GPS

Kevin Zuber Kevin at k-wh.de
Mon Jul 7 12:07:22 CEST 2008


Hi,

I did some tests yesterday and here are my experiences with Freerunner's
GPS:

First I put the Freerunner on a window sill in a penthouse, no high-rise
around. I got a fix after ~20 minutes. Disappointing.
Later on I walked along in a big city between a lot of high-rises. After
15 minutes, the neo found the first satellite, but nothing more, no fix,
only a gps-time. After 30 minutes I activated the power management
(first dim, than lock) so the neo could be sleeping again, I gave up.
Very Disappointing.
A few hours later I sat down on a canvas chair at a bank of a more than
510 meters wide river in the same city, the last high rises are more
than 100 meters behind me. I woke up my neo, unlocked it, started tango
gps, looked at it. Nothing. I looked to the left and to the right,
looked again at my neo. Stop. looked again at my neo and couldn't
believe my eyes. There are 5 satellites in the display, the gps chip is
using 3 of them. I GOT A FIX in less than 20 seconds! Very amazing!
After one more minute, the neo was using 8 of 11 satellites. That was a
really great experience. After the neo got the fix, I nearly could do
everything I want, it won't loose it. I walked back in the city, worn
the neo inside the pouch in my trouser pocket, around me all this big
high rises. It was still tracking me and drawing this very interesting
red line in my tango gps map. I couldn't believe it.
I went to the home of a friend, gone up the staircase and took the neo
out of my pocket in the apartment. I couldn't believe my eyes again.
Tango gps was still drawing the red line, even in the staircase.
It looked like:
___________
| _________|
| |_
|
Unfortunately I missed the chance to take a screenshot, but it was
really amazing. 

My conclusion is that the chipset can't be so bad :)
But the question is: Why is it so hard to get a fix, when there is
something around? Is there any space for software improvements? Maybe
different search algorithm or is it all in the hardware or driver-side?
(All without an external gps antenna)

Kevin


Am Montag, den 07.07.2008, 09:40 +0200 schrieb Kai Römer:
> Hi Al,
> 
> Sounds really convincing, but how do you explain the constantly fast
> fix via external antenna then. I really think its an antenna issue.
> 
> Also the difference of the GPGSV values support this idea.
> 
> Tomorrow evening i will ask a specialist to check the antenna signal
> qualities. Maybe a cable is broken or there is a short circuit on the
> main board.
> 
> I ll report about the results.
> 
> CU Kai
> 
> 2008/7/6 Al Johnson <openmoko at mazikeen.demon.co.uk>:
> > From what I've seen on the wiki the version of the Antares4 on the GTA02
> > doesn't have the memory needed to store almanac and ephemeris, last known
> > position or time. This means that every start is a true cold start, unlike
> > every other reasonably modern GPS we're comparing it to. It starts up
> > thinking the time is midnight on 30th November 1999 and seems to need a fair
> > bit of decent signal to convince it otherwise, contributing to the long
> > startup time.
> >
> > It looks like there is a way around this if you look at the documentation for
> > the assist. The AID-INI message needn't be supplied by a remote server; we
> > can generate it locally to provide the sort of data that's stored internally
> > most of the time. At the very least we have a fair idea of the current time
> > and date. We should also be able to store location, almanac and ephemeris
> > when we shut down the GPS, and provide it at the next startup. We can also
> > have a stab at current location, based perhaps on cell ID or wifi data as
> > discussed by some of the other threads, or on user input.
> >
> > I'll try to patch together something to do this based on the example perl
> > client and server code, and see how much difference it makes.
> >
> > On Friday 04 July 2008, Kai Römer wrote:
> >> I can affirm this for 6 opemoko devices. i guess its an internal
> >> antenna issue. as soon as you connect a external antenna to it works
> >> like a charm. but fur me thats no solution.
> >>
> >> TTFF with external antenna (perfect condition): 40 to 60 seconds
> >> TTFF with internal antenna AGPS (perfect condition): more than 1:20
> >> minute but not always. its like gambling.
> >>
> >> I guess a miss design of the internal antenna.
> >>
> >> CU Kai
> >>
> >> 2008/6/23 Peter Kraker <peter.kraker at volja.net>:
> >> > This timings are insane unless you don't even have a valid almanac, which
> >> > is rare. This doesn't look right.
> >> >
> >> > Yorick Matthys pravi:
> >> >
> >> > Marcus Bauer said:
> >> >
> >> >
> >> > My experience with the Freerunner is ~12 minutes TTFF (time to first
> >> > fix) without use of agps and ~4-8 minutes TTFF with agps from
> >> > agps.u-blox.com using the software from openmoko.
> >> >
> >> > The Neo1973 (GTA01) had a TTFF without agps assistance of ~2 min.
> >> >
> >> >
> >> > 12 minutes without AGPS and 4-8min with AGPS??
> >> > I hope there was a thunderstorm inside the basement where you tested
> >> > this...
> >> >
> >> > :)
> >> >
> >> > Seriously, these just don't seem realistic.
> >> > Compare them for example with some other devices from 2003:
> >> > http://www.pocketgpsworld.com/ttffcomparisons.php.
> >> > Or from ublox: http://www.u-blox.com/technology/assistnow/ (table at the
> >> > bottom of the page)
> >> >
> >> > Surely there must be something wrong with your
> >> > software/settings/hardware/environment...
> >> > (or maybe they still have a lot of work to do on the GPS :))
> >> >
> >> > y
> >
> > _______________________________________________
> > Openmoko community mailing list
> > community at lists.openmoko.org
> > http://lists.openmoko.org/mailman/listinfo/community
> >
> 
> _______________________________________________
> Openmoko community mailing list
> community at lists.openmoko.org
> http://lists.openmoko.org/mailman/listinfo/community





More information about the community mailing list