another GPS idea - speed-sensitive voicemail (Power usage)
OpenMoko at mauve.plus.com
Mon Jul 30 19:05:08 CEST 2007
Visti Andresen wrote:
> On Mon, 30 Jul 2007 16:59:51 +0100
> Ian Stirling <OpenMoko at mauve.plus.com> wrote:
>>Visti Andresen wrote:
>>>While I do find most of the GPS based ideas intriguing I do have one comment.
>>>I wonder if you have realised that the GPS consumes large quantities of current.
>>>(45 mW according to http://wiki.openmoko.org/wiki/Neo1973_Power_Management#Approximate_power_draw_of_various_subsystems)
>>>The 1200mAh battery will be drained in 1200mAh/(45mW/3.7V) = 98h ~= 4 days
>>>And this is for the GPS receiver alone and no power conversion losses,
>>>but GPS reception also requires CPU power..
>>This is not quite correct.
>>The GPS can be turned off and on.
>>45mW is for continuous positioning, 1s/45s will give completely fine
>>positions, though not as accurate as if it was on for longer.
>>If the 'Resume in under 1s' from sleep entry on the table at the front
>>of the above page is done, then the numbers change again.
>>If it takes 350mW for 1s to wake the device, and then 80mW for another
>>second (CPU in slow mode and GPS on) that's 430mWs.
>>One position per 45s, that's some 9mW average, or around 5% of the
>>battery per day.
>>It seems likely from the basic figures that this can be stretched to 1s
>>wake every 3 minutes or so, for around 1.5%/day.
>>After this, you need to wake up for more than a second to keep the
>>position current, so after 3 minutes, the next halving in battery use
>>might end up at 4 hours or so.
>>This can also do bluetooth and WiFi polling of course (at a slight power
> But does a GPS not require large amounts of time to get a "new" fix?
Look at the Hammerhead/Protocol page on the wiki.
As a very brief recap, not if you have a sufficiently good knowledge of
where the satellites are, and the local clock has not drifted by more
than 1/3ms or so.
(I'm involved in the reverse engineering of the GPS chip)
More information about the community