# Practical way to generate 1500V?

In designing the power supply for an radiation detector, I need to supply up to 1,500V at 2 mA from mains (170 Vdc).

I've largely considered a flyback converter, but cannot find a flyback transformer with a voltage gain of N=~16 (for a voltage gain of 8 at 50% duty cycle) rated for 1,500V at the output winding.

Looked into charge pumps/voltage multipliers some, but not yet in depth.

Also just recently learned about CCFL inverter transformers, but still need some time to understand them better.

This project requires a small size and weight, so using a large transformer is not preferred.

What are other ways to step up voltage which I may look into?

• What about a Villard cascade? They were used in ancient televions for the cathode ray tube.
– Bart
Commented Mar 8, 2018 at 12:48
• Consider an inverter with a convenient transformation ratio, then apply the ac to a Cockcroft-Walton multiplier to obtain your target voltage. Commented Mar 8, 2018 at 12:51
• Note that the auto-transformer in your car routinely produces much higher voltages to arc across a gap in the spark plugs. Commented Mar 8, 2018 at 13:42
• @A.S. you should consider formally accepting the answer that most meets your needs and upvoting other answers that prove useful. This is the "currency" on stack exchange. Commented Mar 14, 2018 at 9:04

Try researching Geiger counter power supplies. They don't produce 1500 volts but they do lend themselves for modification such as this design: -

As you can see, the output stage is a Cockcroft-Walton voltage multiplier, so you can add more stages and get more output voltage. Alternatively, you build two of these (with fewer added CW stages) and make a bipolar supply that spans +/-750 volts.

The circuit above runs from 5 volts, but the principle is the same for any DC supply voltage; you make a (circa) 50 kHz oscillator and amplify it to produce a large peak-to-peak voltage swing, then use the CW multiplier to make a larger DC voltage.

More Geiger Muller tube power supply images

• A hybrid voltage multiplier topology is better than a plain Cockcroft-Walton voltage multiplier. A demo converter in this paper: A Hybrid Cockcroft-Walton / Dickson Multiplier for High Voltage Generation delivers 50 Watts from 0.928A and 60 volts DC into 200k ohms at 2,250 volts and 90.9% efficiency using a hybrid Cockroft-Walton / Dickson topology for a voltage gain of 37.5. So it's a high efficiency resonant converter feeding into a 4x4 hybrid voltage multiplier. Commented Oct 4, 2023 at 13:47

There are two main choices

a) Villiard aka Cockcroft-Walton aka Greinacher multiplier

Each stage is stacked on the previous stage, so capacitors and diodes only need to be rated at the input voltage, not the output voltage. However the output impedance increases as the number of stages squared.

b) Dickson multiplier

Each stage returns its pump and storage capacitors to ground, so capacitors need to be rated for the output voltage. However the output impedance increases only as the number of stages, so it can use smaller values of capacitance for the same output impedance as the other cascade.

Given a specification for output voltage, output current, and output impedance, you would need to fully design one of each type to determine which will give you the best cost or volume. Very high voltages would naturally favour type (a), but with only 1500V, both types are contenders.

• You may be interested to know that there is a new hybrid MxN multiplier written about in this paper: A Hybrid Cockcroft-Walton / Dickson Multiplier for High Voltage Generation. They develop a model which they prove by building a 4x4 hybrid resonant converter at 90.9% efficiency for 50W from 10MHz 180Vac p-p input to 2kV output into 100kOhm. Commented May 7, 2021 at 5:10
• @MicroservicesOnDDD Obvious idea, but a very nice paper. I especially approve of figure 3, that's the sort of clear diagram I strive for in my publications. Commented May 7, 2021 at 7:12

There are many companies which will build transformers to your specifications. When building switching power supplies it is more common to use a custom transformer than it is to use an off-the-shelf component.

Books such as Switching Power Supply Design by Abraham Pressman (ISBN: 978-0071482721) and Switchmode Power Supply Handbook by Keith Billings (ISBN: 978-0070053304) go into great depth on how to design custom transformers for switching power supplies.

To prototype your design you can use ferrite cores from companies such as Magnetics, Inc. and Ferroxcube

As for other methods of producing high voltage I would consider a simple boost converter. Properly designed and provisioned for safety, I believe it would also be the smallest and cheapest. (Considering that you can easily buy fast diodes rated >2kV these days)

The reason I would not recommend a voltage multiplier is that the requirement is 1500v * 2mA = 3W. It is unlikely that a voltage multiplier will practically work.

• You assume that "[It is unlikely that a 3W voltage multiplier will practically work.]", but you are incorrect, and your assumption is probably only about the one topology, namely Cockroft-Walton. Commented Oct 4, 2023 at 13:30
• I say this because a demo converter in this paper: A Hybrid Cockcroft-Walton / Dickson Multiplier for High Voltage Generation delivers 50 Watts from 0.928A and 60 volts DC into 200k ohms at 2,250 volts and 90.9% efficiency using a hybrid Cockroft-Walton / Dickson topology for a voltage gain of 37.5. Commented Oct 4, 2023 at 13:32