[time-nuts] Sub Pico Second Phase logger

[Joseph M Gwinn]

Bruce,


time-nuts-bounces at febo.com wrote on 12/08/2008 05:53:08 PM:

> Joseph M Gwinn wrote:
> > People used passive mixers driving electromechanical stripchart 
recorders 
> > to compare high-stability oscillators in the good old days.
> >
> > One assumes that there is a purely analog approach to measurement of 
> > picosecond changes in delay at 10 MHz using a single oscillator, but I 

> > have not seen any methods described, probably because the relevant 
> > articles appeared many decades ago.
> >
> > Can anyone suggest some articles to read?
> >
> > Thanks,
> >
> > Joe Gwinn
> > 
> Joe
> 
> Although one could in principle do this with a single diode double
> balanced mixer used as a phase detector all one may end up measuring is
> the effect of ambient temperature changes on the mixer phase shift.
> Lower mixer phase shift tempcos are possible if the RF port is 
> unsaturated.

Single diode?  Why wouldn't one use a standard (MiniCircuits or the like) 
four-diode two-transformer double-balanced mixer as the phase detector? 
Many mixers have IF response down to DC.


> A classical dual mixer system is probably better in that with matched
> tempco mixers maintained at the same temperature the differential phase
> shift tempco should (with careful matching) be lower.

Dual mixer as in DMTD (dual mixer time difference) would certainly work, 
but is pretty complex and temperature sensitive.

I did use a loaner Symmetricom 5120A (a full digital DMTD implementation) 
to make some measurements six months ago, and after a few days of 
continuous operation it had settled to the point that one could see 0.01 
pS changes.  (And touching one of the BNC connectors caused a 1-3 pS 
jump.)  This instrument costs about $30K, and is intended more for 
measuring phase noise and allan variance than delay changes.

Anyway, I have to wonder what people did before DMTD was invented.


> Other than the numerous classical papers on dual mixer systems and the
> occasionl NIST paper that have some mixer phase shift tempco data
> (albeit sparse), I am not aware of any specific papers.

I've read many or most of the classical DMTD papers, and have seen various 
passing estimates that diode-ring mixers have a temperature sensitivity of 
8 to 10 pS per degree C.  (I recall your figure was 10 pS/K.)  I assume 
that the DC offset also varies with themerature and drive signal 
amplitude.

 
> A purely analog approach to phase shift measurement has to be more
> difficult than a hybrid one using a pair of low frequency ADCs (eg high
> end sound card).

Is the sound-card approach workable at the millidegree to microdegree 
level, if the change is spread out over an hour?  One picosecond at 10 MHz 
is 3.6 millidegrees of phase.

Joe