# Energy in forward state IGBT

What is the formula to find Ef for an IGBT? Can I calculate it from datasheet parameters?

E is for energy dissipation and I found the graph from here: https://www.researchgate.net/post/How-can-I-calculate-the-losses-of-an-IGBT-using-datasheet-parameters

• Can you include a link to the page where you got that picture, or if it's from a book, quote the surrounding text? It looks like the plot itself is the power dissipated in the IGBT, and each $E$ in there is the energy dissipated in the IGBT for that phase of its operation. If indeed that's the case, $E_f$ would just be $V_{CE} \cdot I_C$ times the time that the thing is turned fully on. Nov 12, 2021 at 15:58

That's not energy, that is power, and it's simply the multiplication of forward voltage and collector-emitter current.

An IGBT stores only a negligible amount of energy (because the insulated gate acts as a capacitor), and that energy is not usable -- quite the opposite: turning the IGBT on or off requires changing the charge of this capacitor, and the energy required to do so is lost as heat. The datasheet should mention the capacitance, as this is an important performance characteristic; but this is not related to what is labeled "E" in your graph.

• Actually I think the author of that graph means for it to be dissipated power, with each region labeled $E_{\mathrm{whatever}}$ denoting the energy dissipated in that phase of operation. It might be what I would do if I wanted to demonstrate why turn-on and turn-off dissipation is important when calculating overall dissipation, or if I wanted to demonstrate why efficiency goes down with duty cycle. Nov 12, 2021 at 16:01
• Right, the integral area corresponds to the energy lost as heat inside the collector-emitter channel. Nov 12, 2021 at 16:20

That depends on what you mean by "calculate", and how close of an answer you want. In general, a datasheet doesn't go into enough detail, and those energy dissipations depend circuit operation. While it's theoretically humanly possible to do the calculations by hand, in practical terms it takes seconds to do on a simulator what it would take months to do by hand.

Most manufacturers have SPICE models of their parts -- I would start there. If you want to use a part that doesn't have a SPICE model, then your best bet is to find a similar part, and tweak the SPICE model for the differences. This will not be trivial for an IGBT model.

For all that -- you can get pretty close for the $E_f$ -- that's just the $V_{CE}$ of the device when it's on times the current flowing through it. Assuming that what's on the transistor is inductor current, then for a typical circuit it has a starting current and an ending current, and the waveform is triangular. $E_f$ is just the integral of $V_{CE}$ and the collector current.

$E_{on}$ and $E_{off}$ are more difficult, because they're the losses that happen when the IGBT are partially on during the turn on/turn off transisents. I have little experience with IGBTs, but with MOSFETs you can come do a rough estimate by assuming that the $V_{SD}$ varies on a straight line while the drain current stays constant (because it's being driven by an inductor). Just replace $V_{SD}$ with $V_{CE}$, and $i_{D}$ with $i_C$, and you're golden -- except that the turn-on and turn-off times are very heavily dependent on how the gate is driven.

To really know the turn-on and turn-off times you need to take into account the gate drive voltage, the output resistance (or nonlinear current-drive capabilities) of the gate driver, whether you have an RC or resistor-diode-capacitor network between the gate driver and the gate, whether you have any impedance between the emitter and the gate driver's common (either intentional resistance or wiring/PCB resistance).

It's this dealing with the complex interplay of the gate drive and the switching characteristics of the part that drives the community to use simulation. You only get to the point of knowing what the turn-on and turn-off time will actually be through experience, and you only get that experience through working with circuits.