Hi!.
I built a basic Pierce oscillator: emitter to gnd, about 33pF at B and
C, xtal from B to C, 180kohm from B to C. Feed thru choke. x10 scope
probe at collector.
With 3, 6 or 10MHz xtals, I sweep Vcc from 1.5V onwards, no problem.
With 24MHz, or 22MHz (actually a 110MHz 5th overtone, oscillating here
at its natural f), at first the amplitude increases gradually with
increasing Vcc, the negative peak being far from saturating the
transistor. Suddenly, at some Vcc the amplitude increases to near
twice Vcc and the transistor saturates at the negative peak. Going
backwards with Vcc, I also find hysteresis in this behavior.
I tried a different transistor model with same results.
Frequency shifts a mere 100Hz when jumping amplitude, so it doesn't
seem the xtal is falling into some spurious mode.
If using 1.8kohm to feed the collector, I get a nasty
superregeneration envelope.
I have read thin quartzs must be driven with lower levels to avoid
mechanical damage. But could it be that high levels produce an
instantaneous (not temperature related) lowering of the series
resistance? (which in turn causes more drive, hence more R lowering,
etc.).
Just in case, I will try a Butler oscillator (tuned at the natural
frequency, not an overtone as usual) because it imposes much lower
drive, but I will be glad to hear if anybody had a similar experience
with the Pierce.
Many thanks!

"lw1ecp" wrote in message
...
Hi!.
I built a basic Pierce oscillator: emitter to gnd, about 33pF at B and
C, xtal from B to C, 180kohm from B to C. Feed thru choke. x10 scope
probe at collector.
With 3, 6 or 10MHz xtals, I sweep Vcc from 1.5V onwards, no problem.
With 24MHz, or 22MHz (actually a 110MHz 5th overtone, oscillating here
at its natural f), at first the amplitude increases gradually with
increasing Vcc, the negative peak being far from saturating the
transistor. Suddenly, at some Vcc the amplitude increases to near
twice Vcc and the transistor saturates at the negative peak. Going
backwards with Vcc, I also find hysteresis in this behavior.
I tried a different transistor model with same results.
Frequency shifts a mere 100Hz when jumping amplitude, so it doesn't
seem the xtal is falling into some spurious mode.
If using 1.8kohm to feed the collector, I get a nasty
superregeneration envelope.
I have read thin quartzs must be driven with lower levels to avoid
mechanical damage. But could it be that high levels produce an
instantaneous (not temperature related) lowering of the series
resistance? (which in turn causes more drive, hence more R lowering,
etc.).
Just in case, I will try a Butler oscillator (tuned at the natural
frequency, not an overtone as usual) because it imposes much lower
drive, but I will be glad to hear if anybody had a similar experience
with the Pierce.
Many thanks!


You might wish to consult the chapter on crystal oscillators in "Vacuum
Tube Oscillators" by William A. Edson. In particular you should read
9.9 Power Output and Crystal Dissipation in the Pierce Oscillator. You
might also wish to read the section on the Miller oscillator. Don't let
the "vacuum tube" title scare you away as the book is really good at
fundamentals. Do a Google search for this book. At one time, someone
had graciously posted copy in Adobe format.

73, Dr. Barry L. Ornitz

Barry, many thanks for suggesting this book!. Don't worry, I have a
great esteem towards vacuum tubes and their underlying time-defying
principles (after all, the book is just 4 years older than myself...).
Power applied to the crystal by a tube oscillator is surely higher
than in an equivalent solid state counterpart, but 50's xtals were
also bigger (e. g. FT-243 format) and with metal contact plates that I
guess contributed to dissipate heat as a side effect.
Honestly, it adresses the crystal dissipation from a frequency
stability viewpoint, not quite my present concern, but I have found
interesting material in different chapters for other projects at home,
such as induction heating, and analysis of intermittent behavior. And
now I admit I didn't understand the principle behind the Miller
oscillator.
Ok, I will do some more experimenting and web search, and will
eventually post the results.