Did you 'build' an optical lockin amplifier? If not, what's
different?
Well, it slightly depends.
-- If the thing you are looking for is synchronized to
your local oscillator, then yes, the thing I described
is an optical lockin amplifier.
-- OTOH suppose it's not synchronized. For instance,
suppose you're looking for some natural resonance.
If there's some nonlinearity the response could be
pulled off the drive frequency. Or if you hit it
with a short (broadband) pulse, it will ring at its
natural frequency and it would be a stretch to call
it "locked" to the local oscillator. In this case
you would just call the device a single-sideband
heterodyne receiver.
Been there, done that: we whacked the system with
a short pulse, and then (due to nonlinearities)
its frequency would change and change and change
during the ringdown.
-- Also: Note that from a black-box point of view,
ye olde analogue outputs from ye twenty-year-olde
analogue lockin are effectively a homodyne design:
the local oscillator of the receiver is at exactly
the same frequency as the probe signal. Perhaps
inside the black box there is a heterodyne IF stage,
but eventually they bring the signal down to DC and
apply analog low-pass filters to it. For serious
signal processing, you don't want that, because 1/f
noise gets kinda big at DC. Instead, you want to
digitize the IF. The only good representation of DC
is a digital representation. Nowadays they do a lot
with DSPs.
My thesis work involved using a heterodyne microwave
bridge. The original design was due to Burgess Johnson,
although there were four of us who contributed to the
design and construction, and operated the thing for thousands
of hours. I hacked it to make it heterodyne and single-
sideband, and was responsible for most of the IF stage,
digitizer, and whatnot downstream from there. We called
it the world's finest >1GHz NMR bridge. That is to say
it was the world's first and (at the time) only >1GHz NMR
bridge. (Hint: If the Larmor frequency is >1GHz, you
need whopping great magnet and/or a lot of help from the
hyperfine interaction. Proton gyromagnetic ratio is <43
MHz/Tesla. Do the math.)
As for the optical version, I've got friends who've built
things like that, actually rather trickier than that.
Sometimes they let me help with the design, construction,
and debugging. In particular, the way you build a laser
with high frequency stability is to lock it to some other
oscillator, which involves looking for optical beats, which
is how this conversation began. Talk about debugging, hooo
eeee, what a nightmare. I remember finding a mirror-mount
that had a mechanical resonance near the IF frequency. It
took us several 18-hour days to sort that out. It's a
miracle we ever sorted it out.