Thanks for that , Andy. It is always useful to get someone else's
practical experiences of techniques.
73 de Pat g4gvw
QTH Near FELIXSTOWE, UK
Thought for today:
What politicians do for science, hopefully science will do for them one
day!
-----Original Message-----
From: Andy Talbot <[email protected]>
To: [email protected]; [email protected]
CC: [email protected] <[email protected]>
Sent: Mon, 16 Jun 2008 20:17
Subject: LF: VE2ZAZ GPSDO findings
Having now completed and mostly tested two VE2ZAZ GPS Controlled
oscvillators, I can say I'm rather more than quite impressed with what
can be achieved. Both GPSDOs were built using good quality single
ovenned OCXOs, one a brand new Morion unit with 20Hz total tuning
range (no other coarse adjustment provided - interesting); the other
with one of the surplus OCXOs from The old Decca Navigator units, made
by Piezo Products. Alan, did you once say these had a particulalrly
good performance in some area?
The PPS input comes from a Trimble Palisade timing unit - one where
all the electronics is in the head unit and communicates to ground
using RS422 balanced signalling. The palisade has a PPS stability
greatly exceeding that of a low cost GPS receiver module, so means a
good GPSDO performance ought to be achieved with averaging times of
minutes to tens of minutes, rather than the hours otherwise required
and necessitating a double ovenned oscillator to ride-out the period.
The Morion OCXO unit, because of its wide tuning range and fast warm
up oven, can be used in a quick start circuit design. From cold
turn-on it locks within 20 minutes, and within an hour a frequency
stability of 3E-9 can be achieved, using an averaging period of 64
seconds (4 samples per average). After several days a few parts in
1E-10 was achieved by lengthening the averaging period gradually.
This OCXO, being new, was continually being adjusted by the controller
and presumably after weeks of operation it will settle down. I
haven't soak tested this GPSDO one any more, keeping it as a quick
start standby. The source of this module, stated that it was a
preferred option in timing equipment that had to be turned on from
cold and be ready quickly - it was supposed to have good retrace
performance.
The unit using the old Decca unit is a different story. It had to
have a stabilised 18V supply as the tuning voltage appears to be
applied wrt. the input supply volts, but once I realised this after
the first day of testing and added a regulator / SMPSU, it is
rock-steady. The VCO tuning is only 1Hz over the full 1 - 10V span,
and now, after three days continuous operation, the GPSDO is only
making fine adjustments from time to time. Over the last 2 days it
has only shifted by 6 fine-adjustment steps in total. Each fine
adjustment is a few parts in 10E-12, so the stability is excellent.
I'm still using a reasonably short averaging time of 256 seconds, but
with this level of performance, can probably set it longer.
Beating the 10MHz output against that from the HP source (the caesium
reference box, but in free run mode) with a frequency accuracy
currently of 3E-10 the phase meter stays pretty stationary if I sit
and watch it, although coming back after a few hours it has moved
noticeably.
This GPSDO has been built into a permanently-on rack mouted unit with
battery back up (a switch mode module needed for the +18V supply,
annoyingly!) and will eventually be used as the shack reference
instead of the HP5061.
If anyone has one of these ex Decca OXCOs, use it!. After more than
20 years of continuous use in the Decca chain it will be thoroughly
burned-in. The GPSDOs themselves were very easy to build and set up
- I made my own PCBs from the article (after some playing with the
image to mirror one side), and of course programmed the PICs myself
from the listing on the website.
Also added an 2x40 char LCD using another PIC to display the data
coming from the monitor port. Had contemplated modfiying the 18F2220
firmware to drive an LCD directly as there were more than enough I/O
lines left and plenty of memory space, but being less familiar with
PIC 18F code as I am with 16F, it was easier to write a separate RS232
decoder sending the data directly to a LCD. I also intend to use this
PIC to decode / display the time and date coming from the Palisade,
and convert to NMEA for external use, also to issue a subset of
commands (such as clear alarm, set short / long averaging) to the 'ZAZ
unit from front panel controls to save connecting up a PC to make the
changes.
--
Andy G4JNT
www.scrbg.org/g4jnt
________________________________________________________________________
AOL's new homepage has launched. Take a tour at
http://info.aol.co.uk/homepage/ now.
|
