Assistance with Transistor Theory

[KrisBlueNZ]

I'm still not convinced that it's meaningful to describe a transistor as voltage-controlled.

It's certainly true that voltage is needed in order for anything to happen - you need a base-emitter voltage source and a collector-emitter voltage source - but that is just due to the nature of voltage and current.

No component (apart from special components that can store energy) can do anything without an applied voltage. No current can flow in a resistor unless a voltage is applied. You can't apply a current to a resistor without also applying a voltage. It's simply natural cause-and-effect in the real world.

I pointed this out back in post #123 but I'll spell it out again.

An LED won't light unless voltage is applied. Yet LEDs are described as current-driven, for two obvious reasons: (a) it's not normal or useful to connect a firmly defined voltage, such as a battery, directly across them (unless the current is limited by the battery's internal resistance, in which case you're applying a current, not a voltage), and more importantly, (b) the primary quantity that varies (brightness) varies roughly in proportion to the forward current, and not the forward voltage.

In this sense, a transistor is the same as an LED: you don't connect a battery directly across the base-emitter junction, and the primary quantity (the collector-emitter current) is proportional to the base current, not the base-emitter voltage.

You could argue that the collector voltage (in a common emitter circuit) is also a primary quantity, but it doesn't vary in proportion with the base-emitter voltage either; it varies inversely with the base current. So that doesn't help the argument that a transistor is voltage-controlled.

You might be able to argue for a transistor being voltage-controlled in an emitter follower circuit, but I would argue that it's the circuit, not the component, that is voltage-controlled.

If there is any difference between how those two criteria apply to the LED and to the transistor, please let me know what it is.

If there isn't, then saying that a transistor is voltage-controlled doesn't tell us anything about transistors specifically, because by that definition, every component is voltage controlled.

What I think is more relevant is the mechanism by which the transistor operates. This is something I don't understand - doping, depletion region etc don't mean much to me. But as I pointed out in post #116 (which no one answered directly), I was interested in the explanation posted by Ratch in post #85. He wrote:

In a BJT operating in the active region, the Vbe is used to control the charge carriers diffusing into the thin base area where they are whisked away into the collector circuit by the collector voltage.

Unfortunately, not all the charge carriers are captured. A relatively small number of charge carriers make it into the base circuit where they become waste current because they do not contribute to the collector current. This waste current is a relatively constant proportion of the collector current and is roughly the inverse of BETA.Ratch

He seems to be implying that I_B is an unfortunate by-product of I_CE and should be considered a fraction of it; the fraction being 1 / h_FE. This differs from the usual explanation that I_C is I_B × h_FE. He seems to be implying that I_B is caused by I_CE.

From my limited knowledge of semiconductor physics, I thought that I_BE was needed to bring charge carriers from the emitter into the base, where they can then be attracted to the collector. If that's true, then I_BE causes I_CE to flow; in that case, I for one would say it's most meaningful to describe a transistor as current-controlled. Given the understanding that voltage is needed to cause current to flow (as it is with other components).

Can someone please set this last point to rest? That is, does I_BE cause I_CE or is it caused by I_CE?

 

[LvW]

I'm still not convinced that it's meaningful to describe a transistor as voltage-controlled.
.......
It's certainly true that voltage is needed in order for anything to happen - you need a base-emitter voltage source and a collector-emitter voltage source - but that is just due to the nature of voltage and current.
................
If there isn't, then saying that a transistor is voltage-controlled doesn't tell us anything about transistors specifically, because by that definition, every component is voltage controlled.
................
Can someone please set this last point to rest? That is, does I_BE cause I_CE or is it caused by I_CE?

Hi Kris, I do not intend to answer from the charged carrier point of view. Instead, I just observe and try to explain the BJT`s behaviour:

1.) Let the collector unconnected and treat the pn junction of the B-E path like a simple diode. There will be acurrent IE if the voltage is in the classical region of VBE=0.65 volts.
Now - connect the collector to a voltage of (let`s say) +10 volts. Now, all (nearly all) carriers forming the current IE are attracted by this large voltage and travel to the collector forming the current IC. However, a small percentage (imagination: very close to the base input node) still travels to the base pin forming a current IB.
Question: Is there any good reason to assume that - instead of VBE - suddenly IB should determine the value of IE resp. IC?

2.) We observe that the current IC is temperature dependent. Of course, it is possible to explain this effect and to calculate this increase. And it was calculated (for constant IB) that this increase can be compensated by a corresponding reduction in VBE. This leads to the known value of -2mV/K. In words: The voltage VBE must be reduced in order to compensate the current increase. And this VBE change was not only measured - it can be explained and was calculated based on carrier physics!

3.) Something similar happens in case of VCE increase. If we keep IB constant and increase VCE we observe an increase in IC. This is the known EARLY effect. Why does IC increase? Because of a similar effect (base width reduction connected with electrical field increase). And the electrical field is produced by the voltage VBE and not by a current.

4.) If you design a simple amplifier stage, you will notice that (for identical quiescent currents IC) the voltage gain is independent on the ratio IC/IB (intentionally I avoid the term "current gain").
Instead it is dependend on the transconductance gm=dIc/dVBE only. BJT as a current amplifier?

_____________
Is there any good reason to believe that the BJT is not voltage controlled but current-controlled?
(Comment to your first sentence: I am not sure if the term "meaningful" is appropriate. There is only one single working principle - but I agree for some applications it is sufficient to treat the BJT as adevice that is current-controlled. But that´s a model only - without physical relevance).

 

[Ratch]

KrisblueNZ,

"I'm still not convinced that it's meaningful to describe a transistor as voltage-controlled."

Perhaps I can help.

"An LED won't light unless voltage is applied. Yet LEDs are described as current-driven, for two obvious reasons: (a) it's not normal or useful to connect a firmly defined voltage, such as a battery, directly across them (unless the current is limited by the battery's internal resistance, in which case you're applying a current, not a voltage), and more importantly, (b) the primary quantity that varies (brightness) varies roughly in proportion to the forward current, and not the forward voltage."

A LED works like any other junction diode except it emits light. The light emitted is related to its current. The current is related to its voltage. So when we are talking about what controls what, are we talking about the diode alone, or the diode circuit that drives the LED?

"In this sense, a transistor is the same as an LED: you don't connect a battery directly across the base-emitter junction, and the primary quantity (the collector-emitter current) is proportional to the base current, not the base-emitter voltage."

Yes, a transistor has to be biased if you want it to operate in its active region. The base current is a more or less direct proportional indicator of the collector current. That is because both the base current and collector are exponentially related to Vbe. So dividing the collector by the base current cancels out the exponential terms and gives a direct proportion we call BETA.

"You could argue that the collector voltage (in a common emitter circuit) is also a primary quantity, but it doesn't vary in proportion with the base-emitter voltage either; it varies inversely with the base current. So that doesn't help the argument that a transistor is voltage-controlled."

What is a "primary quantity"? How can you say the base current affects the base emitter voltage when the base-emitter voltage is kept constant?

"You might be able to argue for a transistor being voltage-controlled in an emitter follower circuit, but I would argue that it's the circuit, not the component, that is voltage-controlled."

If you are going to determine what controls a component, then you have to ignore its application in a circuit. The application or configuration of a component does not change its control-effect relationship.

"If there is any difference between how those two criteria apply to the LED and to the transistor, please let me know what it is."

What criteria are those?

"If there isn't, then saying that a transistor is voltage-controlled doesn't tell us anything about transistors specifically, because by that definition, every component is voltage controlled."

No, a gas discharge tube is not. Neither is a magnetic amplifier.

What I think is more relevant is the mechanism by which the transistor operates. This is something I don't understand - doping, depletion region etc don't mean much to me. But as I pointed out in post #116 (which no one answered directly), I was interested in the explanation posted by Ratch in post #85. He wrote:

"He seems to be implying that I_B is an unfortunate by-product of I_CE and should be considered a fraction of it; the fraction being 1 / h_FE. This differs from the usual explanation that I_C is I_B × h_FE. He seems to be implying that I_B is caused by I_CE."

I cannot discern all the "subs" present in your above sentence and what they mean.

"From my limited knowledge of semiconductor physics, I thought that I_BE was needed to bring charge carriers from the emitter into the base, where they can then be attracted to the collector. If that's true, then I_BE causes I_CE to flow; in that case, I for one would say it's most meaningful to describe a transistor as current-controlled. Given the understanding that voltage is needed to cause current to flow (as it is with other components)."

As I explained in an earlier post, diffusion is the primary cause of charges moving from the emitter to the base. When the uncovered charges set up enough back-voltage, the diffusion stops. Applying a Vbe voltage lowers the back voltage and allows charge movement by diffusion to continue. So Vbe controls diffusion which brings charges into the base. Because diffusion is the mechanism for charge movement into the base area, and control of diffusion is controlled by Vbe, the relationship of charge into the base area will be exponential. Once in the base area, the charges are swept into the collector circuit. The collector voltage variation has little influence on the collector current because the emitter-base circuit controls what is available to the collector.

"Can someone please set this last point to rest? That is, does I_BE cause I_CE or is it caused by I_CE"

I hope I have

Ratch

 

[Ratch]

"Merlin3189, post: 1637531, member: 20438"]At the risk of further red herrings, can I ask about something which has long puzzled me about these holes.
Ratch said,
"A lot of folks seem to think that a hole is simply a missing electron in the conduction band. The hole in the valance band is every bit as real as a electron. It has its own mobility, effective mass, drift velocity, diffusion coefficients, etc."
(Put aside his distinction between conduction band and valence band holes.)


"If this were true, how would you explain the behaviour of a hole?
When a hole moves from A to B,
B looses the charge of one electron and looses the mass of one electron."

When a hole and electron meet, they annihilate each other.

"So it would seem that a hole has a charge of +1.602176565(35)×10−19 C and a mass of -9.10938291(40)×10−31 kg"

At the quantum level, particle physicists talk about "effective mass". There are several effective mass values depending on what is being calculated. They allow classical equations to be used in approximate calculations and greatly simplify the work.

"When in an electric field, a hole moves - but which way?
Due to its positive charge there is a force away from positive and towards negative.
As a massive particle I guess it obeys Newtons laws, so a = F / m
Since m is negative, then a has the opposite sign to F and so the acceleration is in the opposite sense to the force and holes should move towards the positive!"

Who says the mass of a hole is negative? What would that mean? No, the mass of a hole is not negative.

"If I understand the hole merchants, this is the opposite of what they claim."

No way.

Ratch

 

[Ian Getreu]

Kri.s,

I'm replying to you because you have asked for a better understanding.
For that reason, this response is ONLY for Kris.

Let's simplify the question. Do you think of a diode as votage driven or current driven?

If you "forward-bias" a diode, do you think of applying a voltage and seeing the current flow?
Or do you think of running a current through it and watching the voltage set up?

If you think of a diode as current-driven you will probably never be able to think of a BJT as voltage-driven.

When you study or learn about semiconductor physics, you can only understand what is actually happening by considering the application of a voltage that raises or lowers the barrier for electrons to flow. Anyone who understands solid-state physics understands that the diode (and the BJT) are voltage driven. Because the voltage causes a current to flow. If you apply a current, the device sets up the appropriate voltage drive that will allow that current to flow.

If you do not understand the physics of how a diode (or a BJT) works, this may appear to be magic - so the result (current) of the cause (voltage) gets lost. I thought that Ratch's response was right on - he pointed out that in a BJT the collector current is proportional to the base current (assuming beta is a constant - which it isn't) because the exponential terms cancel - he did a masterful job on that. [However, it will only be understood by people who realize that a diode current has an exponential relationship to the applied voltage. Current-driven advocates may have difficulty explaining why]. This is why it appears to the uninitiated that the device is current-driven. But that totally ignores how the device actually works.

It's as if you observe from the outside that a car moves when there is a mechanical push on a gas pedal so you therefore conclude that cars are mechanical-gas-pedal driven, not driven by gasoline-powered engines. Just like you can be happy thinking that the car works by a push on the gas pedal, until you have to fill up with gas (petrol to you), you can take the current-driven approach only to the point at which the device actually starts to influence the behaviour of the circuit. The analogy is not that great, but it does get the main point across - what you see from the outside does not necessarily explain what is going on inside.

So, to summarize (and to repeat Ratch's point): if you understand how the BJT (and diode) truly works, it is voltage-driven. If you don't understand how it works, it may be easier to asssume it is current-driven - which will not ultimately help you in designing with those devices. The best advice I can give you is - go read a book that explains how a diode works or talk to an expert you trust. Then decide for yourself. Start with the diode - it's easier and it is a major stepping stone to understanding the BJT.

Since this topic has been driven into the ground ad nauseum, I do not intend to discuss it again. The current-driven advocates may continue to attack. But, as I said in the beginning, I only replied to you because you said you really want to learn. (And it's good to help a fellow-Australasian).

Ian

 

[BobK]

If I roll a rock up a hill, it's potential energy increase by mgH. So, is rolling a rock up a hill weight driven or height driven?

Bob

 

[BobK]

Answer: It depends on whether I have a rock of a certain weight or a hill of a certain height.

Bob

 

[Ratch]

To the Ineffable All,

For those interested, here is a discussion I had with a staff scientist of a big corporation. Judge for yourselves. Start with post #31 of the link below.
http://www.electro-tech-online.com/...-controlled-device.134889/page-2#post-1130919

Ratch

 

[KrisBlueNZ]

Question: Is there any good reason to assume that - instead of VBE - suddenly IB should determine the value of IE resp. IC?

I can see two reasons why it's helpful to describe it that way: (a) V_BE causes I_B and (b) I_C is approximately proportional to I_B and is not approximately proportional to V_BE.

Is there any good reason to believe that the BJT is not voltage controlled but current-controlled?

I see your point that voltage is required in order for current to flow. I've said this many times during this thread. This is true of most components, including diodes and LEDs, and of course resistors. It's a result of the natural relationship between voltage and current - you can't make current flow without applying a voltage, so therefore, you could say, voltage comes first. But saying this doesn't tell us anything useful about the transistor as compared to an LED or a resistor.

I believe it's more useful to describe the transistor as current-controlled because the quantity that's being controlled, I_C, is roughly proportional to I_B. Certainly V_BE affects I_C but the relationship is nothing like a straight line; it's the same shape as the I_B vs. V_BE curve.

Pointing out that a V_BE voltage must be provided before any current can flow doesn't tell us anything specifically about transistors, since it's common to most components.

A LED works like any other junction diode except it emits light. The light emitted is related to its current. The current is related to its voltage.

The brightness is roughly linearly related to the current, but only exponentially related to the voltage. Which quantity is more useful when describing the important function of the LED?

So when we are talking about what controls what, are we talking about the diode alone, or the diode circuit that drives the LED?

I don't know what you mean. The "diode circuit that drives the LED"?

Yes, a transistor has to be biased if you want it to operate in its active region. The base current is a more or less direct proportional indicator of the collector current. That is because both the base current and collector are exponentially related to Vbe. So dividing the collector by the base current cancels out the exponential terms and gives a direct proportion we call BETA.

So are you saying that it's a coincidence that both I_B and I_C are related to V_BE in the same way? Obviously not. My interest is in whether I_B is an indicator of I_C, as you say it is, or whether I_B causes I_C. This relates to the semiconductor physics involved, which I don't understand.

It does seem to me that some I_BE is required in order to cause movement of electrons from the emitter into the base, where they can then be taken by the collector; if that's true, I would describe I_B as causing I_C.

What is a "primary quantity"?

I suppose it is the quantity that we wish to control. In a common emitter circuit, I_C would be a "primary quantity". In an LED, brightness is the "primary quantity".

How can you say the base current affects the base emitter voltage when the base-emitter voltage is kept constant?

I hope I didn't say that.

"If there is any difference between how those two criteria apply to the LED and to the transistor, please let me know what it is."

What criteria are those?

I guess mainly I meant that the primary quantity is roughly proportional to the current flowing, and only exponentially proportional to the voltage. In the case of the LED, we specifically regulate the current, with less concern for the actual forward voltage. To some extent this is also true of transistors.

"If there isn't, then saying that a transistor is voltage-controlled doesn't tell us anything about transistors specifically, because by that definition, every component is voltage controlled."

No, a gas discharge tube is not. Neither is a magnetic amplifier.

I don't know about magnetic amplifiers, but that's obviously a special case. Gas discharge tubes aren't ohmic, but their behaviour depends on the applied voltage (and other internal factors), so I'd say they are voltage controlled. What do you say "controls" them?

Is there any type of component that you would describe as current-controlled?

As I explained in an earlier post, diffusion is the primary cause of charges moving from the emitter to the base. When the uncovered charges set up enough back-voltage, the diffusion stops. Applying a Vbe voltage lowers the back voltage and allows charge movement by diffusion to continue. So Vbe controls diffusion which brings charges into the base. Because diffusion is the mechanism for charge movement into the base area, and control of diffusion is controlled by Vbe, the relationship of charge into the base area will be exponential. Once in the base area, the charges are swept into the collector circuit. The collector voltage variation has little influence on the collector current because the emitter-base circuit controls what is available to the collector.

OK, I don't understand the physics, so I'll just say that it's not clear to me that I_BE isn't necessary to cause electrons to flow into the base where most of them can then flow on to the collector.

Let's simplify the question. Do you think of a diode as votage driven or current driven?

I think of a diode as an element that has a defined (exponential) relationship between voltage and current. It doesn't have any obvious controlled quantity, or output quantity, or "primary quantity" as I called it earlier; it doesn't light up; it doesn't have an output current or voltage; it doesn't move; and so on.

It exists simply to provide a useful relationship between voltage and current. In some circuits it's useful to think of it as converting a current to a voltage; in some circuits, the opposite. In some circuits, either, or both, or it's not clear, I guess.

If you "forward-bias" a diode, do you think of applying a voltage and seeing the current flow? Or do you think of running a current through it and watching the voltage set up?

Again it depends on how it's used in the circuit.

If you think of a diode as current-driven you will probably never be able to think of a BJT as voltage-driven.

How about the case where one or more diodes are used to set up an approximate reference voltage? Either in the forward direction, or for a zener diode, in the reverse direction. In that case an approximate current is fed through the diode and an approximate voltage appears across it.

Sure, you need a voltage source to make the current flow, and that voltage source is what provides the voltage that appears across the diode, but that's true of most other components as well, as I've pointed out many times! So if you want to say that physics means components are voltage-driven and current is just a by-product, then fine, but you aren't saying anything interesting about diodes or transistors!

Is that what you're saying?

When you study or learn about semiconductor physics, you can only understand what is actually happening by considering the application of a voltage that raises or lowers the barrier for electrons to flow. Anyone who understands solid-state physics understands that the diode (and the BJT) are voltage driven. Because the voltage causes a current to flow. If you apply a current, the device sets up the appropriate voltage drive that will allow that current to flow.

Sure, I see your point. So all semiconductors are voltage-driven; voltage is required before anything can happen. What about the controlling quantity? Is it helpful to describe a transistor as voltage-controlled when the controlled quantity (I_C) is roughly proportional to I_B, and only exponentially related to V_BE?

If you do not understand the physics of how a diode (or a BJT) works, this may appear to be magic - so the result (current) of the cause (voltage) gets lost. I thought that Ratch's response was right on - he pointed out that in a BJT the collector current is proportional to the base current (assuming beta is a constant - which it isn't) because the exponential terms cancel - he did a masterful job on that. [However, it will only be understood by people who realize that a diode current has an exponential relationship to the applied voltage. Current-driven advocates may have difficulty explaining why]. This is why it appears to the uninitiated that the device is current-driven. But that totally ignores how the device actually works

I know the V-I relationship for a diode, and a base-emitter junction. I also know that h_FE isn't constant, but the I_B vs. I_C graph is a lot closer to a straight line than the V_BE vs. I_C graph is. Doesn't that mean anything? I think it means that it's more useful to think of a transistor as current-controlled, despite the fact that voltage is required to cause current to flow.

It's as if you observe from the outside that a car moves when there is a mechanical push on a gas pedal so you therefore conclude that cars are mechanical-gas-pedal driven, not driven by gasoline-powered engines. Just like you can be happy thinking that the car works by a push on the gas pedal, until you have to fill up with gas (petrol to you), you can take the current-driven approach only to the point at which the device actually starts to influence the behaviour of the circuit. The analogy is not that great, but it does get the main point across - what you see from the outside does not necessarily explain what is going on inside.

I get the analogy. But let me turn it around. You need to push the gas pedal to cause (or in this case, to allow) the fuel to flow, just as you need to apply voltage to cause current to flow. So the "my car is powered by the gas pedal" explanation is now the voltage-driven explanation!

It would help me if you distinguished "controlled" from "driven". And give me an example of a component that you would describe as current-controlled.

Thanks Ian.

Kris

 

[Arouse1973]

I have never been very good at explaining things but what about this analogy. Imagine a river bank with people on both sides. With no base voltage the river is to wide to cross. So when we apply a base voltage the river bank shrinks and people can jump across.

But the people are running all over the place bumping into each other. Some of these people as they get bounced around see an opening in the base terminal and are attracted to that. They fall into the river and swim towards the base.

But its hard going so not many people try this. As the voltage is increased further more people are bounced around and more see the base door, so more try it. But its still hard going so not many more get through.

The majority make it over to the other side. Now this is what happens in an NPN. The people come out of the base and not into it. So its the voltage that closed the river bank and not the people falling into the river.

Sorry if that was a crap analogy

Adam

 

[Ratch]

KrisBlueNZ,

Most of the points you asked about were covered and explained in the thread. You simply have to get a grasp on the physics of the BJT to see how it controls controls its Ic in the active region. Then you can see that Ib is an indicator of the Ic, not the control.

After a gas discharge tube (GDT) starts to glow, there is a wide region where the voltage is constant while the current can vary widely. The current is usually controlled by external resistors. In a magnetic amplifier, a DC sets the point where the transformer core saturates and thus determines the current amplification of the transformer.

Ratch

 

[KrisBlueNZ]

Most of the points you asked about were covered and explained in the thread. You simply have to get a grasp on the physics of the BJT to see how it controls controls its Ic in the active region. Then you can see that Ib is an indicator of the Ic, not the control.

OK. Thank you, and Ian, very much for trying to explain it. I don't understand the physics so I'll try not to make any more judgement calls on whether BJTs are current-controlled or voltage-controlled!

 

[LvW]

"I can see two reasons why it's helpful to describe it that way: (a) VBE causes IB and (b) IC is approximately proportional to IB and is not approximately proportional to VBE."

Hi Kris, I do not want to explain things again - however, I am afraid you are a "victim" of a deep misunderstanding:
I think, in such a technical discussion about device physics it does not matter
- if one of two different sights is "more helpful", or
- if one of two possible equations looks simpler than another one (proportional vs. exponential).
The semiconductor physics has its own rules that cannot be influenced.

Question: Is it really "more helpful" to follow a false explanation which is in contradiction to the observed properties of the most important BJT application: A BJT gain stage ?

The gain of a simple common emitter stage (with RE feedback) is A=-g*Rc/(1+g*RE) (g: tansconductance Ic/Vt).
You can assume beta values between 50 and 200 - the gain is independent on beta. Do you see any indication of current-control?

Kris, in order to understand your position, it would be helpful if you could show/explain during which design steps your view would be "more helpful".

LvW

 

[Laplace]

I'll try not to make any more judgement calls on whether BJTs are current-controlled or voltage-controlled!

That may not be so easy to do because BJTs are both current-controlled and voltage-controlled. As a circuit designer one does not use semiconductor physics to design circuits; one uses linear circuit analysis & synthesis techniques with a BJT circuit model. That make semiconductor physics theory irrelevant to circuit design. There are BJT models that are based on current gain and others that are based on transconductance. A transistor circuit designer needs to choose a BJT model that is appropriate for the particular circuit and the specified operating conditions. It is quite possible that performing a design with either model will arrive at the identical circuit. But ultimately it is the circuit designer who chooses the model to use that determines whether the transistor is current-controlled or voltage-controlled.

 

[KrisBlueNZ]

"I can see two reasons why it's helpful to describe it that way: (a) VBE causes IB and (b) IC is approximately proportional to IB and is not approximately proportional to VBE."

Hi Kris, I do not want to explain things again - however, I am afraid you are a "victim" of a deep misunderstanding:
I think, in such a technical discussion about device physics it does not matter
- if one of two different sights is "more helpful", or
- if one of two possible equations looks simpler than another one (proportional vs. exponential).
The semiconductor physics has its own rules that cannot be influenced.

OK. Can you answer these questions.

Can you name any components (or subcircuits, for that matter) that you consider to be current-controlled?

Is some base current, I_BE, required to cause movement of electrons from the emitter into the base, where they can then be taken by the collector?

Question: Is it really "more helpful" to follow a false explanation which is in contradiction to the observed properties of the most important BJT application: A BJT gain stage?

The gain of a simple common emitter stage (with RE feedback) is A=-g*Rc/(1+g*RE) (g: tansconductance Ic/Vt).
You can assume beta values between 50 and 200 - the gain is independent on beta. Do you see any indication of current-control?

If you're talking about a common emitter amplifier with an emitter degeneration resistor, that's an interesting example, because it's only the emitter resistor (and the negative feedback caused by the collector-emitter current flowing through it) that allows the circuit to amplify linearly.

If you removed the emitter degeneration resistor and applied the input signal directly across the base-emitter junction (at an appropriately low level, of course, and with appropriate bias), the signal at the collector would be horribly distorted, because the transconductance, if you can call it that, isn't even approximately constant.

The transconductance (i_C / v_BE) of a transisistor in common emitter configuration, with base bias and a resistor load for the collector, corresponds to the angle of the exponential i_B vs. v_BE curve at the operating point at the V_BE value in question, right? As the signal shifts this point along the exponential curve, the transconductance is constantly varying, and the output signal (which is proportional to i_C since it appears across the collector resistor) will be horribly distorted, right?

A diode junction has a V-I curve, but calculating its "resistance" at a particular voltage or current isn't usually useful. The i_C vs. v_BE curve of a transistor has the same shape, so describing a transistor's transconductance at one point on that curve is misleading, since it seems to imply that the relationship between v_in and i_out is linear, when it isn't. At least that's how I see it.

You could say it has a nominal transconductance at a particular operating point, and you can even estimate the incremental transconductance in the same way you estimate the incremental resistance of a diode junction, but in a normal common emitter amplifier, transconductance is so variable as to be meaningless. Right?

Things improve when you add back the emitter degeneration resistor. This simply adds negative feedback to the circuit; I_C causes a proportional drop across that resistor, and the transistor is effectively used as a servo controller for I_C. The larger the emitter resistor's value, the less difference the exponential "transconductance" has on the transfer function, and the less distortion you get.

You can also make the common emitter amplifier more useful by adding a resistor in series with the base. The higher its value, the more of the input signal is converted from a voltage into a current, and the more linear the amplifier becomes. This is the way to linearise the amplifier without using negative feedback - by converting v_in into i_in, so the natural proportionality of i_C to i_B provides the linearisation.

Kris, in order to understand your position, it would be helpful if you could show/explain during which design steps your view would be "more helpful".

Yes, I've tried to do that!

 

[LvW]

That may not be so easy to do because BJTs are both current-controlled and voltage-controlled. As a circuit designer one does not use semiconductor physics to design circuits; one uses linear circuit analysis & synthesis techniques with a BJT circuit model. That make semiconductor physics theory irrelevant to circuit design. There are BJT models that are based on current gain and others that are based on transconductance. A transistor circuit designer needs to choose a BJT model that is appropriate for the particular circuit and the specified operating conditions. It is quite possible that performing a design with either model will arrive at the identical circuit. But ultimately it is the circuit designer who chooses the model to use that determines whether the transistor is current-controlled or voltage-controlled.

Although I have learnd that the forum moderation does not want to continue this thread (https://www.electronicspoint.com/threads/question-about-transistor.272142/#post-1637925) I think it is important to understand the position of the "current-control party".
Laplace - therefore, I kindly ask you to answer my question:
Please, can you give me one single circuit example that can explain what you mean? ("semiconductor physics theory irrelevant to circuit design, designer needs to choose a BJT model that is appropriate for the particular circuit").
I cannot believe that you really think that "semiconductor physics theory is irrelevant to circuit design" - as long as the basic control mechanism is concerned.

 

[LvW]

Kris - my answers are in bold within your text.

OK. Can you answer these questions.
Can you name any components (or subcircuits, for that matter) that you consider to be current-controlled?

We speak about components only! Of course, there are circuits (e.g. Current-feedback amplifiers, CFA) which react upon a current through the controlling nodes).

Is some base current, I_BE, required to cause movement of electrons from the emitter into the base, where they can then be taken by the collector?

No - it is not. This effect is not "required", but cannot be avoided because nothing is ideal (same applies to FET`s)

If you're talking about a common emitter amplifier with an emitter degeneration resistor, that's an interesting example, because it's only the emitter resistor (and the negative feedback caused by the collector-emitter current flowing through it) that allows the circuit to amplify linearly.

No - the resistor RE only improves linearity (at the cost of gain).

If you removed the emitter degeneration resistor and applied the input signal directly across the base-emitter junction (at an appropriately low level, of course, and with appropriate bias), the signal at the collector would be horribly distorted, because the transconductance, if you can call it that, isn't even approximately constant.

No - "at an appropriate level" the distortion can be made equally small. (However, what has this property to do with our question?)

The transconductance (i_C / v_BE) of a transisistor in common emitter configuration, with base bias and a resistor load for the collector, corresponds to the angle of the exponential i_B vs. v_BE curve at the operating point at the V_BE value in question, right? As the signal shifts this point along the exponential curve, the transconductance is constantly varying, and the output signal (which is proportional to i_C since it appears across the collector resistor) will be horribly distorted, right?

The transconductance is the SLOPE of the Ic=f(Vbe) curve (in the selected operational point). And - yes - because this curve is not linear, we have to live with distortions.
So what? May be we do not like this property - that is the motivation for signal feedback.

A diode junction has a V-I curve, but calculating its "resistance" at a particular voltage or current isn't usually useful.

??? On contrary, I think there are many cases where the static resistance (and the corresponding voltage drop) as well as the differential resistance (and the corr. voltage drop) are important design parameters.

The i_C vs. v_BE curve of a transistor has the same shape, so describing a transistor's transconductance at one point on that curve is misleading, since it seems to imply that the relationship between v_in and i_out is linear, when it isn't. At least that's how I see it.

Misleading? But the voltage GAIN depends on this transconductance. And somebody who thinks that this transconductance is valid for all signal levels will soon learn that he was wrong (see distortion aspects mentioned above.

You could say it has a nominal transconductance at a particular operating point, and you can even estimate the incremental transconductance in the same way you estimate the incremental resistance of a diode junction, but in a normal common emitter amplifier, transconductance is so variable as to be meaningless. Right?

Yes - because of linearization effects. Was discussed above; no new argument.

Things improve when you add back the emitter degeneration resistor. This simply adds negative feedback to the circuit; I_C causes a proportional drop across that resistor, and the transistor is effectively used as a servo controller for I_C. The larger the emitter resistor's value, the less difference the exponential "transconductance" has on the transfer function, and the less distortion you get.

You are mentioning "negative feedback". Do you realize that a VOLTAGE is fed back to the emitter node?

You can also make the common emitter amplifier more useful by adding a resistor in series with the base. The higher its value, the more of the input signal is converted from a voltage into a current, and the more linear the amplifier becomes. This is the way to linearise the amplifier without using negative feedback - by converting v_in into i_in, so the natural proportionality of i_C to i_B provides the linearisation.

Using such a resistor you do nothing else than to create a voltage division between this resistor and the base-emitter path. But it does not change transistor physics.

My summary is as follows: In all of your statements you say (more or less directly/indirectly) that you are not happy with the fact that the relation Ic=f(Vbe) is not linear.
I think, this applies to all of us. But - is this the question? Of course, we would appreciate if we could design an amplifier with high gain and low distortion based on a device that is more linear. However,....

Regards
LvW

Remark: Perhaps it would be helpful to read again my post#33. Here I have listed the main steps for designing a common-emitter gain stage. And, Kris, it would be helpful for me if you could answer (a) if you agree or if you normally are applying other design steps, and (b) at which point/step a current-control model would be easier to use .

 

[Ratch]

KrizBlueNZ,

Just a couple of comments. The nonlinearity of a BJT is the reason that transistor design textbooks aver that the input signal should not be more than 20 mV P-P. This small input approximates a straight line response. Anytime you use a current generator to drive a BJT, you are introducing a large external resistor and creating a transistor circuit. Nothing wrong with that, but you cannot then say it proves the BJT is a current amplifier. The circuit itself might be a current amplifier, but the BJT device is still a transconductance amplifier.

Ratch

 

[LvW]

The circuit itself might be a current amplifier, but the BJT device is still a transconductance amplifier.

As another example: The classical opamp (undoubtly voltage driven) very often is used to realize a current-to-voltage converter (current in, voltage out).

 

[Laplace]

I cannot believe that you really think that "semiconductor physics theory is irrelevant to circuit design" - as long as the basic control mechanism is concerned.

Semiconductor physics are not relevant to me because I do not have a fab facility at my disposal, nor is a deep understanding of physics necessary to read a manufacturer's datasheet. But whichever model one uses to design a particular circuit, the goal is always to make the performance dependent on the passive circuit components and minimize the effect of non-linear characteristics in the active device. While it was a long time ago I seem to remember being taught to use current gain to set the bias point and use transconductance to analyze gain performance, but my preference as a hobbyist is to design with IC's and use the BJT only as a current switch. So tell me how to drive a transistor into saturation using anything other than the base current.