Related
This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

GLONASS ON/OFF battery use and position accuracy

CloggyAbroad
over 7 years ago
I just left my 935 to record an 'activity' outside in a fixed position for a couple of hours to look at the difference in battery use between GPS and GPS+GLONASS. I have done this during an actual run before but I got strange results which I was unable to explain, possibly due to changing exposure to the sky and other variables. By leaving the watch in a fixed position with good view of the sky all round and at a fairly stable temperature I hoped to be able to see just the effect on battery use caused by GLONASS.

During this test everything else was switched off, so no OHR, no phone connections, all ANT sensors disabled (switched off in the watch) and recording set to 1 sec.

The resulting 'activity' is here:

https://connect.garmin.com/modern/activity/1862203704

Battery level and calculated discharge rate were written to the FIT file by my battery monitoring data field, which itself uses an unknown amount of power.

It is clear to see the point around 1H40 at which I enabled GLONASS. Before the discharge rate was around 4%/hour, after that the discharge rate went up by more than 1%/hour

Switching on GLONASS had a very noticeable effect on position stability though. In the first 1H40 the recorded distance (ie drift) was 200 meters, in the next 1H30 with GLONASS on only 70 meters. The data points were also much more tightly grouped with GLONASS on.

So based on this not overly scientific test I would say leave GLONASS enabled provided you can life with the additional 1%/hour battery drain.

Next test is going to be 1s recording vs smart recording.
  • 0 over 7 years ago
    Interesting.

    Why was the discharge rate zero for the first 20 mins or so?

    When GLONASS first got introduced to Garmin watches, I picked it added 20% battery drain which seems to tie in with your figures.

    I can't recall seeing any real life demonstrations of why GLONASS may help when out on a run so useful to see it appears to have an effect here.

    Smart recording in "window sill" tests can be misleading as the smart interval can be many many minutes not a few seconds when on the move. As such it appears to give much better battery life than you might actually get in a proper activity. That said, provides an upper bound on what you can hope for.
  • 0 over 7 years ago
    Discharge rate is zero at the start because I cannot calculate it accurately until the level is below 99% (as the charge/discharge curves appears to flatten above that level) and the watch has reported a number of battery discharge steps below 99%. My app checks the battery level every second and records the new level and a time stamp whenever there is a change. It then uses linear regression to calculate the discharge rate using up to the last 8 samples. That's why the reported rate does not jump up instantly when I enable GLONASS.

    I agree with your observation that the window sill test is the best case scenario for smart recording. I might have to update my battery monitor app first as that records every second at the moment and that might obscure any effect from selecting smart recording
  • 0 over 7 years ago
    Nice, thanks for doing this!

    1% extra per hour of activity is not too bad... Especially if accuracy is indeed improved. I now usually run with GPS only btw and that is plenty precise already, but can imagine glonass adds it's value in high building and forest environments for example...

    But will see if I can do some tests today as well, just leave the watch running in the window for an hour or so. Not ideal but hey ;)
  • 0 over 7 years ago
    1% extra per hour of activity is not too bad... Especially if accuracy is indeed improved. I now usually run with GPS only btw and that is plenty precise already, but can imagine glonass adds it's value in high building and forest environments for example...


    I did those GLONASS yes/no tests already with my 920XT and 630. I run in heavily covered forest environments (south-west Germany) and I see no real difference if I use GLONASS or I don't use it.

    I usually did 2 weeks with and 2 weeks without GLONASS and then I visually checked the tracks in known places. Well it showed the usual derivation, I couldn't tell if GLONASS was on or off just by the track or the reported distance.

    For the 935 I came to the conclusion that at least it doesn't hurt (battery is no issue for me because I don't run ultras) and therefore I keep it activated.

    Of course that's just for my region, I can't tell you how that test would look like in Australia or in NYC.
  • 0 over 7 years ago
    So tried some testing myself as well, 1 hour in the window (OHR disabled btw):

    GPS Smart: https://connect.garmin.com/modern/activity/1863861545
    GPS+Glonass Smart: https://connect.garmin.com/modern/activity/1863863315
    GPS 1sec: https://connect.garmin.com/modern/activity/1863867514
    GPS+Glonass 1sec: https://connect.garmin.com/modern/activity/1863865568

    So basically:

    GPS Smart: 3.09% with 720m drift
    GPS+Glonass Smart: 4.16% with 510m drift
    GPS 1sec: 3.57% with 730m drift
    GPS+Glonass 1sec: 4.26% with 490m drift

    So indeed glonass adds a bit of more 'stableness' but drains some more battery (guess the 1% extra is about up to par). About 1sec vs smart, not sure what to think of it. Seems to add just a bit as well (makes sense, writing something every 1 second instead of every few) but seems way less. About 0.1%/hr (but wiht the GPS only the difference seems almost 0.5%)

    ​​​​​​​Interesting to see other results ;)
  • 0 over 7 years ago
    This is really great info to see. I am with hedistocker in regards to battery life not being an issue for me, but i'm gonna guess the Ultra-people (I think those people will like that correlation) will like to see data like this as a approx measure to just what they can do or how far they might go. 4.26% drain per hour (w/1sec) amounts to approx 24hrs of juice which sounds good to me. I don't know though if it would actually get that since I don't know if there is an impact on battery life with the use of BT/external sensors (and no OHRM, just connected to a chest strap), backlight (if any) and whatever other thing that could be a battery drain. Still, glad to see you took the time to get this up here.
  • Former Member
    0 Former Member over 7 years ago
    Interesting test results. I guess the difference may not be that noticeable in real use scenario but two different tests and similar results seems to speak for some advantage for leaving GLONASS on.

    I find interesting though that some people like Fellrnr report they get worse results with GLONASS. Could it be that different locations have less favorable satellite coverage or just day to day differences. Impossible to say. So far I'm not sure if I've really seen any notable difference in track or distance running the same routes.

    Another thing worth noting with current Garmin watches is that while on the move the total distance seems also affected by secondary speed sources like foot pod or wrist dead reckoning. This means the GPS only plays part of the total picture.

    I've seen the total distance hold together even when GPS has been a mess or when starting a run without GPS lock. This makes direct comparison more difficult and a stationary test like this may actually tell something we don't see on the move as here the watch doesn't get aid from any other sources.
  • 0 over 7 years ago
    I personally disregard everything fellrnr says about GPS watches. The man seems to generate his own EMI. Nobody gets results anywhere near as bad as his.

    I think the interesting thing to note from my test results is the difference in the drift pattern in the first half of the test (GLONASS off) compared to the second half. There are some largish jumps in the first half which don't seem to happen in the second half. My guess would be that this is were individual GPS satellites go in and out of view and the receiver determines a new position based on the new satellite constellation. It is at those moments that GLONASS helps fill in the gaps and the overall result is more stable. However, this doesn't automatically give a noticeable difference when you are running as other sources of movement information come into play as JTH9 mentions.

    I did this test outside with a clear view of the sky all around, rather than on the window sill where part of the sky is going to be obscured by the building you're in. When I tried a window sill test earlier the difference between GLONASS off and on was much less clear.
  • 0 over 7 years ago
    Gonna do some new tests today, 1.5hr and OHR enabled for the 4 situations... Will post the results at the end of the day :)
  • 0 over 7 years ago
    So tried some testing myself as well, 1 hour in the window (OHR disabled btw):

    GPS Smart: https://connect.garmin.com/modern/activity/1863861545
    GPS+Glonass Smart: https://connect.garmin.com/modern/activity/1863863315
    GPS 1sec: https://connect.garmin.com/modern/activity/1863867514
    GPS+Glonass 1sec: https://connect.garmin.com/modern/activity/1863865568

    So basically:

    GPS Smart: 3.09% with 720m drift
    GPS+Glonass Smart: 4.16% with 510m drift
    GPS 1sec: 3.57% with 730m drift
    GPS+Glonass 1sec: 4.26% with 490m drift

    So indeed glonass adds a bit of more 'stableness' but drains some more battery (guess the 1% extra is about up to par). About 1sec vs smart, not sure what to think of it. Seems to add just a bit as well (makes sense, writing something every 1 second instead of every few) but seems way less. About 0.1%/hr (but wiht the GPS only the difference seems almost 0.5%)


    What are these "drift" figures you refer to?
    Visual examination of these 4 suggests the ones with Glonass have a much greater "circle" of locations than the ones with GPS only.
    So to me GLONASS here makes things worse not better?