IGBT selection for Large DRSSTC

[AeraCura_]

Hi all,

I'm gathering components to build my first DRSSTC.
My design specifications are:

Resonant Frequency - 40kHz
Height - 2m tall
Power Input - Max 30A 220VAC Single Phase
Driver - UD2.7C

My question is with IGBT selection. I would like to use a single full-bridge IGBT brick. I've followed Kaizer's IGBT selection guide https://kaizerpowerelectronics.dk/tesla-coils/drsstc-design-guide/igbts/, and populated the following spreadsheet with relevent specifications.



From what I can gather, the IGBT i'm considering seems to be functionally similar to the CM300 brick. So my question is this,

Why does it seem that no one is using a full-bridge module? is this a simple value calculation where cheeper bricks offer equivelent performance?
Does the APTGLQ200H120G seem to be a viable brick for the design specifications i've outlined?
Why can't i locate the "Collector-Emitter Breakdown Voltage" for this brick, is it commonly referred to by another name?

Thanks in advance.

[AstRii]

Collector-Emitter breakdown voltage is the V_CES parameter. Any voltage above this can and will destroy the IGBT.

From a datasheet i found on this transistor, i would not push it as hard as a CM300 brick.

It's 12.5nF gate capacitance (4x lower than CM300) is quite small, that indicates not as wide transistor channel which makes it less robust.
Also it's only a single brick occupied with 4 IGBTs, which means faster heating up.

It's been confirmed that you can push CM300 bricks up to 1200A+. For this one i would likely stick with the datasheet values and would not try to overdrive it too much (maybe 900Apk).

I think the main reason why nobody uses single brick full bridge modules is the simple fact that they are rare compared to single IGBT or half-bridge modules.

Good luck with your coil

[acobaugh]

Also to consider is heat dissipation of the H-bridge output. With a single full bridge brick, you have two terminals on the same brick/die. With two half bridge bricks, you're spreading out that heat even further. Taken to the max, you have 4 separate bricks. In my next CM600 full bridge, the bus bars that form the H-bridge output will be water cooled to help with heat dissipation. Some builders go even further and parallel multiple single igbt modules, or choose bricks that already have multiple output terminals (eg. FZ1200).

Also to consider is price, availability, and proven designs. My DR82 coil pushes 10+ foot arcs running 1100A peak with a pair of Mitsubishi CM300's switching at ~38kHz, not too different from yours.

If you expect huge arcs, I would reconsider the overall height of your system. 2m tall might not put your topload high enough off the ground and far enough away from the primary to avoid constant primary/strike rail/ground strikes. My big coil stands ~9ft tall at the breakout point (2.7m) and still hits the ground pretty frequently, with the random strike rail hit, but that's also with a pretty wide toroid (45") and a tall secondary.

That being said, if this isn't your first DR coil, don't hesitate to experiment. But if you're looking for a reliable performer, it's hard to go wrong with a proven design and make improvements and changes from there.

[Mads Barnkob]

As AstRii points out, the gate capacitance is a sign of how large a die the IGBT has. These multi-die full-bridge modules (both single phase with 4 switches and three phased with 6 switches) usually does not have more room inside them, than to accommodate what is essentially TO-247 sized dies.

The only advantage of such a module over single TO-247 devices would be the direct bonding to the copper base plate through a insulating layer. So the junction to case thermal resistance must be much better than a TO-247 package.

As it has been seen with something like the old 40N60 SOT-227B mini block IGBTs vs. the newer 60N60, the old one rated for less current could take almost twice the peak current than of the new version. While the datasheet might say the new parts are better, the old parts was often over-engineered and had much large safety margins, virtually making them better for out purposes.

With a 40 kHz resonant frequency, you are not really limited by switching speeds and you should aim for a solid beast of a large brick full-bridge to push such a large coil to its limits