[GPS_Standard] Efratom 105243-001 OXCO

[Dave Platt]

Well, after a couple of years of getting parts together and
waiting for some free time, I've started work on my VE2ZAZ
frequency standard.

The OXCO I'm using for this is an Efratom 105243-001, salvaged
from an eBay-acquired board that I understand was used as
a cellphone-site frequency standard.  The oscillator is on a
pretty little sub-board which can be snapped free from the main
assembly.

Since there doesn't seem to be much information available
about this oscillator (and a number of people looking for it)
I thought I ought to pass along what I've determined by powering
up the main board and 'scoping the oscillator connections.

The sub-board has half-a-dozen connectors at the edge which
are wire-jumped over to the main board.  Going down from the
top:

W12 - 14 VDC supply to the oscillator itself.  Draws around
       15 mA in use.

W17 - variable DC voltage which "pulls" the oscillator.
       Full range not known.  Actual voltage on my board was
       2.97 VDC.

W3 -  oscillator output - near-square-wave, 5 volts peak
       amplitude (presumably CMOS-level?)

W13 - this appears to be an open-collector or open-drain
       signal, which is actively pulled down to ground once
       the OXCO has reached proper operating temperature.
       It's diode-isolated from the main board.

W10 - 16 volt supply to the OXCO heater.  Draws 300 mA
       during initial warmup, and drops down gradually
       to around 150 mA once the OXCO is warm and operating
       normally (oscillator in open air, sitting on a desk
       in a room at around 65 degrees F).

W15 - Ground

I disconnected the 2.96-volt feed from the main board, and
connected a potentiometer fed from W12 and ground.  Adjusting
the resulting voltage at W17 between 2.5 and 3.5 volts shifted
the oscillator frequency by 5 Hz.

This seems like a huge amount of pull-range, compared to the
"one Hz tuning range" cited in VE2ZAZ's article.  This leaves
me assured that the controller will be able to pull the oscillator
far enough... but makes me think that I might get better accuracy
by padding down the turning range somewhat and using a manual
adjustment to center it right around 10 MHz.  This could be
done by a simple op-amp/resistor/pot circuit, and there's
plenty of blank space on the top of the sub-board to mount
a little daughtercard for this purpose.  I'll probably build
the standard without a tweak-and-pad sub-board and then see
how the system performs.

I'm currently testing out the oscillator board with lower
supply voltages.  Everything seems to work OK at 12 volts,
and I have hopes that even a 9-volt supply will be OK.
More on this later.