Understanding Transistor

[crutschow]

You have to delve into the relationship between Ic, Ib, and Vbe to determine what is cause and effect.

Ratch

For a pedantic physicist, maybe. But as a circuit design engineer I'm more interested in how a device appears to the outside world, not what its esoteric inner quantum-mechanical "cause and effect" is, and for most common low-frequency analog and switching design applications, a BJT looks like a current operated device with a low (albeit non-linear) input impedance. To tell a newbie otherwise, just confuses them as to how to design the transistor into a circuit.

 

[Ratchit]

crutschow,

For a pedantic physicist, maybe. But as a circuit design engineer I'm more interested in how a device appears to the outside world, not what its esoteric inner quantum-mechanical "cause and effect" is, and for most common low-frequency analog and switching design applications, a BJT looks like a current operated device with a low (albeit non-linear) input impedance.

You asked how to tell a chicken from an egg, so I showed you how. I have always said a BJT mimics a CCCS even though it is really a VCCC.

To tell a newbie otherwise, just confuses them as to how to design the transistor into a circuit.

No it doesn't. You have to give newbies more credit than that. It is really not that confusing or complicated. They should know how a BJT works before they try to design something.

Ratch

 

[Winterstone]

To tell a newbie otherwise, just confuses them as to how to design the transistor into a circuit.

Hi Crutschow,

since I have already delivered some time ago one or two contributions in this thread I like to jump once more into the discussion with one single question:
Can you please explain to me WHY it should confuse a newbie to hear that a BJT is voltage controlled rather than current controlled (that is question Nr. 1)?

I was engaged in educating engineers for a couple of years - and according to my experience it is simpler, more logical and the only correct way to explain the behaviour of the BJT based on voltage control.
This has absolutely nothing to do with pedantry or "quantum-mechanical cause-and effect" thinking - it is simply the physical truth.
(By the way: Is "pedantry" a bad property as far as physical-techical questions are concerned? I think, the opposite is true. Pedantry in terms and units and mathematics is absolutely necessary in order to avoid errors and misunderstandings)

Prior to my educational activities I was working as a design and system engineer, but I never had problems to match the physical correct description of the BJT principle
with my observations "how the device appears to the outside world".
Therefore question Nr. 2: Can you please verify inhowfar the BJT appears to the outside world as current-controlled rather than voltage controlled?
Or do you just refer to several books which simply state Ic=h21*Ib.

In this context: I think in most cases a transistor stage working as an amplifier is biased with a low-resistive voltage divider - thus approaching the behaviour of a voltage source.
In such a case, why should I try to explain its function using a current-controlled scheme?
Thank you.
W.

 

[crutschow]

Hi Crutschow,

since I have already delivered some time ago one or two contributions in this thread I like to jump once more into the discussion with one single question:
Can you please explain to me WHY it should confuse a newbie to hear that a BJT is voltage controlled rather than current controlled (that is question Nr. 1)?

I was engaged in educating engineers for a couple of years - and according to my experience it is simpler, more logical and the only correct way to explain the behaviour of the BJT based on voltage control.
This has absolutely nothing to do with pedantry or "quantum-mechanical cause-and effect" thinking - it is simply the physical truth.
(By the way: Is "pedantry" a bad property as far as physical-techical questions are concerned? I think, the opposite is true. Pedantry in terms and units and mathematics is absolutely necessary in order to avoid errors and misunderstandings)

Prior to my educational activities I was working as a design and system engineer, but I never had problems to match the physical correct description of the BJT principle
with my observations "how the device appears to the outside world".
Therefore question Nr. 2: Can you please verify inhowfar the BJT appears to the outside world as current-controlled rather than voltage controlled?
Or do you just refer to several books which simply state Ic=h21*Ib.

In this context: I think in most cases a transistor stage working as an amplifier is biased with a low-resistive voltage divider - thus approaching the behaviour of a voltage source.
In such a case, why should I try to explain its function using a current-controlled scheme?
Thank you.
W.

Nr. 1. If you are educating engineers then I agree, they need to understand the quantum-mechanical operation of a transistor as a voltage controlled device. But I'm not talking about newbie engineers. I'm talking about of the newbies on this forum who have little or no training in electronics and need the most basic information possible to understand how to use a transistor. Saying that a transistor is voltage controlled but it appears to be current controlled just adds additional confusion. But that's my opinion.

If you look up the meaning of pedantic, I think you will understand my use of the term. I'm not against the precise use of technical terms and mathematical descriptions, where appropriate. It's the "where appropriate" that is the issue.

I worked as an analog design engineer for 47 years and I always used the current-controlled characteristics of a BJT when designing with them or analyzing their circuit operation (of course I never did RF or IC design where I realize viewing it as voltage controlled device may be more appropriate). For low frequency and switching applications trying to view a BJT as a voltage controlled device when it basically is acting as a current-controlled device just adds a layer of unneeded complexity to the design effort.

Nr. 2. I don't quite understand your question about "verify inhowfar the BJT appears to the outside world as current-controlled rather than voltage controlled"? I thought that was one of the main points of this whole discussion -- whether a BJT appears to be a voltage controlled or current controlled device.

My simple criteria of this is how the device primarily appears to the outside world. If it has a relatively high input impedance and the output is basically proportional to the input voltage than it appears to be a voltage controlled device. If it has a relatively low input impedance and the output is basically proportional to the input current, then it appears to be a current controlled device. You can quibble with that definition if you like, but that's what I use for the purposes of this discussion.

If a BJT is "biased with a low-resistive voltage divider - thus approaching the behaviour of a voltage source", is seems that is because you are trying to convert an (external) current-controlled device to be more like a voltage-controlled device. When you calculate those bias resistor values, (and I assume you have done that yourself at some point) it is the value of the current-gain (β or Hfe) that is used to determine appropriate resistor values and ratios, not the variation of Vbe with current (as a voltage-controlled device). It would seem a very convoluted task to design a proper BJT bias circuit without reference to its current gain.

In conclusion, I have no particular quibble with the facts of how a BJT transistor really works internally. It's something that any good engineer should know. But for newbies, who may barely understand Ohm's Law, going to that level of detail is likely confusing. They want to know how it appears to them when they use it in a circuit and, for many typical applications of a BJT, it appears primarily as a current-operated device.

 

[Claude Abraham]

crutschow



You have to delve into the relationship between Ic, Ib, and Vbe to determine what is cause and effect.

Ratch

Anybody with a signal generator & a scope can easily demolish the cause/effect theory. Ib/Ie change before Vbe, & Ic responds to changes in Ie. Many critics of current control (herein called critics) constantly build their case by showing that Ib does not cause Ic. Let me emphasize that nobody ever stated, nor even implied, that Ib is the cause of Ic because it is not. Ic is controlled by Ie, the emitter current. Emitter ejects electrons (npn) towards the base, which transit through base due to being ultra thin, then get collected by collector. Ic is controlled by Ie. If the emitter injects 1,000 electrons in 1.0 picosecond, the collector cannot collect 1,010 e- in that same picosecond (unless stored charge was involved).

Let's say Ie is increased via a signal increase from a transducer (microphone, radio signal, etc.). An increase in current at the b & e terminals takes place, then as charges transit through junction, Vbe increases as a consequence. Vbe is incidental it does not control anything. A scope & generator is easy to use to observe this action. The only way that Vbe could literally control Ib/Ie/Ic would be for a constant voltage source to be connected directly across the b-e terminals, which is never done. A bjt is thermally unstable when voltage source driven. The scenario you describe, Vbe controlling Ib/Ie/Ic can not be implemented in practice because p-n junctions have a thermal coefficient that results in more conductivity as temp increases.

Regarding saturated switch operation, all physics texts use charge control model, and so do I. It is the only model that works. Sotrage time, fall time, rise time, & delay time, as well as reverse recovery charge are the important parameters when bringing a diode, or bjt out of saturation. Charge control is also used for FETs, it works great.

Current control model is best for linear region, low speed operation (below ft limitations), charge control is best otherwise. For 60 plus years that is how they have been modeled, & with superb results. Trying to rewrite the book is futile. Every argument put forth in favor of VC collapses on its own weight. These arguments were thoroughly refuted in the above thread. Peace to all.

 

[Mr RB]

... But the cause is really the Vbe affecting both the Ic and Ib.
...
As long as the BJT is operating in the active region, its Ic and Ib are dependent on Vbe, according to Sedra and Smith, and others. That is the physics of a BJT in the active region.
...

Well, I'm not a physicist that works with the insides of transistors, I'm just an electronics guy that connects them to things.

BUT I have noticed that a transistor with 0.6v Vbe and 1mA Ib has Ic of 100mA.

Then when the transistor gets hot it's Vbe might drop from 0.6v to 0.55v... But the Ib is still 1mA and Ic still 100mA.

In my simple non-physicist "guy who connects stuff" way that is saying the transistor is a current controlled device and the Vbe is somewhat irrelevant.

 

[atferrari]

Chiming in once

Why is the proverbial "newbie" considered upfront unable to understand the actual thing? Not all of them are barely understanding the Ohm's law.

To be honest, I find that kind of intelectual paternalism. Or an improper feeling where "the only one who truly understands things is me"?

It is true that some will never pass from the cookbook level but why to consider the rest will be the same?

By now you know (after myself posting for so long in these forums) I am not a professional of this game. Far from that. And I consider myself a newbie, to the point that I rarely explain something here (and ask questions, a lot). Yes, go to my profile.

Countless times I tried to understand things with varied success, as everybody in life.

Whether happy because I managed to grasp it or sad because it was above my capability I know I had the chance.

I tend to prefer a complete explanation in the beginning even if a tad on the complex side.

Sure later, as somebody posted before, I will assemble my own personal model of "how thinks work" oriented to practical applications.

But OK, that is me.

 

[Ratchit]

Claude,

Anybody with a signal generator & a scope can easily demolish the cause/effect theory. Ib/Ie change before Vbe, & Ic responds to changes in Ie.

Not if the frequency is low and the storage capacitance is small or missing. How can you not take into account the Sedra and Smith equations I posted earlier?

Ic is controlled by Ie, the emitter current. Emitter ejects electrons (npn) towards the base, which transit through base due to being ultra thin, then get collected by collector. Ic is controlled by Ie. If the emitter injects 1,000 electrons in 1.0 picosecond, the collector cannot collect 1,010 e- in that same picosecond (unless stored charge was involved).

Since the emitter current contains around 99.44% of the collector current, saying the emitter current controls the collector current is a circular explanation. Although storage capacitance is present in BJTs, their operation does not depend on it. The BJT depends on the diffusion of the charges from the emitter to base, and Vbe is used to lower the opposing voltage from the uncovered charges. If Vbe were not present, the diffusion and Ic will stop when the diffusion and opposing voltage are in equilibrium. For that reason, Vbe is what controls Ic by lowering the opposing voltage caused by the uncovered charges.

Let's say Ie is increased via a signal increase from a transducer (microphone, radio signal, etc.). An increase in current at the b & e terminals takes place, then as charges transit through junction, Vbe increases as a consequence. Vbe is incidental it does not control anything.

No, Vbe is what controls Ib and Ic, and thereby Ie. That is easily shown by setting up the transistor in the active region with a voltage supply and varying the Vbe.

A scope & generator is easy to use to observe this action.

As I said before, let's us keep in the DC realm so that we don't have to worry about capacitive storage effects.

The only way that Vbe could literally control Ib/Ie/Ic would be for a constant voltage source to be connected directly across the b-e terminals, which is never done.

No, it is not. But even driving it with a current, the Vbe shows a one to one correspondence with the Ic and Ib.

A bjt is thermally unstable when voltage source driven. The scenario you describe, Vbe controlling Ib/Ie/Ic can not be implemented in practice because p-n junctions have a thermal coefficient that results in more conductivity as temp increases.

Vbe does control Ib,Ie,Ic whether the BJT is current driven or voltage driven. The variation of conductivity is present whether current driven or voltage driven.

Regarding saturated switch operation, all physics texts use charge control model, and so do I. It is the only model that works. Sotrage time, fall time, rise time, & delay time, as well as reverse recovery charge are the important parameters when bringing a diode, or bjt out of saturation. Charge control is also used for FETs, it works great.

As I said before, I am going to stick in the active region of the BJT.

Current control model is best for linear region, low speed operation (below ft limitations), charge control is best otherwise. For 60 plus years that is how they have been modeled, & with superb results. Trying to rewrite the book is futile. Every argument put forth in favor of VC collapses on its own weight. These arguments were thoroughly refuted in the above thread. Peace to all.

That is because Vbe vs Ic is nonlinear, so putting lots of resistance in be base circuit linearizes the response, and makes the base circuit a current driven circuit. That does not take away the fact that the BJT is a voltage controlled current sourse, where the Vbe is controlling both Ib and Ic. No one is rewriting the book on how BJT's work, that was done a long time ago. The fact that Vbe controls both Ic and Ib is shown in Sedra and other books, and is also explained by the physics of the BJT. This explanation was not refuted by any means in the above thread, although many tried mightly to do so.

Ratch

 

[Ratchit]

Mr. RB,

Well, I'm not a physicist that works with the insides of transistors, I'm just an electronics guy that connects them to things.

Transistors will work the same whether you are a tinker or physicist.

BUT I have noticed that a transistor with 0.6v Vbe and 1mA Ib has Ic of 100mA.

Then when the transistor gets hot it's Vbe might drop from 0.6v to 0.55v... But the Ib is still 1mA and Ic still 100mA.

As you must already know, or have come across, the Vbe of a junction diode drops about 2.5 mv of every degree C increase. So current comparisions have to be made at the same temperature or adjusted for the difference.

In my simple non-physicist "guy who connects stuff" way that is saying the transistor is a current controlled device and the Vbe is somewhat irrelevant.

False conclusion. There is still a one to one relationship between the Vbe and the Ic according to Sedra and Smith which I referenced in a previous post. Of course you have to adjust for temperature difference.

Ratch

 

[crutschow]

Why is the proverbial "newbie" considered upfront unable to understand the actual thing? Not all of them are barely understanding the Ohm's law.

To be honest, I find that kind of intelectual paternalism. Or an improper feeling where "the only one who truly understands things is me"?

It is true that some will never pass from the cookbook level but why to consider the rest will be the same?

......................

I don't know "why' but if you've read many of the newbies (or perhaps more correctly, novice) posts you will see that many (but not all) do not understand the fundamentals of circuits or transistors. That's not being paternal, just realistic. For these, I think the simplest explanation possible on how a BJT works externally is the best. And it's only for rare design applications, that you need to understand that a BJT is fundamentally not a current-controlled device but voltage-controlled one. Few newbies' designs fall in that category.

If you can give me a design example in these forums where it would be easier (or better) to do the design if you treated the BJT as voltage-controlled rather than current-controlled device then I will reconsider my opinion. (None of my designs have even fallen into that category and I've done a few). Otherwise I am staying with my approach of treating a BJT as externally behaving as a current-controlled device. That may not correspond to its internal physics but, as a designer, I'm more interested in what's the easiest way to view transistor operation as a black-box rather then what's in the black-box.

 

[Winterstone]

Hi Crutschow,
I must confess that I am not quiet satisfied with your answer - sorry. let me explain:

Nr. 1. If you are educating engineers then I agree, they need to understand the quantum-mechanical operation of a transistor as a voltage controlled device. But I'm not talking about newbie engineers. I'm talking about of the newbies on this forum who have little or no training in electronics and need the most basic information possible to understand how to use a transistor. Saying that a transistor is voltage controlled but it appears to be current controlled just adds additional confusion. But that's my opinion.

I agree with you up to 100% - provided that the BJT really "appears to be current controlled". I never want to make things more complicated as they are - however, in this case, I really cannot see the advantage of telling newbies that BJT are current-controlled. Why should I ? Is the mechanism of current control easier to understand than voltage control? I do not need any quantum-mechanical secrets for explaining the (simplified) operation of a BJT.
Remember my 2nd question: Why do you think the BJT "appears to be current-controlled"? Because of the formula Ic=beta*Ib ? Or which physical mechanism is the cause of this claim?

I worked as an analog design engineer for 47 years and I always used the current-controlled characteristics of a BJT when designing with them or analyzing their circuit operation (of course I never did RF or IC design where I realize viewing it as voltage controlled device may be more appropriate).

Question: Does this apply also for feedback due to a emitter resistor? For my opinion - and I am in accordance with all authors (as far as I know) - this is a kind of current-controlled voltage feedback.
And - as a consequence - the equivalent closed-loop block diagram contains at its input the difference between V(basis( and V(emitter).
How do you explain this Re-feedback scheme based on the current-control principle? I never have seen a corresponding block diagram showing a current-differencing unit at its input.
If the current Ib really controls the amount of Ic such a diagram should be possible!

My simple criteria of this is how the device primarily appears to the outside world. If it has a relatively high input impedance and the output is basically proportional to the input voltage than it appears to be a voltage controlled device. If it has a relatively low input impedance and the output is basically proportional to the input current, then it appears to be a current controlled device. You can quibble with that definition if you like, but that's what I use for the purposes of this discussion.
.

OK, I understand what you mean. But this sounds to me as if the rule Ic=beta*Ib is the only justification for the assumption of current-control. And. for my opinion, that's not enough.
Do you think that it is really more complicated and hard to understand when I say: Similar to the pn-diode there is an exponential relation between Ic and Vbe - and the base current (as an unwanted but unavoidable current) is a nearly fixed percentage of Ic ?

If a BJT is "biased with a low-resistive voltage divider - thus approaching the behaviour of a voltage source", is seems that is because you are trying to convert an (external) current-controlled device to be more like a voltage-controlled device. When you calculate those bias resistor values, (and I assume you have done that yourself at some point) it is the value of the current-gain (β or Hfe) that is used to determine appropriate resistor values and ratios, not the variation of Vbe with current (as a voltage-controlled device). It would seem a very convoluted task to design a proper BJT bias circuit without reference to its current gain.

Yes. of course - by determining the bias resistors I use the desired Ic to calculate the corresponding Ib. And - as the next step - the impedance level of the resistive divider is fixed (as low as possible with respect to the total input resistance) in order to enable a control voltage in conjunction with the emitter resistor Re. I think this goal (voltage control between B and E) is the only reason for a relatively low-resistive divider.
If not - larger bias resistors would be chosen (because of current consumption and total input resistance). But by doing this I see no necessity to assume current-control. I only use the relation Ic=beta*Ib. That's all.

In conclusion, I have no particular quibble with the facts of how a BJT transistor really works internally. It's something that any good engineer should know. But for newbies, who may barely understand Ohm's Law, going to that level of detail is likely confusing. They want to know how it appears to them when they use it in a circuit and, for many typical applications of a BJT, it appears primarily as a current-operated device.

If newbies accept W. Shockleys equation describing the pn junction of a diode they also will accept the same mathematical formula for a voltage controlled BJT. This is my experience (not very surprising).

Let me come to an end with the following: I have encouraged my students always to ask "WHY" (in case I have claimed something) because they should not believe but understand.
Now, assuming that I have claimed that the BJT would be a current-controlled device. What should I answer if they again ask "WHY"?

Regards
W.

 

[Ratchit]

atferrari,

Why is the proverbial "newbie" considered upfront unable to understand the actual thing?

I have run into that attitude on this and other forms. It is just like manufacturers arguing with regulators like "This new way of disclosing fat and other bad things in our product is going to confuse our customers who cannot understand it". Or perhaps like politicians in your country and mine saying, "This is going to be the best for the people".

To be honest, I find that kind of intelectual paternalism. Or an improper feeling where "the only one who truly understands things is me"?

Don't worry, those to espouse such an attitude are expelling blue smoke.

It is true that some will never pass from the cookbook level but why to consider the rest will be the same?

By now you know (after myself posting for so long in these forums) I am not a professional of this game. Far from that. And I consider myself a newbie, to the point that I rarely explain something here (and ask questions, a lot). Yes, go to my profile.

Countless times I tried to understand things with varied success, as everybody in life.

Whether happy because I managed to grasp it or sad because it was above my capability I know I had the chance.

I tend to prefer a complete explanation in the beginning even if a tad on the complex side.

Sure later, as somebody posted before, I will assemble my own personal model of "how thinks work" oriented to practical applications.

But OK, that is me.

Good for you.

Ratch

 

[Winterstone]

Anybody with a signal generator & a scope can easily demolish the cause/effect theory. Ib/Ie change before Vbe, & Ic responds to changes in Ie.
......................
Ic is controlled by Ie, the emitter current. Emitter ejects electrons (npn) towards the base, which transit through base due to being ultra thin, then get collected by collector. Ic is controlled by Ie. If the emitter injects 1,000 electrons in 1.0 picosecond, the collector cannot collect 1,010 e- in that same picosecond (unless stored charge was involved).

Let's say Ie is increased via a signal increase from a transducer (microphone, radio signal, etc.). An increase in current at the b & e terminals takes place, then as charges transit through junction, Vbe increases as a consequence. Vbe is incidental it does not control anything. A scope & generator is easy to use to observe this action. The only way that Vbe could literally control Ib/Ie/Ic would be for a constant voltage source to be connected directly across the b-e terminals, which is never done. A bjt is thermally unstable when voltage source driven. The scenario you describe, Vbe controlling Ib/Ie/Ic can not be implemented in practice because p-n junctions have a thermal coefficient that results in more conductivity as temp increases.

Regarding saturated switch operation, all physics texts use charge control model, and so do I. It is the only model that works. Sotrage time, fall time, rise time, & delay time, as well as reverse recovery charge are the important parameters when bringing a diode, or bjt out of saturation. Charge control is also used for FETs, it works great.

Current control model is best for linear region, low speed operation (below ft limitations), charge control is best otherwise. For 60 plus years that is how they have been modeled, & with superb results. Trying to rewrite the book is futile. Every argument put forth in favor of VC collapses on its own weight. These arguments were thoroughly refuted in the above thread. Peace to all.

Hi Claude,
I do not know how to respond to this contribution. Hopefully, some others do.
Only one question: You claim The emitter ejects electrons towards the base.

Of course, everybody will agree. But my question is "why"? You forgot to mention the physical cause.

Regards.
W.

 

[Winterstone]

Then when the transistor gets hot it's Vbe might drop from 0.6v to 0.55v... But the Ib is still 1mA and Ic still 100mA.
In my simple non-physicist "guy who connects stuff" way that is saying the transistor is a current controlled device and the Vbe is somewhat irrelevant.

Hi Mr. RB,

I don't know if you have some other arguments to support the assumption of current-control.
But the above explanation is not correct.
It is not true that "Vbe might drop from 0.6v to 0.55v... But the Ib is still 1mA and Ic still 100mA.
I am afraid that you have interpreted the corresponding diagrams incorrectly.

The correct sequence is as follows:
* With rising temperature (and constant Vbe) the current Ic goes up.
* If the value of Ic shall be held constant the voltage Vbe must be reduced correspondingly (extern!). This leads to the well-known value of approx. -2mV/deg.

For my opinion, this temperature phenomenon even supports the voltage-controlled property of the BJT. And - of course - Vbe is not "irrelevant".

Regards
W.

 

[Winterstone]

If you can give me a design example in these forums where it would be easier (or better) to do the design if you treated the BJT as voltage-controlled rather than current-controlled device then I will reconsider my opinion. .

Hello crutschow,

For my opinion, I already gave you such an example - signal and dc feedback (negative) caused by an emitter resistor Re.
And you certainly will agree that this is not an exotic example (other examples are the current mirror and Barry Gilbert's translinear loops).
Or can you explain the results of this feedback without using the relation between Ic and Vbe ?

W.

 

[crutschow]

I agree with you up to 100% - provided that the BJT really "appears to be current controlled". I never want to make things more complicated as they are - however, in this case, I really cannot see the advantage of telling newbies that BJT are current-controlled. Why should I ? Is the mechanism of current control easier to understand than voltage control? I do not need any quantum-mechanical secrets for explaining the (simplified) operation of a BJT.
Remember my 2nd question: Why do you think the BJT "appears to be current-controlled"? Because of the formula Ic=beta*Ib ? Or which physical mechanism is the cause of this claim?

Yes, in my opinion it is easier to view the BJT as current controlled device since that is how it appears externally to a first order. I've repeatedly said that this is an an external characteristic, not a description of the actual physics of the device


Question: Does this apply also for feedback due to a emitter resistor? For my opinion - and I am in accordance with all authors (as far as I know) - this is a kind of current-controlled voltage feedback.
And - as a consequence - the equivalent closed-loop block diagram contains at its input the difference between V(basis( and V(emitter).
How do you explain this Re-feedback scheme based on the current-control principle? I never have seen a corresponding block diagram showing a current-differencing unit at its input.
If the current Ib really controls the amount of Ic such a diagram should be possible!

You said it yourself. It's voltage feedback from the emitter resistor. It has nothing to do with whether the transistor is current or voltage controlled (the same basic feedback equation works for a BJT or a FET). To a first approximation this is normally the dominate voltage, and the Vbe term and the Ibe term are ignored in the feedback equation.

OK, I understand what you mean. But this sounds to me as if the rule Ic=beta*Ib is the only justification for the assumption of current-control. And. for my opinion, that's not enough.
Do you think that it is really more complicated and hard to understand when I say: Similar to the pn-diode there is an exponential relation between Ic and Vbe - and the base current (as an unwanted but unavoidable current) is a nearly fixed percentage of Ic ?

Yes, I believe that in the analysis of simple BJT circuits it's easier to understand circuit operation from a current point-of-view. The base current may be unavoidable but it is always there and must always be accounted for. The small change in Vbe with base/collector current, on the other hand, can often be safely ignored in many designs.


Yes. of course - by determining the bias resistors I use the desired Ic to calculate the corresponding Ib. And - as the next step - the impedance level of the resistive divider is fixed (as low as possible with respect to the total input resistance) in order to enable a control voltage in conjunction with the emitter resistor Re. I think this goal (voltage control between B and E) is the only reason for a relatively low-resistive divider.
If not - larger bias resistors would be chosen (because of current consumption and total input resistance). But by doing this I see no necessity to assume current-control. I only use the relation Ic=beta*Ib. That's all.

Yes, you need to use the current-mode equation Ic = beta*Ib and that's all you need. There's no reason to even think about ΔVbe (of course the Vbe fixed bias voltage must be included). The main reason for a low impedance divider is so that the base current does not duly affect the bias point. Otherwise it could be arbitrarily high impedance since the input signal appear directly at the base, unaffected by the bias resistors. It has nothing to do with (voltage control between B and E).



If newbies accept W. Shockleys equation describing the pn junction of a diode they also will accept the same mathematical formula for a voltage controlled BJT. This is my experience (not very surprising).

Let me come to an end with the following: I have encouraged my students always to ask "WHY" (in case I have claimed something) because they should not believe but understand.
Now, assuming that I have claimed that the BJT would be a current-controlled device. What should I answer if they again ask "WHY"?

Again, I have no problems with teaching your students the correct internal physical operation of a BJT transistor. But, if those students are going to design circuits they need to understand that the BJT acts like a current-controlled device for many (but not all) practical circuits (current mirrors and the Gilbert cell are notable exceptions). If you've ever done much circuit design, I would think you would understand that. But the tenor of your posts implies not. Perhaps that accounts for the difference between an academia and an engineering point-of-view.

My answers are in blue.

 

[Winterstone]

Quotation (post 76): Again, I have no problems with teaching your students the correct internal physical operation of a BJT transistor. But, if those students are going to design circuits they need to understand that the BJT acts like a current-controlled device for many (but not all) practical circuits (current mirrors and the Gilbert cell are notable exceptions). If you've ever done much circuit design, I would think you would understand that. But the tenor of your posts implies not. Perhaps that accounts for the difference between an academia and an engineering point-of-view

Hi chrutschow,
I am a bit disappointed about the last three sentences of your above reply. Why do you use such formulations? Shouldn't we try to discuss on a fair technical basis?

Anyway, I appreciate that you now agree that the physical truth is "voltage-control". And I can assure you that you are in full accordance with reputable persons like Bob Pease, Jim Williams, Barry Gilbert, Paul Brokav.
Not to forget: W. Shockley with his equation describing the current-voltage relation across the pn junction.

That means: The remaining question is only if we should confront the beginners/newbies/students with the physical truth (voltage-control) or with another view that is based solely on the relation Ic=beta*Ib (current control). But - in the latter case - one must be aware that the question comes up: Why is it so? Can you explain?

For my opinion, we could end the discussion now because no new arguments are to be expected
(except perhaps some new sights from Claude Abraham).

Nevertheless, I like to comment some parts of your replies:

Quotation (post 64): My simple criteria of this is how the device primarily appears to the outside world. If it has a relatively high input impedance and the output is basically proportional to the input voltage than it appears to be a voltage controlled device. If it has a relatively low input impedance and the output is basically proportional to the input current, then it appears to be a current controlled device. You can quibble with that definition if you like, but that's what I use for the purposes of this discussion.

This is your definition? And who set's the threshold - and at which input resistance?
That means: According to your "definition" a BJT "appears" as voltage controlled for a small Ic with a large h11 of - let's say - 100 kohms ? And suddenly - for higher currents - it appears as current controlled?
Are you really satisfied with this "definition"? I hope not.
And because you earlier spoke about confusion on beginners side: Don`t you think it confuses somebody to hear that sometimes the BJT "appears" as voltage-controlled (large input resistor) and sometimes as current controlled?

Quotation (post 76): It's voltage feedback from the emitter resistor. It has nothing to do with whether the transistor is current or voltage controlled (the same basic feedback equation works for a BJT or a FET). To a first approximation this is normally the dominate voltage, and the Vbe term and the Ibe term are ignored in the feedback equation.

It has "nothing to do with...current or voltage control"? And the "Vbe term" is "ignored"? Who ignores this term? I cannot follow you.
Is it really necessary to explain the mechanism of voltage feedback via Re?
When a low-resistive voltage divider keeps the base potential Vb constant and when the emitter voltage Ve across Re goes up (due to an increase of Ie) the voltage Vbe across the B-E junction reduces correspondingly.
That is how the feedback works: The reduction in Vbe works against the Ic increase. How can you say Vbe would be "ignored". You will have noticed that it was not at all necessary to mention the base current Ib during description of this feedback principle.

Quotation (post 76): Yes, in my opinion it is easier to view the BJT as current controlled device since that is how it appears externally to a first order. I've repeatedly said that this is an an external characteristic, not a description of the actual physics of the device

Repeatedly I ask : Please explain what do you mean with "appears externally". Up to now your only justification was the formula Ic=beta*Ib. Or do you refer to measurements or simulations (external characteristics")?
Of course I can verify the equation Ic=beta*Ib with some measurements. But the same applies to Ic=f(Vbe). Thus, this cannot be your argument for "extern appearance".

Quotation (post 76):Yes, you need to use the current-mode equation Ic = beta*Ib and that's all you need.

No, again I disagree. Assume you have a simple BJT amplifying stage with given values for all resistors and power supply. Of course, the exact BJT parameters are unknown.
Now the question: Is it possible by hand calculation to estimate the current Ic ? Answer: Yes, it is - and the only assumption you need is Vbe=0.6...07 volts.
For sufficient Re feedback it even is not important if you assume 0.6 or 0.7 volts.
But the relation Ic=beta*Ib does not help at all.
_______________________-
Thanks and regards
W.

 

[crutschow]

Quotation (post 64): Again, I have no problems with teaching your students the correct internal physical operation of a BJT transistor. But, if those students are going to design circuits they need to understand that the BJT acts like a current-controlled device for many (but not all) practical circuits (current mirrors and the Gilbert cell are notable exceptions). If you've ever done much circuit design, I would think you would understand that. But the tenor of your posts implies not. Perhaps that accounts for the difference between an academia and an engineering point-of-view

Hi chrutschow,
I am a bit disappointed about the last three sentences of your above reply. Why do you use such formulations? Shouldn't we try to discuss on a fair technical basis?

I apologize if I offended you with my remarks. I was just trying to say that the view of an engineer is how a device acts as a black box when used in a design whereas a teacher emphasizes the device's inner workings (as it should be).

............................................

That means: The remaining question is only if we should confront the beginners/newbies/students with the physical truth (voltage-control) or with another view that is based solely on the relation Ic=beta*Ib (current control). But - in the latter case - one must be aware that the question comes up: Why is it so? Can you explain?

I believe you should confront the student/engineer with both views. The voltage-mode initially for analysis since that is the physics of the device. Then, when teaching circuit design, cover the current-mode model also, since that is a much easier way to view the transistor for many BJT design synthesis calculations. Any good student should be able to handle that duality. But for the beginner hobbyist (not engineer), who's mainly interested in how the device acts externally in a simple circuit, the current-control mode is easier to start with and understand. Fair enough?

................................................

Quotation (post 64): My simple criteria of this is how the device primarily appears to the outside world. If it has a relatively high input impedance and the output is basically proportional to the input voltage than it appears to be a voltage controlled device. If it has a relatively low input impedance and the output is basically proportional to the input current, then it appears to be a current controlled device. You can quibble with that definition if you like, but that's what I use for the purposes of this discussion.

This is your definition? And who set's the threshold - and at which input resistance?
That means: According to your "definition" a BJT "appears" as voltage controlled for a small Ic with a large h11 of - let's say - 100 kohms ? And suddenly - for higher currents - it appears as current controlled?
Are you really satisfied with this "definition"? I hope not.
And because you earlier spoke about confusion on beginners side: Don`t you think it confuses somebody to hear that sometimes the BJT "appears" as voltage-controlled (large input resistor) and sometimes as current controlled?

I have no particular input impedance point where I call it a current-controlled device or a voltage-controlled device. Obviously a FET is a voltage controlled device. For most typical beginner circuits the collector current is high enough and the input impedance low enough for me to consider a BJT a current-controlled device. But even at 100k ohms input impedance you can still view it as a current controlled device, just as you can view a device with a 1k input resistance as a voltage controlled device.

I think it would be very difficult to try to do a BJT circuit design without using the idea that it appears more like a current-mode device than a voltage-mode device for many of the design calculations.

Quotation (post 76): It's voltage feedback from the emitter resistor. It has nothing to do with whether the transistor is current or voltage controlled (the same basic feedback equation works for a BJT or a FET). To a first approximation this is normally the dominate voltage, and the Vbe term and the Ibe term are ignored in the feedback equation.

It has "nothing to do with...current or voltage control"? And the "Vbe term" is "ignored"? Who ignores this term? I cannot follow you.
Is it really necessary to explain the mechanism of voltage feedback via Re?
When a low-resistive voltage divider keeps the base potential Vb constant and when the emitter voltage Ve across Re goes up (due to an increase of Ie) the voltage Vbe across the B-E junction reduces correspondingly.
That is how the feedback works: The reduction in Vbe works against the Ic increase. How can you say Vbe would be "ignored". You will have noticed that it was not at all necessary to mention the base current Ib during description of this feedback principle.

Well, I'm a little confused by your argument, also. The Re resistor is generally selected so that any normal β changes (or voltage gain changes if you like) of the transistor have only a small effect on the stage voltage gain (so the stage gain can be calculated as Rc/Re). This feedback voltage is much larger than the small Vbe change due to the signal. So for normal amounts of feedback from Re, both Vbe and/or β can be ignored. But for precise calculations they can be included, of course.

This does bring up the point that the Re feedback makes the amp look voltage controlled. But, as I previously said, this is independent of whether you view the transistor as voltage-controlled or current controlled.

Quotation (post 76): Yes, in my opinion it is easier to view the BJT as current controlled device since that is how it appears externally to a first order. I've repeatedly said that this is an an external characteristic, not a description of the actual physics of the device

Repeatedly I ask : Please explain what do you mean with "appears externally". Up to now your only justification was the formula Ic=beta*Ib. Or do you refer to measurements or simulations (external characteristics")?
Of course I can verify the equation Ic=beta*Ib with some measurements. But the same applies to Ic=f(Vbe). Thus, this cannot be your argument for "extern appearance".

OK. I think Ic = beta * Ib, as an external measured characteristic with the input voltage appearing as a forward biased diode, is a good start for me to view the BJT as a current-controlled device, but here are some additional subjective reasons:

1. The BJT has a relatively low input impedance, more characteristic of a current operated device than an obvious voltage operated device (such as a MOSFET).

2. The current gain is relatively linear with changes in current. The voltage gain is very non-linear. It's easier to work with linear equations, than non-linear.

3. You use beta, not voltage gain to determine a BJT bias network's value.

4. You use beta to determine the input required to saturate a BJT transistor as a switch.

Quotation (post 76):Yes, you need to use the current-mode equation Ic = beta*Ib and that's all you need.

No, again I disagree. Assume you have a simple BJT amplifying stage with given values for all resistors and power supply. Of course, the exact BJT parameters are unknown.
Now the question: Is it possible by hand calculation to estimate the current Ic ? Answer: Yes, it is - and the only assumption you need is Vbe=0.6...07 volts.
For sufficient Re feedback it even is not important if you assume 0.6 or 0.7 volts.
But the relation Ic=beta*Ib does not help at all.

Au contraire. You are looking at the academic analysis of the circuit, not the engineering synthesis. To design the bias network you need to know the base current so the maximum equivalent resistance of the bias network can be determined for good bias stability. And you need the beta of the transistor for that. The transistor voltage gain is largely immaterial.

W., I've enjoyed our discussion and I realize I will likely never convert you completely to my position. The whole point of my argument has been to show that, for the simple types of circuits that novices work with, the current-mode operation of a BJT is easier to understand than the voltage-mode. But engineers must be initially taught the inner physics of the transistor as a voltage-controlled device and it must be viewed as a voltage-controlled device for applications such as current-mirrors or amp stages with little feedback. But for design purposes, the current-mode viewpoint should also be taught, otherwise the circuit synthesis can become rather difficult. I was taught the voltage-mode operation of the BJT and understood it at the time, but have mostly considered it to be a current-controlled device in my design calculations ever since. Of course all my designs were verified by simulation (when available) and breadboard which, of course, includes the actual transistor characteristics.
_______________________-

My responses are again in blue.

 

[Mr RB]

...
The correct sequence is as follows:
* With rising temperature (and constant Vbe) the current Ic goes up.
* If the value of Ic shall be held constant the voltage Vbe must be reduced correspondingly (extern!). This leads to the well-known value of approx. -2mV/deg.
...

For you and for Ratchit, I think you have it reversed, as you are probably also claiming for my argument.

If the base is supplied with a fixed current it causes control of the collector. There is a simple proof in my argument that for X base current there is Y collector current, and this occurs regardless of Vbe voltage characteristics.

So there is a simple case where Vbe can be two different values, BUT for both values of Vbe, Ib ALWAYS sets Ic. So Ic is current controlled.

As a simple analogy I could say that the steering wheel can control the car steering, but the actual thing that steers the car is the direction the wheels are pointing. There can be heat or slop in the steering mechanism so the steering wheel can be in two different known positions, but in BOTH cases the direction the car wheels are pointing ALWAYS controls the car.

You could argue that the steering wheel must be first or must be necessary etc, but the fact is the direction of the wheels is what actually controls the car, in the same way in the transistor the base current is what actually sets the collector current.

 

[Ratchit]

Mr. RB,

For you and for Ratchit, I think you have it reversed, as you are probably also claiming for my argument.

No, the temperature coefficient of Vbe is -2 to -2.5 mv/°C https://www.edaboard.com/thread156499.html . That cannot be denied.

If the base is supplied with a fixed current it causes control of the collector. There is a simple proof in my argument that for X base current there is Y collector current, and this occurs regardless of Vbe voltage characteristics.

So there is a simple case where Vbe can be two different values, BUT for both values of Vbe, Ib ALWAYS sets Ic. So Ic is current controlled.

As I showed previously, the equations from the reference Sedra and Smith showed both Ic and Ib are exponentially related to Vbe. In other words, Ic does not depend on Ib for its value and vice versa. Ib and Ic will be related to each other through Vbe, however. This relationship will be a constant called β. So, a BJT has a built-in scaled ammeter in the base circuit that indicates what the collector current will be, but this built-in ammeter does not control Ic. So you are beguiled or fooled into thinking that the indicator in the base is controlling the collector current. It's not.

Within the active region of the BJT, the Vbe will be the same for the same collector current and junction temperature no matter what the base current or base supply voltage is. Vbe will not have two different values as you aver above.

As a simple analogy I could say that the steering wheel can control the car steering, but the actual thing that steers the car is the direction the wheels are pointing. There can be heat or slop in the steering mechanism so the steering wheel can be in two different known positions, but in BOTH cases the direction the car wheels are pointing ALWAYS controls the car.

You could argue that the steering wheel must be first or must be necessary etc, but the fact is the direction of the wheels is what actually controls the car, in the same way in the transistor the base current is what actually sets the collector current.

The analogy senario is true, but it does prove that Ib is controlling Ic. Ib is an indicator of Ic, not a controlling factor.

Ratch