Re: [gpsd-dev] PPS support

[Eric S. Raymond]

Sanjeev Gupta <address@hidden>:
> I started using gpsd three years ago, to interface with NTP servers I am
> running for the pool.  The issue I had, repeatedly, was to check if my
> devices (all cheap consumer stuff) reported a PPS, and if this was being
> passed to NTP in a usable fashion.
> 
> Unfortunately, there seemed to be no way to even know reliably if the
> device was sending a PPS.  The list provided lots of advice, including
> looking at SHM segments, tuning, calibration, etc.  Note the problem: not
> that PPS did not work, but I did not know if it was working.

Hm. It appears that I erred in writing ppscheck and then dropping it into 
the contrib directory rather than documenting it as a fully-fledged tool.
Fortunately - as you note - we fixed this by teaching gpsmon how to display
PPS bars.

> Over the last couple of months, PPS display support has become a
> non-brainer.  Thanks to the developers (I am looking at you, Mr Miller), I
> have been meeting local enthusiasts, plugging their devices into my laptop,
> and in every case, seeing both a PPS display in PPS, and NTP locking onto
> it.  Note that all these devices are "unbranded", bought off eBay.

Huh? What? We've had the devil's own time finding inexpensive PPS-capable
devices.  Tell us more; where can we dig these up?

> killall gpsd
> gpsd -n /dev/ttyUSB0
> /etc/init.d/ntp restart
> gpsmon
> (go for a smoke)
> ntpq -pn

Gary, that looks to me like an improvement on this:

$ su - (or sudo -s )
# killall -9 gpsd ntpd
# ntpd -gN
# sleep 2
# gpsd -n /dev/ttyXX
# sleep 2
# cgps

But the technical content in the Time Service HOWTO is your baby and I don't
want to change that without clearing with you.

> Could someone who has a wider range of devices see if this item still needs
> to be in the TODO ?
> ==
> **** In gpsmon's PPS Offset field
> 
> Presently PPS Offset is displayed in direct mode for NMEA, SiRF, and
> UBX devices.  Others should probably do likewise, notably the
> Motorola Oncore and Garmin drivers.

It does.  Garmin support is on Gary's to-do list.

> Thank you, thank you, thank you

You're welcome. I've been consciously steering the project in the direction
of supporting time service better; nice to see it's having an effect.
-- 
                <a href="http://www.catb.org/~esr/";>Eric S. Raymond</a>