BJT Operation

Thread Starter

The Electrician

Joined Oct 9, 2007
2,971
I'd like to continue this thread:

http://forum.allaboutcircuits.com/showthread.php?t=24059

in this forum so that it will be possible to reference ideas not directly and specifically involving BJT's.

Ratch asserts that it is Vbe that controls BJT's.

I notice that this discussion usually ends being a discussion mainly of interpretation and philosophy, so I want to go there immediately.

The question I direct to the reader is this:

What controls drain current in a MOSFET, the gate voltage or the gate charge?

Obviously, the gate voltage and gate charge are related, so why should a person choose one over the other as the controlling variable?

Would it not be fair to say that the gate voltage is an indicator of the gate charge?

Yet most descriptions of MOSFET operation say that a MOSFET is a voltage operated device.
 

mik3

Joined Feb 4, 2008
4,843
MOSFETs are said to be voltage controlled devices because the current flowing through the gate is very small (except in transients).

BJTs are said to be current controlled devices because the base current varies significantly but the base-emitter voltage stays almost constant. However, I agree with Ratch who says that BJTs in reality are controlled by voltage because is the voltage which causes electrons to cross the potential barrier of the base-emitter junction and form the current. Current cannot exist without voltage.
 

steveb

Joined Jul 3, 2008
2,436
Captain Kirk said:
Then, ... what is it?
The Guardian of Forever said:
A question! Since before your sun burned hot in space, and before your race was born, I have awaited a question.
Captain Kirk said:
What are you? Are you machine, or being?
The Guardian of Forever said:
I am both, and neither. I am my own beginning and my own ending.
Spock said:
I see no reason for answers to be couched in riddles.
The Guardian of Forever said:
I answer as simply as your level of understanding makes possible.

OK, just being silly. Really, I don't like the word "control" without a definiton. As an engineer interested in controlling the MOSFET with input signals, I usually make a voltage controlled model of the MOSFET, never a charge controlled model. I sometimes even make a current controlled model for high frequency transient analysis, although one can think of this as trying to influence charge.

If I develop a sophisticated nonlinear dyanmic model, I use gate-source voltage as an input and MOScapacitor voltage as the state variable. You can probably reformulate this a model with gate charge as a state variable, but then it's harder to relate to the typical circuit variables (voltage and current).

Still, it can be argued that the drain current is a function of charge, just as well as we can say it is a function of voltage. I don't see any way to defend a position that one is the controller and the other is along for the ride.

Others can argue whether or not the statement that a MOSFET is charge controlled is a case of the "dog wagging the tail", or "the tail wagging the doll". I don't need to think about this question to understand how the device works. A more pressing question for me is whether I control my wife, or does she control me? Or, do we just live together in alternating periods of discord and harmony? :D
 
Last edited:

Thread Starter

The Electrician

Joined Oct 9, 2007
2,971
MOSFETs are said to be voltage controlled devices because the current flowing through the gate is very small (except in transients).
Just because the current is very small doesn't mean that the current isn't the controlling variable. I'm not saying that it is or that it isn't; I'm saying the reason you give is good for an engineering point of view that voltage is the controlling variable, but not for a deep understanding of what's controlling in reality.

Can you give a stronger reason than that the current is very small? Something more connected with the physical electronics of device operation.
 

mik3

Joined Feb 4, 2008
4,843
Current is caused by voltage. Without voltage no charges move and thus no current exists.

For a MOSFET, the tiny current through the gate (at steady state) is due to the finite resistance of the insulating material (SiO2) between the gate and the channel. For example, an n-channel MOSFET has its drain and source regions n-type doped and the region near the gate p-type doped (substrate). To turn it on, electrons must be attracted from the substrate towards the gate as to invert the p-type region near the gate into n-type region (form an n-channel and thus its name) and allow current from drain to source. You can see that you need voltage to achieve this. Also, during the movement of the electrons current exists (transient)-like a capacitor. After, the gate charges up (electrons have moved due to the applied voltage as to balance this voltage the only current is due to the leakage current of the SiO2.

You can see that it is the movement of electrons (charge) which causes the MOSFET to turn on/off. This charge movement is caused by a voltage and is also called current. :cool:

To conclude, you can see that you need both voltage and current to control both MOSFETs and BJTs. MOSFETs are said voltage controlled because voltage is the quantity which varies more and BJTs are said current controlled because current is the quantity which varies more. It is just terms defined by engineers but in both cases the job is done by charge movement which is caused by voltage.
 

Thread Starter

The Electrician

Joined Oct 9, 2007
2,971
For a MOSFET, the tiny current through the gate (at steady state) is due to the finite resistance of the insulating material (SiO2) between the gate and the channel.
I'm trying to get beyond imperfections of real devices and get to the fundamentals of an ideal device. Theoretically one could have a MOSFET with no leakage. Then after the initial transient of charging the gate by applying a voltage source, there would be no gate current.

You can see that it is the movement of electrons (charge) which causes the MOSFET to turn on/off. This charge movement is caused by a voltage and is also called current. :cool:
I'm not asking how it turns on, having been off, or how it turns off, having been on.

I want to know, in an ideal device with no gate leakage or other parasitics, once the gate is charged by the application of a voltage source, and then that source is removed, leaving the gate charged, what is controlling the drain current?

At the most fundamental level, is it the gate voltage or the gate charge?

The fact is, with a good quality power MOSFET, you can charge up the gate with a 9 volt battery and disconnect the battery, leaving an ohmmeter connected to source and drain, and you will find that the MOSFET will remain enhanced for days. So, even in an imperfect, real device, what is it that keeps the channel enhanced? Is it the gate voltage or the gate charge?
 

mik3

Joined Feb 4, 2008
4,843
They are the positive charges (lack of electrons) in the metal of the gate which attract the electrons in the substrate near the gate to form the n-channel. At the fundamental level they are the charges but you can say it is the voltage too because charges form the quantity we call voltage.
 

mik3

Joined Feb 4, 2008
4,843
Where did I say that? I haven't yet said what I believe.

I'm just trying to get some of the people who got so involved in the other BJT discussion to show up here.

Maybe some of them don't believe you. :(
I was kidding. I didn't say you said you don't believe me. I am sorry for the misunderstanding.
 

beenthere

Joined Apr 20, 2004
15,819
You have set the conditions for your ideal MOSFET -
I'm trying to get beyond imperfections of real devices and get to the fundamentals of an ideal device. Theoretically one could have a MOSFET with no leakage. Then after the initial transient of charging the gate by applying a voltage source, there would be no gate current.
As there is no current, no charge moves. Or are we to assume from
I haven't yet said what I believe.
that you were not seriously proposing your ideal MOSFET? If that is the case, what is this about?
 

steveb

Joined Jul 3, 2008
2,436
Does anybody else have anything to say? Ratch? Steveb?
Well, I already put in my 2 cents above, but I could comment further by looking at a more simple device.

Let's say I invent a simple switch. It's a simple continuously sliding switch with very small friction, but it stays in whatever position I move it to. Let x equal the distance from the left most position, and it slides continuously from 0 to 1. If 0<=x<=0.5, then the switch is off, if 0.5<x>1.0, then the switch is on.

Here are my claims about this switch. The output state (resistance) of the switch "DEPENDS" on the position state of the switch. Also, the "CONTROL" input to this switch is the position. In other words, the internal state and the control input are the same exact thing, and the device is a position controlled device with an output resistance dependent on the position state.

Now imagine that I modify the device by putting a spring which pulls the slider to x=0. Now the switch is off, but we can still say that the output state (resistance) "DEPENDS" on the device state (position). However, the "CONTROL" input is force and not position. Hence, we have a force controlled device where the output resistance "DEPENDS" on the position state. Note that this force controlled device can still be controlled with position, if I implement a position controlled feedback loop. Hence, I can still control this device with position. However, the force and position are now intimately related and you can't force one without forcing the other. It almost becomes a matter of opinion (or perhaps perspective) as to whether the device is controlled by force or position. Internally, the position seems to be more fundamentally important, but is this control, or dependence? To me this is just semantics.

Now I modify this device once more. I put a mass on the slider and cover the switch with a cover. Now I have a motion controlled switch which is controlled by acceleration. The device is still fundamentally dependent on the internal position state, but is externally controlled by acceleration.

This basically highlights why I think it is important to define what the word "control" means in a given context. If an engineer is talking about control of an input to get the desired output, then he is talking about something very different than someone trying to get an understanding of the internal physics of the device. In the other thread Ratch expressed the opinion that "it is a distinction without much difference", but to me there's a big difference. It's a difference of inputs versus outputs. It's a difference of design versus understanding. It's a difference of things I can precisely change versus things I can only indirectly influence. Often, the internal states are not directly controllable and not directly measurable. In such cases, the designer must either create an "observer" to estimate the internal states, or exercise feedback control by measuring outputs and feeding back to the inputs. There is an entire branch of engineering called "Control Theory" which is concerned with one interpretation of the word "control", and not the other. Certainly, this important word is due special consideration in a forum like this.

http://en.wikipedia.org/wiki/Control_theory
 
Last edited:

Thread Starter

The Electrician

Joined Oct 9, 2007
2,971
You have set the conditions for your ideal MOSFET
The conditions were:

Theoretically one could have a MOSFET with no leakage. Then after the initial transient of charging the gate by applying a voltage source, there would be no gate current.
As there is no current, no charge moves.
Yes, after the initial transient there is no current; but during the initial transient, there is.

True, there is no more movement of charge on the gate; only through the channel if a voltage source is applied there.



Or are we to assume from that you were not seriously proposing your ideal MOSFET? If that is the case, what is this about?
I was seriously proposing it. I asked what controls the drain current after the gate of the ideal MOSFET is charged up (with respect to the source, of course), and the voltage source that initially charged is removed. Not that its removal changes anything; Vgs will still be some non-zero value.

This is going to lead to some more discussion of BJT operation.
 

beenthere

Joined Apr 20, 2004
15,819
By the way, your example -
The fact is, with a good quality power MOSFET, you can charge up the gate with a 9 volt battery and disconnect the battery, leaving an ohmmeter connected to source and drain, and you will find that the MOSFET will remain enhanced for days. So, even in an imperfect, real device, what is it that keeps the channel enhanced? Is it the gate voltage or the gate charge?
pretty well describes a cell in a EPROM. Under influence of voltage, charge is injected into a well and keeps the FET turned on for years. The charge present maintains the voltage that keeps the electric field in effect (Field Effect Transistor, remember?), thus maintaining the FET in the on state.
 
Last edited:

Thread Starter

The Electrician

Joined Oct 9, 2007
2,971
OK, just being silly. Really, I don't like the word "control" without a definiton. As an engineer interested in controlling the MOSFET with input signals, I usually make a voltage controlled model of the MOSFET, never a charge controlled model. I sometimes even make a current controlled model for high frequency transient analysis, although one can think of this as trying to influence charge.
The BJT discussion got so philosophical that I just had to join in, but the question I raised about MOSFET's digressed too much for the forum the discussion was in. It probably should have been here, anyway.

Imagine some physical process(es) where A causes B which causes C which causes D, D being some effect whose cause(s) we're investigating.

I think it's perfectly reasonable to say that A causes D (ultimately, at least), but it's also worthwhile to wonder if we can say that there is a final, fundamental cause.

Anyway, to add to the noise level, but hopefully to improve the SNR, here's a page from the 1959 General Electric Transistor manual. I looked through a number of older texts on the physical electronics of transistors, and the GE manual put it most succinctly:

A transistor is a charge-controlled device.

But, as a general proposition, it seems to me that any charge-controlled device is also reasonably considered to be a voltage-controlled device. Any electrode that you attempt to supply with charge is going to have capacitance to reference, and once it receives its charge, a voltage will be produced. In fact, the usual way to add charge to an object is with a voltage source.

So, herewith two images. One from the GE manual, the second from one of the SEEC Volumes, the one titled "Physical Electronics and Circuit Models of Transistors".

In the second image, at the bottom of the page is a sentence referring to a lot of previous development of the theory of transistor operation, and the notion that Vbe is the controlling variable for collector current. The sentence is "This is a powerful point of view, closely related to the operating mechanism of the transistor, and we will find much employment for it."

Notice that they don't say that it is the operating mechanism, just closely related to it.

What they say in excruciating detail earlier is that applying a voltage, Vbe, produces stored charge in the base region, and the gradient in that charge controls the collector current. It's only certain defects in BJT operation like recombination that result in steady base current. Recombination causes the stored charge to gradually disappear unless replenished by a base current. The MOSFET gets around that; there's no recombination mechanism in the gate region.

So, it seems to me that Vbe is one step removed from the final cause of collector current.

I especially prefer charge control as the "final cause" in these devices, because charge is something you can put in a container and carry around, unlike voltage or current. Voltage is a consequence of separation of charges, and current is a consequence of movement of charge. Charge just seems more fundamental to me.

Sorry about the images needing to be rotated. I didn't do it because I didn't want to add more noise to the jpgs. I'll see about doing them over tomorrow.
 

Attachments

Last edited:
Top