Battery use

786 views
Skip to first unread message

Hooloovoo

unread,
May 22, 2012, 9:36:30 AM5/22/12
to Osmand
Hi,

I love OSMAnd and use it all the time. One thing that causes me
problems is OSMAnd's high battery use.

As an example, I often go for long walks (5 hours+) with OSMTracker
running the whole time and it has very little effect on the battery.
If I use OSMAnd, I run out of battery very quickly.

To give some hard numbers, the last few days of long walks, I've had
OSMTracker running the whole time and only used OSMAnd for around a
minute every 15mins or so, to check I was on track. At the end of each
day, the Android battery usage thing says that OSMTracker has used
around 12% and OSMAnd around 80%. I have the "run OSMAnd in the
background" turned *off* and after I've checked my position, I 'back'
right out off OSMAnd to make sure it isn't running in the background
(I even checked in the list of running applications).

Does anyone have any ideas?

Regards,

Aaron

Robert Grant

unread,
May 22, 2012, 10:25:08 AM5/22/12
to osm...@googlegroups.com

Screen off as much as possible and switch to airplane mode to prolong buttery life. You can leave background service on continuously.  I achieve 7+ hour run times with this method.

sanderd17

unread,
May 22, 2012, 10:44:28 AM5/22/12
to osm...@googlegroups.com
You can't enable airplane mode when tracking with OsmTracker.

I think it's normal OsmAnd uses more battery: it has to render the maps (hence use a lot more CPU or GPU power). Or download the tiles (so extra power to the antenna). It has to calculate a route (more CPU power).

OsmAnd isn't a light application but I do agree that the numbers are a bit high. I get the same numbers: tracking with OpenGpsTracker and constantly following on the map with OsmAnd (not even routing): 82% of the battery goes to OsmAnd and 12% to OpenGpsTracker.

On the other hand, these statistics might be a bit wrong. I don't know how it's calculated when both apps need the GPS antenna.

Maybe I have to search on this.

Shorty66

unread,
May 22, 2012, 10:57:24 AM5/22/12
to osm...@googlegroups.com
Use vector maps, use native rendering and dont show magnetical compass bearing to reduce battery usage. Also, do not restart osmand every time, but leave it in the background to use LESS battery: Osmand needs some time to start up and doesn`t do anything in the background if background service is turned off. Start up uses more battery than having it in the ram.

Robert Grant

unread,
May 22, 2012, 10:58:39 AM5/22/12
to osm...@googlegroups.com
Why not try using your phone in airplane mode with only Osmand background on and screen off? If you must have tracking, use the feature provided by osmand or another GPS only app like Strava. That's what I do for long battery. Just see if that helps.

Robert
--
Sent from my Android phone with K-9 Mail. Please excuse my brevity.

Horst Müller

unread,
May 22, 2012, 11:26:56 AM5/22/12
to osm...@googlegroups.com
Still, I think it's a good question if OsmAnd can do some things within itself to save some battery. 

For example: Would it be possible to produce some kind of pre-rendering?
That means, I know the area/route I want to follow. Still at home, I can pre-render all of the necessary area and store it in a subdir. Following the route OsmAnd simply reloads these pictures instead of rendering them again. 

Things like that. It wouldn't change the overall power need, but adjust it better. At home I can plug in the power cord and don't have to worry about energy consumption. Later on, I do.
I don't know if this would save battery or is programmable at all, I'm no developer, sorry. But I think my main intention gets clear.



2012/5/22 Robert Grant <rgran...@gmail.com>

Victor Shcherb

unread,
May 22, 2012, 12:22:14 PM5/22/12
to osm...@googlegroups.com
My rough numbers are : 
1. 50% is screen
2. 30% GPS 
3. 20% CPU (offline rendering). For sure if you use tile maps this percentage is much less.

To drain your battery less you certainly need to use Background service.

I definitely didn't see a problem taking into account consumption of : screen, GPS and mobile network. 

As I remember Android has special statistics based on GPS, Screen... feel free to provide more accurate numbers. These number highly depends on the device. 

Best Regards,
Victor

2012/5/22 Horst Müller <lungen...@googlemail.com>

Hardy

unread,
May 22, 2012, 2:30:43 PM5/22/12
to Osmand
Here are my numbers:
- Use background service only, no screen
- Location provider GPX
- Positioning interval 30 sec
- Wait for fix 90sec.
Battery consumption is 6% per hour in this fashion, giving me a full
day of tracking on my expeditions


On 22 Mai, 18:22, Victor Shcherb <victor.shch...@gmail.com> wrote:
> My rough numbers are :
> 1. 50% is screen
> 2. 30% GPS
> 3. 20% CPU (offline rendering). For sure if you use tile maps this
> percentage is much less.
>
> To drain your battery less you certainly need to use Background service.
>
> I definitely didn't see a problem taking into account consumption of :
> screen, GPS and mobile network.
>
> As I remember Android has special statistics based on GPS, Screen... feel
> free to provide more accurate numbers. These number highly depends on the
> device.
>
> Best Regards,
> Victor
>
> 2012/5/22 Horst Müller <lungenfil...@googlemail.com>
>
>
>
> > Still, I think it's a good question if OsmAnd can do some things within
> > itself to save some battery.
>
> > For example: Would it be possible to produce some kind of pre-rendering?
> > That means, I know the area/route I want to follow. Still at home, I can
> > pre-render all of the necessary area and store it in a subdir. Following
> > the route OsmAnd simply reloads these pictures instead of rendering them
> > again.
>
> > Things like that. It wouldn't change the overall power need, but adjust it
> > better. At home I can plug in the power cord and don't have to worry about
> > energy consumption. Later on, I do.
> > I don't know if this would save battery or is programmable at all, I'm no
> > developer, sorry. But I think my main intention gets clear.
>
> > 2012/5/22 Robert Grant <rgrant51...@gmail.com>
>
> >> Why not try using your phone in airplane mode with only Osmand background
> >> on and screen off? If you must have tracking, use the feature provided by
> >> osmand or another GPS only app like Strava. That's what I do for long
> >> battery. Just see if that helps.
>
> >> Robert
> >> --
> >> Sent from my Android phone with K-9 Mail. Please excuse my brevity.
>
> >> sanderd17 <sander...@gmail.com> wrote:
>
> >>> You can't enable airplane mode when tracking with OsmTracker.
>
> >>> I think it's normal OsmAnd uses more battery: it has to render the maps
> >>> (hence use a lot more CPU or GPU power). Or download the tiles (so extra
> >>> power to the antenna). It has to calculate a route (more CPU power).
>
> >>> OsmAnd isn't a light application but I do agree that the numbers are a
> >>> bit high. I get the same numbers: tracking with OpenGpsTracker and
> >>> constantly following on the map with OsmAnd (not even routing): 82% of the
> >>> battery goes to OsmAnd and 12% to OpenGpsTracker.
>
> >>> On the other hand, these statistics might be a bit wrong. I don't know
> >>> how it's calculated when both apps need the GPS antenna.
>
> >>> Maybe I have to search on this.- Zitierten Text ausblenden -
>
> - Zitierten Text anzeigen -

Rob

unread,
Nov 5, 2012, 3:35:14 PM11/5/12
to osm...@googlegroups.com
I've found OsmAnd is great for navigating when cycling but does seem to be using a lot of battery.  In my case I am just viewing cached map tiles, panning and zooming maps whilst stationary, as well as using the location and direction indicator ('where am I').

It seems like a large part of this is due to the GPS.  PowerTutor suggests 400mW for GPS, 300mW LCD, and 150mW CPU.  Disabling GPS completely for the phone roughly halves consumption.

Turning off GPS isn't a solution for me though: the location indicator is very useful, and I am also using Strava to log my rides (Strava uses maybe 100x less power over the course of a 5hr ride, but is just logging every 3s without LCD).

Possible improvements might be:
1) use a passive location provider if available (in my case Strava will already have a recent location, presumably the same would be true of OsmTracker?)
2) only fetch one location (when not tracking or giving route directions) and then remove gps updates, or use a much longer update time (different defaults for car, cycle, foot modes?)

The magnetometer may also be using a significant amount of power (but it isn't listed by PowerTutor).  I use 'North is up' mode, but also enable the compass when no heading is detected (I would actually prefer the location indicator to always use the magnetometer since the current indicated direction is quite unreliable).  Perhaps increase this sensor delay too, and only update direction for a few seconds after the 'where am I' button is pressed?

Since other people's use cases will be different some of the above power savings could be automatically disabled when the phone is charging (eg. from a car).

thanks,
Rob
Reply all
Reply to author
Forward
0 new messages