[time-nuts] Designing and building an OCXO and GPSDO

Bruce Griffiths bruce.griffiths at xtra.co.nz
Wed Aug 13 21:12:39 EDT 2008


John Miles wrote:
> If you don't want pushbutton convenience, you can measure the close-in phase
> noise with not much more than a $5 mixer and $2 opamp.  It will take a lot
> of "sweat equity," and you will need to build two of whatever you're
> measuring, or buy/borrow a known-cleaner source at the same frequency.
>
> TSC analyzers are great but they are not the only way to go.  Actually their
> biggest advantages lie in their size/weight and the fact that the reference
> doesn't have to be at the same frequency as the DUT.  Other than that, their
> performance is not necessarily better than a homebrew single-mixer
> quadrature PLL or an 11848A.
>
> -- john, KE5FX
>   
John

For compatibility with the PLL technique a trimmer and EFC is required.
Using back to back varicaps is supposed to minimise the effect of 
varicap noise on the oscillator phase noise.

The next step below using a cross correlation technique is to use a low 
phase noise frequency multiplier to decrease the system phase noise floor.
A good multiplier can have a significantly lower phase noise floor than 
a mixer in the flicker phase noise region.

Another candidate for evaluation would be an oscillator using a low 
phase noise modular amplifier with 50 ohm input and output impedance.
A splitter is used on the amplifier output to extract a useful signal.
Lumped component 1/4 wave lines are used on the input of the modular 
amplifier and the feedback output of the splitter to achieve a match to 
Rs/2, where Rs is the crystal series resistance.
A full wave diode clamp can be incorporated to limit the crystal 
dissipation.

A phase noise of -156dBc/Hz @100Hz offset has been achieved with such an 
oscillator.


Bruce



More information about the time-nuts mailing list