Super Transistor?

I'm trying to come up with little group of components that can be used in place of a single bipolar transistor for added performance ? sort of a "super transistor" with high current gain, low input capacitance, low output impedance, etc. Flipping through my trusty Horowitz & Hill I found two techniques that look promising. First is a FET/bipolar combination in a Darlington type configuration (section 6.07). They discuss it as means of lowering a FET's output impedance, but I think it can just as easily be seen as a way to raise a bipolar's input impedance. Second is a FET cascode that exhibits very low dynamic input current (section 6.16).



Would a combination of these two techniques give me the super transistor I'm looking for? Am I wrong in thinking that this circuit can be used exactly like a single bipolar transistor with a very high beta and a gate-emitter voltage of about 0.6V?

Maybe some other combination would work better?

Thanks!
Thomas

Would a combination of these two techniques give me the super transistor I'm looking for? Am I wrong in thinking that this circuit can be used exactly like a single bipolar transistor with a very high beta and a gate-emitter voltage of about 0.6V?

Click to expand...

The topologies you posted won't be an ready-to-go drop-in replacement for a single BJT since the operating point will be different (other DC-'situation'), but most circuits will be adaptable to it, given a not too low supply voltage.

Question of course is whether circuits will benefit. 'More' & 'universal' will not necessarily always mean 'improvements'. The topology-merging you did looks straightforward though, could be of use.

A generally better compound pair for many purposes would place a PNP with collector to source, emitter as the new composite "drain", and the drain driving the base with a base-emitter R to determine drain current (for example 650 ohms for about 1mA for the FET). Again, this cannot be substituted will-nilly---life isn't that easy---and performance improvements will require circuit understanding and most often adjustment of other components.

There is also an interesting stage using three bipolars of the same sex that Tektronix used for a while---it's discussed in one of Williams' compendia of articles by analog geniuses, and I believe the author was the great vertical amp designer John Addis. I'll find the ref. if you like. It behaves (with a few other parts) like a supertransistor with essentially 0 Vbe.

Brad

> a gate-emitter voltage of about 0.6V?

Only if you can trick the FET into running at exactly zero V Gate-Source.

That is supposed to happen if you replace the base-emitter resistor with a matched FET with gate shorted to source. I had a Heath o'scope which used this technique. It was a $20 scope and I was not looking to make it "better", but that FET buffer gave me so much trouble in offset and signal handling that I ripped it out and put in an IC.

> a "super transistor" with high current gain, low input capacitance, low output impedance, etc

If there were a universally better drop-in, somebody would be making it.

There are some "close" supers. The Darlington connection is old as the hills and has been sold almost as long. I'm sure you see why it is not a zero-thinking drop-in for every circuit, but often the mods are simple. BTW, in 4-leg form it can be very low Miller Capacitance (though we sometimes ADD Miller Cap for stability). A Diff-Pair is a more radical change but gives "zero" offset voltage and cancels 2nd harmonic: this has been a standard IC for 30 years. More recently cascodes and current-mirrors are available as small parts. And the (MOS)FET-BJT pair is sold as a big-power transistor as "IGBT".

FWIW: the $2 op-amp IS the "universal transistor replacement". It isn't drop-in, but gives better results.

Up to this point I've been pretty much an Op-amp guy. I'm just starting to get into discreet designs. And I have a very fuzzy understanding of how FET circuits actually work, so I appreciate the guidance.

Brad, yes I'd love to see that design!

Btw, what would happen if I substituted MOSFETs for the JFETs? Would that work more like a BJT?

Thanks guys!

Look up IGBT it is kind of a mosfet gate BJT collector emitter.

Say, PRR
I've been wondering about IGBTs. Does anyone use them in power amp outputs? Would they be a good route to go? Pros/cons?

Thanks

IGBTs being a mosfet gate have lower sustained drive need. IGBT can have a large gate input cap that need alot of push or pull current if you are looking for fast turn on and off times.

IGBT at high voltages and current can have less loss because it is more the saturation voltage of the BJT part.

Power BJTs with there low Hfe need lots of current to drive them at high currents

Power MOSFETs have high gate cap and a rating called the RDS on resitance Drain to Source. Lower voltage mosfets can be as low as .025ohm RDSon Higher voltage ones tend to have higher RDSon.

You need to know what power output you want, How linear the power transfer to the speaker, every 1/2 the ohm load should be X2 the power before. Then you need to calculate the driver stage for the output stage.

Different power supply volatges and power wanted can give a different "better" device for the output stage. All power devices need good drivers.