[hpsdr] Low noise oscillators

[Chris Bartram]

Hello John

> I suspect it's quite possible to
> build a noisy oscillator, even with a great crystal.

That's a simple matter of engineering principles. It's undoubtedly possible, 
even easy(!) to build poor oscillators  - in terms of phase noise - with good 
crystals, simply by making an oscillator with low loaded resonator Q.

> I'd be really interested in testing some prototypes built using the
> principles in Chris's VLNO.pdf paper, especially if someone has two of them
> that can be locked in quadrature with each other with a varicap.  I can
> 'see' down to -160 dBc/Hz or so with reasonable confidence, depending on
> the offset.

I've just laid-out a PCB for the oscillator detailed in that note, and I have 
just received some (15USD...) 125MHz crystals from a supplier (who wouldn't 
usually be on my preferred list). They are surprisingly good with a measured 
Qu of ~80K. Work permitting, in the next few weeks I intend to make four of 
the oscillators, and I'm tempted to send two of them on a world tour to 
people I know have the facilities to measure them... If you want to 
volunteer, I'd be very grateful! 

So far my measurements of the close-in performance are of Allan Variance using 
a counter, and of course, that's are not directly translatable to phase 
noise. However the figures are very encouraging. Not unsurprisingly, I have 
problems with my current (autocorrelation/delay line) PN test set-up when 
looking at very close-in noise, although it's good enough for most of the 
commercial work I do. It was very satisfying to see the AV measurements 
confirmed when I had a look using a newly acquired HP53310A modulation domain 
analyser! The next prototypes will have voltage tuning, and will hopefully 
allow me to make some realistic measurements. 

Vy 73

Chris
GW4DGU
--