[LinuxPPS] Questions using Linux PPS version 5.x.x-rc1

[Thiago Tiedtke]

Hi,

I'm use LinuxPPS + NTP in a aplication that require a precision between 30us
and 250 to 350us relative to reference clock, and this combination (Linux
PPS + NTP) solve my problem.

In a recent visit in site:
http://wiki.enneenne.com/index.php/LinuxPPS_support, i can see which the
version of PPS is updated to 5.x.x-rc1, ok.

In my application (3 o 4 months old),  i configure the linuxpps with the
patch for linux kernel: linux-2.6.21.

I try use the new version, following the istructions on Rodolfo's site, and
not obtain the correctly work of my NTP using PPS + NTP drivers of
references clocks.

A description of my system, actually run in a previous version of linuxpps
5.x.x-rc1:

GPS:
  SuperStar II
    - 1 PPS Time mark +/- 200ns accuracy with time GPS
    - Serial Port (NMEA and binary format)

  PC:
   - Via Epia M - 600 MHz
   - 2 serial ports RS232

  SO:
   - Debian Linux 4.0 r1

But, i want update to 5.x.x-rc1  :).

Using the LinuxPPS page, following step by step, my system is obtaining the
PPS and the NMEA from GPS, see:

The pps0 is a parallel port, but i using a serial port in pps1.

tiedtke at timesync-debian:~$ sudo ppstest /dev/pps1
trying PPS source "/dev/pps1"
found PPS source "/dev/pps1"
ok, found 1 source(s), now start fetching data...
source 0 - assert 1191102137.613937290, sequence: 3 - clear
1191102138.612803506, sequence: 4
source 0 - assert 1191102138.613802949, sequence: 4 - clear
1191102138.612803506, sequence: 4
source 0 - assert 1191102138.613802949, sequence: 4 - clear
1191102139.612668091, sequence: 5


tiedtke at timesync-debian:~$ tree /sys/class/pps/
/sys/class/pps/
|-- pps0
|   |-- assert
|   |-- dev
|   |-- device -> ../../../devices/pnp0/00:09
|   |-- echo
|   |-- mode
|   |-- name
|   |-- path
|   |-- power
|   |   `-- wakeup
|   |-- subsystem -> ../../../class/pps
|   `-- uevent
`-- pps1
    |-- assert
    |-- clear
    |-- dev
    |-- device -> ../../../devices/platform/serial8250
    |-- echo
    |-- mode
    |-- name
    |-- path
    |-- power
    |   `-- wakeup
    |-- subsystem -> ../../../class/pps
    `-- uevent

8 directories, 17 files

I apply the patches for this location:

http://ftp.enneenne.com/pub/misc/linuxpps/refclocks

for NMEA e ATOM (driver 20 and 22) to source of ntp. I'm compiling the ntp,
ver 4.2.4-p4, with the options:

./configure --disable-all-clocks --enable-ATOM --enable-LOCAL-CLOCK
--enable-NMEA --enable-SHM --enable-parse-clocks

And the ntp.conf:

server  127.127.1.0     # local clock
fudge   127.127.1.0     stratum 10

server 127.127.20.0 prefer mode 2 minpoll 4
fudge 127.127.20.0 flag3 1 flag2 0

server 127.127.22.0 minpoll 4 maxpoll 4
fudge 127.127.22.0 flag3 1 flag2 0 time1 0.000

driftfile /usr/local/etc/ntp.drift
logfile   /usr/local/etc/ntp.log


But, waiting 30 min, the PPS reference clock appear not work, and the
estimated error has a great value, beyond very unstable, see:

tiedtke at timesync-debian:~$ ntpq -p && ntpdc -c kerninfo
     remote           refid      st t when poll reach   delay   offset
jitter
==============================================================================
 LOCAL(0)        .LOCL.          10 l   22   64  377    0.000    0.000
0.001
*GPS_NMEA(0)     .GPS.            0 l    9   16  377    0.000  -12.802
1.134
 PPS(0)          .PPS.            0 l    -   16    0    0.000    0.000
0.001

pll offset:           -0.012358 s
pll frequency:        172.479 ppm
maximum error:        0.019243 s
estimated error:      0.000603 s
status:               0001  pll
pll time constant:    4
precision:            1e-06 s
frequency tolerance:  512 ppm

What I am making of made a mistake?

Thanks for any help!!

Regards,

Thiago Tiedtke dos Reis
Geomatic Engineer
Unesp - Universidade Estadual Paulista
Presidente Prudente - SP - Brazil
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://ml.enneenne.com/pipermail/linuxpps/attachments/20070930/9966376d/attachment.html