MOSFET Guide for Idiots?

[Dylfish]

Hey Guys,

i was wondering what the best way of learning how a MOSFET amplifies? Ive read the GEOFEX article but I was wondering if there was anything that was nice and simple to explain the principles before i jumped more in-depth.

Cheers

[carboncomp]

This guys got some really easy to understand videos

Check out:

Field Effect Transistors, Part 1 and 2
What is a transistor? How does a transistor work? 1 and two
How does a transistor amplify?



It should help you understand GEOFEX article a little better keep in mind JFET use PN junction to provide high input impedance where as MOSFET use isolated gate to provide high input impedance.

[R.G.]

The simplest way to think of them is that they are transconductance devices. That is, a change in the input VOLTAGE causes a change in the channel CURRENT.

For big power MOSFETs, the transconductance is about one ampere per volt of change on the base. For weeney little TO-92 things, it's often about 1/10 to 1/2 that, or about 100ma to 500ma per volt.

There is an initial gate voltage hump to get over. You have to raise the gate voltage above the source voltage by something like 2-5V before anything at all happens. This is the "threshold voltage". Once you're over that, you get the transconductance value of current per voltage.

In the case of a MOSFET with Ygs (they use "y" for transconductance a lot) of 100ma/V and a 2.5V threshold, nothing happens til the gate is up to +2.5V, then raising it from 2.5 to 3.5V willl let 100ma flow in the drain. Assuming, of course that the external power supply and resistors, etc. will *let* that 100ma flow. But the MOSFET will.

To get voltage gain, we use this neato device that converts current flow into voltage. One of these with a constance of, say, 1000 V per ampere will convert a change of 1ma into a change of one volt. We call these resistors.

The gain of a transconductance stage is the transconductance times the load resistor as long as nothing outside restricts it. A source resistor adds negative feed back, and may lower the gain, but cannot increase it.

There is **always** a parasitic diode from source to drain in today's MOSFETs as a leftover from the manufacgturing process.

[Dylfish]

Thank you guys very much. Its only 7am at the moment so ill have to have a look when ive woken up a bit cheers!

[Dylfish]

Hey Guys,

after 2 days of intensive reading, watching and hair pulling this is whats I've learnt...

Adding Positive voltage to the gate kicks in after getting past the threshold point of the mosfet, allowing current to flow from the drain to the source.

The Voltage between the Gate and the Source is the gate voltage minus the threshold voltage, and from there the YGS will flow per volt. E.G. YGS = 100 | Gate = 3.5v | Threshold Voltage = 2v | Vgs = 1.5v | Current flow from drain to source = 150ma

I can see how it would be used as a switch but I'm still struggling to see how the AC (or DC for that matter) gets amplified and by what factor. Also without getting into hardcore maths.

Can anyone dumb it down for me a little bit more? I've been getting through R.G's booster design text but im still losing myself, but im way too damn stubborn to give up.

any help is more than welcome.

Thanks Guys

[amptramp]

The simplest design of amplifier uses a JFET and a drain resistor.  You have a transconductance given for a JFET which can be as low as 500 micromhos (or microsiemens) to over one mho (or siemen).  Transconductance is the ratio of current change at the drain output to voltage change at the gate-source junction.  If you add a drain resistor, the output voltage is current change through the resistor.  Let's say you have a JFET with 2000 micromhos transconductance or 2 mA drain current change per volt change at the input.  If the drain resistor is 5000 ohms, then one volt change at the input gives you a 2 mA change with the 5K resistor or a change of ten volts across the resistor.  This gives you a gain of 10 for that stage because a 1 volt signal in gives you a 10 volt signal out.  The rest of the design effort is biasing the gate to a voltage where you are in the active region and stabilizing the gain over variations in JFET parameters (which are typically all over the place).

[Dylfish]

hmmm ive been reading exclusively about mosfets so its a tad different from what ive been reading. thanks anyway. I think my confusion stems from setting the source from the gate, how does the signal go against the direction of the current up towards the drain resistor. Maybe im looking at this the entirely wrong way?

[R.G.]

You're looking at it the wrong way - I think.

The gate of a MOSFET is completely isolated from the drain-to-source channel by a layer of very pure glass. The channel is made so that there are conduction carriers available, but not easy to get to, so the thing won't conduct until you do something to the channel.

The gate-to-source voltage is that "do something". Putting a few volts positive (for an N-channel; reverse for a P-channel) on the gate with respect to the source causes an electrical field to be set up. Remember how static electricity can attract dust and hair? In this case, the electrical charge attracts those charge carriers into the conduction channel, and it then conducts.

There is a certain minimum voltage/pressure that is needed to start pulling them over, and that's the threshold voltage. Once you get over that, the harder you pull on charge carriers by gate voltage, the more charge carriers come into the channel and the more current flows.

[defaced]

Someone recommended this book on the forum a year or so ago.  The first 15 pages or so explain how field effect transistors work.  The once you get past that chapter, the book deals with applications and how to select a FET/understanding the device characteristic, and a host of other topics.  I find it pretty easy to read as far as technical text goes.    
Designing with Field-Effect Transistors, Second Edition, Siliconix, Revised by Ed Oxner.  It's worth the 10 bucks + shipping.  

-----

hmmm ive been reading exclusively about mosfets so its a tad different from what ive been reading. thanks anyway. I think my confusion stems from setting the source from the gate, how does the signal go against the direction of the current up towards the drain resistor. Maybe im looking at this the entirely wrong way?

Electron flow vs conventional current flow. Electrons flow from negative to positive.  Conventional current (also called "holes" or positive charge carriers) flow from positive to negative.  It's not so much the electrons are flowing against the direction of current, it's the one terminology is looking at the cars driving down the road while the other is looking at the movement of the spaces between the cars.   You may want to back step a little and learn the fundamental operating principals of vacuum tubes - triodes specifically. They share alot of basic operational concepts with FETs.  

As a tip, you may want to select when you use electron flow or when you use current flow.  On a circuit level, I find it easier to think in terms of current.  On a device level, I find it easier to think in terms of electrons.  

[Seljer]

Curse whoever it was that decided electrons should have negative charge!  I'm pretty sure my semiconductor devices class would have been a week shorter without that convention.

[tca]

If you can find a second hand copy of  Design of V. M. O. S. Circuits: With Experiments (Blacksburg continuing education series) Robert T. Stone and Howard M. Berlin, that is a very good reading. Also take a look at this: http://www.muzique.com/schem/mosfet.htm

Cheers.

[PRR]

> how does the signal go against the direction of the current

Water flowing in a hose. Kink the hose. Water stops flowing. On _both_ sides of the kink.

Or more to the point: get a pump going and hose its output back to its input. Kink the hose. All flow stops.

It really does not matter what the charge polarity is. Semiconductors have both Pos and Neg charges moving. While common audio vacuum tubes have only negative charges, gas tubes use positive ions also. Metallic (wire) conduction is usually explained with loose electrons, but there is NO home experiment to prove that it is the electrons moving; and they move MUCH slower than the audio signal.