[Elecraft] K2 PLL frequency optimization

[Eric Swartz WA6HHQ - Elecraft]

Hello fellow K2 optimizers,

Huff and Puff is way too complicated for add on here and is probably be too
slow to handle the T/R PLL frequency shifts required when sending CW or
running split. Even non split CW needs to shift the PLL by the amount of
your sidetone each time you transition between TX and RX. Wider splits make
it an even more difficult problem for PLL lock in time. Anything done in the
PLL needs to have zero impact on lock time to avoid introducing chirp.

The discrete jumps introduced by huff and puff may also be a problem for
folks running PSK31 and similar narrow band digital modes. Those modes can
easily track gradual changes, but they have trouble if the frequency change
is abrupt.

We are testing a much simpler mod that compensates the varactors in the PLL
reference circuit. They dominate the temperature response of the PLL. The
12.096 crystal is not a major contributor, especially since we went to a
single crystal in Rev B K2s.

Stay tuned!

73, Eric   WA6HHQ
------

Lyle Johnson wrote:
> 
> Hello Tony!
> 
> > >This has promise as long as the introduced phase noise is low
> > >enough.  One the the K2's reasons for great signals is its
> > >very clean oscillators...
> >
> > Huf-n-puf does not introduce phase noise. It is simply a counter which
> > decides whether to gently increase/decrease voltage across a
> > varacter which is in parallel with the LC cct of the VFO -
> > in this case an xtal.
> 
> The original huf-n-puf introduced corrections at about a 10 Hz rate, with
> filtering.  For phase noise purposes, this is equivalent to a PLL with a 10
> Hz reference.  You will get sidebands at +/- 10 Hz, +/- 20 Hz, etc., so the
> filtering is very important.  I have seen plots of frequency lock for
> huf-n-puf over the years, but not for phase - but I just downloaded a lot of
> additional material I had not reviewed previously and may find some info
> there.
> 
> The later Fast huf-n-puf circuits have a much higher update rate, which is
> like a PLL with a higher reference frequency.  Again, the filtering in the
> control voltage to the varactor is key.
> 
> With clean power and good filtering, the added noise will be very small.
> 
> 73,
> 
> Lyle KK7P