Building a Geiger clicker is somewhat easy assuming the maker can procure two essential components: a geiger tube and the high voltage supply. As most detectors are portable and made to operate on batteries, the high voltage supply is usually conceived as a high frequency inverter, that boosts the DC low input voltage up to the high voltage required to operate the Geiger Muller tube.
In my circuits I usually go with Russian Geiger tubes, and most of them use 400V. So this is why I've designed the circuit presented here, as a high efficiency 3V to 400V inverter, with extremely low ripple.
The circuit is based on a modified Armstrong oscillator, a small ferrite transformer with a feedback coil for the oscillator, a blocking transistor commanded by an array of zener diodes. The diodes dictate the output voltage and are part of the regulating block, responsible for outputting precisely the selected voltage. When building Geiger counters, the voltage control is important, in order to protect the tube which can be affected negatively by uncontrolled voltage. This topology also ensure a high efficiency, while it only requires a couple of simple components, unlike the more complicated inverter circuit presented a while ago.
The circuit details
Download the Eagle files and the PCB pdf, here: geiger_high_voltage_3V_400V_supply
The ferrite core transformer
This is also easy to make, although some people are really scared about making their own transformers. You'll need 16 turns for primary, 16 for the feedback coil, and 650 turns for the secondary. We are not using any voltage multipliers in this circuit, so the high ratio is needed to be able to make the big jump from 3V to 400V. The wire diameter is 0.2mm for the primary and feedback and 0.08mm (very thin!) for the secondary. Careful to add proper insulation, as 400V is enough for internal arcing that could ruin hours of work. Also note that the transformer will work in flyback mode, so a tiny piece of paper must be added between the two ferrite pieces for a minimum of distance.
I improvised a coil former using paper, on a hex key, matching the internal diameter of the ferrite core. Cutting some edges I got a nice support, where the wire will not fall down while winding. I scrap the wires from old relays, but I advise you to use new wire with a good insulation if possible. What I had worked perfectly anyway.
The core I've used is an old 18mm diameter, A22 type core, with AL=2800 .
The T1 transistor (MPSA18) is absolutely vital for this circuit. I usually design devices allowing considerable freedom in changing some of the parts, but for this design, you absolutely must use the MPSA18. Here's an output comparison list, with various transistors used instead of T1, on an input of 3V DC:
T1: MPSA18 - 3V in gives 384V, maximum selected by the zener diodes, clearly the transformer puts out a few tens of volts more
T1: 2n5551 - 3V gives only 152V, the regulator section doesn't even fire
T1: 2n4401 - 3V gives 328V, nice, but not enough
T1: 2n2222 - 3V gives 284V, again too low
T1: BC548 - 3V gives 220V, only
As you can see the MPSA18 makes a huge difference. With the MPSA18, the circuit will oscillate even at 1.5V, outputting 236V!
The Zener diodes connected in series will control the output voltage. For my boards I used 3x 120V Zeners + 2x16V and got 384V output , considering there's a little voltage drop across them. YOu can use any configuration needed. You can also replace the rectifier diodes with a multiplier block in case you need higher voltages, or simply increase the input voltage. The way this works is that when the output voltage is greater then the cumulative zener threshold, the diodes will open the T2 transistor that will block T1, stopping the oscillation, so the voltage will drop. This feedback mechanism, coupled with the output filter capacitors, C3 and C6 will assure we have a constant output voltage, regulated to the value allowed by the Zener diodes.
The R1 resistor, controls the output max current, but also the consumed current. The lower the value, the more current will flow through the T1 and the transformer. Higher current will result in heating, so change this carefully and only if needed.
Selecting the voltage
As presented above, changing the Zener diodes will impact the output voltage. You can use 4x 100V zeners connected in series for 400V output (D5, D6, D7, D8 in the diagram). For my circuit I used 3x120V + 2x16V and got an output of 384V, just perfect to operate Russian Geiger tubes.
The following pictures show the output voltage and the input current:
384V output, 128mA with a neon bulb load (high consumer), 83mA with no load all at only 3V input. These values were collected with R2 set at 4.7K . For the current value presented in the schematics (6.8K), the consumption is even lower, it gets down to 50mA. This can be decreased even further if needed, by changing R2 up to 10K. There's a limit though, as setting R2 too high , will stop the oscillation and the inverter won't work.
The following are screenshots from my o-scope, CH1 (yellow) shows the output voltage and the nice constant value, while the CH2 (cyan) has been connected to T1's collector in the first picture (showing the driving frequency), and to T2's collected in the second picture (showing the regulation frequency).
The MSPA18 is running close to its limits and this can shorten its life. We can replace it with the more robust T1:C1008, while also changing the R1 to 2.7K
With this modification, the circuit draws 70mA in idle mode and 130mA with the neon bulb connected.
A quick test circuit, probably the simplest Geiger clicker circuit, with just a tube, a limiting resistor and a Piezo speaker.
Careful, most of the energy stored in the capacitor will be dumped in the tube, and that will come as a high current burning the quenching gas. This circuit was meant just for a few seconds demo, do not build a counter this way.