[time-nuts] On some pitfalls of the dual mixer timedifferencemethod of horology

[Tom Van Baak]

> Note also that, based on my limited experience, most
> commercial mixer implementations use a much faster
> beat note: 10, 100 Hz, even 1 kHz. A faster beat note
> may help your concern #1 above, and #2 below.

Ulrich,

One thing I forgot to mention earlier -- there is another
advantage in using a higher beat frequency; that is, you
can average many more samples in less time. In all
your examples you seem to imply a 1 Hz beat and an
'instant' measurement in just one second. I think in the
real world phase comparators use both a higher beat
frequency and a longer measurement reporting time.

If you instead use a 100 Hz beat it seems to me you'd
get 100x more zero crossings and any white noise would
then average down by sqrt(100), or 10x. Furthermore, if
you wait 10 or 100 seconds for a final result instead of
1 second, that's another order of magnitude in sensitivity.

I would guess a lot of your instrumentation noise is white
so you'd get good leverage here. I could be wrong about
all this, but as you continue to experiment, please try
several different beat frequencies and averaging periods
and let us know what you find.

/tvb