High Voltage SMPS Circuit Design WIP - Please can you take a look and comment

Thread Starter


Joined Aug 21, 2013
Hi All

I have designed a -30kv high voltage SMPS based on designs of power supplys in my workshop and powersupplys on the internet. It has been designed to be through hole, simple and modular.

I have researched each individual section and put together a circuit diagram.

- Voltage and Current need to be Adjustable on Faceplate
- Frequency and Duty cycle need to be adjustable on PCB
- Analog 50uA Ammeters as the KV and mA output meters
- IRG4PC40W IGBT's are switched at approx 100khz (adjustable) Low side PUSH PULL config
- ETD 49 Core High Frequency Transformer T3 - 320 Vac Primary - 4000Vac Secondary 100khz
- 4 stage cockroft walton voltage multiplier takes 4000Vac 100khz IN for -30000Vdc @ 30mA OUT (http://www.blazelabs.com/cw-brm-java.asp)
- The 4 stages of the CWVM will be potted in resin
- Totem pole driver for Gate Isolation Transformer T2
- All 15Vdc components will be on a seperate PCB

I have some questions:

- Ive used a voltage divider on the 30kv output 200M ohms / 13.3K ohms to feed the TL494 with 0 - 1v dc. Is there a way to increase the accuracy by increasing the voltage range? The TL494 seems like 1V / 2V on the comparator?
- Will adjusting VR6 (10K voltage adjust pot) cause problems for the analog reading ? Do I need a op amp in here?
- Is the gate drive circuit for Q1 and Q8 acceptable to stop both IGBT turning on at same time?

I am going to add:

- 12v DC relay to switch ON/OFF the 320V DC after the capacitors so I can turn off the high voltage output with a panel mount switch
- Current limit op amp to drive a warning LED
- Current limit op amp to trip the HV output if current goes above a set limit

I would like some feedback on the current design please and any changes you suggest.

Thanks in advance




Joined Sep 20, 2005
What core and number of turns are you planning for you transformer?
In the schematic I don´t see how any return point for the output current back to the multiplier. Also I don´t see how the output voltage and current sensing is supposed to work.
The primary of the gate transformer should be AC coupled to prevent saturation.

Thread Starter


Joined Aug 21, 2013
Hi kubeek

Not sure of the gate isolation transformer core or turns yet. I was looking for 18v ac on the secondary
The multiplier return is through "current sense" R3 10W 10ohm resistor to ground

How do I AC couple a transformer?


Joined Sep 20, 2005
You would use a capactitor in series with the primary winding.
The core type and number of turns was meant for the power transformer.


Joined Aug 11, 2017
I looked at your schematic I found that you have connected output control pin(13) to ground , this mode is for single ended ,for push pull mode you must connect pin(13) to Vref pin(14)

And for your volt meter with resistor values for VR6 ,R6,VR4,R7AND R10 The voltmeter wont go beyound 19uA even at 30KV output
the error amplifiers in TL494 can take input voltages as high as vcc-0.3v but well about Vref so you may increase your feedback voltage to have better output control


Joined Oct 28, 2019
I'm no expert in high voltage design but a couple of observations.
You may need more than R20 and R19 as feedback resistors. All resistors have a maximum voltage rating (irrespective of wattage rating) its usually around 500V. 100M is a very high resistance. Use many more resistors in series.
Stability vs accuracy is going to be a challenge. The voltage multiplier has a very slow response and this will affect the system speed (to react to output voltage changes). How good is your maths in order to calculate system response?
Is your health insurance paid up? One slip and its off to hospital.