Re: [gpsd-users] Uncertain system time

[Gary E. Miller]

Yo Massimiliano!

On Thu, 3 Oct 2019 19:41:56 +0200
Massimiliano Fago <address@hidden> wrote:

> Il giorno mer 2 ott 2019 alle ore 19:48 Gary E. Miller
> <address@hidden> ha scritto:
> 
> > Note two things:
> >
> >     1) it is built in the directory that html is served from.
> >     2) all the options are in optionfiles.
> >
> > Here are the two options files:
> >
> >      # cat /var/www/localhost/htdocs/day/optionfile
> >     --clip
> >     --name=Kong
> >     --nice
> >     --outdir=day
> >     --period=1
> >     --terminal=pngcairo
> >  
> 
> perfect, work fine.

Good.

> In your experience, which type of GPS is best suited to do this type
> of application without spending a lot?

No need for "the best", that can be very expensive.  The granularity
of most CPU clocks is around 50 ns.  Worse than any modern PPS serving
GNSS receiver.

> Now I have purchased an ublox neo-m8n. When I get there I see the
> difference with the current one.

The NEO-M9N works great, and cheap (under USD$20).  No need for anything
better unless you have an oscilloscope and a TICC.  If you do, head
over to time-nuts.

> Sometimes someone approaches and tells me: "hey man, the time you
> show is not correct" showing me the one on the mobile phone with a
> gps app.

Hehehe.  Battery life is the current darling metric.  To get better
battery life, the phones play games with CPU sleeping, clock rate, etc.
All that derades the timing.

You can get NTP apps for phones that show you how bad your phone clock
is.  I use "ClockSync" on Android.  My experience is that the cell phone
clock quickly drift off by seconds once set accurately with NTP.

> How accurate can the mobile phone be, and do mobile phones
> have PPS signal management internally?

The cell phone standard require very accurate clocks in the base stations.
Pretty much every one has a Rubidium (Rb) standard!  Look on ebay and
you can but used one cheap (USD$200).  I have an SZ.22c.

Every cell phone gets to transmit only during its very narrow allocated
time window.  Each generation shrinks that window.  But that accuracy is
not passed on to the CPU.

> Now that everything seems ok, I can answer these man: "hey, there are
> no comparisons between the two systems" ...

It used to be you could use the hourly tone from a radio or TV station
to get the TOS.  But now everything is digital, buffer bloat everywhere,
so that no longer works.

RGDS
GARY
---------------------------------------------------------------------------
Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97703
        address@hidden  Tel:+1 541 382 8588

            Veritas liberabit vos. -- Quid est veritas?
    "If you can't measure it, you can't improve it." - Lord Kelvin

pgpyNwbBFa6ml.pgp
Description: OpenPGP digital signature