Nokia N900

[Peter Dobratz]

This may be stating the obvious, but it tripped me up when I was
trying out my Garmin Etrex.  You have to be still in order to get a
fix.  I had the Garmin attached to my bicycle and I was riding up and
down my driveway while waiting to go off and map the neighborhood.  No
fix until I *stopped* for a minute or so.

Peter

On Sun, May 16, 2010 at 8:42 PM, David Rysdam <david at rysdam.org> wrote:
> On 05/13/2010 08:19 AM, David Rysdam wrote:
>> On 05/12/2010 06:13 PM, Joshua Judson Rosen wrote:
>>>     Cold tests
>>>
>>>          were performed when both Pocket PC and GPS receiver were
>>>          powered off (if GPS receiver has a separate power source, the
>>>          GPS receiver was unplugged) for a period of between 8 to 12
>>>          hours requiring a cold start.
>>>
>>>     Warm tests
>>>
>>>           Warm tests were performed when both Pocket PC and GPS
>>>           receiver were powered off (if GPS receiver has a separate
>>>           power source, the GPS receiver was unplugged) for a period
>>>           of 30 minutes
>>>
>>>     Hot tests
>>>
>>>           Hot tests were performed when both Pocket PC and GPS
>>>           receiver were powered off (if GPS receiver has a separate
>>>           power source, the GPS receiver was unplugged) for a period
>>>           of 15 minutes
>>
>> You missed the colder-than-cold test:
>> -
>> Factory is where the receiver has no knowledge whatsoever of Almanac
>> data in turn to locate the satellites and retrieve Ephemeris data, and
>> for a full Almanac to be downloaded can take approx 12.5 mins, hence
>> most companies suggest a factory start of 15 minutes.
>> -
>>
>> So we already know the N810 has to be worse than it should be, due to
>> multiple reports even on this list of >15 minute TsTFF.
>>
>> Also, I apparently "cold test" my dedicated GPS 1-2 times per day.  It's
>> always read to be a GPS before I'm ready to be a driver.  Perhaps that
>> says more about how long it takes me to start driving than about how
>> fast my GPS is.
>>
>> That said, I'll try to test it.  I'll assume the N810 needs a factory
>> start by this point and then start doing some cold timings.  Assuming it
>> doesn't take so long that I have to terminate the test.
>
> OK, so my tests are complete.  I controlled the GPS with MaemoMapper,
> which shows how many satellites you've heard from (I guess) and what the
> signal strengths are.
>
> I'm basing average performance on these tests:
> http://www.pocketgpsworld.com/ttffcomparisons.php
>
> 1) Until Saturday, I hadn't had a fix for many months, so I was going to
> count that as a "factory" start.  Factory start allows for around 12-15
> minutes.  I got a fix in under 4 minutes.  Then I realized that although
> I was outside, I was still WiFi'd in to the house so it was using AGPS.
>  Result: Invalid.
>
> 2) A little over an hour later, I tried from a WiFi-free location.
> 60-90 minutes is more than 30 minutes (warm test) but much less than
> 8-12 hours (cold test).  According to the baseline, cold tests should
> take ~1 m, warm tests ~30s.  N810?  5 minutes.
>
> 3) As soon as I had a fix for a few seconds, I disconnected and retried.
>  This would be a hot test and should take under 10-20 seconds.  N810?  5
> minutes *again*.
>
> 4) Over 12 hours after that, I disabled WiFi at home and tried again.
> This should be a strict cold test and again should take 1 minute.  N810?
>  13 minutes.
>
> Furthermore, I think it would have taken longer if I'd had more
> patience.  It went from 0-5 satellites (I thought 4 or 5 was enough, but
> it didn't complete then), then wandered down to 1, then back up to 5.
> This was at the 12 minute mark.  So I pointed it up at the sky (it had
> been resting on the ground) and suddenly all the signal strengths leapt
> up and I got a fix within a minute after that.
>
> Conclusion: The N810 GPS hardware and/or software definitively sucks.
> _______________________________________________
> gnhlug-discuss mailing list
> gnhlug-discuss at mail.gnhlug.org
> http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
>