What's new

Michel Dumais Electrostatic Speakers topic reply
Laurine Wynn Fried new topic
Kynix Semiconductor Hongkong Limited Dynamic Speakers topic reply
Kynix Semiconductor Hongkong Limited Dynamic Speakers new topic
Kynix Semiconductor Hongkong Limited Dynamic Speakers new topic


American Power Design
American science
C and H Sales Co
ESL Information eXch.
Fair Radio Sales
Herbach and Rademan
Metrum Acoustics


On the top of this page you can find an overview of all brands that supply high voltage power supplies.

There are several ways to create a high voltage power supply for use in an electrostatic loudspeaker. A popular one is to build a voltage multiplier. As the name implies, are circuits designed to multiply the input voltage. The input voltage may be doubled (voltage doubler), tripled (voltage tripler), quadrupled (voltage quadrupler), etc. Voltage multipliers are also power converters. An AC input is converted to a higher DC output. These circuits allow high voltages to be obtained using a much lower voltage AC source.

Typically, voltage multipliers are composed of half-wave rectifiers, capacitors, and diodes. For example, a voltage tripler consists of three half-wave rectifiers, three capacitors, and three diodes. Full-wave rectifiers may be used in a different configuration to achieve even higher voltages. Also, both parallel and series configurations are available. For parallel multipliers, a higher voltage rating is required at each consecutive multiplication stage, but less capacitance is required. The voltage capability of the capacitor limits the maximum output voltage.

Voltage multipliers have many applications. For example, voltage multipliers can be found in everyday items like televisions and photocopiers. Even more applications can be found in the laboratory, such as cathode ray tubes, oscilloscopes, and photomultiplier tubes.


Cockcroft - Walton generator

The Cockcroft-Walton (CW) generator, or multiplier, was named after the two men who in 1932 used this circuit design to power their particle accelerator, performing the first artificial nuclear disintegration in history. John Douglas Cockcroft and Ernest Thomas Sinton Walton used this voltage multiplier cascade for most of their research, which in 1951 won them the Nobel Prize in Physics for "Transmutation of atomic nuclei by artificially accelerated atomic particles". Less well known is the fact that the circuit was discovered much earlier, in 1919, by Heinrich Greinacher, a Swiss physicist. For this reason, this doubler cascade is sometimes also referred to as the Greinacher multiplier.

A similar circuit is the Marx generator, which has the same "ladder" structure but consists of resistors, capacitors and spark-gaps. The Marx generator produces short pulses, whereas the CW generator produces a constant DC.


This Cockcroft-Walton voltage multiplier was part of one of the early particle accelerators responsible for development of the atomic bomb. Built in 1937 by Philips of Eindhoven it is now in the National Science Museum in London, England.


The CW is basically a voltage multiplier that converts AC or pulsing DC electrical power from a low voltage level to a higher DC voltage level. It is made up of a voltage multiplier ladder network of capacitors and diodes to generate high voltages. Unlike transformers, this method eliminates the requirement for the heavy core and the bulk of insulation/potting required. Using only capacitors and diodes, these voltage multipliers can step up relatively low voltages to extremely high values, while at the same time being far lighter and cheaper than transformers. The biggest advantage of such circuits is that the voltage across each stage of the cascade is equal to only twice the peak input voltage, so it has the advantage of requiring relatively low cost components and being easy to insulate. One can also tap the output from any stage, like a multitapped transformer.

Operational characteristics

In practice, the CW has a number of drawbacks. As the number of stages is increased, the voltages of the higher stages begin to 'sag', primarily due to the AC impedance of the capacitors in the lower stages. And, when supplying an output current, the voltage ripple rapidly increases as the number of stages is increased. For these reasons, CW multipliers with large number of stages are used only where relatively low output current is required. These effects can be partially compensated by increasing the capacitance in the lower stages, by increasing the frequency of the input power and by using an AC power source with a square or triangular shaped waveform. By driving the CW from a high frequency source, such as an inverter, or a combination of an inverter and HV transformer, the overall physical size and weight of the CW power supply can be substantially reduced.

CW multipliers are typically used to develop higher voltages for relatively low current applications such as bias voltages ranging from tens or hundreds of volts to millions of volts for high-energy physics experiments or lightning safety testing.


CW multipliers are also found, with a higher number of stages, in laser systems, high-voltage power supplies, X-ray systems, LCD backlighting, traveling wave tube amplifiers, ion pumps, electrostatic systems, air ionisers, particle accelerators, copy machines, scientific instrumentation, oscilloscopes, TV sets, CRT sets, bug zappers and many other applications that use high-voltage DC.


James Wong (December 2001):Along with my audio hobby, I am also an amateur radio operator, and hold anadvanced Canadian license. One of the more popular home projects radiohams often build is a high power linear (RF) amplifier. Usually, theseunits use vacuum tubes in the output stage, and consequently, requirea very high voltage power supply : often in the 4000 to 10000 volt DC range. Schematic drawings and parts lists can be found in amateur radio publications, or you can get assistance in your quest by any well informed ham operator.


High Voltage Power Supply schematic (Audio Design Guide (by Andrea Ciufolli))

Schematic for regulated constant High-Voltage-supply (Chris Neuhaus)

Electrostatic Bais Supply (Sheldon D. Stokes) (PDF)

Schematic for high voltage power supply (Marc Schroeyers) Easy to build, very compact, adjustable HV Supply. Shortly a printboard layout will be added.(PDF)


ePanorama.net, Power Supply Page (Simon-Thijs de Feber, january 2001)

Questions and answers

Hans Engelen (January 2001): Is there anyone who has tried the DC-DC converters (series VV) from Pico Electronics?? On their site they decribe several converters from low DC inputs (5, 12, 15, 18 etc) to variable HV DC up to 5 kV. The working frequencies are around 20 kHz. I asked information via e-mail but they didn't reply.

Mark Rehorst (May 2002): There is a design for a variable HV bias supply using similar DC-DC converter modules made by Emco (competitor of Pico) on my web page: http://www.rehorst.com/esl DC-DC converters work fine and are a lot safer to deal with than HV transformers that plug into the wall (I have first-hand experience!). MR.

Please contact The Audio Circuit if you can provide an answer.

Hans Zeeuwe (December 2001): Jan had made the power supply on a small circuit board 60 x 160 mm full with holes. Around the holes there is a small ring off copper to solder the components on. Over here they are called 'Euro-circuit boards DIN 41612'. I suspected that the board would lose a lot of Voltage because the copper rings were very close together. The space between them was less then 1 mm. I bought a small drilling device and milled away all copper rings that weren't needed. This gave a considerable improvement. The ESL's sounded louder now and had a little more bass. But they were still very far from good.

I decided to build a completely new power supply. I got the description from the book of E. Fikier and encountered my first problem. How to make a circuit board. I surfed the net for a bit and found some sites with information on circuit boards. Making one myself would be a costly business. Luckily I found the RES site (Robs Electrosoft)(no longer online). Rob made me the boards for about $ 10,- a piece, all holes drilled and in perfect shape. I bought the components at Conrad and started soldering everything in the right place. I used the following components:

HV power supply

Above the layout of the circuit board. Enlarged plans are Available on exchange basis

  • 2 circuit board trafo's (220/12 V)
  • 27 capacitors 0.047 uf / 630 V
  • 27 diodes 1N4007 (1000 V)
  • 8 resistors 1 M Ohm

Soldering the board was easy, I had had to make sure that the solder was nice round shaped with no points sticking out. After an hour or two I finished the board. I soldered the connections to my first set of ESL's and the audio transformer. I turned the power and waited a while for the ESL's to charge up. I put a CD in the player and turned the volume knob. The sound was much, much better now.. There was more bass, they sounded much louder and the sound itself was very clear. But they still were absolutely no match for the Quads.

As I thought that I did the construction part of the ESL's reasonably well I decided to look at the power supply again. I had no way to measure the high Voltage output so I Bought a high Voltage probe. It's an expensive device ($ 90,-) but very useful as well. With the help from MartinJan Dijkstra I managed to measure the high voltage output. The unit delivered 4200 V. That was about 3000 V short to the 7500 V I needed!. I contacted Roger Sanders and told him about my problems. He suspected that the capacitors weren't large enough. I build a new power supply with the following components:

  • 2 circuit board trafo's (220/12 V)
  • 27 capacitors 0.1 uf / 630 V
  • 27 diodes 1N5408 (1000 V)
  • 8 resistors 1 M Ohm

This resulted in a power supply that delivered 5200 V. I had read somewhere that I should cover the entire circuit board with acid free silicone lute in order to diminish internal losses. I bought a tube of the stuff and covered the whole board with it and make it look good as well I evened it out. To do this I put some liquid soap on my finger because this works very good with silicone lute. Unfortunately I forgot that liquid soap is also very conductive. Some of the soap must have found its way to the circuit board. The power supply now delivers only about 2700 V.

Gary Winter (December 2001): I just read your project description and may have some help for you. I am an electronic engineer and have produced many products. It appears to me that your 7500 Volt power supply has a few problems. If I am reading all of your information correctly, you are using voltage multiplication techniques to produce the 7500 volts. If true, then coating your circuit board with silicone compound is a mistake. You were given bad information. Silicone compounds are hygroscopic. That is, they absorb and hold moisture. Also, they employ an Acetic acid/carbon dioxide cure that will corrode your connections over time.

So, on these boards, cleanliness is the key. After completing soldering, clean the boards thoroughly with Ethyl, Methyl or Isopropyl alcohol. Make sure that the capacitor type used can withstand this cleaning. Next, use a freon based solvent (usually sold as a circuit board cleaner) to spray the board down well to remove any residue.

After this step, DO NOT TOUCH THE BOARD SURFACE, ANYWHERE NEAR ANY COMPONENT! Handle only by the edges. If you want to coat the boards after this, use a product called "conformal coating" that is designed for high voltage use. I would be happy to send you a can of this if you cannot find it locally (and if it is allowed for me to send internationally). I would suggest testing the boards individually first to see if they now produce the required voltage. If you follow these instructions and the circuit still does not produce the required voltage then you may have one of two problems:

  1. leaky capacitor (inadequate voltage rating?),
  2. inadequate capacitor size to supply the current needed for the load and to compensate for losses.

If you test each board separately to see how much voltage each produce, one board may show up as bad. This suggests a leaky capacitor. If they all test with low voltage, this suggests inadequate capacitor value or wrong type of capacitor.

Try doubling the capacitor value. Capacitor type is also important. Polypropylene capacitors would be the best to use in your application. Make sure that the voltage rating is high enough for each doubling. The first capacitor must withstand 2 times the input voltage. The last capacitor must withstand 2 times the output voltage from each board. For example: if your input voltage is 100 Volts peak, the first capacitor must be rated at 200 Volts minimum. Each subsequent capacitor must be rated at twice the previous capacitor's voltage rating. Your last capacitor must be rated at at least 4000 volts.

I hope this information helps you. I wish to tell you that I admire your courage and perseverance on your project. I hope you make it work. Maybe one day I will be buying your speaker products. Regards, Gary Winter

Jarno Juhani Laine, December 2001: Don't know if this is an already solved problem, but have you ever tried to build a hi-voltage source withOUT any 12V transformer ? The ladder-network comes straight to mains. It's good to use capacitors 1250V minimum. And those resistors.. I'm not quite sure how you have connected those or where, but 22nF is totally enough for capacitors if the securing resistors are for example 20x1Mohm between the mains zero and phase and another 20x1MOhm between diaphragm and HV source. Although, it's not the most legal way to form a 5kV here in Finland. Possibly you have it more 'free' there in Holland. :-)

Wagner in his ESL-book says that '5kV is the optimum voltage to polarization. If voltage goes higher it doesn't actually HELP and so on'. I too have noticed that 4.3-4.8kV is OK. Better sensitivity is achieved of course, thinning the isolators. (Is that proper English?!) We have already built 3 prototypes from 25 x 100cm to 33 x 200 cm. The transformers have been 1:50. They are extremely insensitive (78dB) but whattheheck. Rotel 991 has the power. :-)

Please contact The Audio Circuit if you can provide an answer.