Fenix stop record GPS data during activity

The Garmin connect is private so I cannote check, but seems it is happen the same of me. 

I've also analyzed the files with fitfileviewer, and is missing long/lat info, even if all the other data are present. What is really strange is how can be calculated the pace and the speed if gps is not working? I think the gps is perfectly working and simply the software does not save lat/long data.

I've also notice that this behaviour is not related to a particular place or action, like a stop or a lap, but happens without a specific condition. Today is happen during warm up, before the first km, other times after many kms.

Also the firmware is not the problem. To me is happen with realase 13, and today with 14.

Today I've talk with support in chat and the obvious answer has been to try a full reset, and then, if the problem persist, to bring back to local dealer. The problem is that guarantee is expired on august, so I don't think I can do much more. 

It's software.  Look, I didn't have this issue in the previous release and this is the first run on 14.

Also, I was looking at the map while running, the trace was there DURING the run, but not in the file. So the FIT file saving is corrupting the contents.

From the accelerometers.

If the track is o.k. during a run and corrupt afterwards, might it be that it is corrupted during the transfer to Garmin Connect (BT or Wifi)?

What happens if you connect the watch to a computer, transfer the .fit file manually and upload it to GC, Strava or anywhere else where you can analyse the track?

As I just received 14.0 on Wednesday, I did only 2 activities with it, but they were without issues, as all activities with 13.0, so I don't think it is a problem of the firmware.

Thanks Alan,

As I mentioned in the 1st post, I took the file straight from the watch, not from GC. And I saw the missing Long/Lat values. So it was during the file saving, in the watch. So no chance to recover the track (I converted into GPX, and manipulated the coordinates, before uploading again to Strava.  Not perfect as the distance ended up being 4% longer, and HR, pace and cadence freeze for the repaired section. Thankfully, it was very steady in all senses)

Garmin support asked me to do a master reset and try again.  I guess I will run with 2 watches for the next couple of weeks.

I've had this problem with the 13.7x betas (.fit files directly from watch had partly missing GPS coordinates, while the position on the watch's map screen had been correct all the time _during_ the activity).

I thought it might have something to do with having LiveTrack active and pausing/resuming the activity, or going to the widget loop/main watch face for > 2 minutes (to read messages) and then returning to the activity.

Probably unrelated: unexpected '...connected' notifications for ANT+ sensors and 'GPS found' notification during activities. Both issues reported to the beta team.

(not suggested as solution) I don't have this anymore now I run with 2 watches and the F5+ (on 14.00) does not do LiveTrack anymore.

I don't use LiveTrack and have not installed any kind of third party app. As jaimegu says, the problem is in the file. All data that depends on GPS tracking is there and is correct. So the GPS works well during the activity. The problem occurs during data transfer in the .fit file.

I'm a developer, so I would be curious to know how the data is managed during the activity, since speed and cadence are calculated in real time and both depend (at least partially) on GPS data. 

The FIT SDK gives insight in the principles on how FIT files are written - but that's not necessarily the same way the watch firmware has that implemented.

In general: with every-second recording on, every second 1 'record message' is written to the FIT file. Interspersed with the record msgs are event msgs and other msgs.

AFAIK speed and distance come from GPS when available, but can fallback to foot pod or the accelerometer in the watch, don't know specifics. 

What I see in my FIT files: mid-activity some record msgs without the LAT and LON data. Distance will stop updating - all record msgs from there on contain identical distance value. After GPS is back, distance stays on whatever it was when GPS was lost, but speed will be fine. Track will be fine except the missing part. FIT file from Fenix 1 (worn during same activity) has no missing GPS data in FIT file.

Also, a few times I noticed all sensors connecting, as if the activity had just started.

Disclaimer: just my silly guesses, what goes through my mind while running, it could be any or none of these things.. spontaneous restart of sensor hub/gps software, workload too high, watchdog timer, too many interrupts, out of memory, bad battery and brownout detection kicking in, watch rebooted and I didn't see it.. Embedded software is just kind of challenging.

I did some queries on my activities and found more (1 out of every 4) with missing GPS than I expected. Mostly 10 to 30 second gaps in the data, a few times per activity. And not just GPS. Sometimes RD pod and Tempe data goes missing for 10 to 30 seconds. These sensor data gaps cluster around GPS gaps. Mostly distance and speed were OK.

Update: Went back reviewing data to when I first got the F5+. It was on SW 9.0 and I see exactly the same behavior in FIT files from that time. Almost every run and walk activity has at least 1-4 GPS dropouts (during ~4-20 records), often also some missing Tempe and RD pod data. Only one run (SW 9.1) was terrible with 20% of the GPS and sensor data missing. Bike activities are generally much better with fewer and shorter GPS dropouts. Walks are worse.

Other units: Forerunner 935: missing GPS - much less often, and only up to ~10 seconds in a row. Forerunner 945: not enough data but appears the same as the 935. Epix and Fenix 1: never any GPS dropouts.

What I'll do next: Activities without RD Pod and Tempe sensors seem to fair better, but I have too few of those, so I'm going to rule out the role of RD Pod and/or Tempe by doing 10 runs without them. Maybe the issue is somehow related to or made worse by these sensors (after all FR935 had GPS dropouts as well, and the same sensors were used all the time). It will take a while before I report back on that.