GPS Accuracy

My One and Only Test

Fenix 5, compared to 735XT; fenix 5 on left wrist and 735XT on left.

fenix 5 - 2.72 Beta, 1s recording, GPS +GLONASS
735XT - 1s recording, GPS + Glonass

http://www.mygpsfiles.com/app/#vi2kCdFC

Just looking at the track post run you might say the fenix 5 track is not bad and from this calculation distance accrual seems good, but that is not the full story.

Take a look at my GC file:

https://connect.garmin.com/modern/activity/1652113535

The distance accrual in GC is off by 6% on a known route compared the the 735, and look at all the GPS dropouts (you can see it in Pace graph). While running my Lap Pace (laps set to 1 mile on each watch) were off by between 45s / mile and 1 minute / mile.

Here is the 735XT file for comparison:

https://connect.garmin.com/modern/activity/1652131733

You will notice np GPS dropouts and distance accrual of 5.01 miles, which I know this route to be from 100s of runs. Just as a comparison I included a run on the same route from 1 year ago using the fenix 3 and here is that GC File:

https://connect.garmin.com/modern/activity/1107272465

With this amount of error in this watch and the dropouts of Stryd ConnectIQ data, it is back in the box and back to Clever Training.

Good luck guys, no more Alpha testing for me :)

Did you do the 15 minute GPS soak on the F5 after upgrading to 2.72 firmware as recommended? Also it is recommended viewing the tracks in another application other than GC for a more accurate representation of the actual path.

CW

First hike with my F5X on 2.72- everything was extremely stable- I am very happy with the watch's performance, elevation tracking accuracy and mapping features.

I was the one leading the group during a big part of the hike as the footpath was on the map and I found it extremely easy to follow it and guide everyone else. I also used my Stryd power meter.

I haven't been able to examine the track accuracy as I'm currently on the train back to London- do let me know what you think!

https://connect.garmin.com/modern/activity/1652098293

Did you have your FR935 with you ?

Back from a run with the F5 under 2.72 (and yes I let it "soak" although it's useless for a modern GPS, see my comments in the v2.72 thread) vs the FR235 and I can't say I see much improvement on the GPS side. The pace was maybe a tad more stable but the laps were still irregular and there's the same poor handling of multipath situations.

I made a screenshot in the same spot as the other day, you be the judge :



v2.40

Did you do the 15 minute GPS soak on the F5 after upgrading to 2.72 firmware as recommended? Also it is recommended viewing the tracks in another application other than GC for a more accurate representation of the actual path.

CW

17 minute soak, and you can see I posted the tracks in MyGPSFiles.

MTB seems good to me so far

Whilst I'm not currently running to give my Achilles a chance to repair, I thought I'd throw in some Mountain Bike ride tracks for comparison (other folks are covering running and walking anyways). The Fenix 3 was mounted on the handlebars, the Fenix 5 was on my right wrist and my iPhone SE with Strava App was in the top lid of my Camelbak backpack. So the Fenix 3 and iPhone are perfectly placed to get the very best direct GPS reception, thus the Fenix 5 is at a disadvantage and already the underdog here (in my past testing I've found always found significantly worse performance from watches worn on a wrist vs mounted on the handlebars). Most of the ride is along tracks or narrow roads lined by tall hedges and trees, but there are a couple of wooded areas (just after 4 km and next at 14.5km). Both watches are on 1 second GPS+GLONASS and my Fenix 5 is on 2.72 beta (with 4.22 GPS) and had a 25 minute static activity soak last night after updating. The Fenix 3 and iPhone were connected to my Scosche Rhythm+ for HR, that was placed optimally on my upper arm, with the Fenix 5 using its inbuilt WHR. I've cropped the tracks slightly in this comparison for my privacy, but that removes less than half a km from each end (all the tracks were spot on in the end sections I trimmed anyway).

http://www.mygpsfiles.com/app/#vtWZUk8F

So where I actually went and how far I went for a baseline, the Fenix 3 has totally nailed it to within 2-3 metres the whole way around and even in the woods; but it was bar mounted which I've always found to produce excellent tracks, even in challenging GPS environments. Bear in mind the Fenix 5 had the worst placement for GPS reception (on my wrist), so I can forgive the slight cutting of the loop in the wood at 4.5km. But the iPhone, with a clear skyward view, had no excuses for going off-piste for the 300-400 metres in the wood after 14.5km. Not GPS related, but I'm greatly impressed by the F5 WHR compared to the Scosche too, given the arm placement disadvantage the Fenix 5 had over the Scosche.

All being well I'm going to do a couple of laps of a local all forest mountain bike trail tomorrow, so I'll mix things up with the Fenix 3 on my wrist too for one lap and have the HRM Run chest strap on for benchmark HR comparison.

Strava links:
F5 https://www.strava.com/activities/923730812/shareable_images/map_based?hl=en-US&v=1491072200
iPhone Strava App https://www.strava.com/activities/923736297/shareable_images/map_based?hl=en-US&v=1491072234
Strava wouldn't let me upload the F3 one as well :(

Did you have your FR935 with you ?

Back from a run with the F5 under 2.72 (and yes I let it "soak" although it's useless for a modern GPS, see my comments in the v2.72 thread) vs the FR235 and I can't say I see much improvement on the GPS side. The pace was maybe a tad more stable but the laps were still irregular and there's the same poor handling of multipath situations.

I made a screenshot in the same spot as the other day, you be the judge :



v2.40

No but I will be running with both tomorrow morning.

Looking forward to the results, will you be going back to Millwall Ward ? That's a great testing place for the F5 (and GPS in general I would say) because of all the buildings on one side forcing the GPS to filter out the reflected signals, something the F5 is currently not doing very well for some reason. Not sure if you use autolap but that's a convenient way of getting notified automatically of discrepancies between devices.

Chimpware, was the Stryd calculating the distance for the F5? With the current incompatibility with Stryd I bet the watch was using it to calculate your pace AND distance. It would explain why GPS calculation is spot on but the GC calculation is off.

I had the same issue today (short by 3%), but not seeing nearly as many drop outs as you're seeing.

Stryd was set to; Speed --> Always and Distance --> Indoors. It would have explained it possibly, and it may be the settings do not work properly, not sure.