[LinuxPPS] (no subject)

[Remco den Besten]

I recently published about using LinuxPPS in conjunction with DCF77 pulses
and some graphs.
A while ago I got a remark about the fudge factor in conjunction with driver
8, related to the precision reported by me (see topic Feedbacks, April 
2008).

Here are the peerstats from my DCF77-PPS experiment:

[/home/remco]> peerstats
       ident     cnt     mean     rms      max     delay     dist     disp
==========================================================================
192.53.103.108   602   -1.076    2.344   19.075   33.836  982.405   29.138
127.127.8.0      655   -0.012    0.265    1.179    0.000    1.102    1.012
127.127.22.0    1320    0.002    0.374    1.634    0.000    0.763    0.565
195.169.124.100  654    0.131    0.162    0.564   14.304   13.106    3.440

I consider the offsets both the PARSE driver (8) and ATOM driver (22) very
good.
The RMS's of these offsets reasonable, but related to the time reference
used (a 'radioclock at approx. 500 km distance)' understandable.

The following lines in ntp.conf may be relevant:

#ATOM (rising edge, flag2 0)
server 127.127.22.0 minpoll 5 maxpoll 5
fudge 127.127.22.0 flag2 0 flag3 1 stratum 0 refid PPSa

#PARSE Conrad RAW DCF77 (time1 0.2374)
server 127.127.8.0 mode 133 prefer
fudge 127.127.8.0 time1 0.2374 flag1 0 time2 0.0282 flag2 0 flag3 1 stratum 
0 refid DCFa

Time1 was determined experimentally (a long time ago) and time2 was taken
from (the experimentally
derived) time1 for the radioclkd2 (shm) DCF77-driver.