[time-nuts] Thinking outside the box a super reference

[jimlux]

On 11/4/16 5:27 PM, Richard (Rick) Karlquist wrote:
> On 11/4/2016 4:04 PM, Poul-Henning Kamp wrote:
>
>> Historically resonance cavities were used so that step/avalance
>> diode multipliers had enough power to excite them.  Today we have
>> semiconductors which work at those frequencies.
>>
>
> A great deal of complexity in the 5061 went into
> exciting an SRD at 90 MHz and getting a sufficient
> line at 9180 MHz to put on a sideband at 9192.
> I spent a lot of time trying to do this during the
> 5071 project and was never able to get anywhere
> near the efficiency that the 5061 waveguide structure
> was able to do.  It was designed by a visiting Korean
> professor, so he wasn't around to mentor me.
> The 10816 also used an SRD, and it was also a
> struggle, although I was able to make it work.
>
> Fortunately, we were able to replace all this with a
> DRO and PLL, and that was 25 years ago.  At this time,
> it is even more of a no brainer that you don't want
> to knock yourself out trying to make an SRD multiplier
> work.  Also, these days, it is harder than ever to
> purchase good SRD's.
>
You can get a nice GaAs VCO that will tune 9GHz with no sweat, and the 
PLL parts to go around it. I'm not sure what the DRO market is these 
days - You can probably build a synthesizer with lower DC power with a 
DRO than using GaAs (although CMOS is getting ever better and getting up 
to that kind of range).  There are probably low cost low precision 
(unlocked) designs like LNBs that use a DRO.

But things like 10 GHz microwave motion detectors these days use VCOs, 
not DROs.

DROs are subject to microphonics and such - yes, the dominant resonance 
is the puck, but the cavity it's in affects it too, so vibration and 
temperature have effects.