Yeah I made one of those when a student - made the neighbours complain that something on fire

Not as complex or as many parts though

There is a terrorist under every bed. Uncle Bush has done his work well.... some of you are terrified and Mr Bush is the only man who can protect you from the dangerous, but invisible, enemy.

Reply to R. Ed Huwa -- Don''t forget this device has a hot common! One side or the other of the power line is connected to circuit common via the full-wave rectifier, even if via the light bulb. You don''t want that coupled to your equipment (or yourself!). Even via suitably rated capacitors, the voltage step from the SCR could cause damage to both the signal source and the device. The safest way to inject a signal from an external source would be via an opto coupler. Remember also that Q1 serves the dual function of noise generator and voltage regulator by operating the emitter junction in reverse breakdown. Anything which defeats this voltage limiting function will allow the normally low voltages across C2 and C3, and thus PUT1, to run away. OTOH, if the voltage at PUT1''s gate gets too low it will stop triggering the SCR, with the same result. You might try connecting a transistor-output opto coupler across R1 (+ side to Q2''s collector) and set R1 to max resistance. That would modulate PUT1''s gate voltage while keeping it within the design limits. Safety considerations: If you want to connect test equipment to this device, you''ll need to power the device from an isolation transformer. I used a small dual-primary power transformer for this and a low-wattage bulb. The circuit behaves pretty much the same independent of lamp load, and the small load suited the small transformer. Keep aware that you have line voltage at your fingertips even if you''ve floated it with the transformer, and that once any part of the circuit is grounded via the scope or other instrument, the rest of the circuit is no longer floating! Keep yourself insulated from ground (e.g., the set screws in your scope knobs), and both hands in sight or one of them in your pocket. I don''t want to discourage anyone from experimenting who understands how to do it safely. It won''t jump off the bench just to bite you, but you do need to be careful.

Project is simple to build and well thought out.
Thanks to the author for sharing the article and the applications with multiple positive extensions.
Consider using a random noise generator modulated with audio spectrum low frequency VCO adjustable to enhance the effects, capacitively coupled to Q1 base.
Not sure of the loading effects from source Z, could take some experimenting.