[time-nuts] mixers for frequency measurement

[Bruce Griffiths]

Chris Albertson wrote:
> I want to get set up to make accurate frequency stability measurements
> at reasonable cost.  I think mixers are the way to do that.   I'll set
> up a single mixer first them later learn to use the double mixer
> technique.   But I will use the simpler single mixer to try out the
> parts.  My goal is a multi-channel double mixer setup.
>
> First question:  Which type of mixers work best for this.   I have
> some SA612 chips in the parts bin.  Or should I be using diodes or
> something else.  Yes I know I could simply buy one from Mini Circuits
> but that defeats the purpose which is learning how to do this
>
> 2nd question.  if the device under test is a square wave oscillator is
> it best to run the square wave right into a mixer or filter them to
> sine waves first. I see pros and cons of each.  What do the experts
> do?
>
>    
Diode mixers (or perhaps, for low frequencies (5-20MHz) a diode style 
mixer <http://tf.boulder.nist.gov/general/pdf/2556.pdf> using diode 
connected (collector base short) 2N2222's as recently espoused by NIST) 
have far lower phase noise than other mixer types.
This paper also has a plot that illustrates the noise performance of 
this custom mixer compared to a couple of conventional diode mixers.
Running the mixers in saturated mode lowers the phase noise compared to 
linear mode operation.

Bruce