* Re: [Thumbgps-devel] cold start of Macx-1 GPS receiver
2012-05-11 4:20 ` [Thumbgps-devel] cold start of Macx-1 GPS receiver Jau-Yang Chen
@ 2012-05-11 15:29 ` tz
0 siblings, 0 replies; 2+ messages in thread
From: tz @ 2012-05-11 15:29 UTC (permalink / raw)
To: thumbgps-devel; +Cc: veronica, tcf, Jau-Yang Chen, gloria, lily, Gary E. Miller
[-- Attachment #1: Type: text/plain, Size: 4032 bytes --]
Cold start should be under 10 minutes with an open sky. It would help
taking the unit on a USB battery pack outside; I usually see 5 minutes or
less - that is what the "65" channels do although not more than 12-14 will
ever be used to track satellites. Usually there are at least 10 sats
visible under such conditions. Once locked and the ephemeris is obtained
for each satellite, it can warmstart back inside quickly if there are
enough satellites even in a partial view.
The first problem is getting a lock - it acts like a scanner initially
searching the entire band. If only a few satellites are visible, the
chances of hitting one are lower. Then it has to maintain the lock for
around 40 seconds to get the complete ephemeris of that one satellite. To
get all satellites might take many hours as their orbits bring them into
and out of the view. It needs 3 satellites for a 2d fix, 4 for a 3d.
Usually there is an algorithm preventing use of those with too low an
elevation. It needs a clearer signal to achieve the initial lock than to
track (it uses doppler so it is easier to maintain or find once you know
the expected value, otherwise you have to tune carefully).
So if you have enough problems, a very restricted sky view, multipath,
intermittent signals (the 40 seconds must be uninterrupted), it might take
an hour to get that for 4 satellites at the same time (as they move in and
out of the aperture or lose signal), get their full ephemeris, and start
solving. But once it has a complete set of ephemeris data and knows the
general location it will start very quickly.
You can usually see this with the satellite display - the circle with
azimuth and elevation and signal strength.
I have to have my antennas right near my window to get a reasonable
cold-start.
On Fri, May 11, 2012 at 12:20 AM, Jau-Yang Chen <cjy@navisys.com.tw> wrote:
> Dear Gary,
>
> Frankly, we are curious why one GR-601W sample took so long time to fix the
> position when you tested it in the very first time. Did you ever issue cold
> start to test it? Is it back to normal (say, to fix position within one
> minute under open sky) afterwards?
>
> Yes, per tz's comment, if GR-601W keeps the ephemeris data collected in
> Taiwan, it won't help to hot/warm start in USA. On the contrary, some old
> GPS engines might mislead the internal algorithm and get deadlock. Of
> course, u-blox6 engine will apply the cold start algorithm automatically to
> re-collect the correct ephemeris data.
> > Also the Ephemeris doesn't help if you move more than about 500 miles
> from
> the last place it was turned on. Until it can get at least a 2d fix and
> new latitude/longitude, it won't know which satellites to look for.
>
> Best regards,
>
> JY
>
> -----Original Message-----
> From: Gary E. Miller [mailto:gem@rellim.com]
> Sent: Friday, May 11, 2012 12:44 AM
> To: Jau-Yang Chen
> Cc: 'Dave Taht'; thumbgps-devel@lists.bufferbloat.net; 'veronica';
> tcf@navisys.com.tw; gloria@navisys.com.tw; 'lily'; Eric S. Raymond
> Subject: Re: [Thumbgps-devel] supercap vs. battery in Macx-1 GPS receiver
>
> Yo Jau-Yang!
>
> On Thu, 10 May 2012 18:38:41 +0800
> "Jau-Yang Chen" <cjy@navisys.com.tw> wrote:
>
> > After being fully charge, the supercap in Macx-1 could last around 4
> > hours, and around one week for Li-ion battery,
>
> Well, that explains my one hour cold start when receiving the sample units.
> One week hold-up is pretty short, and 4 hours hardly worth the effort.
>
> Luckily the thumbgps project is for always on, but for weekend use neither
> is good. Most mouse GPS have batteries are good for a month or more of
> standby.
>
> RGDS
> GARY
> ---------------------------------------------------------------------------
> Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97701
> gem@rellim.com Tel:+1(541)382-8588
>
> _______________________________________________
> Thumbgps-devel mailing list
> Thumbgps-devel@lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/thumbgps-devel
>
[-- Attachment #2: Type: text/html, Size: 5021 bytes --]
^ permalink raw reply [flat|nested] 2+ messages in thread