[time-nuts] FE-5680A tuning vs resonant peaks

[Jeff Woolsey]

I wonder if my FE-5680A is locking to the wrong resonance, or that I've
tweaked it too far off the main one.

Setup:

FE-5680A Rubidium oscillator with passive support/breakout board,
ostensibly regulated with an LM1084-ADJ. Trimble Thunderbolt GPSDO,
feeding an HP-53131A universal counter as external reference.

Powered up from cold, the Rb eventually locks around 9999999.95644 Hz. 
Turn the bench supply voltage down about 1.5V, and it becomes
10000000.0089 Hz fairly quickly, later settling around 9999999.9988 Hz. 
While 2mHz accuracy is sufficient for my needs (let's face it, it's
overkill, and nominally better than any other 8-digit counter I have),
this is time-nuts after all.

I'm wondering if it isn't locking on a nearby secondary resonant peak. 
There's a graph at http://leapsecond.com/images/cfield.gif showing
resonant peaks for Cesium. I haven't yet found one for Rubidium. 
Knowing how far apart the peaks are, and the base oscillator (~50.5MHz)
that gets multiplied by 136 up to 6.8+GHz, perhaps one peak accounts for
the 50mHz difference I see.

Before coming up with this theory, I had tweaked the non-programmable
(per ka7oei.com) FE-5680A offset using a Windows-only program, perhaps
too far.

The other theory is that Lady Heather reports the TBolt OSC value
typically in the high tens of ppt (with similar variance tick-to-tick)
and in the hundreds of ppt a little less often, usually after a change
in satellites (logs available).  I wonder how that affects the counter
accuracy.

-- 
Jeff Woolsey {{woolsey,jlw}@jlw,first.last@{gmail,jlw}}.com
Nature abhors straight antennas, clean lenses, and empty storage.
"Delete! Delete! OK!" -Dr. Bronner on disk space management
Card-sorting, Joel.  -Crow on solitaire