Need help designing a single shot igbt driver

[PoyntExH]

Hi all. I am using a 1200v 400a igbt to drive a large pulsed soleniod with a single pulse with a pulse width of about 3 milliseconds. When in use, the solenoid is only pulsed once every few minutes. Current in the pulse is usually about 400A.

What I want is to make or use a simple driver with enough output current that can switch it on fast enough to minimize the switching losses that I know I must be having with the setup I'm using.

Here is the datasheet for my IGBT:

http://www.newrock.com.cn/PDF/IGBTMD/EUPEC/DN2SERIES/BSM400GA120DN2.pdf

I do not know much about mosfet and IGBT gate driving, other than the basics. I know the driver needs to have enough available current in order to charge up the gate capacitance quickly enough to switch it on without significant losses. I know that this current is only needed for a few nano seconds or so, and I know that the current required is based on the drive voltage and the gate resistor. Since my gate resistor is 2.7 ohms, if I want to drive it with 15 volts I'll need around 5.5 amps to charge up the 26 nF gate.

Right now I am firing the gate with 17 volts from an lm7555 timer in one-shot configuration. (this is the improved cmos version of the classic lm555 timer) The problem with the 7555 is that it can only output 100mA of current, so I know that I am causing a much slower turn-on than what the igbt would normally like. (I can also see it on my oscilloscope) I was thinking of designing some sort of output buffer to put on the output of that 7555, maybe using a fast high current op-amp buffer, (if I can actually find one) though I know this is probably isn't a very good way to do this.

I know that there are a multitude of igbt gate drive IC's out there, but I know next to nothing about them and I've found no information on how to set up and use one for a one-shot firing of the igbt.

Thanks,
-Rob

 

[LiteYear]

It might help to think of the IGBT passing through the linear region before saturating. Your goal is to limit the time spent in that linear region, so perhaps think of the IGBT as acting like a BJT amplifier - the greater the base current the greater the pulse current. How do you increase current through a BJT amplifier? You cascade two of them (in a Darlington pair arrangement).

In other words, you could try simply using the output of your LM7555 to drive the base of a BJT, and have the output of the BJT drive your IGBT. This is one of the features built into gate driver ICs.

 

[PoyntExH]

Yeah I was doing some more reading today and I found some application note talking about how the linear region of the IGBT is extremely lossy, and we want to avoid it at all costs. It makes sense.

Would it be better to try to have the LM7555 driving a small mosfet like a IRF510 to buffer the gate? I'm thinking I might have a hard time finding a BJT that can pass 5.5A and still have a high enough beta to become saturated by the 7555's 100mA output. Also would this BJT have a rise time faster than 100ns? I guess if I’m worried about the beta I could also try a darlington pair for my 7555 buffer, but I’m still worried about it not being fast enough in that case.

-Rob

 

[LiteYear]

If that's what your calculations show, then sure, use an IRF510. You're talking serious amplification, so nothing wrong with using serious transistors.

 

[PoyntExH]

haha. Well it does so happen that I have a few lying around here and the irf510's gate capacitance looks low enough that I should be able to switch it on fast enough with the 7555 so unless anyone else has a better idea, when I get back to it I'll give it a go with the irf510.