[time-nuts] question on apparent offset between two Endrun CDMA clocks

starlight at binnacle.cx starlight at binnacle.cx
Sun Sep 11 17:23:00 UTC 2011


I have a minor mystery and if any ready explanation
is available I would be curious to learn of it.

Have two Endrun CDMA time sources:

   Praecis Cntp stand-alone time server
     internal time module is Ce or variant thereof
     kernel 2.2.13-DOC
     ntpd 4.1.1c-rc2 at 1.866
     10-BASE-T HDX network connection

   Praecis Cf timing module
     single 2.4 GHz P4 system
       old Dell Dimension 2400 (incredibly reliable)
     CentOS 4 kernel 2.6.9-100.EL
     ntpd 4.2.4p4 at 1.1520-o
       Palisade mode
     1000-BASE-T FDX network connection
        Intel 82546GB e1000, IRQ coalescing disabled

Both are attached to the same SMC GS24 gigabit
switch with short-run cables.

Both are synced to the same CDMA source (Verizon
cell, primary B) with the antennas a few feet
apart.  'cdmastat' on the Cntp reports SNR of
about 5 and zero errors.  SPSTAT query to Cf
reports SNR of about 8 and zero errors.

Both 'ntpd' daemons consistently claim to track
their Praecis device to +/- five microseconds or
less.  Jitter on the Cntp is 8 microseconds, on
the P4 is 4 microseconds.  Polling interval is
16 seconds, everywhere.

'ntpq -pn' and 'ntpdate -q' queries from both systems
and from other client systems consistently show
the Praecis Cntp between 30 and 80 microseconds
ahead of the Praecis Cf device.  The median
offset seems to be about 60 microseconds.

Is this just some minor network timing asymmetry?
Or is it possible that the two devices have
slightly different takes on the CDMA time?
Network jitter is generally reported to be between
10 and 50 microseconds.  Round trip time to the
Cntp is about 960 microseconds and round-trip time
to the P4/Cf is about 60 microseconds.  'ntpdate -d'
shows that the Cntp takes a total of 3600
microseconds to service a request where the
P4 takes 65 microseconds.


More information about the time-nuts mailing list