There is an Analog Devices ADI/Maxim Application Note describing a low power consumption circuit for simply and economically generating the high voltage, at very low current, required by a Geiger-Müller tube. The PDF file may be downloaded from the link I have given to the ADI website.   

To maintain simplicity in a very basic portable Geiger Counter I needed a simple output stage. In addition to the audible clicks I wanted an LED flash with each click as a particle was detected. The conventional approach would be to use an audio amplifier for the loudspeaker, and a monostable driving the LED through a resistor.  

But an audio amplifier has some standby current, and a monostable is an extra IC, or at least two gates - plus interfaces between the G-M tube and the monostable and another (probably just a resistor) to drive the LED.  

The power supply uses all six Schmitt input inverters in a 74HC14. Five of them in parallel drive the high capacity gate of an NMOS transistor. But since this design was published suitable NMOS devices with lower C_in have become available[1] and I considered that I could steal two inverters for my part of the system without affecting the PSU performance[2].

When a particle is detected by a G-M tube the gases in the tube ionise and discharge the anode/cathode capacitance. When the voltage drops below that necessary to sustain ionisation the discharge stops and the capacitance recharges through the series resistance R3 (which is usually in the range 1-10MΩ - see the tube's data sheet). A charge of the order of 20-80 pC flows in a few μS.  

The input capacitance of an inverter in the 74HC14 is ~5 pF so if the charge is at the high end it would charge to 16V, nearly three times the absolute maximum input voltage. Fortunately the 74HC series devices have input clamps which switch very quickly and can carry 20 mA, so the voltage does not rise outside the danger level, but this is a use of 74HC logic with inputs outside data sheet limits. It is not a very interesting abuse, though, and such clamping is quite common and never a problem. The abuse for which this article was written is elsewhere.  

The time constant of 5 pf / 10 MΩ is 50 μS so the output of the inverter fed from the G-M tube is a negative pulse of 50-100 μS. This pulse discharges C2 via the 1N4148 diode, so the OP of the second inverter goes high, turning on the NMOS device in series with the moving coil loudspeaker. C2 recharges through R1 and the NMOS turns off (quickly - the 74HC14 is a Schmitt input device) after about 5 mS. The loudspeaker coil is, of course, inductive so there is a negative flyback voltage and the LED conducts for a short high intensity flash.  

This use of the voice coil of a moving coil loudspeaker as a flyback inductor is the major abuse for which this circuit is submitted.  

An output stage which steals two inverters from the original HT power circuitry and adds just five components:- a capacitor, a diode, an NMOS FET, and two resistors, to make a combined speaker and LED driver might well also be considered an abuse of the original circuit.

[1] I shall not suggest a better choice of NMOS as Murphy's Law would inevitably cause it to become obsolete, expensive, or politically incorrect. Read my article or RAQ on choosing transistors at (URLs in links).

[2] The PSU design is clever and deserves consideration - but it is not mine, so I shall leave it to the original author to discuss here if he should wish to. I would, however, point out that there is an abuse in the design - it uses an operational amplifier as a comparator. This is rarely a good idea - see my RAQ article and application note (URLs in links). I have not yet built it as designed but have done something similar using an Analog Devices ADCMP350 comparator, with integral reference, in place of the op-amp/comparator and separate reference...