General Question About FETs

[Paul Marossy]

I know that FET's have a high input impedance, but what about the output impedance?

In general, do data sheets for transistors/opamps indicate input or output impedance? If they do, I don't recognize it by the terminology that is used... I do see stuff like input capacitance, output capacitance, etc.

[gez]

Output impedance has more to do with the circuit type than the actual device (with the possible exception of source-follower circuits).

[Doug H]

In general, NPN's can source more current than JFETs. In that sense, JFETs are more like pentodes than triodes, Paul.

Doug

[Paul Marossy]

So, you are saying that it depends on what is after the FET? (or opamp)

[Paul Marossy]

Ah, more like a pentode than a triode. That kind of makes sense when you put it that way.

[Jay Doyle]

Paul,

The quick and dirty answer is that FET input impedence is so high that the bias resistor swamps the parallel resistance formula and sets the input impedence. The resistor on the terminal you are taking the signal from, sets the output impedence.

Those are generalizations, but for calculations and first order tuning they work.

In reality, well, input impedence is just a little less than the bias resistor. That one is easy.

The output however is not so straightforward when thought through. Consider the case of a JFET source follower with a 10k resistor on the source. As the signal is going up, the FET is essentially reducing it's resistance/impedence (allowing more current through) to bring the signal toward V+. In this case the output impedence is very low as it is the parallel resistance of the FET (very low) and the 10k resistor. The lower resistance always is the dominant factor in the parallel resistance equation.

BUT as the FET "pushes" the signal toward ground, it has to increase its resistance (let less current through) and the output impedence approaches 10k as the source resistor then becomes the dominant factor in the parallel resistor formula.

This of course is not exactly how a FET works, but in terms of defining the output impedence correctly it is the best guide.

Also, see my "Sweetness" schem in Schematics Two for a way to use this assymetric output impedence as an effect.

Hope this helps,

Jay

[R.G.]

You usually have to distinguish between the output impedance of the device and the circuit you use it in.

MOSFETs, JFETs, and bipolars all have quite high output impedances for the device itself.  That is, beyond a certain low voltage region where things are funny, they are all best described as a constant current source. The input parameter sets a current through the device, and as long as the external parts and power supply will let that current flow, it flows. Much more like a pentode than a triode, as got mentioned.

The parameter you're looking for in data sheets is hoe for bipolars, don't remember for FETs. Probably listed as the reverse, output conductance. In all cases, the output impedance increases as the "flatness" of the characteristic curves increases.

The circuit these devices operate in make a low impedance out of this high impedance in a couple of ways. One is with parallel resistors. The collector resistor on a bipolar is effectively in parallel with the output current, forming a Norton equivalent circuit, with the output impedance being equal to the resistor.

The other easy way is with feedback. Overall negative feedback will lower the output impedance that external devices see by controlling the current source to make the right voltage appear. This can make quite low output impedances out of devices that have high impedances in and of themselves. Opamps and OTAs are good examples, although the voltage feedback connection of a single bipolar transistor is simpler to see.

[Paul Marossy]

Thanks RG. That really tells me what I wanted to know. I do remember seeing specs for output conductance on spec sheets before.