[time-nuts] Racal-Dana 1992 switches

Pete peterawson at earthlink.net
Sat Mar 29 15:43:31 EDT 2008


Jeff,

There's nothing magic about 1KHz as the heterodyne
frequency. I picked it to avoid power line pickup &
permit a tuned circuit to limit the noise bandwidth.
But, any tuned circuit in the signal path will have
temperature driven phase variations & possibly
signal level phase variations, as well.

The opamps contribute some noise & they also limit
the output slew rate; yielding approx. 400ps rms jitter.

Using rather stable L & C components limits the
total impact of these effects to less than +/-1E-9
@ 1KHz (equivalent to +/- 1E-13 @ 10MHz).

These issues speak against extending this scheme
to much longer times.

The synthesizer PN does contribute to the counter
measurements. But, even with 2s gate time you are
averaging 1000 samples & things quiet down nicely.
Some synthesizers are better used at 9.999MHz as
opposed to 10.001MHz, but I've no solid data on
this. Either frequency is OK in this scheme.

Pete Rawson



More information about the time-nuts mailing list