[time-nuts] FE-5680A's suitability for use as a 10 MHz reference for microwave transverters
C. Turner
turner at ussc.com
Mon Feb 13 21:03:16 UTC 2012
Last week I noted that the FE-5680A's "barefoot" output was found to NOT
be a suitable 10 MHz reference for microwave transverters.
Specifically, I tested it on two different 10 GHz transverters and found
there to be objectionable levels of "grunge" on signals caused by
low-level phase modulation internal to the '5680A and at 10 GHz the
result of this phase modulation was a racket of audible and subaudible
noises on CW carriers that made it difficult to find zero beat! In
comparison, the 10 MHz outputs of the Z3801, Isotemp VCXO and LPRO-101
yielded results at 10 GHz that were quite clean. Related observations
were also made by N8UR in his web page comparing various units.
In order to clean up the output of the FE-5680A I did the obvious thing,
disciplining a homebrew VCXO to its output - details are found here:
http://www.ka7oei.com/10_MHz_Rubidium_FE-5680A.html
While the comparison frequency is fairly high (1.25 MHz) the loop gain
and bandwidth are quite low so it's pretty much the Butler oscillator
VCXO that determines the phase noise of the 10 MHz output and I can't
detect any audible artifacts from the '5680A at all. At the moment I
don't have the means of generating a "pristine" test signal at 10368
MHz, but from what I can determine, the resulting CW notes from the
transverter (being locked to the "regenerated" output of the the '5680A)
compared to the other 10 MHz sources sound the same. At some point I
hope to do a more-detailed analysis.
Had I a low-noise "canned" 10 MHz VCXO around, I'd have probably used
that rather than go through the hassle of building the oscillator, but
none of the 10 MHz VCXOs that I *did* have on hand produced as good a CW
note as the Butler built around a cheap microprocessor-type crystal. I
also had on hand some 10 MHz ovenized VCXOs which would have worked
fine, but not only were these too large to fit in the box, they would
have added even more current consumption to an already power-hungry
frequency source - an important consideration when operating from a battery!
There are, no doubt, a number of ways one could do this same thing, but
it's clear that this simple of a circuit will do an admirable job of
extracting the frequency stability of the FE-5680A without the
synthesis-related artifacts. Of course, the regenerated 10 MHz output
will have a degree of variable phase offset with respect to the '5680A's
"barefoot" 10 MHz output over varying conditions (such as temperature)
but when used only as a frequency reference these rather slow changes
are unimportant.
73,
Clint
KA7OEI
More information about the time-nuts
mailing list