RadioBanter - Electrothermal squigging in a regenerative radio receiver.

On Sun, 1 Apr 2012, Lemon Tree wrote:

I am conducting ongoing experiments with regenerative radio
receivers. One problem I have repeatedly encountered is low frequency
squigging/motorboating. Particularly when using somewhat higher CE
transistor currents. Oddly, changing the values of any of the
capacitors in the circuit did not change the frequency of the
squigging. There was nothing wrong with the decoupling and even
adding additional decoupling capacitors had no effect. Almost
certainly the unwanted low frequency oscillations are due to
electrothermal effects.
Directing hot air from a hairdryer at the transistors for less that a
second stopped the low frequency oscillation. After stopping the
hairdryer the circuit started squigging again after a couple of
seconds.
The fact that regenerative receivers are operated with the amount of
positive feedback (regeneration) so close to the threshold of
oscillation makes the extremely sensitive to the gain of the
sustaining amplifier.
Even slight changes in the gain of the transistors (caused by self-
heating or the environment) can then have a big effect. This is
almost certainly the basis of the low frequency squigging in the
circuit.

But that's "normal".

Howard Armstrong noticed it early on, and never pursued it. Then later, I
think when he was doing some work on regeneration preparatory to the court
battle over the patent, he noticed it again. And then he pursued it, and
got superregeneration. So his original regen circuit could oscillate at
ultrasonic frequencies as well as at radio frequencies if the regen was
turned up too much. It was just incidental, later he tried an external
ultrasonic oscillator to "modulate" the regen receiver, but for most of
the time ever since, the same active device has been both the RF
transistor and the ultrasonic oscillator. It was all very black box for
most of that time, as it got relegated to secondary status, the
explanations dropping off and few pursuing it much.

It's really a scheme similar to reflex (where one stage is both amplifying
radio frequency and audio frequency) or "autodyne" circuits, where the
same stage is both a rf mixer and rf oscillator.

One project in QST and later their VHF manual had a receiver chain that
used converters feeding an IF strip and then a regen detector, the regen
detector configured so it could go into ultrasonic oscillation too so it
could be either regen (for CW) or superregen.

Turn down regen. That may fix the problem. Or try another circuit.
Realistically, it's not the sort of thing that people report, so it would
seem plenty of regen circuits, tube and transistor, don't have a problem.

Michael VE2BVW