Experiment to see how little voltage is needed to ring a telephone bell. runs on 1.2V to 3V DC
Transistors are all fairchild PN100, they sink 300mA without any problems and cost about 8c each.
Transformer is just a fat 1 inch diameter toroid I pulled out of an old power supply.
The bell (blue coil) is out of an old PSTN telephone, it's solenoid has 8.8k ohm resistance. To ring the bell normally takes 48V AC at 20Hz from the phone exchange.
Circuit is made up of three parts, from left to right:
Capacitor charging;
Voltage detector;
Capacitor dischargen
When the switch closes, the charging circuit transistor Q1 oscillates and back emf spikes get fed into the 1uF cap.
The voltage divider on the storage cap holds Q2 off until the cap hits about 60V. Then Q2 switches on. This switches Q3 off and sends Q3's collector up to 3V. Feedback from here goes back to Q2's base to hold it high.
This turns Q4 on, which lets the cap discharge through the coil for a short while. The 80k feedback resistor holds Q2's base high and Q3's base low to allow the capacitor to discharge a bit.
Unfortunately the bell requires AC to make it ring continuously to push the solenoid in both directions. The solution I found was to mount the bell vertically and let gravity handle the downward strike.
Tweaking the resistors on the base of Q2 gives all sorts of crazy ring patterns.
Next step is to work out how to feed the excess charge in the cap back into the button, so idiots that like to ring the Doorbell for more than a few seconds get "behavioural adjustment" feedback
