I'm building a solid state tesla coil, which requires me to switch very high voltages at high speed to achieve resonance. The problem: I can't find any IGBTs that can switch the voltage I need.

full bridge driver

In theory, I will have four IGBTs in a full-bridge driver configuration. I plan on using the IRG7PK35UD1PbF (datasheet), which has a max voltage of 1400 volts. However, I need to drive the primary coil at 6500 volts.

How can switch a load this large without getting obscenely expensive IGBTs? Would it be possible to put them in series/parallel to do this?

  • 1
    \$\begingroup\$ If you need to drive the primary from 6500V, you might be better connecting it to the secondary of another transformer : step up 6.5:1 from 1000V, where you CAN get IGBTs. (Or 10:1 from 700V which is easier to generate) \$\endgroup\$
    – user16324
    Commented Feb 28, 2015 at 17:42
  • 1
    \$\begingroup\$ You should do more research into the various different classes of SSTC design. They don't generally involve switching kilovolts with transistors. \$\endgroup\$ Commented Feb 28, 2015 at 18:53

1 Answer 1


Yes with ... considerations.

Can you stack them in series? Yes especially in theory. In theory the devices in a series stack would have the same leakage (in fact no leakage) and thus share the voltage perfectly. In theory they would all switch at the same rate (in fact instantaneous switching ) and in theory the devices in the stack would receive the signal at the same time

In practice however... its a different story

Aspects that need to be considered and mitigated

  1. Each device does not share the same voltage. This can be improved via the use of a high-impedance resistor across each Collector-Emitter. The resistance needs to be high (Meg's). This will produce a voltage divider to help mitigate the differences in leakage and thus standoff voltage sharing concerns.

  2. Each device will have their own switching characteristics and as such one could start switching before the others and thus the entire voltage appears across one device == BAD. An R-D-C snubber across each collector-emitter can help produce a repeatable switching characteristic

  3. Even if you had ideal IGBT's, the source of the signal will not be. You have a microcontroller somewhere, two independent isolators (opto's) and different propagation through each gate-drive. To mitigate this the GATE lead of each IGBT can be magnetically coupled via a CM choke to improve the gate current flowing into and out of each IGBT at the same time and rate

Have a look at this whitepaper on the topic



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.