[time-nuts] noisy varactor diodes

John Miles jmiles at pop.net
Thu Dec 20 02:40:16 UTC 2012

> -----Original Message-----
> From: time-nuts-bounces at febo.com [mailto:time-nuts-
> bounces at febo.com] On Behalf Of lists at lazygranch.com
> Sent: Wednesday, December 19, 2012 4:56 PM
> To: Discussion of precise time and frequency measurement
> Subject: Re: [time-nuts] noisy varactor diodes
> But putting the diodes in series should increase the noise. They are
> noncorrelated, so I would say increase the noise in a RMS fashion.

I see the idea as creating a 'virtual' varactor with the same C/V
characteristic as a single diode, but with a much larger junction.  I'd
expect a reduction in flicker noise from making the effective junction area
larger relative to its surface perimeter (if that makes any sense).   The
same trick might apply to other types of diodes -- Zeners, bandgap
references, and the like. 

Also, each diode in a series-parallel configuration has a smaller share of
the bias voltage.   Lower voltage leads to lower Q in a varactor, but Q is
only part of what determines the noise contribution from a varactor.  People
have reported worse noise performance in oscillators after swapping out Si
varactors for GaAs parts with higher Q.  My understanding is that this is,
again, due to surface-noise effects that might improve as the junction gets

I don't think anything in that patent would come as a surprise to
semiconductor process gurus.

> HP has made at least one radio with switched varactors in the prefilter.
> made a SIGINT rack for I presume three letter agencies, though it was in
> catalog in the 80s IIRC. The radio was offered to me from a destruct house
> that wasn't clear on the definition of destruction. It was even GPIB.

The 85685A preselector might have done that.  I thought I had a schematic
for that one around here somewhere but can't seem to find it at the

-- john
Miles Design LLC

More information about the time-nuts mailing list