I am busy trying to repair a IGBT H-Bridge inverter that forms part of an AC spot welder. The problem I am having is that it works perfectly with a resistive load, but when I connect the step down transformer to the output the output turns on again without a gate signal after turning off.

This is the output measured with the transformer connected (No load on the secondary) I have circled in blue the unexpected "turn on" of the IGBT. 1

This is the output with only resistive load connected. 2

Here is the power circuit: The bit in blue is used for the current limit circuit. The resistor RL (circled in red) is the resistive load I referred to above. It seems to be there to ensure that the H-Bridge doesn't go open circuit as the transformer is pluggable in the field. The right half of the bridge circuit is identical to the left half. The IGBT's used are INFINEON FF300 range 300A 1200V 62mm modules. AH and AL form one module as do BH and BL.


And finally, here is the firing circuit which is repeated for each IGBT gate. This circuit is identical to other similar systems that I have seen in the industry with only the driver chip changing. (Currently Toshiba TLP250)

4 This system has been operating like this for a number of years, but is prone to failure every couple of months. Generally the system will operate 18hr per day.

I have not shown the capacitor bank for the DC bus, but it does exist and consists of both electrolytic capacitors as well as RC circuit and is directly across the 320V DC bus that is supplying the H-Bridge. The firing circuit is microcontroller based.

I hope I have been specific enough in the information provided and would appreciate any assistance.

My suspicion is that the reverse EMF produced by the transformer when turning off is somehow causing the IGBT to turn on again, but I cannot prove this. I have also tried different gate resistors and external freewheel diodes over the IGBT Collector and Emitter.

Thank you in advance.

  • 2
    \$\begingroup\$ Can you be sure that it isn't back-emf from the transformer causing what you see. \$\endgroup\$
    – Andy aka
    Commented Jul 6, 2023 at 11:59
  • \$\begingroup\$ I suspect it is, do you mean that the IGBT is not actually turning on, but I am just measuring something else? \$\endgroup\$ Commented Jul 6, 2023 at 13:31
  • \$\begingroup\$ I've formalized an answer. \$\endgroup\$
    – Andy aka
    Commented Jul 6, 2023 at 14:32

2 Answers 2


do you mean that the IGBT is not actually turning on, but I am just measuring something else?

Transformer primary leakage inductance

Every transformer primary has leakage inductance and, every time you drive a transformer primary with nothing connected to the secondary, that leakage inductance effectively becomes the full primary inductance. In other words there is always something to cause a back emf when you disconnect the primary from a voltage supply.


Whether it's MOSFETs or IGBTs, all four devices will disconnect every cycle of PWM and, when this happens, the leakage inductance does what it always does and finds a conductive path to flow current in. That current doesn't last long but, it creates a back-emf.


The diodes in parallel with each device in the H-bridge will conduct the leakage inductor current to the supply rails and, the back emf generated is approximately the supply rail.

This is common.

  • \$\begingroup\$ Thank you Andy, you have been a great help. \$\endgroup\$ Commented Jul 6, 2023 at 16:49

Andy made valid points, however, I respectfully disagree. Firstly, when the secondary of the transformer is OC (open circuit), the transformer would behave like a simple inductor. When a positive voltage is applied, the current would increase in the positive direction, and when the voltage is disconnected, the induced voltage would be in the opposite direction. This is due to Lenz's law, which states that the induced voltage will oppose the change in current. Another issue I noticed and am confident about is your driver schematics. Your circuit uses 0 voltage across the IGBT gate to turn it off, which is not a reliable method. To turn it off, a negative voltage like -5 must be applied across the gate, which can be achieved by changing the drivers. If you want to ensure that the IGBT is not turning on in the mentioned areas, you can monitor the gate-emitter voltage. If the gate voltage remains at 0, the IGBT is definitely OFF.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.