Trigger order of the IGBTs

[PEDRO HENRIQUE]

Hello, I want to do a DRSSTC with a full bridge of IGBTs, but I ddon't know the trigger order of IGBTs, I know that have a phase where essentially we turn on Q4 and Q1, and after we turn off Q4 and Q1 and turn on the Q3 and Q2, but I dont know if we have to do that in other order.
Someone can tell me the right order to turn on the IGBTs?

Use the image to understand the text:
https://lh3.googleusercontent.com/proxy/78CZP5tAZFLAQ6Y0wFEPNGoW2ychi8gBEOOgniqwBk9mDqXOALIYoXK86dNGX0v8OG5IFscOn7wB-Gesp7FTtC9-nQPt

Sorry if have something wrong in the text, it's because I'm Brazilian and I don't master English completely.

Thanks.

[AeraCura_]

Your explantion of how Q1-Q4 are switched is correct, with the exception that the switching is done simultaneously. I'll add that this is done from a GDT (Gate Drive Transformer) with the wires reversed for one pair such that one half of the h-bridge turns off when the other turns on. Conveniently Mads DRSSTC 1 schematic uses the same naming convention so you can see how they're wired to the GDT.



A GDT is used to isolate the controller from the high voltages present on the DC Bus as well as allowing the winding ratios to be customized.

[PEDRO HENRIQUE]

So if I just switch the IGBTs like I said, it will work perfectly?

I have this question because someone said to me, that the right order is:

time 1: Q1 and Q4 are on.
time 2: just Q1 are on.
time 3: Q1 and Q3 are on.
time 4: just Q 3 are on.
time 5: Dead time
time 6: Q3 and Q 2 are on.
time 7: just Q2 are on.
time 8: Q4 and Q2 are on.
time 9: just Q4 are one
time 10: Dead time.

I think it's wrong, but I'm not sure.

[AeraCura_]

I can only assume you would switch the h-bridge in that manner to control for the asymmetry between on and off times common in semiconductors.

To be honest this is something i've never considered, but it does bring up an interesting point.
If for example the off-time for a particular semiconductor is 100ns longer than its on-time than the h-bridge will have a 100ns short. My intuition says that 100ns isn't long enough to create any problems.

Are you considering creating your own driver circuit with such precise timing?

[PEDRO HENRIQUE]

No, I'new in DRSSTC, and I want to learn how to build one.
In other forum some person said to me that these times that I sent to you is the correct for, and if I switch the IGBTs otherwise they will explode and I don't want that my IGBTs explode.
I just want learn what is the correct way to do that.
Please explain the correct for to me.

[TMaxElectronics]

If for example the off-time for a particular semiconductor is 100ns longer than its on-time than the h-bridge will have a 100ns short.

That is why there is usually a dead time control in IGBT drivers. It delays the switch-on by the time needed for the other one to switch off to prevent a shoot through. That is the reason why you sometimes see a reverse diode across gate resistors, as that makes gate discharging faster than charging, thus adding a little more margin to the switching times.

And 100ns can potentially cause some problems already:
If we assume a bus voltage of 500V and a bus inductance of 1uH (which I would consider to be very large) you would already have 50A short circuit current flowing (assuming the IGBTs actually switched with a 100ns offset, without rise time etc.). That would only get worse with better bus layouts that have a lower inductance, and you could easily get >100A in those 100ns. The shock heating that causes could kill IGBTs.

[PEDRO HENRIQUE]

So, if I take 4 IGBTs in this positions:
Q1   Q3
Q2   Q4
And the IGBTs have 100ns of turn-on time and 150ns of turn-off time.
And I turn on Q1 and Q4.
After I start turn-off time off Q1 and Q4, now I wait 50ns and start the turn-on time of Q3 and Q2, after start turn-off time off Q3 and Q2, now I wait 50ns and start the turn-on time of Q1 and Q4 and I repeat it forever, it will be ok, or have something wrong?

And what is rise time?