Ok, so I’m chatting with an engineer pal who is also the self anointed Grand Poobah of Asperger’s and the super simple oscillator is mentioned.
I won’t bore you with the minutia of it’s many interesting subtleties but we were both left somewhat ‘triggered’.
It’s the leg! Cutting that middle leg from a 9018 transistor’s belly is essential to making it oscillate (we know this to be evident) but neither of us knew why, and while my pal was able to list at length all the states of a reverse avalanche he didn’t know, couldn’t ‘see it’, and I couldn’t tell him.
Not knowing sends my friend down a dangerous and anti-social theoretical rabbit hole and not being able to help him triggers me.
So, with the leg cut off, what’s changed?
Any and all thoughts, maths*, physics or tales of old would be appreciated.
Best to all!
*It’s maths with an s, the shortened form of mathematics.
It seems that this is a common misconception. Quoting Matthew Skala:
The story circulates among synthesizer users that an input connected to nothing at all will “pick up noise” by a physical effect similar to how a radio antenna works, and so if one is not deliberately normalling an input to something other than 0V, then it should always be normalled to 0V. That story is not literally accurate. Radio antennas, with some few exceptions not relevant here, need to be on the same order of magnitude in size as the waves they will recieve. They will not pick up any other signals. An open-circuited input jack certainly cannot pick up any audio-frequency noise; the waves in question, and therefore the wire that could receive them, would be many kilometres long. The PCB trace and other wiring between a Eurorack front-panel jack and the circuitry it connects to are usually only about 1cm to 10cm long, which corresponds to a frequency in the range of about 3-30GHz; frequencies on that order of magnitude are the only ones it can pick up. If you have enough microwave energy floating around your environment for an audio input circuit to be affected by receiving gigahertz-range signals through a disconnected trace like an antenna, then you already have serious interference problems through other routes anyway.
Not an expert on transistors, but I should think it acts like an antenna and picks up noise (likely in the 10gHz range), this noise allows charge to “leak” through the junction so that the avalanche threshold is not met.
Would the same be necessary if the form was SMD, in an array, parallel (doubled up - v old school robotics) or some other form other than TO92 through hole?