NMEA and PPS - NTP

This is a discussion on NMEA and PPS - NTP ; Hi the rtfm is confusing on the subject of NMEA and PPS. I have an external RS-232 GPS rx with NMEA output feeding /dev/gps0 and 127.127.20.0 in ntp.conf (on intel/linux platform). With the correct use of fudge time1 I can ...

+ Reply to Thread
Results 1 to 4 of 4

Thread: NMEA and PPS

  1. NMEA and PPS

    Hi
    the rtfm is confusing on the subject of NMEA and PPS. I have an
    external RS-232 GPS rx with NMEA output feeding /dev/gps0 and
    127.127.20.0 in ntp.conf (on intel/linux platform). With the correct
    use of fudge time1 I can get my ntpd to act as a stratum 0 server and
    will track pool ntp servers to within 100ms drift . I wish to improve
    the precision/accuracy and would like to use the PPS kernel/drivers. Is
    this possible with a simple NMEA device i.e. does the PPS kernel patch
    give a better stability or do I need a PPS gps clock ?


  2. Re: NMEA and PPS

    In article <1152047875.382440.265180@75g2000cwc.googlegroups.c om>,
    "DJ" wrote:

    > use of fudge time1 I can get my ntpd to act as a stratum 0 server and


    You mean stratum 1. A value of 0 indicates unsynchronised. The reference
    clock will be stratum zero, but the NTP protocol will report statum
    1.

    > will track pool ntp servers to within 100ms drift . I wish to improve
    > the precision/accuracy and would like to use the PPS kernel/drivers. Is
    > this possible with a simple NMEA device i.e. does the PPS kernel patch
    > give a better stability or do I need a PPS gps clock ?


    You require a time transfer receiver. Your problems are not fundamental
    to NMEA, but navigation receivers will generally not pay attention to
    accurate timing of the time messages and even time transfer receivers
    may assume that their PPS output is accurate, so the NMEA doesn't need
    to be. I don't know if there are any time transfer receivers without
    PPS, although, if there are, they would, presumably have low jitter
    NMEA outputs.

    The ntpd PPS code is of no benefit unless the receiver generates an
    accurate PPS signal.

  3. Re: NMEA and PPS

    Dufus - I meant stratum 1 !!!

    Thanks for the info. NO need to build a PPS kernel then !

    David

    David Woolley wrote:
    > In article <1152047875.382440.265180@75g2000cwc.googlegroups.c om>,
    > "DJ" wrote:
    >
    > > use of fudge time1 I can get my ntpd to act as a stratum 0 server and

    >
    > You mean stratum 1. A value of 0 indicates unsynchronised. The reference
    > clock will be stratum zero, but the NTP protocol will report statum
    > 1.
    >
    > > will track pool ntp servers to within 100ms drift . I wish to improve
    > > the precision/accuracy and would like to use the PPS kernel/drivers. Is
    > > this possible with a simple NMEA device i.e. does the PPS kernel patch
    > > give a better stability or do I need a PPS gps clock ?

    >
    > You require a time transfer receiver. Your problems are not fundamental
    > to NMEA, but navigation receivers will generally not pay attention to
    > accurate timing of the time messages and even time transfer receivers
    > may assume that their PPS output is accurate, so the NMEA doesn't need
    > to be. I don't know if there are any time transfer receivers without
    > PPS, although, if there are, they would, presumably have low jitter
    > NMEA outputs.
    >
    > The ntpd PPS code is of no benefit unless the receiver generates an
    > accurate PPS signal.



  4. Re: NMEA and PPS

    DJ wrote:
    > Hi
    > the rtfm is confusing on the subject of NMEA and PPS. I have an
    > external RS-232 GPS rx with NMEA output feeding /dev/gps0 and
    > 127.127.20.0 in ntp.conf (on intel/linux platform). With the correct
    > use of fudge time1 I can get my ntpd to act as a stratum 0 server and
    > will track pool ntp servers to within 100ms drift . I wish to improve
    > the precision/accuracy and would like to use the PPS kernel/drivers. Is
    > this possible with a simple NMEA device i.e. does the PPS kernel patch
    > give a better stability or do I need a PPS gps clock ?
    >



    It all depends if your gps has a pps output - rtfm(receiver) - some (like the gps 18lvc do, others
    like the gps18serial don't). Having pps is independent of being an NMEA device (both of the
    examples above are NMEA).

    If you do have pps on your gps then you need to figure out how to get it to ntp and that will
    involve ether a kernel rebuild (FreeBSD) or the user space pps code (Linux 2.6) or ppskit (linux
    2.4) google for your os and pps will normaly yield info.

    John

+ Reply to Thread