Reason for GPS problems found!

[Brian C]

Andy Green wrote:
> Yes GPS chip is very "sticky", even when depowered for some time it
> seems to hold state and not act in a deterministic way.  It makes it
> hard to know if the last thing you changed is responsible for the
> behaviour staying the same or changing, or if it was going to do that
> anyway.  That makes debugging it very confused.
> 
> -Andy

While the issue you mention may have affected my test, here's the results:

1. Upgraded to today's kernel:

root at om-gta02:/# uname -a
Linux om-gta02 2.6.24 #1 PREEMPT Fri Jul 18 02:16:41 CEST 2008 armv4tl
unknown

2. Tested GPS TTFF without SD card (from an indoor position not near a
window): 306s

3. Shutdown and insert SD card

4. Tested GPS TTFF with SD card (from same indoor position not near a
window): 484s

Previously from this same spot with SD card I could let it run all night
long and never get a fix, so maybe the fix in step 2 is helping the fix
in step 4, but the impression I get is that the most recent kernel has
improved things.

Brian

Note: I get much better TTFF times outdoors, but didn't have opportunity
to check.