minutes it would stop because it lost the speed reading so thought I had stopped riding. I had to press the start/stop button to manually restart it. MAN was I pissed at the end of that ride. I thought I was going to have to return it, but then I realized I might have the sensor too low, so I adjusted it and everything started working peachy-keen.

The sampling rate might have something to do with it -- the rate changes the longer your ride. I'm interested to know if you get a response from Polar, too.

When you review the file, before you upload, the data is different that what is uploaded? Thats odd...

When you review the file, before you upload, the data is different that what is uploaded? Thats odd...


That sums it up correctly!

Let me upload tonights ride, but I'll write down what the watch says first....


As far as a sampling error, or distance from transmitter to reciever, none of that would cause whats happening.

When you first got the Polar, and you configured the watch, did you do it manually by pressing the watch buttons, or did you enter the data on the computer and upload it to the watch??

I checked the data in file F6, and wrote it down, then uploaded it. It came out the same. Are you checking the correct date? I went to the calendar and selected the big orange box "tuesday" then went up to edit, and selected "edit workout data" The figures were the same.

And yes, Sunday was a hard day....

http://img88.exs.cx/img88/2677/polar1.jpg

I'm sure that I uploaded and reviewed the same date/exercise. I have performed this test 5 times and to be sure I am comparing the same data I erased and reset everything.

However, I notice you are using PC Coach, while I am using the Polar software. I wonder if the difference is the software.

Do you like the PC Coach better?

I' go back to what I said before about the sampling rate. I think (because I haven't got the manual handy) that the sampling rate for short exercise intervals is something like once every second or two. When it gets above an hour (or something like that) the sampling rate changes. Over, say, 10 hours, I think the sampling rate is around two minutes. This all for the 520 model.

As the sampling rate changes, the data that may be been instantaneously recorded at, say, one-second intervals, may be lost as the sampling rate changes to two, five or 10-second intervals. I know for instance, that the HRM on the wrist will show maybe a top speed of, say, 72km/h, and a max HR of 162bpm. But when the data is downloaded and brought up on screen, the max speed might be 65km/h and the max HR 158, simply because the original max readings have been displaced by the sampling rate.

Dr Stern probably knows the answer to this, and I could very well be wrong (not for the first time).

Same here. The 720i is a very good product. I've been very happy with mine since day 2. Day 1 was frustrating, getting it set up. Since that time, it's been quite good. I have the power option, and it's pretty accurate, considering it's sampling rate.

How much did you get the power option for?? Did it include the cadence and speed sensor that is placed at the rear?

I don't think it is a matter of uploading correctly. There are slight differences in the information, and I believe there is a difference between the ave speed. I get about .5 km/h more after the information is uploaded. No difference in ave HR and only about .1 difference in distance.

I don't think it is a matter of uploading correctly. There are slight differences in the information, and I believe there is a difference between the ave speed. I get about .5 km/h more after the information is uploaded. No difference in ave HR and only about .1 difference in distance.


So are you suggesting the monitor has two different sets of data? One that gets uploaded and the other for the display?

I never used the Polarsoftware....

I have a S710i, and it is constantly wrong. It is exactly as you say - what you upload is not what you see on the watch. One that I constantly check is the Average speed for a ride. What is shown on the watch at the end of the ride is not the same as what is shown in the PPP4 software.

It is always out....it is not like it is an intermittan thing or anything. What is shown on the watch is always higher than what the PPP4 software reports.

In this case, I have checked with other riders on the same ride, and their readings generally agree with the PPP4 software, so I am assuming that the S710i is incorrect.

I *think* small differences between the watch and computer are due to how the HRM calculates things on the fly. With all the sensors running the watch has to process a fair amount of data... I have always wondered if the Polar engineers programmed the watch to use fairly low precision routines for generating calculated quantities.

The PPP software meanwhile has the full power of your PC to grind through 5 sec. interval-recorded data.

However, you've got me thinking now. I will compare my cyclocomputer, watch and PC over the next few rides.

I'm going to put my Astrale back on and compare. Based upon my comparisons with other riders in my group last night it seems the average and max speeds are low on the upload...

Keep us posted on your results.

I the meantime, I have sent Polar 3 messages through their customer support website asking them about this abberation with no response....

I've noticed the same and it's why I think the HRM is doing "cheap" math on the fly to preserve processor time. Let's all experiment and see what we can come up with.

As far as a response from support if you've tried the polarusa.com site then try polar.fi I sent a msg to polar.fi's support requesting the HRM download file specifications and I had a reply (and the documentation) within a week and a half.

Okee-dokey... just got back from my 17 km climbing loop. No comments about average speed, OK... I'm recovering from a back injury.

I'm not seeing much of a problem here. The timing difference between the Cateye and the Polar is due to me stopping a few times (base of climb, top of climb, base of climb, home) to stretch. The polar will count for 5 seconds before auto-stopping.

http://foosoftware.com/polar.gif

Talked to a Polar rep. today. The Polar monitors are more accurate than the data that is uploaded to the computer. The sampling rates will have about a +-9% deviation.

Doesn't seem accurate enough for me.

I've noticed the same and it's why I think the HRM is doing "cheap" math on the fly to preserve processor time. Let's all experiment and see what we can come up with.

As far as a response from support if you've tried the polarusa.com site then try polar.fi I sent a msg to polar.fi's support requesting the HRM download file specifications and I had a reply (and the documentation) within a week and a half.

Could you forward the HRM file specification to me at garyg (at) shastasoftware (dot) com ? I've puzzled most of it out, but I still have some questions about it. In particular, I've had problems reliably getting distance from the hrm file (it seems to be stored in more than one place in the file, but they're not always the same number!?).

FWIW - I'm trying to understand the Polar hrm file format because I want to be able to support importing rides in .hrm format to my CycliStats software ( http://www.CycliStats.com ). It currently supports direct import of the .hrm files, but I think the program is still sometimes misinterpreting the data.

[QUOTE=SSP]Could you forward the HRM file specification to me at garyg (at) shastasoftware (dot) com ? I've puzzled most of it out, but I still have some questions about it. In particular, I've had problems reliably getting distance from the hrm file (it seems to be stored in more than one place in the file, but they're not always the same number!?).
[QUOTE]

It is now available from Polar's site... see the PDF link at the lower right

http://developer.polar.fi/developer.nsf/main

When I requested it there was no Developer's program at Polar so it was a "special" request.

I'm writing my own file edit/merge utility for use by the guys I train with... it'd be nice to have the Polar DLL to work with the unit but it's not worth $500 to me...

[QUOTE=SSP]Could you forward the HRM file specification to me at garyg (at) shastasoftware (dot) com ? I've puzzled most of it out, but I still have some questions about it. In particular, I've had problems reliably getting distance from the hrm file (it seems to be stored in more than one place in the file, but they're not always the same number!?).
[QUOTE]

It is now available from Polar's site... see the PDF link at the lower right

http://developer.polar.fi/developer.nsf/main

When I requested it there was no Developer's program at Polar so it was a "special" request.

I'm writing my own file edit/merge utility for use by the guys I train with... it'd be nice to have the Polar DLL to work with the unit but it's not worth $500 to me...

Thanks for the link.

Good luck trying to get the data from those .hrm files. The Polar engineers don't seem to understand that Distance, Ride Time, and Average Speed are related to one another. I've spent a fair amount of time trying to parse these fields out of their .hrm files and here's what I've found:

1) Distance
Their documentation is in error for the [TRIP] block. The documentation indicates that the first entry in the [TRIP] block is the ride distance in km or miles. But, from the hrm files I've seen the distance in the [TRIP] block is always shown in km and never in miles, even when the user sets up the watch for US units of measure.

2) Ride Time
The .hrm file does not seem to contain a value for the actual time spent riding. Instead, it contains "exercise duration", which means that it includes time you spent stopped at stoplights, fixing a flat, etc. This makes calculating an accurate Average Speed using the .hrm file data very difficult. The only way I've found to get close to the actual ride time is by parsing the entries in the [HRData] block, and subtracting off any intervals for which the reported speed is zero.

3) Average Speed
Oddly enough, the Average Speed reported in the [TRIP] block does show the correct average speed in mph or kph, but only to 1 decimal place of accuracy.