Dave,

At present I have an issue with a Butler emitter follower oscillator at
150MHz.

I have Matthy's "Crystal Oscillator Circuits" and Randall Rhea's
"Oscillator design and computer simulation" books. Also EMIRFD,
"Introduction to RF design by Wes Hayward", "Complete Wireless design by
Sayre", "RF design by Boswick", ARRL handbook, VHF/UHF handbook and a
few more fundamental electronics and antenna books but try as I might I
just cannot seem to ever extract enough information from all these books
combined to get a useful result.

The Butler design I am looking at has a circuit for 100MHz. I assume
this would be close enough to use the same typology but adjust the
tuning components for 150MHz.

The book mentions tuning near by not at resonance. I have no idea how
far "near" is and should it be above or below resonance or does it not
matter.

He then states the holder capacitance of the crystal is 4.2pF and that
he tunes this capacitance out with a 410nH inductor. If I calculate
this, I get 600nH inductor to tune out 4.2pF at 100MHz. Why would he use
410nH instead of the larger value ? No mention in the text.

He mentions that tuning "near" resonance is achieved with L and Cap from
base to emitter. He shows a 8-15pF cap. If I assume the trimmer set
about half way at say 12pF and use his value of L of 120nH then the
circuit is actually tuned to 132MHz. (I am not sure if this 30% higher
frequency is what e really means by "near resonance". He also does not
take into account the Base emitter capacitance or the cap to ground of
the c-tap. Is the assumption that these do not matter ?

I made the circuit with changes in values to try to achieve 150MHz and
it worked but had extremely critical tuning (very small adjustment in
the tuning C stopped the oscillation). It also appeared to run slightly
lower than the stamped marking on the crystal so I assume the either the
crystal is cut 100ppm lower for parallel mode or I have not tuned out
enough Co or something ?
This is where I now need more detail to understand why the circuit
behaves like this and what to do to fix it.

I would also like to change the bias current as the 19mA current draw is
a lot for my application. 5mA would be good but without properly
understanding how to calculate Rout of the emitter follower, I am not
sure the resistance is getting too high toi drive the crystal.

I have tried common base butler designs and they seemed to be easier to
get going and understand but several text point out this design is prone
to have instabilities and small ranges of inductance in the tuned tank
that they will run over.

I need to find some text that would cover issues like this but not so
deep mathematically that I need a degree to use the math.

Thanks




Dave Platt wrote:
Dave,

I do have EMIRFD but find almost all the circuits are based around rf
transformers for matching and most of them are low frequency (3MHz etc).

For the primary matching stuff, that's true.

You might want to look at the sections on crystal filters... if I
recall properly, there are some filter arrangements which use an LC
match at one or both ends.

You might want to dig into the ARRL books on UHF and microwave, and
perhaps a copy of the old ARRL VHF manual (out of print but available
used).

I think, though, that you're probably out on the bleeding edge of
amateur experimentation. Not a lot of people are homebrewing VHF gear
these days, so there may not be a single book which goes into the sort
of circuitry you're interested in with an eye towards from-scratch
design work.

I'm looking right now at the ARRL VHF Manual's description of a
2-meter portable design, circa 1972 - AM transmitter and
superregenerative receiver. It's all solid-state - no tubes.
However, as with most such designs I've seen, it has quite a few
transformers in it - typically part of single- or double-tuned
resonant circuits between the stages. The design seems to do both the
tuned filtering, and the impedance matching with the same components -
not an unreasonable approach. The oscillator strips in a
commercially-built 2-meter FM repeater's transmitter and receiver I
tore apart recently use a similar approach.

If you really want a transformerless design, I'm sure it can be done,
but the necessary L/T/pi matches may end up being more difficult to
design, and trickier to tune than a more conventional
transformer-based design.