Hello Sveta,
Perhaps you'd like to avoid transformer-based circuits to save size and weight. That's a fine goal, and it explains why in many applications their use has been declining. Unfortunately, inductor-less "charge pump" dc-dc designs are usually limited to inverting a dc voltage or multiplying it a few times. Magnetic components are virtually always used in a high-voltage application such as yours. The reason reminds me of Scotty from Star Trek, who would famously say, I can't change the laws of physics! Here, the law is Faraday's Law of magnetic induction, from which we get two effects that help your cause immensely:
- The voltage "kick" an inductor supplies in response to changing current, as in v = L di/dt
- The voltage multiplication at a magnetically coupled winding with a higher number of turns
An inductor gives you the first benefit, a transformer can give you the second, or both.
The good news is that typical Geiger counter circuits often have a 10 megohm resistor in series with the tube. This means your 500V source may only need to deliver only about 50uA or so. One thing magnetic components do really well is trade input current for output voltage, so let's see what can be done:
The boost converter uses an inductor's "kick" to provide a voltage as much as 5 to 10 times its input, but that's not quite enough for your needs.
Adding a second winding to the boost inductor produces a transformer, the basis of the flyback converter shown here:
This takes advantage of both magnetic effects described above, which actually makes it easier to deal with. That's because we're pushing neither the inductive kick nor the turns ratio to their limits. In this circuit the MOSFET is driven to switch ON for just a few microseconds at a time. While ON, energy builds up in the transformer's core (which usually includes a tiny air gap). When the MOSFET switches OFF, that energy dumps into the secondary circuit and into C1. The output diode D1 must be a fast-recovery type. The resistor R3 represents your load. C2 and R2 form a snubber which damps out unwanted resonances ("ringing") of the transformer's windings with stray capacitance. R1 damps X1's gate circuit, and there should also be a bulk capacitor (not shown) at the input supply. The transformer has a 1:20 turns ratio, which is not particularly critical.
We can see this thing work in a simulation: this photo shows a few cycles at start-up:
The yellow trace shows X1's drain waveform with its switching cycles, the red trace shows the output increasing with each one. If we "leave it running", we see the output surpassing 500V:
Of course, you would like to have a control IC that would adjust the MOSFET's switching to regulate the output voltage. In your application, I suspect the output will be nearly unloaded except when the Geiger tube is subjected to unusually high radiation levels, Thus, the control circuit needs to throttle back the switching when output reaches 500V. It could either reduce the MOSFET's ON time, or have it run at a slower frequency. I'm leaning toward the latter method, because it would result in the least wasted power during minimum-load operation. This "constant ON time, variable frequency" operation is far less common than the "pulse width modulation" (PWM) method usually found in control ICs. I took a quick look at ADI's controller offerings, they are mostly for low-voltage, high current applications. Perhaps someone from ADI could help you find one. Or (gasp!) you could look elsewhere. You could also design your own. Nowadays some microcontrollers may be fast enough to run this thing directly (you'll need to buffer the drive output to switch the MOSFET quickly).
A Google search of "DIY Geiger counter" yields many converter designs using a non-regulated 555 timer driver and rather large transformers -- fine for hobby use but not exactly professional grade. Unfortunately, switching transformers tend to be custom made for each application, thus are hard to come by. To breadboard something quickly, perhaps you could take apart some no-longer used cell phone chargers -- it seems like everyone has a drawer full of them. Running the transformer from one of these backwards would be fairly close to what's needed.
Best regards,
Bob