"ralph sansbury" wrote in message
...

So if the intermediate frequency doesn't change for 1000 seconds
then you say the frequency accuracy is .001 Hz.
But what about the accuracy in the normative sense that the received
frequency,
or this intermediate representation of the received frequency,
is sometimes zero at different times than the local oscillator whose
frequency and phase are fine tuned to most closely match this intermediate
frequency?
If for example half of the time there was no correspondence between the
predicted frequency zeros and the observed frequency zeros then the
accuracy would be 2.292 GHz plus or minus 1 GHz.

If the PLL is in lock then there are exactly the same number
of zero crossings in the actual signal and the LO signal in a
33 minute period. However, each zero crossing of the LO might
be slightly earlier or later than the corresponding event in
the actual signal by a fraction of a cycle.

George