transistors: so confusing!!

I thought this would be the time I finally understood it, but after reading John's answer, and my head exploding, I think I am going to give up again for another year. (Not a bad answer I just couldn't understand it.)

My feeble understanding of a transistor is like this:

Take a wire, cut it in half and leave a small gap. Current won't flow across this gap. In the magical gap where you cut it, insert another wire. As you add voltage to this new wire, it makes the gap between the old wires conductive. It's a tiny relay?

The other version is a "normally on" relay. Current does flow acrosst he gap. Add voltage to the extra wire and the gap becomes non conductive.

I think in a regular gap there is always some current leaking between the extra wire and the ground, but in a MOSFET there isn't?

I have no idea the actual why of it.

(snip)

Please, for my benefit, if not your own, post a copy of my explanation, up to the point where your head exploded, so I can take a crack at improving the explanation. By the way, the part at the end about the sidewalk, the curb and the slope down to the street was inspired by the famous photo of the cross section of that paving being explained by William Shockley:

More paving cross sections here:

I think a good explanation of terms is the first area I need to understand. I don't mean just definitions, but the whole concept. (Analogy: More than a piston goes up&down and a crankshaft spins, but that a piston is pushed downward with great force from a controlled explosion and this linear force is converted into a spinning force which is also used to reset the system...)

I've read about PN junctions before and tried to grasp them (diodes). I've read about "holes" and electrons but these things don't really mean anything to me.

I'm not certain that I understand why the emitter, base, and other wire are called those names. I think I always confuse the two. The digrams for transistors as used in circuit layouts don't make sense to me. The arrows going in or out, etc. In those diagrams it isn't intuitive for me which is the emitter, which is the base, etc.

I wonder if understanding a vacuum tube would help, but I've not quite got that one either I don't think.

I've read over this stuff maybe half a dozen times over the past 10 years and at times thought I understood it, but I decide later that I do not.

What does reverse biased mean? What is a doped electron? What does it mean that holes find each other? Holes are pushed? I thought they were a stationary place for electrons to gather? Are these made of silicon or what?

I think the general lack of understanding is why the behavior at a different temperature is further beyond me.

Okay, my explanation was not tailored for your particular mind. I'll try again.

(snip)

Then the rest of your questions must be put on hold till we get past basic semiconductor physics. Maybe we will get back to the transistor, later.

(a little ramble on solid state electronics)

A solid material is conductive only if there are movable charge carriers in the material. In metals, these movable charge carriers are electrons (usually, one per atom) that do not take part in holding the material together as a crystalline form (electrons that are shared between two neighbors, and locked into that bond) nor are buried in the electron cloud of each atom, (and are not shared with any other atom) but are shared fairly equally with several, neighboring atoms. The total negative charge of this sea of shared electrons is exactly balanced by an equal number of protons in the nuclei of the atoms sharing these electrons. If a very slight electric voltage is applied across a chunk of such material, the whole sea of conduction electrons drifts toward the more positive end of this voltage, with electrons leaving the metal at the point of positive application, and new ones entering the metal at the point of negative connection. There is no forward bias or reverse bias, only bias in some direction, or none at all. The ratio of the voltage across the chunk to the current passing through the chunk (volts per ampere) is called the resistance of the chunk (ohms).

Insulating materials, in contrast use all the electrons of each atom to hold the material together, except for the electrons that are trapped around their individual nuclei. When a voltage is applied across a chunk of insulator, almost no electrons jump from atom to atom to contribute to any current, so the volts per ampere (resistance) is very high.

Semiconductors are made of materials that, if very pure, are essentially insulators. But if you jam in a few atoms, here and there that have one extra electron in the outer interacting layers, compared to the semiconductor (dope the semiconductor) all but one of their external electrons will link into the semiconductor crystal structure, leaving one dangling electron that is easily dislodged, because it is not part of any bond. The semiconductor atoms can be encouraged to accept one of these extra (in the sense of bonding requirements, but not in the sense of not being balanced by a nuclear proton) with only a little coaxing. The doping atoms have to be close enough to each other that their protons don't miss their wondering electrons so much that they pull them back, because someone else's electron is getting closer at the same time theirs is wondering away.

So these doped in electrons act a lot like the way shared conduction electrons act in metals. This combination of semiconductor and electron donor doping creates N type semiconductor.

If, instead of doping with electron donor atoms, you jam into the semiconductor, atoms with one less electron in their outer shell than the semiconductor, you get a crystal that is electrically neutral (equal number of protons and electrons, so no net charge imbalance), but semiconductor atoms with electrons that are in a position to share with a neighbor to form a bond, but a neighbor who has no electrostatic need for that electron, because it has one less proton in its nucleus. This unfilled crystal bond position is called a hole. This hole acts a lot like a particle with a positive charge, even though, it is really the absence of a particle with a negative charge.

With very little encouragement (additional energy), a semiconductor electron can be nudged into a crystal bond with that doping atom neighbor, to take the place of the electron the doping atom does not provide, in spite of the net electrical imbalance it produces, because being in a bond association is a low energy state for an electron. This ability of the doping atom to hold on to an extra electron, by virtue of it being part of a crystal structure that makes such bonds a low energy state for an electron, is the reason we call these kind of atoms acceptors.

However, once the semiconductor atom gives up one of its bonding electrons for this purpose, it is missing one of its electron bonds with some other neighbor. In effect,the hole has changed atoms. But just as with donor atoms, as long as the acceptors are close enough to each other that a hole moving away from it is compensated for by some other hole moving toward it, it takes relatively little energy to make this sea of holes drift in response to an applied voltage. And since the mathematics of movements of missing negative charged particles in one direction is is equivalent to the mathematics of positively charged particles moving in the other direction, we refer to the holes as positive charge carriers, as if they were particles. Semiconductors that have the ability to conduct current by the movement of holes are called P type. If we connect a voltage across a chunk of P type semiconductor with metal contacts, the holes drift toward the more negative connection, where they are filled with electrons, as they arrive. At the positive connection, electrons are pulled out of the semiconductor crystal, creating new holes that head away as they form.

If we put chunk of N type semiconductor in intimate contact with a chunk of P type semiconductor, we have an unsymmetrical arrangement, so we can describe two possible relationships of applied voltage to this pair. If negative voltage is applied to the N-type material and positive voltage is applied to the P type material, we say the junction is forward biased (each side of the junction is being injected with the kind of charge carriers it normally would use in order to conduct current). Electrons are being pushed into the N type, repelling the inherent doped in electrons toward the junction, and electrons are being sucked out of the P type material, creating holes that repel the doped in holes toward the junction. At the junction, electrons from the N type side are pouring into the P type side, filling holes in short order, and holes are being pushed into the N side, where they are swallowing up electrons in short order. Continuous conduction of current (partly made up of holes moving, and partly made up of electrons moving) ensues.

If the voltage is connected up the other way, we say the junction reverse biased. That is, the more positive voltage is connected to the N type material, attracting its electrons toward that terminal (and removing some of them), and the more negative voltage is attracting the holes in the P type material (and filing some of them with electrons). At the junction, electrons are being pulled away on the N side, and holes are being pulled away on the P side, leaving semiconductor that acts as if it is made of pure silicon with neither donor electrons or acceptor holes, but like an insulating crystal. Once the pull of the external voltage is balanced by the internal pull of the separated electrons and holes across the insulating junction, all current ceases and a stand off is created.

These two bias cases (forward and reversed) define the conducting and nonconducting states of a junction diode.

Has your head exploded, yet?

So does this disqualify all elements or compounds that when bonded precisely "use up" all of the residual electrons in their

respective orbits? My chemistry skills are not as sharp, but I assume that hydrogen(2) compounds would insulate as would

carbon(4) since the outer orbit is saturated? I presume carbon would since it lies in the same column as silicon and I

assume silicon's outer orbit has the same limit of 8 electrons as does the 2nd orbit of carbon?

Does is matter the type of bond? For example, I remember covalent bonds the most from high school, but recall there are

other types.

Tell me if I am mistaken. I see on the periodic table that copper has a "need" for two more electrons. Is this why it

makes a good carrier medium for electricity?

Still with you.

So now we've made a circuit, right?

If I cram two extra electrons per atom into this mass of atoms, does that comprise voltage X? If I cram 4 times more

electrons than this into this mass, does that comprise a voltage higher than X? Do these electrons become less stable in

that mass and more anxious to "jump away" because the atom is that much less stable? Are these electrons will to jump a

greater distance, or arc, because of the opposite positive force of the protons against them? Does the light from the arc

occur because there is a quantum leap taking place, or am I way off here?

I'm not sure I understand. Is this another way of saying that a negatively charged sea doesn't have any natural propensity

to go one way versus another? It will just exit to whatever positive hole is created in the same way a sea of water doesn't

care where it goes so long as it is downward and will take the opportunity to go down into any hole introduced?

I've long understood resistance as an unwillingness to permit current to flow, however, I don't understand this mathematical

relationship. It would see that if I increase voltage, the numerator goes up and I have changed the resistance of the

circuit. I thought resistance was a static thing that existed by nature and we didn't have the ability to change this

figure.

If I recall correctly, H2O is a good insulator?

If I follow correctly, you occasionally disobey the cystalization (vocabulary?) rule within this compound and intentionally

leave one of the crystal structures incomplete. This incompleteness has to occur with exactly the surplus of 1 satisfied

electron? This imperfect crystal is a "doped" crystal?

Would this be like an arrangement of 3 silicon atoms and 1 phosphorus atom? If not, can you give an example of the normal

formula and the doped one?

If I follow, then you have created an electron highway?

Hmmm. So any lone doped atom won't work, but if you put a whole bunch of them in a row, the charges interact enough that

your electron can hover over all of the atoms? (Like those superconductor levitation examples?)

Do you not mean the doped atoms rather than doped electrons? (Or maybe doped compounds? Doped crystals?) I think what you

mean is that by forcing in an atom with an extra electron than what is needed for the bond (higher mass, not an anion), you

have created a material that would typically have insulated if done properly, but instead now can conduct?

I want to be sure I understand this one.

I think I am following along. We've made a crystal but done the opposite of before. I assume this time we used something

like aluminum which is 1 unit lighter. I'm guessing it acts like a cation (+) not because of charge but because it won't

complain too much about being host to an extra electron. If not aluminum, then what is it?

I think I follow.

So the new electron in town has really shaken up the dynamics of the group? But they are fickle enough to "network" and move

around and be content since the environment is so condusive? By way of quantum "osmosis" everything balances out?

Whoa, this way of thinking seems all backwards to me. I guess it makes sense. In my mind, the first scenario was like a

stage diver being passed around overhead of a crowd. This example, in my mind is more like water pails being passed along an

old fire line swinging from one person to the next.

I think I see how neither of these are conductive if there is no voltage present to them. If there is voltage applied, it

takes some activation energy to get things started, but once it does, you have a sort of chain reaction. Is this break down

voltage? Is this why diodes don't turn on until 0.7 volts?

So an N type and a P type would take the opposite state from each other if given the exact same voltage? Is this right, or

do they both become conductive but by different internal mechanisms as already described?

Hold up. Does this mean "forward bias" means that the two semicondutor types will trade electrons all day long?

In order for this to happen, we had to put negative voltage to the N type... I may be lost here. Where is the circuit? Is

there a wire on each end of the N, and then two other wires of a separate circuit on both ends of the P type? Perhaps the

two are connected in series as part of the same circuit? I am not sure what it looks like.

What is the negative voltage going to cause to happen to the N? If voltage is a difference, what does it matter if it is

"postive" or "negative" it's relative right?

So "forward" is like they are looking towared each other for electrons (eye-to-eye), and "reversed" is like they are looking

"away" from each other for electrons (back-to-back)?

It's about to, but this time it is due to a relapse of a sinus infection.

Ryan

To a first approximation (maybe 0.1-th approximation):

Electricity flows from positive to negative. Electrons and holes don't exist, there's just this "stuff" that goes from a place where there's a lot of it (positive) to where there's not much of it (negative). Stay with me, it works, and I blame Franklin.

Most of the things that this stuff interacts with are either metals, where the stuff flows freely, or non-metals, where the stuff doesn't flow at all.

Some guys found a material where the electricity sort of half-way wants to flow and also discovered that they could fabricate it in such a way (doesn't matter how) where electricity prefers to flow in just one direction through the material. Pretty neat trick.

When one side of the material with the special fabrication is connected to the positive electrical source and the other to the negative side, then electricity will flow through it from positive to negative; flip it around and there is no flow. Doesn't matter why, really, it just works that way.

For convenience, call the side that would connect to the positive electrical source when there is a flow of current 'p' and the side next to the negative one 'n'. That's because they were engineers and liked to keep things simple; if they had been scientists they'd probably have used some Hittite script or something. Be thankful for little things.

So, to review, if the 'p' side of the material is more positive than the 'n' side, there is a current flow. If the 'n' side is more positive than the 'p' side, it's backwards and there is no current flow.

Okay, these guys had a new toy and wanted to fool around with it. Turns out that they noticed that you can stack three layers in the order n-p-n and, pretty much as expected, if the 'p' bit is more positive than the 'n' bit there is a current flow into the 'p' and out the 'n'. Ho hum.

Then, when they stuck the positive side onto the top 'n' bit and stuck the other 'n' bit onto the negative side, so that the current would have to go from 'n' to 'p' and then to 'n', well ... oops, no current flow. No surprise, either, since the top bits are backwards n-p, of course.

But remember that the middle 'p' bit can be connected on the outside so that when it's enough more positive than the bottom 'n' bit then there

The effect is kind of like grains of sand in a funnel (okay; big grains, small funnel, stay with me here) where the grains bunch up so tightly that they plug the funnel and nothing gets out. But if you ran a small thread down through the funnel and kept pulling the thread out, the relatively tiny "flow" of the thread would keep a much larger flow of sand moving. Stop pulling the thread, the funnel clogs. Pull it, you get a pile of sand.

Pretty much, that's all that's happening: a small current flowing into the base (the middle 'p') and out the emitter (the bottom 'n') controls a much larger current that flows from the collector (the top 'n'), past the base, and then out the emitter.

An aside: For some transistor types (called bipolar; the ones drawn with the outside lines at an angle) the collector and emitter (always the one with the arrow) are optimized for their roles and can't usefully have their roles reversed. There are other types (called MOS; drawn with the outside lines parallel like the sides of a box) where it's sometimes possible ignore which is the collector and emitter and use either one in either role; the base is always the base for both bipolar and MOS transistors. Note that in MOS transistors the terms collector, base, and emitter are replaced by drain, gate, and source. Don't let it worry you, life is perverse.

Okay, back to the transistor / funnel. The ratio of the amount of current that you get to flow from the top (collector) out the bottom (emitter) versus the amount of current going into the middle (base) and out the emitter varies from the 10's to the 100's depending on how the transistor is made, how much current is flowing through it, how hot it is, etc., etc. By themselves, there's so much variation that transistors aren't of much use in applications where precision is needed.

But resistors can be used to control the amount of current into the base, or into the collector, or out the emitter. They can be selected so that the gain of the "stage" (i.e., the ratio of output to input for the transistor and all of its hangers-on) depends more on the ratios of the resistors and less on the gain of the individual transistor. Now the stage gain is predictable. This is technically known as A Good Thing.

Resistors are mundane and not very exciting. They have the virtue of being relatively stable and can be selected for specific values and tolerances. Transistors are, by comparison, exotic creatures with widely varying characteristics. When a transistor stage is anchored by resistors, the effects of item-to-item variation are much reduced and circuit behavior is much more predictable. Lets us build a thousand widgets that all work the same, instead of having to hand-tune every one of them (and worry about what happens if the widget ends up in Saskatoon or in Salt Lake City).

The different ways of hooking-up transistors and resistors result in stages that function as switches, voltage amplifiers, or current amplifiers. Which you use depends on what you want.

Remember: The one with the arrow (irrespective of the direction of the arrow) is the emitter. The middle one is the base; and the other one is the collector. Current flows from positive to negative (the p to the n) so if the arrow points from the middle to the outside then the middle must be the one that's more 'p' and the outside one more 'n' and so you have an n-p-n transistor.

There are p-n-p transistors, also, where the arrow on the symbol (always the emitter) points from the outside to the inside. They are used in the same way, by setting up a small current in the p-to-n direction between the base and emitter to control a larger current between the emitter and the collector.

The next question "Where to put the resistors to 'bias' a transistor?" is answered: What do you want the transistor to do?

I think it does. The exception might be ionic solutions and plasmas, where whole atoms (minus an electron) move, as well as electrons. But those aren't solids. The electrons under discussion are the outside ones, normally referred to as valence electrons. The rest are so close to the nucleus that they never leave that atom, under normal solid state situations.

Frozen hydrogen is an insulator, but if you compress it enough, it switches to a metal.

Carbon in some crystalline forms is fairly metallic (graphite) and on others is a semiconductor (diamond). It depends on what those electrons are busy doing and what the energy levels of those tasks are.

Yes it matters. I think all semiconductors are held together with covalent bonds, but this is getting out of the area of knowledge I use often.

I think copper is a metal that contributes two electrons to the conductive cloud. But don't quote me on that. This is the kind of thing I have to look up when I need it. My explanation was not intended to be exactly right in all respects, as it was to be a fly over of the basic effects involved in conductivity.

Right.

If you cram even one extra electron per atom into a crystal of metal, it explodes with great violence from the incredible repulsion of all that negative electric charge. Al conductors except diffuse plasma are almost perfectly neutral, meaning that in order to jam an electron in somewhere, you have to pull one out, somewhere else.

You are not talking about electro statically balanced matter any more, but a collection of ions.

Right. Isotropic. The same in any direction. No inherent sides or grain.

Right. Water flows downhill.

Same with electrons in metals.

Many materials, metals included, hold a quite constant ratio of voltage versus current for a wide range of both. This property is what makes the production of linear resistors possible. But an ohm is still a volt per ampere. A hundred ohms requires 100 volts across it to push an ampere through it.

So is ice.

You got it.

That would be a very highly doped crystal. 1 phosphorus atom per

They don't have to be that close. Their effect reaches out over quite a few atomic diameters.

I mean electrons provided by the doping atoms. But I get tired typing that long phrase. I appreciate your consideration.

Yes.

Me too. It is hard to picture.

Boron, aluminum, gallium, or indium. ( I Googled [acceptor atoms].

Sounds good to me.

You will have to ask someone else about that. But I think you are very close.

The second, I think. They conduct effectively with opposite polarity of charge carrier. Remember that holes (spots that can be encouraged to hold an electron) get passed around exactly as if they were particles with positive charge. They move at a different speed and have a different effective mass than electrons, so P type conductivity has a unique description from an electrical perspective than N type conductivity. The general concept is the same, but they are somewhat different.

...will spill oppositely charged carriers into each other all day.

There is a wire from the negative terminal of a voltage source connected to the N type material and a wire from the positive side of the voltage source connected to the P type material, and the block of N type material in in intimate contact (continuous crystal structure) with the block of P type material. A diode with a battery connected across it.

That's it.

It will pour electrons into it, forcing electrons out the other side. Total charge balance is maintained, or very nearly so.

Right. voltage difference across the chunk of material is all that matters. Inside the block, the atoms have no way of knowing what the average potential of the rest of the universe is.

Make that electrons and holes.

Works for me.

I hate it when that happens. )-;

: Something that has been confusing me no-end, and I just cant seem to : grasp, is how a TRANSISTOR works!!!

: I cant grasp WHY they are so extremely important - probably because I'm : finding it hard to understand their basic operation !

Other people have given you explanations that are way too complicated (There might've been some in there that were to the point, but I might've missed them.)

My explanation/answer to your two questions above will focus on MOS transistors, but can be applied to BJTs as well.

A simple explanation of HOW a transistor works is that it is a device that can implement 2 elementary functions:

1. A switch 2. A constant current source

Breaking this down a little bit more, a transistor can be in one of three modes of operation, each of which can be modeled with a basic circuit element:

1. An open switch 2. A closed (but non-ideal) switch 3. A constant current source.

Which mode of operation depeds upon the relationships between the voltages at each of the transistor's 3 terminals (actually, a MOSFET has 4 terminals, but let's assume that the Bulk is connected to the Source so that Vbs = 0 all the time, to make things easier.)

When I say non-ideal switch, I mean that the switch has some resistance between it's terminals. A non-ideal switch behaves like an open circuit when it is open, and a resistor (NOT a short-circuit) when it is closed.

Furthermore, let's also assume that, for the moment, we are only interested in examining how the transistor behaves for DC voltages. Therefore, NO CURRENT flows into the gate.

Let's look at an NMOS transistor.

Vgs = Gate Voltage - Source Voltage Vds = Drain Voltage - Source Voltage

When Vgs < Vt (The threshold voltage -- an intrinsic property of the transistor, usually equal to about 0.5V - 1.0V in modern processes) the transistor is in the CUTOFF region of operation, and behaves like an OPEN SWITCH. No current flows from drain to source.

When (Vgs > Vt), AND (Vds < Vgs - Vt) the transistor is said to be in the LINEAR (or TRIODE, or NON-SATURATED) region of operation, and behaves like a CLOSED, NON-IDEAL SWITCH. A current will flow from the drain to source. The amount of current that flows is DIRECTLY PROPORTIONAL TO BOTH Vgs AND Vds.

When (Vgs > Vt) AND (Vds > Vgs - Vt) the transistor is said to be in the SATURATION region of operation and behaves like a CONSTANT CURRENT SOURCE. Current will flow from the drain to source. This current is DIRECTLY PROPORTIONAL to Vgs, but INDEPENDENT of Vds (so long as Vds > Vgs

- Vt).

That's the HOW. Now, the WHY. Any digital circuit is essentially a collection of voltage-dependent switches. A binary voltage represents a boolean value. Therefore, transistors can implement any boolean function, which enables the enormous microprocessors and other marvels of digital logic that are commonplace today. On the flip-side, most useful analog circuits require constant current sources. Amplifiers are essentially 2 constant current sources that fight each other. Sure, vacuum tubes can also implement these functions, but transistors can implement them and are MICROSCOPIC. Therefore, transistors are important because they allow for the building of both digital and analog INTEGRATED CIRCUITS of enormous complexity that are still barely visible to the naked eye.

Hope that helps in answering your questions....

Joe

You are being deluged with information. Start *really* simple, then build on it.

A bipolar transistor has 3 leads: emitter, base and collector. The schematic diagram always has the base in the middle, between the emitter and the collector. The emitter is always drawn with an arrow. The arrow points toward the base for PNP transistors, and away from the base for NPN transistors. Base-emitter current controls current across the emitter-collector. When the arrow points toward the base (PNP) the base must be more negative than the emitter (by roughly .6 volts) to cause current across the emitter-collector. When the arrow points away from the base (NPN) the base needs to be more positive than the emitter (again, by roughly .6 volts) to cause emitter-collector current.

That's the basics to learn for now. Once you learn that to where it is second nature to you, you can build on it. There are other types of transistors, and other materials (which affect that "roughly .6 volts" figure) but hold off on worrying about them. Hold off on holes, and electrons, and valence rings etc until you understand the symbol and the concept of controlling the c-e (collector-emitter) current via the b-e (base-emitter) current.

Ed