Time out by 10 seconds

Yes have set it to 1 second recording since day 1

aweatherall i have powermeter as others (who has problems) have too. something not ok with the timer for recording in the FIT file, only if we do this "turning on-waiting4lock-turnoff-turnon-waiting4lock and go" method. maybe You know why.

question: the source of the timer in the FIT file is the same as the "time of the day" datafield? or a different variable? because when the 1st gps lock happens, this "time of the day" changed to accurate (if it had any delay) but in the FIT file it stays in delay. and only the 2nd try (or the turn off/on) is changing it to the accurate value

Let me take a look at this.

Just to summarize this, the data written out in the FIT files uses a timestamp that is offset from the actual time if the unit made a time adjustment when it got a GPS lock? The offset is the same as the time adjustment.

Is the activity timer being started before the unit got a GPS lock?

aweatherall

"Just to summarize this, the data written out in the FIT files uses a timestamp that is offset from the actual time if the unit made a time adjustment when it got a GPS lock? The offset is the same as the time adjustment."  i dont know how much is the exact difference between internal timer and gps time when the unit switched on and still before a gps lock (just i see for a short time it is different comparing to my other Garmins on the handlebar and to my mobile app's gps time with X seconds, i'm using Fenix3 (gps+glonass), Edge 820 (gps+galileo) , Xiaomi MI8 (GPS logger app) and the Edge 830 (gps+galileo), there is a track example in my old thread, check the 1st page of this thread) but i assume it is the same as the offset in the FIT file. i never experienced larger difference in the FIT file offset than 15-18sec, what is weird a bit (because if the unit is not turned on for a long time, the internal timer offset to gps time could be more minutes too...) , but i usually skipped 1 or maximum 2 days between 2 trainings. (so maybe it is the reason why the difference is not larger than this 15-18secs but never know...)

"Is the activity timer being started before the unit got a GPS lock?"   no, i always wait circa 2minutes for sure outside (clear sky, the lat/lon values are almost standstill, within 1-2 meter to the saved starting location) before pressing the start button and ride.

here are 3 consecutive rides, i dont remember for all the circumstances but i think i didnt make the double on/off method on the last day for sure, the +-1sec difference is acceptable between 2 devices i think (these are the 830 (black as base) and 820 (red))

but i found now a 40sec diff at my friend's log (the unit was off for 7days, so it is the reason for this higher diff, the daily slowdown of the internal timer is apprx 5-6sec as i see ), they rode next to each other, my friend has 530 (gps+galileo, black as base), the other guy has 830 (maybe gps+glonass as default and it was his 2nd or 3rd cycling activity on the day, it was late afternoon and he is commuting too ). so the diff<16sec theory is busted :)  (one of my idea was the galileo time to UTC 16sec difference somehow a problem but as we see here, it is not)

aweatherall probably now i could determine the bug:

If the unit was in sleep mode (after auto-sleep or just with manual short power button press to go to sleep), the offset in the FIT file is the same (20sec) as it was immediately after the switch on (i tried it yesterday, my unit was in sleep mode for 4 day after the latest recording , with the daily ~5sec slowness it is just the 20sec difference) but the "time of the day" datafield and the clock on the topright after the gps lock (i waited apprx 3mins on clear sky again after the gps lock and start the recording) was spot on (within 1sec) with my other reference devices before i pressed the start button.  (black=edge 830 in late, red=reference#1 unit) 

so in this "waking up from sleep mode" could be something different than in the "waking up from turned off mode" .

last week i tried it what happens if the unit is switching on after a completely turned off-state after many days (so not in sleep mode) and then there were no offset issue in the FIT file (and after the switch on there was a huge time-offset issue  obviously in this case too until the gps lock )

today i went out again, the 830 was in sleep mode since yesterday's ride , switched on, the time offset was apprx 5-6secs (apprx the Edge was in 18hs sleep mode), waited for gps lock , got, the "time of the day" and the clock in the topright corner was spot on comparing  to the 2 reference devices' timers, waited for another 3mins , and started the recording on all 3 devices: the time offset is 5sec in the FIT file in the Edge 830 (the situation is the same with 530 too due to the common code).

so 1000%  that the source of the problem is "the waking up from sleep mode" 

Luko79 said:

so 1000%  that the source of the problem is "the waking up from sleep mode" 

I'm pretty sure I switch the 530 off completely after every ride and I notice the delay sometimes.

And the theory of turning it on, waiting for GPS lock then shutting it down then switching it back on before starting your ride doesn't work - couple weeks ago I noticed the time was out by about 10secs after doing this so I have given up on switching it on then off first.

But have noticed the past 10 rides or so the time seems to be spot on with my friends devices (520, Wahoo) and I haven't been doing the on/gps lock/off/on.

Seems to be intermittent whether you get the lag or not.