SRM PC8 Strange KMs difference

Back by popular demand, the general all-things Road forum!

Moderator: robbosmans

Post Reply
fbonde
Posts: 161
Joined: Sun Feb 14, 2016 10:26 pm

by fbonde

Hallo everyone,

I have noticed that on my PC8 screen when I finish the ride, it displays a total number of the overall distance (like the picture below) But, when I upload the file to TP or Strava it shows a completely different overall distance. Did any of you experience the same? Or has an explanation why it occurs?

example fro today. on the PC8 screen it says 142,7KM on TP/Strava it says 138 KM that a big margin (4,7 km) :/ scary..
Attachments
IMG_0863.jpg
Screen Shot 2016-05-21 at 22.05.52.png
Screen Shot 2016-05-21 at 21.22.43.png
insta: @fbonde

TurboKoo
Posts: 651
Joined: Wed Oct 29, 2008 7:55 pm

by TurboKoo

Maybe you should check your wheel settings. Seems like difference between gps and wheel sensor setting.
Cannondale SuperSix
Shimano 9270

by Weenie


Visit starbike.com Online Retailer for HighEnd cycling components
Great Prices ✓    Broad Selection ✓    Worldwide Delivery ✓

www.starbike.com



fbonde
Posts: 161
Joined: Sun Feb 14, 2016 10:26 pm

by fbonde

TurboKoo wrote:Maybe you should check your wheel settings. Seems like difference between gps and wheel sensor setting.


Thanks for your replay Turbokoo. I'm not using ANT+ speed sensor at all. I only use GPS. I mean, the wheel size setting some alive, when you choose ANT+ sensor for speed.
Attachments
Screen Shot 2016-05-22 at 07.48.21.png
insta: @fbonde

sungod
Posts: 1702
Joined: Fri Dec 24, 2010 9:37 pm

by sungod

gps 'distance' depends to the calculation used to join the dots of the sample points - simple straight line interpolation will under report on any twisty/hilly course, fitting a curve takes more calculation but should be more accurate though steep climbs/descents still will introduce more error due

there's no one 'right' way to do this, so over a long course it'd be easy for different methods to arrive at significantly different answers

if you want accurate speed/distance, a wheel sensor (with accurate circumference measurement) is better than gps

fbonde
Posts: 161
Joined: Sun Feb 14, 2016 10:26 pm

by fbonde

sungod wrote:gps 'distance' depends to the calculation used to join the dots of the sample points - simple straight line interpolation will under report on any twisty/hilly course, fitting a curve takes more calculation but should be more accurate though steep climbs/descents still will introduce more error due

there's no one 'right' way to do this, so over a long course it'd be easy for different methods to arrive at significantly different answers

if you want accurate speed/distance, a wheel sensor (with accurate circumference measurement) is better than gps


I agree, but surely we can agree on one thing, that a margin like this, is way above the "limit". I have never on my garmin or my Polar GPS watch experienced that high error on GPS points.
insta: @fbonde

sungod
Posts: 1702
Joined: Fri Dec 24, 2010 9:37 pm

by sungod

but unless you *know* the calculation method used by each, you cannot know what the likely error bounds are, on the route involved it may be reasonable... if strava is attempting to apply a terrain database to calculate surface distance over gradient and the gps is using a 2d algorithm, they could give very difficult figures (both of which could be equally wrong)

if the garmin/polar use the same algorithm, they will show similar results, it doesn't mean they are any more accurate

it comes back to: if you care about accuracy, do not use gps, use a wheel sensor

fbonde
Posts: 161
Joined: Sun Feb 14, 2016 10:26 pm

by fbonde

sungod wrote:but unless you *know* the calculation method used by each, you cannot know what the likely error bounds are, on the route involved it may be reasonable... if strava is attempting to apply a terrain database to calculate surface distance over gradient and the gps is using a 2d algorithm, they could give very difficult figures (both of which could be equally wrong)

if the garmin/polar use the same algorithm, they will show similar results, it doesn't mean they are any more accurate

it comes back to: if you care about accuracy, do not use gps, use a wheel sensor



To be honest, it seams very strange, that the SRM device agent read one number, and SRMX read a completely different number. To me, thats not good at all if we are talking "Gold standard" like SRM. I dont reallt care about Strava or TP, its more the SRM software it self. Nothing is perfect iknow, but margins like this is simply unacceptable.
insta: @fbonde

sungod
Posts: 1702
Joined: Fri Dec 24, 2010 9:37 pm

by sungod

fwiw i get an exact match between pc8/srmx/golden cheetah, i use a wheel sensor, if you are seeing a difference it does suggest the cause is in the gps data interpolation

if you want a definitive answer to any pc8/srmx discrepancy i think you need to contact srm support and ask about it

fbonde
Posts: 161
Joined: Sun Feb 14, 2016 10:26 pm

by fbonde

sungod wrote:fwiw i get an exact match between pc8/srmx/golden cheetah, i use a wheel sensor, if you are seeing a difference it does suggest the cause is in the gps data interpolation

if you want a definitive answer to any pc8/srmx discrepancy i think you need to contact srm support and ask about it


Thanks for that information.

Is it possible to use a speed sensor for speed and distance, and let the GPS in the PC8 visualize the route afterwood ?
insta: @fbonde

User avatar
toiyuet
Posts: 174
Joined: Wed Dec 14, 2011 2:47 pm
Location: Hong Kong

by toiyuet

fbonde wrote:
sungod wrote:fwiw i get an exact match between pc8/srmx/golden cheetah, i use a wheel sensor, if you are seeing a difference it does suggest the cause is in the gps data interpolation

if you want a definitive answer to any pc8/srmx discrepancy i think you need to contact srm support and ask about it


Thanks for that information.

Is it possible to use a speed sensor for speed and distance, and let the GPS in the PC8 visualize the route afterwood ?

Yes, that is what we did.

fbonde
Posts: 161
Joined: Sun Feb 14, 2016 10:26 pm

by fbonde

toiyuet wrote:
fbonde wrote:
sungod wrote:fwiw i get an exact match between pc8/srmx/golden cheetah, i use a wheel sensor, if you are seeing a difference it does suggest the cause is in the gps data interpolation

if you want a definitive answer to any pc8/srmx discrepancy i think you need to contact srm support and ask about it


Thanks for that information.

Is it possible to use a speed sensor for speed and distance, and let the GPS in the PC8 visualize the route afterwood ?

Yes, that is what we did.


Perfect - Any setting I should be aware of?
insta: @fbonde

sungod
Posts: 1702
Joined: Fri Dec 24, 2010 9:37 pm

by sungod

it's easy to do in pc8 device agent, settings are in the bike profile...

once you've got the speed sensor, you pair it to the pc8 and set the wheel circumference (best to measure it, not rely on any default values), you also set speed to ant+ instead of gps, and save the config to the pc8

fbonde
Posts: 161
Joined: Sun Feb 14, 2016 10:26 pm

by fbonde

sungod wrote:it's easy to do in pc8 device agent, settings are in the bike profile...

once you've got the speed sensor, you pair it to the pc8 and set the wheel circumference (best to measure it, not rely on any default values), you also set speed to ant+ instead of gps, and save the config to the pc8


Sounds perfect. So by setting the speed to ant+ dosent turn off the GPS automatically? if so, the "problem" is solved. How long time do you guys whait for the GPS signal ? I wait op to 7 minutes every single time.
insta: @fbonde

by Weenie


Visit starbike.com Online Retailer for HighEnd cycling components
Great Prices ✓    Broad Selection ✓    Worldwide Delivery ✓

www.starbike.com



sungod
Posts: 1702
Joined: Fri Dec 24, 2010 9:37 pm

by sungod

setting speed to ant+ only affects the speed/distance method, gps data is still recorded for position (unless you turn gps off completely)

i usually wait until the signal indicator shows 2-3 bars, just a minute or so if there is nothing tall nearby - though even with full signal i noticed the gps data can sometimes be off by 10-15 metres if there is heavy tree cover or similar obstruction on a route

Post Reply