[GPS_Standard] VE2ZAZ

[Dave Platt]

Bob Bownes wrote:
> Indeed, I'm shocked at how temperature sensitive this oven is. If,
> after trying some of Dave's methods and tests, I find the oven is
> good, I suspect I will be placing everything in a very temperature
> stable environment, possibly thermally bonding it to a Very Large
> thermal mass, aka a 12" thick concrete slab ~90" below grade. :)

Maybe you could adopt the approach of the Icecube neutrino
observatory... a cubic kilometer of Antarctic ice ;-}

> What are words of wisdom on a power supply? I have dozens of switchers
> about that will do the job, but what of any noise?

If you're going to use a switcher, I'd recommend having additional
filtering and (if at all possible) linear pre-regulation, before
you let the power near any of the components in a frequency standard
of this sort.  Don't put switcher-regulated power directly into
any of the boards or oscillator, for just the reason you cite -
noise.

My own box does use a switcher (originally for a Sony laptop)
which delivers up to around 2 amps at 16 VDC.  Its output seems
relatively clean, but I don't really trust that.  Immediately on
input to the box, it goes through a multi-stage filter to snub out
RF and some lower-frequency noise (a monolithic cap directly
across the power socket pins, several turns of wire wrapped through
a ferrite toroid, and a big 'lytic).  Power for the VE2ZAZ board
and the GPS/serial-port "schmoo" adapter board then goes through
a 7809 preregulator and is fed out to these boards.  The filtered
+16 supply is also fed to an oscillator interface board, which has
the two 7812 regulators for the oscillator and oven, and the
DAC-control-voltage coarse-adjust/padding circuity and 7805
voltage regulator/reference.

The telco board I salvaged the oscillator from, seemed to use
a somewhat similar approach - it took in +24 VDC from a power
bus (which I strongly suspect was switcher-regulated) and
used 7812/7815 regulators to power the oscillator and oven.

The 78xx regulators seem to have pretty good supply-side
rejection, and so should be attenuating whatever residual
hash is present on the switcher-generated "raw" DC power
fairly well.