Advice on the art of radio design, local oscillators and filters etc
 

hmmm number 1 rule...make sure you're not chasing problems faced by
poor test equipment!

Well got rid of the scope probe and used a direct connection with some
50R coax, suddenly I can see the actual oscillator waveform (was
completely killed by all 10 sets of probes I pulled out of the box,
some are 100Mhz rated). So I now know when its oscillating or not.
Also, the 200 odd MHz reading on the frequency meter was caused by an
impedance mismatch doubling the frequency. With 50R input on the
frequency counter and a suitable output resistor on my oscillator
buffer amp I get the correct frequency.

Also I'd made a silly mistake which was throwing me (causing no low
mixer output)...I connected a 455kHz IF can filter can across the
output of the 612, but the impedance was too low, so it was just
hammering the output. With a suitable 1.5kR ceramic filter I can see
tuned carriers. (My excuse is that I was suckered into it by having
used a similar tuned can on the input of the 612 in the past!)

So...a bit of progress. Still looking out for a good HP spec with
300Hz res band option! Everytime I see a skip I look in
hopefully....

Thanks to everyone who provided their thoughts on this...its helped a
lot

Advice on the art of radio design, local oscillators and filters etc
 

Hi,

I've been recently going through a similar process. Designing a UHF
VCO for a hobby project, mainly as a self-education project. As in,
rather than buying the thing from Minicircuits or whoever and be done,
learn how to design and build one myself, so that I never have to buy
them. And to get them much cheaper... a few cents a piece, as long as
you don't mind the effort.

Anyway, when I couldn't get it quite right for a while, with tuning
range, harmonics, output power, etc, here's what I did: Build a
prototype, measure the frequency and output level. Try with a few
different tuning cap values in place where the varicap would be. Got a
proper RF simulation software*. Spice is probably not the best tool
for that. See if simulated oscillator frequencies match what was
really measured. I found out that after I included all parasitics
properly (eg. 0805 capacitor might have about 2nH inductance), and
also the PCB interconnections and the right transistor model, I get
almost exact match. A 1.4GHz frequency in the simulation, within a few
% off the real measured frequency. So now I know I got the simulation
right, and I can go ahead and experiment with different varicap types,
circuit topologies, etc... and I only get to the next PCB+soldering
iron exersize when the simulated performance is what I need. Saves
time.

* I realise that getting 'proper' RF simulation software can be tricky/
expensive...

-- Cheers,
3flp