[time-nuts] 10 MHz -> 16 MHz clock multiplier

Bill Fuqua wlfuqu00 at uky.edu
Fri Jan 4 23:10:06 UTC 2013


At 07:51 PM 1/4/2013 +0000, you wrote:
>My question is about the phase noise of the final 16MHz signal.  Do
> > crystal filters "clean up" the signal.  It seems that after several
> > 16MHz crystals in series the output should look a lot like an XO.
> >
>
>For offsets out to 100 Hz or so, using a crystal filter will cause
>the signal to have the same flicker noise that an oscillator built
>with that crystal would have.  Thus don't try to use some junky
>clock crystals to make a crystal filter as described in numerous
>ham radio articles about receiver IF filters.  If you have a
>residual phase noise measurement system like the Agilent E5505A
>and a very low flicker noise source, you can actually measure your
>filter crystals.  Of course, the crystal time base in the source
>has to be better than the crystals you are measuring.  You also
>have to avoid overdriving the crystal.  This will require a low
>noise buffer amplifier to bring the signal back up to a high
>level.
>
>Now after considering all that, crystal clean up filters don't
>sound like such a great idea unless you have no alternative.
>
>Rick Karlquist N6RK
    Flicker noise (1/f ) noise would be introduced by an amplifier and not
by the filter.  I am only suggesting ways to multiply the frequency.
You could use LC filters or a crystal filter. Using 3 doublers would
do the job just as well. Naturally if you are concerned about flicker noise
you could simply make the 2 MHz signal higher in amplitude before selecting
the 8th harmonic.
   I was not saying your going to clean up a good crystal oscillator with a
crystal filter. I though you were talking about generating 16MHz from 10MHz
in a clean way. Using a microcontroller or even most synthesizers 
techniques would make
it even worse.
   The PTS synthesizers have fairly good phase noise when they use 
frequency multiplication,
division, mixing, comb generation and filtering on the most part to achieve 
low phase noise.
The later models use a DDS at the lower frequency levels but do have 
greater phase noise
close to the carrier.  In fact in the SGA unit the reference input goes 
thru a transistor ( to distort it)
and then into a series 10MHz crystal filter so that it can accept either a 
5 or 10 MHz input.
    The crystal also helps filter out any birdies that may be on the 
reference signal.
The filter should be fairly high Q since it has 47 Ohm drive impedance and 
100 Ohm load impedance.
73
Bill wa4lav





More information about the time-nuts mailing list