[time-nuts] Digital Pots, wiper noise and fine tweaking oscillators

[Attila Kinali]

On Mon, 27 Jan 2020 11:34:07 -0500
Bob kb8tq <kb8tq at n1k.org> wrote:

> One advantage that a digital pot has is small size. If you want to ovenize a pot to improve
> it’s temperature performance, that’s a good thing. Indeed a lot of modern oscillators have
> digital pots in them to set this or that during the production process. 

This is the most prevalent use of digital potentiometers: adjusting
a circuit after production. They have basically replaced the pots
of old, because they can be adjusted digitally. I.e. you can build
the adjustment into the post-production test without manual intervention.
That's also the reason why you can find quite a few digital-pots that
have non-volatile settings.

There are a few key points with digital-pots that make them unsuitable
for EFC use:

* low resolution. 10bit is the best you can get, AFAIK.
* high glitch energy
* high resistance (>=10k)

You need somewhere 18-22bit resolution for a GPSDO, the 10bit are very
far from that. The glitch energy will make your OCXO jump every time you
program a new value into the digital-pot. The high resistance can be a
problem with some OCXO that have a low specified minimum EFC input
resistance. Though usually in the 100kΩ range, I've seen it spec'ed as low
as 10kΩ. Also, the dominating noise will still be the reference voltage.
Even if you use something like a LTZ1000 with high current.

You are better off getting an DAC like the AD5542A + LTC6655 + some
low noise opamps to amplify and filter than using a digital-pot, which
you would be using like a DAC anyways.

			Attila Kinali
-- 
It is upon moral qualities that a society is ultimately founded. All 
the prosperity and technological sophistication in the world is of no 
use without that foundation.
                 -- Miss Matheson, The Diamond Age, Neal Stephenson