Re: Attempting to access the GD-8200 GPS

[Greg Troxel]

Stuart Blake Tener <stuart.tener@bh90210.net> writes:

> I took your advise and took the laptop outside, right at about the 13
> minute mark satellites started to pop in on cgps' display. One, then
> two, then, bang 9 of them! I did not realize it could take 12 to 30
> minutes to obtain the almanac and ephemeral data, next time I will be
> far more sensitive to that fact. Moreover, the last outdoor test I did
> I was driving around with the laptop sitting in my front passenger
> seat. This time, I sat outside by the pool in my backyard with laptop
> sitting still on the table.
>
> It is remarkable that whence I returned into the house the number of
> satellites that the GPS receiver could hear diminished briskly and
> cgps even put "na" where the longitudinal and latitudinal data was for
> a few seconds then refilled it back in (this kept happening back and
> forth). This may be due to a poor antenna connection on the GPS,
> perhaps a problem with the receiver, or just that part of my house is
> a Faraday cage (I used to get zero mobile coverage in my master
> bedroom nor could I hit any ham repeaters in there, but step outside
> the open door and things improved). The indoor GPS testing was done in
> my dining room, where my mobile phone has always worked fine. I also
> will make sure my configuration specifies that this receiver is based
> on a SiRFstar III chip too.

Older receivers have few channels (12 is typical) and don't search for
satellites so well.  The almanac is long, so the sequence is
  - search for any signal
  - listen to it long enough to get almanac
    NB: this requires a strong enough signal to receive it without bit errors!
  - if there is a last location, use that to compute what is visible
    and assign receivers to those
  - if not, just search
  - when a satellite is found, decode ephemeris (30s)
  - eventually, compute a position

If alamanac, approx time, and position are retained, this is faster.

Newer receivers have ~150 and can find satellites from nothing much
faster.  Newer receivers are also more sensitive (lower noise figure).

Early on, GPS just didn't work indoors.  Now, with stronger signals and
mostly better receivers, it tends to work in wood-frame houses.  But if
you lose 15 dB, it doesn't really work.  It's clear that your particular
receiver and your particular house do not play well together.

> Putting PPS to the side, what do you recommend as a newer more quality
> USB GPS to use today? By the way, my computer does have a DB-9 serial
> port on it too, though they make USB to DB-9 serial cables these days
> too. What do you think of the "GlobalSat BU-353-S4" USB GPS?

The standard approach these days for cheap/decent is a u-blox 8 series.
7 is ok too, but you should be able to find an 8 in USB for < $20.  The
point is that it is among the best documented and best supported in
available code.  For cm to dm level positioning, the F9P is interesting,
but it's an entirely different price class.

You're asking about time sync, so you should read thoroughly:

  https://gpsd.gitlab.io/gpsd/gpsd-time-service-howto.html

before continuing.

Note that this is focused on high accuracy.  If you don't care about the
last 10 ms or so*, using time based on nmea sentence arrival works, once
you calibrate it with online peers.  I think this is not well-covered in
the howto.

* IMHO this is a reasonable position; while I like everything precise
  and accurate, most of the value of time sync on a computer is achieved
  by being below 1s, in terms of timestamps in logs being good enough.

signature.asc
Description: PGP signature