[time-nuts] Close-in phase noise question...

[Grant Hodgson]

Brian

Thanks for the details.

There are a number of problems with crystal filters.  A crystal filter 
will have a finite bandwidth, and at offsets less than half the 
bandwidth the filter won't have a significant effect on the phase noise. 
  For example the 40MHz xtal filter used in the 8662A has a bandwidth of 
6kHz (according to the service manual), so phase noise at less than 3kHz 
offset will not be affected.

For your application you would need a filter with a very narrow 
bandwidth indeed - maybe 50Hz or so.  Ignoring the manufacturing issues 
of such a device, there could be problems with temperature sensitivity 
and microphony.  These problems could of course be overcome, but I'm not 
convinced that a crystal filter would give an overall benefit in this 
application, given that the primary goal is to achieve leading-edge 
close in phase noise.

The MSA1105 has better phase noise than I thought, but I'd still be 
tempted to change one or both of the passive doubler/buffer stages to an 
active doubler, if the figures for the NIST active doubler are correct.

regards

Grant