[Test-Equipment] HP 8656A question

[John Miles]

> > Thanks for the number - I'm guessing it's right in the ballpark
> for what I'm
> > detecting by ear.
> >
>
> The spec for the 8640B is < 5 Hz (for a differenct frequency range and
> bandwidth).  I should have included that for a bandwidth of 0.3 to 3kHz
> the residual FM for the 8656A is < 15 Hz RMS.  I don't really know how
> this is measured.
>
> This is something I have never paid much attention to, since the 8640B
> seems stable enough for anything I have used it for.

They are both good generators, and a bargain considering what they're
sellling for these days.  They complement each other well.  Sometimes you
want a rock-stable, precisely (re)settable, but not necessarily pristine
source; that's the 8656A in a nutshell.  Other times you want a clean source
with continuously-adjustable frequency and amplitude, and that's when you
use the 8640B.  Horses for courses.  I keep them both on my bench.

The best of both worlds is the 8662A, but those are still rather spendy.

You measure residual FM with either a modulation-domain analyzer, or a
spectrum analyzer in zero-span mode, tuned halfway up the slope of the
desired resolution-bandwidth filter.  The idea is that deviations from the
desired frequency should appear as vertical excursions on the display, where
they can be measured relative to the filter's shape factor.  Sounds more
complicated (and less obvious) than it is.

The analyzer's own LO has to be cleaner than the signal source being tested.
For example, with a Tek 490-series analyzer, I can "see" the close-in noise
of my 8656A, but not my 8662A or 8640B.

-- john KE5FX