GPS Problem (maybe antenna?)

[Sebastian M.]

Hi list,

I have some gps problems on my Freerunner, but it seems like they do  
not origin from bad soldering of the internal gps connector (removed  
the plug with a small watchmakers screwdriver and examined it with my  
eyes [don't have much more at my dorm] and it seems to be ok (inner  
pin is not malformed, all solderings *look* alright - but could still  
be bad)).

My setup is: Om2008.8 with default sources plus angstrom distribution  
source. gpsd is installed and configured (/etc/default/gpsd stuff),  
agps is set up as well, I use a script /usr/bin/agps which contains:

> agps-onlinec -c full -u "xxxmailxxx" -k "xxxpassxxx" -la 49.362960  
> -lo 7.773482 -al 400 -p 500.00

Those coordinates are from Google Maps, verified at Wikipedia (as well  
as the altitude, 400m). Accuracy is 500 km (if I'm not wrong!).

I do all my testing with the Freerunner pointing towards the sky (like  
OpenMoko Staff did for their measurment) and standing on the edge of  
my balcony (50% sky visible).

Then I run those commands by ssh over WiFi (Mokos IP is 192.168.178.26):

On Moko: /etc/init.d/gpsd start
On Laptop: xpgs 192.168.178.26

This shows the current GPS data on my laptop - which is nothing (not  
even seeing a satellite) and does not change even after 5 minutes  
(remember: no SD Card inserted).
I once had a usb GPS device which saw satellites in the very same  
position at an instant and took about 40 seconds to get a fix.

Now I run agps:
root at om-gta02:~# agps
connect to agps.u-blox.com
aid data 2664 bytes received
aid data raw written ok

This correctly "injects" satellite data into the gps as xgps now shows  
some satellites with a Signal-to-Noise Ratio of 0, but the gps still  
does not see any of them (and judging from the plotted data and that I  
know where north is it should be able to see at least 5 satellites).
Letting the MoKo run for 40 minutes doesn't make it any better.

I tried this with Om2007.2 (on SD), Om2008.8 (on Flash), FSO Milestone  
2 Stable (on SD), FSO Milestone 12. August from totalueberwachung.de  
(on SD).
All systems were "opkg update && opkg upgrade"-d, apgs configured and  
gpsd configured.
Using something else than xgps (tangogps on moko, agpsui on moko,  
gpspipe -r) shows the same results.
Well, gpspipe shows those "I have no clue where I  
am"-position-datasets and after using agps it shows the injected  
satellites => Software should be alright alright.

This leaves either the internal connector or the antenna itself.
Because the connector *seems* to be ok I think there could be also a  
problem with the antenna. The soldering one can see after removing the  
rearcover looks not very good, but I can't measure any of them right  
now.

As soon as I get to my parents (propably on monday) I can use my  
fathers equipment (multimeter for solderings, a nice high frequency  
analyzers for the antenna :-D) and knowledge to search for the  
problem, but in case someone has any hints and/or ideas what I could  
try at my dorm (not much hardware here besides some screwdrivers, a  
linux laptop and a wireless network ;-)) I would happy to try that.

Any help would be appreciated.

Good night, Sebastian




(I hope my english is understandable - I'm quite tired because I have  
been working on this the last 7 hours, and before that was at my  
system admin job for about 6 hours and I am somewhat addicted to my  
Freerunner so I got not much sleep last night :-D)

-- 
Sebastian M.
Student of Computer Science at the University of Kaiserslautern.