[LinuxPPS] [PATCH]

[Bernhard Schiffner]

On Saturday 02 February 2008 16:19, Udo van den Heuvel wrote:
...

> Curently I only have $GPRMC messages from the gps-18.
> We migth also need $GPGSA, right?
> Any better messages?

My receiver, most NMEA-messages on
(Standard)
(Example)
(Explanation)

$GPGGA,hhmmss.ss,Latitude,N,Longitude,E,FS,NoSV,HDOP,msl,m,Altref,m,DiffAge,DiffS*cs
$GPGGA,134420.041,5101.3710,N,01346.1144,E,1,04,3.1,74.3,M,,,,0000*3C
FS: Fix - 1 (normal)
NoSV: SV used - 4

$GPGLL,Latitude,N,Longitude,E,hhmmss.ss,Mode*cs
$GPGLL,5101.3710,N,01346.1144,E,134420.041,A*31
Mode: A - 2D/3D-fix

$GPGSA,Smode,FS{,sv},PDOP,HDOP,VDOP*cs
$GPGSA,A,3,11,19,20,28,,,,,,,,,4.3,3.1,3.0*34
FS: Fix - 3  3D

$GPRMC,hhmmss,status,latitude,N,longitude,E,spd,cog,ddmmyy,mv,mvE,mode*cs
$GPRMC,134420.041,A,5101.3710,N,01346.1144,E,0.22,12.87,030208,,*33
status: A - vaild position

For me:
1.) GGA NoSV>=4 (best)
2.) GSA : 3D-Fix (good enough)
3.) RMC: valid position (weak)
4.) GLL: mode (weak)

> So first we start parsing these as well.
> Then we must 'fail' the pps reception if the amount of sats is below a
> settable number. (ntp.conf?)
>
> Is this correct?

Depends on your demands. I try to compare sources:

wallclock (mostly unmaintained) 5minutes

LCD-TV 2000 ms (because of image processing)
analog radio 500 ms (your hands are slow)

NTPD with analog radio (DCFa) 200ms (demodulation is "special")
NTPD by modem 80 ms (remember: max error = half roundtrip)
NTPD with analog radio (DCFa) 5ms (adjusted against better source)
NTPD over LAN 0,5ms

GPS 1 SV (no other information, receiver initialisation) 30ms
GPS 3 SV (guessing surface of earth, 2D-solution, happens sometime) <0,030  ms
GPS >3 SV (3D-Solution) <0,001 ms

My favorite GPGGA$ with NoSV >=4

Bernhard