I'll try that but I'm not seeing corners being cut but "wobbling" problems and differences between the way out and the way back which explain why I systematically get more distance on the F5 than the FR235 that doesn't display that problem. It seems to happen mostly when there are buildings on one side of the road, a classic "GPS multipath" problem as you don't want to use signals that have bounced off that building. Why they haven't figured out how to handle it properly on the F5 versus the FR235 is hard to explain though...
I've done 3 runs so far. 1st one I forgot to turn on GLONASS and track was okay, no major errors, but total distance was a little over 1% short.
Last 2 runs with GLONASS on and tracks have been excellent.
I can't remember who is testing devices on different wrists, but I'd recommend switching wrists (and changing appropriate settings on the watch). I discovered that I get worse tracks on my right wrist when I had two 920xt's to test.
I can't remember who is testing devices on different wrists, but I'd recommend switching wrists (and changing appropriate settings on the watch). I discovered that I get worse tracks on my right wrist when I had two 920xt's to test.
I was just going to suggest WEBVAN swap the wrists the 235 and F5 are on the next time he does the test over the section he's seeing the track goes 'off piste' (needs to be the same direction of travel of course). I similarly have noticed which wrist a GPS watch is on can make a big difference to offsets and cut corners in certain places I run past.
Yes the swap was done last week by "accident" and the results were better...but it was a fluke as the next day it was back to the same type of error. Come to think of it there's no logical reason for the performance to be any different since it's the same 6 o'clock side that's pointing upwards.
At this point I see two options :
1. there's some tweaking they can do to the GPS firmware to handle multipath better
2. the design (metal, barometer, screen) means that the signal they're working with is too degraded compared to the one of the "plastic" FR235 to do any better.
Given Garmin's experience in handling GPS data I'd say #2 is more likely...unless they're using a new GPS chipset that the supplier can still tweak but AFAIK no info has filtered about that. If I remember well it was a Mediatek chipset on the FR235, can't see why they would have changed...but it could be a new version of course.
2. the design (metal, barometer, screen) means that the signal they're working with is too degraded compared to the one of the "plastic" FR235 to do any better.
So basically, we're punished for buying a premium watch? I used the Polar M400 for the last 2 years and as soon as I switched to F5 last week, I started seeing differences in the measuring of the length of my usual routes.
I have a route which I pretty much run on a daily basis and it's exactly 6.22km according to Google Earth and according to the tens of times I ran it using the M400. Today I ran the same course using the F5 and I got 6.25km (GPS only) and last week I got 6.17km (GPS + GLONASS). It's really disappointing that a "cheap" watch like M400 was able to get the routes correctly each time (using its "cheap" GPS) while my F5 fails to do so due to it being made out of premium components...
About 1% short compared to the mapped course. GPS tracing is good with no glaring errors. It nailed some portions that other watches have struggled with, but had me on the wrong side of the street for the 1st mile or so. Instant GPS pace is pretty good as well.
Oh and I ALWAYS prefer a watch that reads consistently short. When I trained with the 910 and earlier watches, I was routinely disappointed with my paces during races. Since the 920xt race paces are more inline with my training. Take what you will from that.
Sorry, but I don't believe your M400 was able to measure 6.22km exactly each run based solely on GPS. The difference made by Fenix was less than 1% which I find quite acceptable.
Now I don't want to appear as a Garmin and F5 "fanboy", I had a lot of issues with GPS tracks on F3, to the degree it was hardly acceptable and I had to return it. There is of course space for improvements on the F5's side, but what I've seen here in this thread so far, it is quite OK and I could live with it.
I know it is quite difficult to accept that such a hi-end device is not at the top of GPS accuracy, but it's not only GPS what makes the watch, I guess. Am I happy with it? Nope. But as I have said, I can live with it.
I'm also waiting for other GPS tracks to be posted here, there weren't enough to draw the complete picture yet.
So basically, we're punished for buying a premium watch?
I used the Polar M400 for the last 2 years and as soon as I switched to F5 last week, I started seeing differences in the measuring of the length of my usual routes.
I have a route which I pretty much run on a daily basis and it's exactly 6.22km according to Google Earth and according to the tens of times I ran it using the M400.
Today I ran the same course using the F5 and I got 6.25km (GPS only) and last week I got 6.17km (GPS + GLONASS).
It's really disappointing that a "cheap" watch like M400 was able to get the routes correctly each time (using its "cheap" GPS) while my F5 fails to do so due to it being made out of premium components...
Well that's just me trying to find an explanation for the F5's inferior performance compared to the FR235, which I must say is the most accurate and consistent GPS watch I've had and my first one was the FR201 in 2003 ;-) Given Garmin's experience it's unlikely they let their flagship product out where they hadn't done their best, especially after the serious GPS problems of the F3 reported here and elsewhere (never owned one myself as it felt like a hockey puck on my wrist!).
Anyway another run today and while the overall distance was very close 16.01 for the FR235 vs 16.02 for the F5, the F5 lost the plot (track has cut corners, early turns as analyzed on GE) for some reason between Lap 1 and Lap 2 where it lost 40 meters...but thanks to its usual "gain" over the FR235 it had caught up by lap 13. The end of lap 1 had four 180°s to change banks over a bridge. I would have thought that its gyroscope would have helped but it didn't. As a result the F5 gave me a pace of 4'59"/km vs the 4'46"/km of the FR235 (correct distance as measured on Openrunner), something that would really have had me worried during a race! So yes some are going to say that you shouldn't follow your watch during a race, that generally there are mile markers, etc...possibly, however why have to deal with this kind of inconsistency when you don't have to?
The two activities below that were recorded on the same track by myself (Fenix 5X) and a friend of mine with a Vivoactive HR at the same time (with a small deviation at the end). My watch was set to record GPS every second with GPS+GLONASS enabled (I believe his was on smart recording);
The two activities below that were recorded on the same track by myself (Fenix 5X) and a friend of mine with a Vivoactive HR at the same time (with a small deviation at the end). My watch was set to record GPS every second with GPS+GLONASS enabled (I believe his was on smart recording);
I sent an email to Garmin support- hopefully they should be able to do something about this..
Your distances were only .01 mile apart. Maybe to solve the problem, you and friend should just hold hands as you cross the finish line! : ) Seriously though, is it the track you are dissatisfied with?
