[time-nuts] Low-Cost Rubidium Performance

[John Miles]

> Important note: there's known to be a lot of variability among these
> units, so my sample size of one or two each isn't a guarantee that your
> unit will perform similarly.

Neat.  Your FE-5680 noise plots look just like the ones I took.   Seems
clear that this particular variant was meant for timing or calibration work
rather than communications.
 
> Next step is to do some longer-term stability measurements (out to 100K
> seconds).  You may see that data in a month or so.

How's the environment where you're running the long-term test?  Any exposure
to HVAC circulation or any other sources of changing temperature?  I've
taken a few longer-term plots with the LPRO, but I've always suspected it
could do better if it were actively temperature-stabilized, or at least
shielded from room air currents.  

New page with my short-term plots here: http://www.ke5fx.com/rb.htm

(These include HP 5065A, LPRO-101, SRS PRS10, and FE-5680A)

-- john, KE5FX