GPS track differences

Andy Terry · Post by **Andy Terry** » 29 Nov 2014 17:14

Mr Weaver said today on Strava:

"Ok, today I rode with Garmin GPS units, the Edge 500 which is the other .gpx file and this one taken from the Edge Touring Plus. Both units were switched on at exactly the same time. Firstly look at the difference in the amount of climbing, 1798ft vs 1365ft? Then look at the times on some of the segments, for example, look at the segment "RRPCC-9am Ride". Remember that both units are sat on the same bike, yet the Edge 500 has recorded the segment time as being 3 seconds faster than the Edge Touring Plus??
2 hours ago"

Going to look at the 2 track files (if he can get them) but by way of illustration, this show that 2 Garmins carried together will not give the same track.

Andy Terry · Post by **Andy Terry** » 29 Nov 2014 17:26

In this example you can see the Edge Touring (orange) has logged 3 trackpoints and the Edge 500 (green) has logged fewer points and effectively cut the corner.

Interestingly on the ride in question the Touring logged 2492 points against the 500's 3676. It seems they were set differently - 2 sec v 3 sec recording interval.

AlanW · Post by **AlanW** » 29 Nov 2014 19:08

Andy, this would suggest that one unit would consistently record a "faster" time than the other. However, and based on today's exercise this is certainly not the case, on a couple of segments the other one recorded a quicker time?

AlanW · Post by **AlanW** » 29 Nov 2014 19:34

Andy Terry wrote: Interestingly on the ride in question the Touring logged 2492 points against the 500's 3676. It seems they were set differently - 2 sec v 3 sec recording interval.

Just out of interest, but I have searched to see if I can establish what the recording intervals are set at for each of the units. The only point that I can establish is that are both set to "Smart" recording intervals, what ever the "Smart" interval duration is?

Apparently the only time that it will alter is when a Power hub is connected and then both devices will default to one second intervals.

EDIT
"Smart" recording can vary from anything from 1 second to 10 seconds, it alters at key points, such as a change in direction, speed or heart rate.

Andy Terry · Post by **Andy Terry** » 29 Nov 2014 19:39

I think that means they will record points to suit what is happening on the ride. More changes of direction = more frequent points basically.

Andy Terry · Post by **Andy Terry** » 29 Nov 2014 19:45

I have looked at the 2 files side by side and all I can say is that they don't record exactly the same points, and they are sufficiently different to account for differences in segment times. Simple as that.

Everything is within expected tolerances so nothing to worry about.

A 4 second difference on a 4 minute segment is only 1.66% after all.

Andy Terry · Post by **Andy Terry** » 29 Nov 2014 19:49

Here's a good example of how the units record at different intervals

AlanW · Post by **AlanW** » 29 Nov 2014 19:53

Andy Terry wrote: Everything is within expected tolerances so nothing to worry about.

A 4 second difference on a 4 minute segment is only 1.66% after all.

I'm not worried, just raising the issue really but 4 seconds is the difference between getting a KOM and only getting a top ten placing! Which is clearly dependant on which unit records the quicker time on the day at that precise moment in time, as I proved this morning.

Looks like I need to go everywhere with both GPS units now......

Andy Terry · Post by **Andy Terry** » 29 Nov 2014 20:20

I suppose it's only natural for people to read segment times as accurate to 1 second, which they plainly aren't.

Look at this:

that's the start of the Plymouth Road segment

the two trackpoints have been recorded as

<trkpt lat="52.304708" lon="-1.945313">
<ele>117.40</ele>
<time>2014-11-29T08:44:39Z</time>
</trkpt>

and

<trkpt lat="52.304690" lon="-1.945384">
<ele>111.00</ele>
<time>2014-11-29T08:44:41Z</time>
</trkpt>

2 seconds difference.

AlanW · Post by **AlanW** » 29 Nov 2014 20:44

Andy Terry wrote:I suppose it's only natural for people to read segment times as accurate to 1 second, which they plainly aren't.

Look at this:

that's the start of the Plymouth Road segment

the two trackpoints have been recorded as

<trkpt lat="52.304708" lon="-1.945313">
<ele>117.40</ele>
<time>2014-11-29T08:44:39Z</time>
</trkpt>

and

<trkpt lat="52.304690" lon="-1.945384">
<ele>111.00</ele>
<time>2014-11-29T08:44:41Z</time>
</trkpt>

2 seconds difference.

Now that is interesting, and as you say a 2 second difference and that's before you have even started the segment!