mpapatheodorou

In the last 2 years I have used for collecting data various device combinations (Garmin 520 with wahoo speed sensor and wahoo cadence sensor, Garmin 520 with wahoo cadence sensor only, Garmin 820 with wahoo cadence sensor, Garmin 820 with Garmin Vector 2 for cadence & power, iPhone 6s with wahoo cadence sensor on strava - strava does not use data from a speed sensor but only the gps' & glonass ones - iPhone 7 with wahoo cadence sensor). Lately I only use Garmin 820 with Garmin Vector 2. Two main reasons, battery life of Garmin vs iPhone and tons of data information collected through Cycling Dynamics. Out of my little experience I can say that on road cycling the accuracy of a gps unit vs speed sensor is almost identical.

Seattle Forrest

And much better after a fire.



Had to go about 40 miles from here ^ before my phone got reception. It wasn't able to augment with wifi networks or anything.

genec

That data is not all that difficult to obtain... just be sure you are sitting on the bike when you do it. Mark a spot on the floor and on your wheel, roll forward until that wheel mark is at the bottom again, and then mark the floor at that new spot... there is your exact circumference.

I have found wheel sensor data to be more accurate than GPS, over a long distance. If that stuff really matters...

corrado33

3 rides on identical routes with garmin GPS.








Three rides, different days, much different speeds (riding with a slow friend on one, riding with faster friends on the other two.) Unsure why one of the rides reported 400ish feet more climbing than the other two, the route was identical for all three rides. I'm wondering if it's due to the barometer in the 910xt . Maybe a low pressure front rolled in as I was riding. (I have elevation corrections turned off)

Note: Occasionally I'll stop the watch when I stop at a stop sign and get something to eat and forget to start it again for a bit, which probably accounts for the discrepancies in the distances.

That's accurate enough for me. +- 0.15 miles over 25 miles. That's about half a mile over a hundred miles.

I do have the speed/cadence sensor on one of my bikes, but it doesn't look like I've ridden this particular route with it. I THINK the garmin prefers the speed sensor over GPS... unsure.

drewguy

Well, to answer my own question (or at least shed light on it), I did something I should have done a while ago, which is overlay the Strava (iPhone 5S) GPS track with the Garmin GPS track. The Garmin (red) is much straighter (although not always on the road!) compared to the Strava (grey) track. I've attached a couple of the more wacky ones, although generally there were places where the Strava track wobbled a lot more, whereas the Garmin seems generally pretty closely matched to the road. The places where Strava seemed the least accurate were in more tree-filled areas, but not exclusively.

If Strava's distance is based on this track, it's no surprise that it's ~1% longer (actually surprised it's not more!).

tyrion

This suggests (to me) either Strava is taking more frequent location readings or the Garmin is more aggressively rounding off the sharp corners. Or both.

nuclear_biker

What's the Garmin sample rate set to?

Seattle Forrest

Try satellite view and see if that affects how well tracks line up with roads.

drewguy

Not sure . . . and not sure I can set it (Edge 25). Judging from some of the tighter turns (and thus more obvious corners) it looks like about 5 seconds (maybe 3).

It looks like both - Strava samples more frequently, but with less accuracy.

nuclear_biker

Sorry I don't have my head in the game and missed that, believe the Edge 25 is 'smart' only. That could definitely also be part of the issue. If you didn't change it, it's almost certainly smart update.

Seattle Forrest

It's probably set to "smart recording" which means the device is constantly reading its position from GPS and writing a new point to the file when a change in your speed or direction requires a point to be written.

You can export the original FIT file and open it in Golden Cheetah, or export it to TCX or GPX which are both text-based XML formats. In any case all of these will have some kind of time stamp, and you can see how often a point is written.

Seattle Forrest

That's almost certainly not it, just for reasons of battery life alone. The Garmin device probably has a better GPS chipset, antenna, and a better position on the bike for view of the sky. It probably also has better filtering. Phones are powerful devices, but they're generalists and the obviously better track came from a specialist device.

gsa103

Garmin devices can be set to automatically calibrate wheel size. The Garmin basically counts revolutions over a half-mile or so of riding and uses that. On a relatively straight segment, the GPS accuracy should be <1%, since GPS is very good at determining the distance between two points.