fast bipolar power transistors

[wjgregg]

I'm looking for something like a TIP48 power transistor, but a bit faster. The TIP family runs about 4 microseconds from rise to fall, on a 10 nanosecond trigger, and I'm hoping to find something that will get me down to 1 microsecond (or less), at 200 to 300 VDC. (I'm throwing single pulses at 1 MHz PZT's, and they tend to stretch the times out a bit.) Anybody know of something faster?
Cheers, Bill

 

[Nigel Goodwin]

Have you tried looking at switching transistors rather than power transistors?, such as those used in switch-mode PSU's and line output stages. You don't really give much information?, but you could try something like BU508 or BUT11 - don't forget you have to provide suitable base drive circuitry to switch fast - check SM PSU examples!.

 

[Papabravo]

Transistors can be optimized for any number of different applications. This is done by changing the sizes, areas, and geometries of the various elements on the silicon. So you can have high speed, or you can have power handling capability, but you can't have both simultaneously. The geometry that provides the one, prevents you from having the other.

 

[Hero999]

It's called a power MOSFET!

 

[Nigel Goodwin]

It's called a power MOSFET!

Not when the request if for a bipolar device!

Certainly its still VERY uncommon to use MOSFET's in switch-mode PSU's, and I don't think I've ever seen one used as a line output device?. The increased cost is probably the reason, and certainly where they are used in SMPSU's they don't have much of a reputation for reliability, certainly no better than bipolar, and probably slightly worse?.

 

[Papabravo]

It's called a power MOSFET!

MOSFETS can be slow as molassas in January when you consider the need to charge and dischrge the gate capacitance.

 

[wjgregg]

Thanks everybody. Always good to have input. I tried switching tr's (like the BU508, and BUT11) before, Nigel, but the TIP family beat them (slightly) in total rise/fall. Power MOSFET's seemed like a good idea too, Hero, but the ones I played with before (like MTD6N20E) ended up slower than the TIP 48. The small signal ones are fast, but they don't fit my application. It's that inverse relationship Papa mentioned. I've also used SCR's before, like the old C231M, and newer ones, but they're slow, and they generate a lot of RF in my particular circuit. Microsemi used to make the GB201/GB301 series SCR's that were in the 200-400 volt range with Tr and Tq in the low nanosecond range; but they're discontinued according to their engineers. Same with the 3902GRF apparently. So I've settled on the TIP48 for now. I DID improve the trigger circuit a bit over the last 2 days, Nigel, with good results. I squared it up better, gave it a bit more amperage, and put in better pulse width adjustment.
Cheers, Bill

 

[Nigel Goodwin]

What are you using to drive the base of the transistor, post the circuit here, bear in mind it's a LOT harder to turn a transistor OFF quickly than it is to turn it ON.

 

[wjgregg]

Hi Nigel. It's a simple trigger. Just an LM555 timer with a 6-position resistor selector across pins 6 and 7 to provide pulse frequencies from 60 pps to 2 pps. These pulses go to a 74HCT123 for pulse width control, using a 10K pot from 10 usec down to nothing. Nothing new.
Cheers, Bill

 

[Nigel Goodwin]

Then you're not helping yourself, your problem probably isn't the transistor, it's the poor drive arrangements. You need to provide fairly high drive current to turn it ON fast, and a higher negative drive VOLTAGE to turn it OFF fast enough.

A common method (for SMPSU's) is to use a capacitor to feed the base, with a zener diode across it, fed from a fairly low impedance source. The capacitor (100uF or so) charges up with the base current through it, with it's terminal voltage being limited by the zener diode (something like 5.6V). When the drive pin goes LOW, the capacitor provides a 5.6V negative going pulse to the base, this turns it OFF a LOT faster than a normal drive circuit.

Something like this:

 

[wjgregg]

Thanks. I was assuming a problem at the other end, because I had slow turn-off when the circuit was under load from the PZT's, but rapid turn-off when I substituted a resistive load. The turn-on time was always fast. I'll try this next week.

 

[wjgregg]

Nigel, I was looking at the datasheets on the BU806. It's down at 200V but is a lot faster than the TIP 48. The original SCR circuit I "inherited" was operating off a 440 VAC supply controlled by a powerstat. The thin 1-MHz PZT's were firing safely at a 63% powerstat setting, but would not fire below 45%. The new TIP 48 circuit gives the same PZT response at only 140 volts, and they fire all the way down to a zero setting (.6 volts). If my thicker low frequency discs end up with the same power requirements, I could probably use the BU806.