[ntp:questions] Re: Configuring dual refclock/PPS sources

John R. Ackermann N8UR jra at febo.com
Fri Jan 16 14:11:55 UTC 2004


I went ahead and tried to configure ntpd to use a pair of Z3801A 
GPS-disciplined oscillators.  I think I can verify that, at least under 
FreeBSD 4.9 with a PPS-enabled kernel, just duplicating the server lines 
doesn't work.

Here's the relevant part of ntp.conf:

	enable pps
	pps /dev/pps0 assert hardpps
	server 127.127.26.0 prefer ttl 1
	fudge 127.127.26.0 time1 -0.966
	fudge 127.127.26.0 refid GPS0

	server 127.127.22.0
	fudge 127.127.22.0 refid PPS0

	pps /dev/pps1 assert hardpps
	server 127.127.26.1 prefer ttl 1
	fudge 127.127.26.1 time1 -0.966
	fudge 127.127.26.1 refid GPS1

	server 127.127.22.1
	fudge 127.127.22.1 refid PPS1

The single-clock version of this worked fine, but this attempt clearly 
doesn't do the job (note particularly the horrible ntptime statistics -- 
with a single PPS source, the stability is more like 0.02ppm):

$ ntpq -p
     remote           refid      st t when poll reach   delay   offset 
jitter
===========================================================================
===
+GPS_HP(0)       .GPS0.           0 l   59   64  377    0.000  -27.719 
6.811
 PPS(0)          PPS(0)          16 l  247   64  110    0.000  -20.462 
0.086
+GPS_HP(1)       .GPS1.           0 l   43   64  377    0.000  -29.444 
7.138
oPPS(1)          .PPS1.           0 l   17   64  377    0.000  -20.465 
0.002
-time-C.timefreq .ACTS.           1 u   44   64  377  102.044  -20.903 
0.134
xntp2.usno.navy. .USNO.           1 u    3   64  377  191.221  -13.602 
4.940
xtick.usno.navy. .USNO.           1 u   13   64  377  198.978  -11.235 
3.321

$ ntptime
ntp_gettime() returns code 0 (OK)
  time c3b26efc.0597f628  Fri, Jan 16 2004  9:09:00.021, (.021850349),
  maximum error 12931 us, estimated error 1 us, TAI offset 0
ntp_adjtime() returns code 0 (OK)
  modes 0x0 (),
  offset -20465.123 us, frequency 412.662 ppm, interval 64 s,
  maximum error 12931 us, estimated error 1 us,
  status 0x2107 (PLL,PPSFREQ,PPSTIME,PPSSIGNAL,NANO),
  time constant 6, precision 0.001 us, tolerance 496 ppm,
  pps frequency 412.662 ppm, stability 31.779 ppm, jitter 3.815 us,
  intervals 712, jitter exceeded 577, stability exceeded 3, errors 688.

So, do I gather that FreeBSD can't support multiple PPS sources?  Or are 
there configuration changes that will fix this?

John



More information about the questions mailing list