[hpsdr] Pheonix synthesiser

[Grant Hodgson]

Brian

Interesting comment about IMD2 and dithering.  The new generation of 
Fractional-N PLLs such as the LMX2486 and the AD4252 also have a 
dithering function; the LMX2486 has two dithering settings - 'weak' and 
'strong', as well as 'off'.  Dithering is provided to reduce fractional 
spurs at the cost of a slightly increased noise level; however my 
measurements to date on a 2.3GHz LO indicate that the level of noise 
increase is much smaller than the reduction in spurs, which makes it a 
very useful feature.

It would be very interesting to measure the IMD2 of a receiver with a 
dithered Fractional-N PLL.

For the Pheonix synthesiser there may have to be an engineering 
trade-off between cost/complexity and performance.  If a single-loop 
microwave PLL is used then the close-in phase noise will be determined 
by the PLL IC, and with current technology this will give a phase noise 
performance worse than the Orion, and possibly worse than the SDR-1000, 
especially on the higher bands such as 28MHz.  But this may be 
acceptable for the Pheonix receiver, and should give a very acceptable 
LO performance.

The real difficulty is with Horton, where the aim is to have a top-notch 
receiver which obviously requires a top-notch LO.  In order to get a 
top-notch LO it will be necessary to adopt something more sophisticated 
such as a dual-loop microwave PLL.  I believe that this approach will 
give a receiver that is better than the SDR-1000 and the Orion.

regards

Grant  G8UBN


  --snip--
> 
>  Now to my real comment.
> 
> Reading the PHOENIX article synthesizer choices, I liked the “Microwave 
> PLL divided down” choice, similar to Radware, but then I  got to 
> thinking. If this is used dithering will be much harder and I still 
> think IMD2 is a bigger problem than the spurs produced by the AD9951.
> 
> As a second consideration, dividing by 4 using a single AD9951 puts an 
> upper limit of less than 50 MHz on the receiver. Using 2 AD9951’s puts 6 
> and 2 meters within range.
> 
> 73 Brian KF6C   G3ZVC