[time-nuts] Would this work as a frequency standard? Would it damage the Stanford Research function generator?

Dr. David Kirkby (Kirkby Microwave Ltd) drkirkby at kirkbymicrowave.co.uk
Sun Jul 19 14:49:33 UTC 2015


I had a plan to construct an "off-air" frequency 10 MHz standard based on
the 198 kHz from the Droitwich radio 4 transmitter, just by hooking up
various bits of test kit and writing a computer program to control the
equipment via GPIB. No electronics, apart from an antenna, would need to be
built at all.  My thoughts were the following - hopefully it makes sense,
if not I could add a drawing.

1) Connect an antenna suitable for 198 kHz into the input terminal of an
EG&G 7260 dual-phase lock-in amplifier.

http://www.signalrecovery.com/download/190163-A-MNL-D.pdf

This is not a very nice lock-in to use interactively, but is okay via GPIB.

2) Set the EG&G lock-in amplifier to use an external reference.

3) Feed the reference input of the lock-in amplifier from a Stanford
Research DS345 30 MHz function generator

http://www.thinksrs.com/products/DS345.htm

which is set to output a 198 kHz square (or sine?) wave. The function
generator has a frequency resolution of 1 uHz, and 12 bits of resolution on
the DAC. This can produce arbitrary waveforms.

This function generator has a 40 MHz crystal, the frequency of which can be
set by adjusting the Calbytes via GPIB - similar to the SR620 time-interval
counter. Obviously if the frequency of the 40 MHz oscillator is high, so
the output frequency will be above 198 kHz. The specification of the
oscillator are

Frequency: 40 MHz (according to the manual, unless a typo)
Type: Ovenized AT-cut oscillator
Stability: < 0.01ppm, 20 - 60°C
Aging: < 0.001ppm/day
Short Term: < 5 x 10-11 1s Allan Variance

4) Use the lock-in as a phase detector, measuring the difference in phase
between the function generator and the 198 kHz from Droitwich. The
sensitivity of the lock-in, as well as the time-constant of the filters,
can be adjusted via GPIB.

5) Write a computer program that tries to set the oscillator in the
function generator to exactly 40 MHz via GPIB, based on the rate of change
of phase between Droitwich and the function generator.

6) Since the function generator has a 10 MHz reference output, which is
derived from the 40 MHz crystal, It should give me a solderless 10 MHz
frequency standard.

Does this seem viable, or have I overlooked something? I'm not looking to
build a lab standard - just have a bit of fun.

Does anyone know if regularly writing to the calbytes of the function
generator would damage it?

Dave



More information about the Time-nuts_lists.febo.com mailing list