> I'm not trying to be a smart-ass,

Nor, am I. But this sounds like a band-aid job to me.
I would think that the circuit that can't endure a 5mV ripple is the real
problem.
That's where I would be investing my effort. Perhaps as ripple, perhaps not,
but 5mV supply irregulariites can be induced or otherwise introduced by many
unexpected means. Instead of having to deal with the snakes that will appear
in the grass when the design is in real-world use, why not solve the problem
at the root?

> > problem: the
> >     circuit is EXTREMELY sensitive to supply noise.  Right
> > now we're using
> >     a $2000 lab supply, which has <2 mV ripple.  It works.  I
> > have managed
> >     to make a 5 mV ripple regulator, but it still isn't good enough.

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics