Author
WSQT Radio
Since I had no source of a PLL chip for the WSQT FM setup, I used an
old amateur radio approach to VHF stability-premixing. A lot of ham
rigs use it even today in HF, as it is the only way to make an SSB
signal for many frequencies and that mode is used a lot by hams.
In premixing, a low frequency VFO is mixed with a higher frequency
crystal controlled signal, essentially duplicating the stability of the
low frequency VFO at a higher frequency. I used a 10.24 mhz crystal
that when put in a MPF 102 oscillator circut goes at 10.3mhz, a
quadrupler followed by a doubler after it , a balanced mixer,and a 5.5
mhz VFO with the varactor as the other mixer input.
The exciter easily passed the "freezer test" of staying close enough
to center for a digital tuner to lock on at the limits of range(exciter
board only no antenna or amps) after being tuned warm and put in the
freezer. In fact, FCC specs for FM band are a 2 KHZ tolerance for
drift, easily achieved by a 5-7 MHZ VFO. The mixer is not multiplying
the oscillator frequency, so it doesn't multiply the drift.
The harder part of this is filtration, to kill the unwanted outputs
5.5 mhz apart. The use of a balanced mixer to keep out the oscillator
harmonics helps. With a two-MPF 102 singly balanced mixer and the low
oscillator as the balanced input, there is no output on teh oscillator
frequency or any of its even harmonics. Odd harmonic output also stays
down as long as the high signal is the stronger of the two. Essentially
this form of mixer is a push-pull AM amp for the high signal, gate
modulated by the low signal, with the output tank tuned to a sideband
only. If "modulation" if over 100%, you get the odd harmonics of the
oscillator at intervals above and below the desired signal. The
"sideband" is of course your desired output frequency, generated the
same way sidebands of an AM signal or an SSB signal are made.
After the mixer, the filter amps do the real work. They must have
the highest Q tuned circuts possible to reject the spurs. The mixer
output is dirty and not that sharp tuning, but the first filter am is
extremely sharp tuning and brings up the desired signal while knocking
down the trash. It is neutralized as a JFET amp will oscillate
otheerwise, just like a tube. Next, a source-follower isolates the
first filter amp and drives the second filter amp, a bipolar stage
built like a multiplier with a loose-coupled tank with 4 turns tapped
at one in the primary coil, and a one turn secondary. With its resonant
caps, that too is very selective. By the time the signal goes through a
resonant driver, resonant 1W amp, and resonant 10W amp, all spurs are
over 40 db down(maybe 46). For a long time that was all the FCC asked
of ham rigs for operating in residential areas at much higher power.
PLL rigs are theoretically cleaner, but there are some bad PLL boards
out there too. I've heard cheap PLL tuned receivers where digital
sounding noise is what you hear with no signal, and I remember that
story of someone trying out a PLL chip-and in their board they got
spurs in close at only -19 dbC.
BTW, never run a cheap PLL through a broadband amp without filtering
it unless you have very thoroughly check it out! A bad PLL can be just
as dirty as a bad premix board, and a good premix board can be plenty
clean. In any case, testing and more testing prior to any on-air use is
the key. I've got a month is coming up with a premix board that works.