[IGSMAIL-3072]: Addendum to IGSMail No. 3071

[kogan at ldeo.columbia.edu]

******************************************************************************
IGS Electronic Mail      25 Oct 08:17:36 PDT 2000      Message Number 3072
******************************************************************************

Author: Mikhail Kogan

It is appropriate to make the following Addendum to the IGSMail No. 3071
by Jim Ray regarding his praise of my "detective work" on Ashtech Z12
receiver with the external frequency standard.

There were two contradicting pieces of evidence on the behavior of the internal 
clock in this configuration:

1) On one hand, at the June 1999 IGS AC workshop in La Jolla, Jim Zumberge 
showed a plot of Gipsy-produced estimates of clock solutions 
for a zero-baseline setup -- TurboRogue and an ordinary Z12 -- 
where both receivers used a maser --the _same_ maser -- as an external 
frequency reference, and both receivers had clock steering disabled.  
That plot clearly showed smooth variation of both receiver clocks 
(it turns out to be at the few-cm level over 5 minutes, as the caption 
on the plot showed), and a slowly-varying difference (tens of cm over hours) 
between the two. 

Unfortunately, this result was not made widely known, otherwise there would be 
no need for my experimenting with configuration (Z12 + Cs clock). 


2) On the other hand, several IGS stations with this configuration showed
poor performance of the internal clock despite the slaving by the external
frequency standard. There was no obvious reason for that. See numerous
examples in http://maia.usno.navy.mil/gpsclocks/finals/1073


Therefore I simply tried to find out which combinations of the receiver settings
might contribute to the unacceptable behavior of the internal clock. 
Dave Stowers (JPL) and Brian Tallman (NIMA) urged me to perform these tests
with the Cs clock kindly provided by JPL to Columbia University. Both Dave and
Brian predicted that the internal clock of Z12 should reproduce the quality
of the external frequency if all is set up properly. Helpful technical
explanations on how the clock slaving is realized in the Z12 were given to me
by  Richard Phelan, and Art Sauer (both at Ashtech/Magellan).


Finally, I should note that Jim Zumberge pointed out to me that for the time 
transfer applications, a Z12 with the Metronome option in his view is still 
necessary to avoid errors on the order of 1 nanosec arising from the 
"internal circuitry" of the Z12 receiver. I cannot comment on this 
because of my humble experience with the subject. Still, an ordinary, 
unmodified Z12 equipped with the cesium clock is capable of producing 
+-0.6 nanosec accuracy (1-sigma relative to the linear drift) which is 
2-orders of magnitude better than reported by several IGS sites having 
this configuration.


Regards,
- Mikhail Kogan


_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/
_/                                                                        _/
_/ Mikhail G. Kogan                                                       _/
_/ Lamont-Doherty Earth Observatory    E-mail   : kogan at ldeo.columbia.edu _/
_/ P.O. Box 1000 / 61, Route 9W        Telephone: 914-365-8882            _/
_/ Palisades, NY 10964 USA             Fax      : 914-365-8150            _/
_/                                                                        _/
_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/