common-source vs. common-gate vs. common-drain.... RG?

[Dragonfly]

I was was just wondering about the advantages and disadvantages of each setup.

From what I can tell, the common source setup provides for more gain but is less stabile. Also, the output is 180 degrees out of phase. Are there any advantages or disadvantages I'm missing ? This seems to be the most commonly used setup with regards to effects pedals, but is it really the best ?

Common gate seems to be far more stabile, but provides for less gain. The output would be "in phase" with the input.  Would, with regards to "real world" usage in a mosfet based preamp, this possibly be the "best" way to setup a circuit? I would think (there I go thinking again...) that even though this circuit inherently has a lower gain, that with a high gain device like a mosfet, that you could still get plenty of gain for most uses and have a better shot at stabilizing the mosfet. Am I wrong in this ?

Common drain seems to be similar to "common source", but the output would be "in phase"...correct?  Any other advantages or disadvantages ?

Also, since the gate-ground resistor provides bias for the mosfet, wouldn't it make more sense to use a trimpot in this position instead of the power-drain resistor?

Sorry for the basic questions...I'm on the "learn as you go" plan

Thanks,
  Andy

[R.G.]

No problem.

Common drain is also called a source follower. The drain is tied to the power supply or bypassed to ground, so no AC signal appears there.  there is 100% negative feedback at the source, so all of the device gain is going towards making the output be equal to the input. It's most stable, most predictable. Input impedance is high, even higher than the raw impedance of the gate, which is high indeed. Output impedance is low. It makes a good unity gain buffer, and that's exactly what it is used for.

Common gate is used to get high frequency response. There is an inherent capacitor from gate to drain. That capacitor feeds the drain voltage back to the input as frequency increases, so it limits high frequency response. By shoving signal into the source, the grounded gate doesn't much care if the drain is  moving around, it's at ground and the feedback is nullified. The source is a low impedance point, so the drain-source capacitance has a far smaller effect, and you can get gain at much higher frequencies. Of course, you really, really have to want to do this because the low impedance at the source is the devil to drive. That fact generally resigns common gate (and -base and -grid) circuits to RF devices. And exception is the cascode connection. In this one, an input device is driven by the input signal at its gate. Its drain feeds the source of second device which has its gate held at some DC voltage (and hence AC grounded). The output comes from the drain of the second device. This is a common-source stage feeding a common-gate stage, and you get high voltage gain and high frequency response. But you really gotta want to go to the trouble to do this for some specific reason.

Common source is analogous to common emitter or common cathode. The input signal goes in the gate. Any impedance in the source lead causes a feedback signal in series with the input (which is really the gate to source voltage here) and the signal comes out the drain. This series feeback sets the effective gain. CS is the usual compromise, as you get high input impedance (though less than CD) and moderate gain (though less than CG) at the same time.

CS, use for gain stages. CD (source followers) use for buffers. CG, don't use until you already know that that's what you need.

[Dragonfly]

thanks RG...i was off searching the archives when you posted this, and i found a bit of this info out...but not nearly all of it, and not explained quite as nicely as you did here. its nice to have the info all in one place

so...with regards to the second "hidden" question, even in a common source circuit, would it not make more sense to use a trimpot for the "gate to ground" resistor than the "power to drain" resistor ? it seems to me that the gate to ground resistor is more instrumental in establishing bias than the drain resistor.

thanks again,
 andy

[R.G.]

so...with regards to the second "hidden" question, even in a common source circuit, would it not make more sense to use a trimpot for the "gate to ground" resistor than the "power to drain" resistor ? it seems to me that the gate to ground resistor is more instrumental in establishing bias than the drain resistor.

It depends on the device.

I was giving you answers for both MOSFET and JFET. Your hidden question is about JFETs, most likely.

For MOSFET biasing, see "Designing and Building Your Own MOSFET Boosters" at GEO (10-06-01), which covers the gate bias trimpot idea for MOSFETs.

For JFETs, the common way of using a single resistor to ground for setting bias by the current flowing through the source resistor, you can't use a trimpot for the gate resistor and have it work. The resistor carries essentially no current, so it has no voltage drop, and varying the resistor does nothing to the bias of the JFET except change the loading on the input signal.

[Dragonfly]

so...with regards to the second "hidden" question, even in a common source circuit, would it not make more sense to use a trimpot for the "gate to ground" resistor than the "power to drain" resistor ? it seems to me that the gate to ground resistor is more instrumental in establishing bias than the drain resistor.

It depends on the device.

I was giving you answers for both MOSFET and JFET. Your hidden question is about JFETs, most likely.

For MOSFET biasing, see "Designing and Building Your Own MOSFET Boosters" at GEO (10-06-01), which covers the gate bias trimpot idea for MOSFETs.

For JFETs, the common way of using a single resistor to ground for setting bias by the current flowing through the source resistor, you can't use a trimpot for the gate resistor and have it work. The resistor carries essentially no current, so it has no voltage drop, and varying the resistor does nothing to the bias of the JFET except change the loading on the input signal.

thanks again...i'd been using the  trimpot for the "drain" resistor, and that seemed to work fine...i had just been wondering about the other way i've gotta stop looking at rangemasters and in books (just kidding)

thanks again....as always, your help is greatly appreciated !

andy