is this the maximum transistor gain possible?

[ibanezts808]

I've pretty much got my mind wrapped around biasing... I thought.  But then I found out there are a ton of configurations for the various things you want out of a transistor: Gain, Battery Life, etc...  So what I'm asking is, if it is gain I want, is the configuration below basically the max possible gain?  I'm sure there is a special way that will make the transistor do cartwheels and generate 5 billion hfe, but for beginning purposes, this is it right? 

Also I've been having trouble understanding this part.  Where does the signal come out of?  When I read about transistors I was under the assumption that the amplified signal came out of the emitter, but then I started building circuits, then it was coming from the collector(I think) then I was reading about signal coming from the collector and emitter, in some type of envelope effect.  I know this is probably a really general description.

I ask these questions assuming it's an NPN transistor(which I just realized I left the arrow off of the pic, sorry).

[R.G.]

There are some things you're missing.

An NPN transistor is like a pair of diodes. The collector-base diode is reverse biased. To make that reverse biased diode conduct, you put current into the base-emitter junction. The amount of current you inject into the base lets a much larger current flow through the collector base - and then on through the emitter.

Hfe is the name we apply to the current gain. There is really no current "gain", it's just that a little current in the base lets a bigger one pass through the collector to the emitter.

But we want voltage gain usually, not current gain. In the picture you have shown, if you put a small wiggling signal *current* into the base, some of it gets eaten by the
220K/220K biasing resistors, and some of it goes into the base. Actually, almost all of it goes into the base because the base looks like a much lower resistance than the 220Ks.

So the signal current goes in the base, and that lets more/less current flow through the collector. The collector current plus the base current flows in the emitter.

If everything else was right (it's not, but we'll get there) then hfe times the base current signal passes through the collector and hfe+1 times the base current comes out of the emitter.

What's the voltage gain? It's zero. As shown, there are no resistors in either the collector or emitter path, so there is no signal voltage developed.

If you put a resistor between the collector and the emitter, then there would be a signal voltage developed, equal to the current times the resistance. At the emitter the signal voltage developed would be hfe+1 times the emitter resistor, and at the collector, it would be hfe times the collector resistor. This is just Ohm's law of V=I*R applied to the currents.

So what's the gain now? Easy - it's Rc/Re. That's because the same current (OK, almost the same current, hfe*ib versus (hfe+1)*ib) flows in both resistors, so the voltage across them is proportion to the resistors, or Vc/Ve = Rc/Re.

The emtter voltage Ve must be the base voltage minus one diode drop, because otherwise the device doesn't work. So whatever voltage you set the base to, the emitter will be one diode drop lower. If that's impossible for external reasons, then the transistor will either not amplify at all, or burn up trying to make it true.

To bias a transistor with an emitter resistor, YOU MUST SET THE BASE VOLTAGE TO BE THE VOLTAGE YOU WANT THE EMITTER TO BE PLUS ONE DIODE DROP with your biasing resistors. So you pick the current you want the collector/emitter to run at, pick the gain you want by the ratio of Rc and Re, then you pick the biasing resistors to put the base at the right votlage to make the selections you made for collector and emitter come true.

But that gives you a gain of Rc/Re. How do we get more gain?

That's easy, and it takes two steps.

First, make Re very small. We either short Re to ground or short it by AC only with a capacitor. The cap trick lets us put DC bias points wherever we want them, but still increase AC gain. When we do that, we do NOT find that the gain is infiinite because Re is now zero. There is an internal "Re" in the base emitter junction that is about equal to 26mV/Ie that is always there. This is called the Shockley resistance and it's generally small, 10 to 100 ohms. So the gain gets higher when we bypass the emitter. To make it even higher yet, we make the collector resistor as big as we can.

If we are tied to using a resistor, we use the largest power supply voltage we can and also the largest resistor we can and still keep the transistor biased with an appropriate trickle of current. If we truly want big gain, we have to use a current source load instead of Rc. Current sources can have effective resistances in the megohms, so the gain gets very big.

Which brings up another problem. With a high collector load, whatever gets the highly amplified signal voltage must not load down the collector, as that would itself limit gain. So we have to use an active buffer - another transistor at least - on the collector to keep the gain we worked so hard to get.

Signals being put into the base come out BOTH the emitter and collector. We can alternatively hold the base very still and wiggle the emitter. In that case, the signal comes out the collector.

[gaussmarkov]

thanks R.G.  very nice summary.