Regulated High Voltage Power Supply: Switchmode control, 30 kV, about 1 mA

Disclaimer: This circuit description is for your information only. Do not attempt to duplicate! Danger! High voltages can kill you!

For many purposes, a regulated high voltage supply can be very handy. Purchasing one is not easy because such supplies are in high demand by all hobbyists that love sparks, and used units are quite expensive.
Fortunately, high voltage transformers are much more commonly available, from TV sets, so-called “Flyback” transformers. These provide, depending on type, up to 30 kV of DC voltage, at a pretty decent power of about 20-30 Watts.

Quite a few circuits are around, to provide drive signals for flybacks. These work, but often only for a short time – after a few sparks, the driver stage transistor blows. And, most of them have no means of adequately controlling the voltage.

Control of the high voltage typically requires measurement of the high voltage, a task that is not easily implemented – requiring expensive and bulky divider resistors.

The circuit described here is time-proven, and eliminates many of these shortcomings.

(1) Snubber networks around the coil and switching MOS-FET eliminate spikes in case of a shorted output (sparks act like shorts on the output).

(2) The primary winding – it is of very low inductance, just a few turns. This makes it easier to control the voltage spikes on the primary, and allows low voltage drive circuits (much reduced risk of electric shock, much better for amateur use). Still this circuit is dangerous! Do not attempt to replicate!

(3) Switchmode regulation with current limit – also this provides added protection against overload. Dead-time control limits the maximum duty cycle (power).

The design, all build around a real classic, a TL494 switchmode regulator.

hvpwr schematic switching regulator

Power is supplied via mains filter, and two transformers. Second transformer is only a needed for the LCD panel meter (Voltage display in kV). Note that the regulator and primary coil is on a floating ground. Earth ground is used for the high voltage coil, with a 1 mA full scale amp meter. Actual output current can be above 1 mA, depite the caption “0.5 mA”. Make sure to adjust the maximum current conservatively, and also the dead time, to keep the output power limited – otherwise, the flyback transformer will suffer and eventually stop working.

hvpwr schematic pwr supply

The most interesting part, the flyback driver and snubber circuits. The snubber circuits were designed with a lot of effort, using a scope to probe the overshoot voltages, etc. – if you change the flyback or primary inductance, make sure the check proper dampening! The BY329-1200 is a fast diode, with rather slow (soft) recovery. This will lead so some extra power losses, but this can be tolerated here. The VDRs add some more protection, but actually, they were added more for peace-of-mind than for any real purpose.
One thing to be improved in further design updates is the gate drive: the gate to source voltage is currently the full voltage of (about 24 Volts), it works, but it is running close to the limit of the IRFP450 device, or even above the limit. Furture circuits will include an independent linear regulator, to run the TL494 from a 15 V supply, derived from the main supply, and maybe a Zener diode added to the gate drive signal path, for added protection agains gate voltage excursions.

The IFRP450 is driven rather hard, via a 47 Ohm gate resistor, and has very small switching losses. No heatsink required, mounting it to the rear panel will provide ample heat dissipation.

hvpwr schematic flyback driver

The flyback itself, a 0100170 equivalent to HR 8409 type. The only coils used are the secondaries. The high voltage coils, for the output, and the 11-12 coil, for the voltage feedback. The voltage derived from this coil by half-wave rectification is a very accurate representation of the high voltage output. This has been checked at multiple voltages and load conditions!
The primary winding is 8-1/2 turns of rather heavy copper wire. You can also twist multiple thinner wires, if no thick wire is handy. I used double-isolated type; the device below is just a lab demonstrator for personal use, for professional cirucits, it is suggested to use PTFE (Teflon) or silicon tubing to provide additional isolation of the primary winding from the ferrite core.

Keep your fingers (and other wires) off the other primary windings!! These carry dangerously high voltages, at significant power, and might be more dangerous than the acutal high voltage secondary!

hvpwr flyback hr 8409 equiv 0100170

The actual unit:
hvpwr front panel

hvpwr inner workings 1

hvpwr inner workings 2

hvpwr inner workings 3

hvpwr inner workings 4

hvpwr rear panel

The rear panel has the high voltage output: a big isolator, machined from HDPE plastics, with a 4 mm receptacle hidden inside.

12 thoughts on “Regulated High Voltage Power Supply: Switchmode control, 30 kV, about 1 mA”

  1. Very interesting article. I built a similar HV generator using the 2SK2611 MOSFET and the flyback transfo HR-46057, able to reach 57 KV. I used the original primary, instead of winding a new one of a few turns, and a 70 V power supply. The generator works fine if the output voltage remains under 30 KV, for higher voltage I burned many transistors, may be because I did not use snubbers. Well, I want to add snubbers to my project, but I can not read the values for some of the components shown in the diagram above. Can you help me about that? Thank you.

    1. Sure, just let me know which components you are uncertain about. Some of the snubber capacitors may need adjustment, because of the larger inductance of your primary coil. Main reason for adding a low inductance primary here is to avoid burning transistors when output is shorted by mistake or by spark.

      1. Simon, can you tell me about the two capacitors just above and connected to the primary winding ? the other two capacitors should be rated for 1000 V ? and also the two VDR for 90-100 V? thank you.

        1. These are 2200 µF 40V (a good low ESR electrolytic would be best), and 2µ2 100 V WIMA MKS-4 (or similar type). These will absorb current spikes and need to be wired close to the primary via a low inductance path. The two snubber caps are 0.022 µF 630 V WIMA FKP1 (high current pulse capacitor). VDRs are 60 V, S14K60 (0.6 Watt VDRs).

  2. Probably one of the best flyback drivers on the internet, have you experienced any MOSFET failures at all?

    1. Hi John,
      No, this has been working ever since. The low inductance primary in combination with the snubber networks take care of any spikes.

  3. Love it! Easily one of the best HV CRT flyback drivers on the internet, its not just over-voltage which destroys the switching device but over current events from core flux integration which the current sensing takes care of.

  4. Hello Simon.
    can you specify … tl494 vcc supply is 15v dc via stabilizer 7815 ..?
    and where does kv dc measurement , take with panel indicator or from 1 pin tl494 ..?
    I’m not lucky to have a stable kilovoltage.


    1. Hi Gin, the tl494 is powered directly from the 24 volts unregulated supply, but I was thinking it would be better to drive it from a stabilized source. The measurment is taken from U_sens via the capacitor and resistor divider as shown in the right hand side lower corner. The resistor divider is adjusted such to make sure the voltage is tracking the meter correctly. If you don’t get stable voltage maybe you need to tap another coil on the flyback that is closer to the secondary.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.