[time-nuts] Setting Rubidium to match GPS source

Bruce Griffiths bruce.griffiths at xtra.co.nz
Thu Sep 27 20:28:47 EDT 2007


Pete wrote:
> Chris,
>
> The 3575A does NOT have adequate phase resolution to set an Rb
> source to match a GPSDO. A few, now rather old, phase meters were
> sold for this purpose, but they used a strip chart recorder to average
> phase shift over 1 to 24 hour measurement times & required great
> investments of time to use.
>
> I proposed a 1KHz heterodyne scheme for this very purpose on this
> list a few months ago. The scheme took a few hits for lack of state-
> of-the-art performance, but I'm still using it without suffering any
> disappointments. It does require you to have a few items of test
> equipment (readily available on eBay) which not everyone has &
> it requires some circuit construction using readily available parts.
> Except you must wind a couple of 5mH inductors on your own.
>
> Contact me off-list if you want more details.
>
> Regards,
> Pete Rawson
>   
Pete

This method requires a source offset by 1kHz from the desired frequency.
Its much easier and simpler to tune to zero beat with a phase detector.
A suitable phase detector with a performance equivalent to that of the
HP K34-59991A linear phase comparators can be built using an AD9901 plus
an opamp or 2.
The performance will be somewhat below state of the art but a resolution
of better than 1E-10/T should be easily achievable, if an ADC is used to
monitor the phase detector output.
This performance should be more than adequate for the application.
Nothing other than the phase comparator and the 2 sources to be compared
is required. However the 2 sources must have the same nominal frequency.

Bruce



More information about the time-nuts mailing list