Maker Pro
Maker Pro

Transistors

Ratch

Mar 10, 2013
1,099
Joined
Mar 10, 2013
Messages
1,099
Over the years, I have locked horns with a lot of folks over whether a BJT is current-controlled (CC) or voltage-controlled (VC). I believe I can argumentatively pound anyone into the ground who thinks that a BJT is a current-controlled device.

Here are some of the points folks make when they aver that a BJT is current-controlled.

1) Look at this Ebers-Moll model of a BJT that uses current-dependent sources.

2) Ic changes when Ib changes in a somewhat linear fashion within a current range.

3) It depends on your point of view or how you preceive it.

4) It is simple (or easier) to explain how a BJT operates if current-control is assumed.

5) Have been designing with BJTs all my life, and current-control works fine for me.

6) It says so in the transistor manual, or textbook, or literature that the manufacturer puts out. They should know, shouldn't they?

7) Look at this circuit, it is obvious that it is a current amplifier.

There are other points that have been made and I invite you all to challenge me if you can think of any. The following are my answers to the above points.

1) Models show what a device does, not how it works. The makeup of the model can have no relationship with the physics of the device.

2) When my bedside clock shows 5:30 AM, the sun rises. Do I infer that the clock is controlling the sun, or is it celestial orbital mechanics? So why do people think that Ib controls Ic instead of considering the possibility that Ib is an indicator of Ic?

3) The way something works is never a perception issue. Arguing about whether the brake drums or the the tire friction with the pavement stops a car is not considering the system as a whole.

4) In this case, the explanation is false. Even if it appears plausible.

5) Design is one thing, knowing how a BJT really works is another. Ib is a good indicator of Ic, so it should be used in design.

6) Often they are talking about a current amplifier transistor model. They seldom go into the physics of a transistor.

7) You can make a current amplifier circuit from a voltage amplifier or vice versa. Then you are talking about the circuit and not the device.

Here are my reasons for what I say. First of all, the correct way to determine whether a device is a CC or VC to understand its physics. A BJT is a diffusion device. It does not depend directly on a voltage produced electric field to drive the charge carriers into the base of the BJT like a copper wire does. When a transistor is first manufactured, the charge carriers from the P-type and N-type material diffuse into each others material. This stops when a depletion region (charge carriers significantly reduced) forms from the holes and electrons annihilating each other. A barrier voltage forms on each side of the boundaries of the region that repels further charge carriers from entering the depletion region. All this happens without any voltage being applied across the PN junction. If a forward voltage is applied, that lowers the barrier voltage, and more charge carriers can diffuse into the base region from the depletion region. So the Vbe voltage controls the flow of charge carriers into the base. In a BJT, the forward bias of the base-collector sweeps the charge carriers from the base into the collector circuit. Therefore, the Vbe controls both the collector current and base current. The diffusion process also explains why the voltage to current function is exponential instead of linear. You cannot explain the above process by current sources. The physics will not allow it. So the BJT is a VC current source or transconductance device, not a CC current source.

Ratch
 

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
Over the years, I have locked horns with a lot of folks over whether a BJT is current-controlled (CC) or voltage-controlled (VC). I believe I can argumentatively pound anyone into the ground who thinks that a BJT is a current-controlled device.
I think you probably can! I'm certainly not going to argue with you - my understanding of solid state physics is very limited.
(...) why do people think that Ib controls Ic instead of considering the possibility that Ib is an indicator of Ic?
Is that how you would describe it? This is interesting.

Recently I did some experiments (well, just LTSpice simulations actually) to try to delve deeper into the operation of current mirrors, and I discovered that the current vs. voltage relationship of a transistor's base-emitter junction depends very much on how much collector current is able to flow.

If the collector is not connected to anything, the base-emitter junction behaves like a boring old diode, as you would expect. But when voltage is made available to the collector, provided that voltage is significantly more than Vce(sat), the base starts drawing current at a lower Vbe than it did without the supply to the collector, and (IIRC) the slope is steeper. And that's only the most obvious and easily described difference.

As Vce is dropped towards Vce(sat), and below it, the base-emitter junction becomes more like a diode.

All of this was the behaviour of LTSpice's transistor model - a 2N3904 I think - and I don't know how closely that model represents reality.

I would like to hear your thoughts :)
 

Ratch

Mar 10, 2013
1,099
Joined
Mar 10, 2013
Messages
1,099
I think you probably can! I'm certainly not going to argue with you - my understanding of solid state physics is very limited.

Is that how you would describe it? This is interesting.

Recently I did some experiments (well, just LTSpice simulations actually) to try to delve deeper into the operation of current mirrors, and I discovered that the current vs. voltage relationship of a transistor's base-emitter junction depends very much on how much collector current is able to flow.

If the collector is not connected to anything, the base-emitter junction behaves like a boring old diode, as you would expect. But when voltage is made available to the collector, provided that voltage is significantly more than Vce(sat), the base starts drawing current at a lower Vbe than it did without the supply to the collector, and (IIRC) the slope is steeper. And that's only the most obvious and easily described difference.

As Vce is dropped towards Vce(sat), and below it, the base-emitter junction becomes more like a diode.

All of this was the behaviour of LTSpice's transistor model - a 2N3904 I think - and I don't know how closely that model represents reality.

I would like to hear your thoughts :)

I have not studied the saturation region of the BJT very much, so I am not going to comment on something I know little about. It is somewhat more complicated than the active region. I do plan one day to peruse it, however.

Ratch
 

LvW

Apr 12, 2014
604
Joined
Apr 12, 2014
Messages
604
I like to contribute some general thoughts to the subject under discussin (current vs. voltage control).
For my opinion, it is really a funny (?) story which has started long time ago.
Let me explain:
I suppose, all students and newcomer in electronics know that for a pn diode there is an exponential relationship between voltage and current.
There is no doubt that this behaviour is caused by the thickness of the depletion layer which is determined by the applied voltage across the pn junction.
I can see no reason why this should be different for the pn junction between base and emitter of a BJT. And the explanation is very similar to the diode case.
More than that, I cannot see why students and other newcomer suddenly should be "overstrained" in accepting such an exponential behaviour (which they know from the pn diode) .
With other words: I cannot see any reason to switch to another model (it is just a model!) which is based on current-control.
Just because there is such a nice and simple relationship between base and collector current? Is this the only justification?
Of course, in most cases it works for designing an amplifier.
Nevertheless, for my opinion - and according to my experience with many, many students - it is not a penalty to know the difference between a (working) model and the physical reality.
 
Last edited:

Harald Kapp

Moderator
Moderator
Nov 17, 2011
13,700
Joined
Nov 17, 2011
Messages
13,700
I respect your attitude. It is a good one when it comes to students of electrical engineering. I too would expect them to develop that kind of understanding for electronic components (not only transistors).

However, I think here we deal more with a hands-on approach. From my experience in this forum I get the impression that most members here are happy to understand a transistor in its basic function using numbers and possibly graphs from a datasheet. Asking for an understanding of the physical effects within the transistor (like the difference between minority and majority charge carriers, bandgap etc.) is asking way too much.
 

LvW

Apr 12, 2014
604
Joined
Apr 12, 2014
Messages
604
[QUOTE="Harald Kapp, post: 1606077, member: 22009"
....From my experience in this forum I get the impression that most members here are happy to understand a transistor in its basic function using numbers and possibly graphs from a datasheet.[/QUOTE]
Agreed. On the other hand, the data sheets contain a graph Ic=f(Vbe) and the slope of this graph is the most fundamental parameter for BJT amplifiers (transconductance gm). I am afraid, those people feel lost without knowing the meaning of this fundamental graph. And we should not forget: Also the "current-control group" makes use of the 0.7V between B and E for designing an amplifier.

[QUOTE="Harald Kapp, post: 1606077, member: 22009"
... Asking for an understanding of the physical effects within the transistor (like the difference between minority and majority charge carriers, bandgap etc.) is asking way too much.[/QUOTE]

Yes - agreed. But you will have noticed that I didn´t mention at all minority/majority carriers etc. All is needed - as a minimum - is the knowledge that a voltage across a pn junction determines its thickness and, thus, the current through the junction (similar to the diode principle).

Final remark: With my former post it was my only intention to express my astonishment that such a discussion takes place nowadays.
(Recently I have pointed also to the fact that even the term "transistor=transfer resistor" was chosen because of the voltage-current relationship.)
 

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
I suppose, all students and newcomer in electronics know that for a pn diode there is an exponential relationship between voltage and current. There is no doubt that this behaviour is caused by the thickness of the depletion layer which is detrmined by the applied voltage across the pn junction.
I can see no reason why this should be different for the pn junction between base and emitter of a BJT. And the explanation is very similar to the diode case.
I'm with you so far. And that was what I was taught.
More than that, I cannot see why students and other newcomer suddenly should be "overstrained" in accepting such an exponential behaviour (which the know from the pn diode). With other words: I cannot see any reason to switch to another model (it is just a model!) which is based on current-control.
Just because there is such a nice and simple relationship between base and collector current? Is this the only justification?
Yes. I don't know of any other justification. And I think it's sufficient justification. I certainly think it's more sensible to explain that Vbe causes Ib with an exponential relationship, then Ib causes Ic with a roughly direct relationship, than to say that Vb causes Ic with an exponential relationship.

If it's true that explaining it that way actually reflects reality more accurately, when you take the physics into account, then that's certainly a good argument for explaining it that way.

I never really understood the physics of transistors, and what I did understand, I forgot long ago. But I consider myself competent to design all of the transistor circuits that I have designed as part of my job. I'm aware that I would need to learn more if I was designing, say, gigahertz circuits using modern germanium transistors, or kilowatt amplifiers, but I don't do that, and nor do most people on this site, so I am content with my level of competence. Everyone simplifies devices to a black box with known characteristics at one level or another.

As for this discussion, I'll leave it to the people who understand the physics to determine whether it's more meaningful to say that a transistor is voltage-controlled or current-controlled, and whether, as Ratch suggested in post #41, that the base current may be an "indicator of" the collector current, rather than the cause of it. From my "black box with known behaviours" understanding, it makes no difference at all.

Nevertheless, for my opinion - and according to my experience with many, many students - it is not a penalty to know the difference between a (working) model and the physical reality.
Agreed. If you (or Harald, or Ratch) feel that it's misleading to characterise a transistor by saying that Vbe causes Ib, and then Ib causes Ic, and you can justify that opinion with an understanding of the real physics involved, then more power to you :)
 

Ratch

Mar 10, 2013
1,099
Joined
Mar 10, 2013
Messages
1,099
I suppose, all students and newcomer in electronics know that for a pn diode there is an exponential relationship between voltage and current.
There is no doubt that this behaviour is caused by the thickness of the depletion layer which is determined by the applied voltage across the pn junction.
I can see no reason why this should be different for the pn junction between base and emitter of a BJT. And the explanation is very similar to the diode case.
More than that, I cannot see why students and other newcomer suddenly should be "overstrained" in accepting such an exponential behaviour (which they know from the pn diode) .

The exponential behavior of the voltage to current of a junction diode is due to the fact that it uses diffusion to transport its charge carriers instead of electric fields like a plain conductor does. Other applications like an ink drop spreading throughout a glass of water also show this same exponential relationship with respect to time. It is not such a mind-bending, mind-warping, out of body experience to know that diffusion is what makes junction devices work. Even if all the details are not understood, it should be realized that Vbe controls Ib exponentially, and Vbe also controls IC exponentially, and that no direct physical coupling exists between Ic and Ib. But, when two exponential quantities are divided, their nonlinearities disappear and Ic/Ib = beta. This is a handy useful fact that should be utilized in design.

Ratch
 

Arouse1973

Adam
Dec 18, 2013
5,178
Joined
Dec 18, 2013
Messages
5,178
Spot on Ratch. I have always thought a BJT was voltage controlled, but then told I was wrong. I still didn't believe them because I knew that if I changed the voltage on the input to a BJT I would be able to change the current through the device. The fact that an NPN base clamps itself to what ever it's Vbe is +Ve doesn't really matter. I am still able to control the current through the device by increasing the voltage to the base resistor. Yes it might be extremely sensitive to this but is still the voltage that is controlling the current.
I don't know if that makes sense.
Adam
 

Ratch

Mar 10, 2013
1,099
Joined
Mar 10, 2013
Messages
1,099
Spot on Ratch. I have always thought a BJT was voltage controlled, but then told I was wrong. I still didn't believe them because I knew that if I changed the voltage on the input to a BJT I would be able to change the current through the device. The fact that an NPN base clamps itself to what ever it's Vbe is +Ve doesn't really matter. I am still able to control the current through the device by increasing the voltage to the base resistor. Yes it might be extremely sensitive to this but is still the voltage that is controlling the current.
I don't know if that makes sense.
Adam

That is why it is beneficial to put some resistance R in the emitter leg. This resistance gets betatized by the negative feedback. Then you are controlling the current by applying voltage across the much larger R*(beta+1), which swamps out the much smaller now irrelevant resistance of the base-emitter junction.

Ratch
 

Arouse1973

Adam
Dec 18, 2013
5,178
Joined
Dec 18, 2013
Messages
5,178
Good point.
Have a look at the Temperature compensating a transistor thread. I think you will find this interesting. I would be interested on your thoughts.
Adam
 

Dylan Wisdom

Apr 10, 2014
26
Joined
Apr 10, 2014
Messages
26
Okay so I've been reading a little more on transistors and I just want to double check something before I go on. The base of a transistor is the factor that decides whether or not the voltage or current is amplified (depending on the type of circuit) and exactly how much it is amplified. Am i wrong, right, or am i close?
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
Okay so I've been reading a little more on transistors and I just want to double check something before I go on. The base of a transistor is the factor that decides whether or not the voltage or current is amplified (depending on the type of circuit) and exactly how much it is amplified. Am i wrong, right, or am i close?

The base is the element (strictly speaking the relationship between base and emitter) which controls what happens between the collector and the emitter. Whilst there are circuits that don't make that immediately obvious (common base is the classic) that's what is going on.
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
Spot on Ratch. I have always thought a BJT was voltage controlled, but then told I was wrong. I still didn't believe them because I knew that if I changed the voltage on the input to a BJT I would be able to change the current through the device. The fact that an NPN base clamps itself to what ever it's Vbe is +Ve doesn't really matter. I am still able to control the current through the device by increasing the voltage to the base resistor. Yes it might be extremely sensitive to this but is still the voltage that is controlling the current.
I don't know if that makes sense.
Adam

Your explanation is that the transistor is current controlled. We don't care *how* you change the current through the base (if that involves an increased voltage on a fixed resistor or a constant voltage through a variable resistor -- or indeed any other way -- it makes no difference).

What you have wrong is that Vbe actually changes slightly as the current changes. It is this *voltage* that controls the behaviour in the transistor. Since the BE junction is a forward biased diode, the only way of increasing the voltage is to increase the current.

It is possible to place a *voltage* on the base of the transistor and control the transistor directly (and ignoring the current flow). However, this is *very* sensitive, and the resulting current can easily destroy the transistor, so it's never done (except in current mirrors).

What we *say* is that an increasing base current controls a larger current in the collector. What actually happens is that a voltage change on the base allows an exponentially increasing current in the collector. Because the relationship between base voltage and base current is similar to the relationship between the base voltage and the collector current, we get a far more linear relationship between base current and collector current. For many uses this is easier to understand and calculate.
 

Ratch

Mar 10, 2013
1,099
Joined
Mar 10, 2013
Messages
1,099
Okay so I've been reading a little more on transistors and I just want to double check something before I go on. The base of a transistor is the factor that decides whether or not the voltage or current is amplified (depending on the type of circuit) and exactly how much it is amplified. Am i wrong, right, or am i close?

The base determines the charge carrier flow between the emitter and the collector by controlling the diffusion process by adjusting Vbe. Once the charge carriers leave the emitter and enter the base, they are swept into the collector by the higher collector voltage. Increasing the collector voltage has little effect on the current because the charge flow from the emitter is controlled by the Vbe, not Vc. That is why you see a horizontal curve on a Ic-Vc plot in the active region. Because the Vc has little effect on the current, it acts like a current generator. And, as explained earlier, this current is controlled by Vbe. That makes a BJT a transconductance amplifier, not a current amplifier. Notice the Ib that leaks out of the base is waste current that has no effect on Ic. Ib is related to Ic by the same exponential ratio throught Vbe, but it does not affect Ic.

Ratch
 

Ratch

Mar 10, 2013
1,099
Joined
Mar 10, 2013
Messages
1,099
Your explanation is that the transistor is current controlled.

No, we both aver that a BJT is voltage controlled.

We don't care *how* you change the current through the base (if that involves an increased voltage on a fixed resistor or a constant voltage through a variable resistor -- or indeed any other way -- it makes no difference).

The difference is that you cannot explain the control of diffusion with a current. The diffusion is controlled by a Vbe voltage that increases/decreases the barrier voltage caused by the creation of the depletion region where a PN junction exists. Decreasing the barrier voltage allows more charge to flow through the depletion region, and increasing the voltage does the opposite. Base current is only an indicator of the charge carriers continuing on to the collector.

What you have wrong is that Vbe actually changes slightly as the current changes. It is this *voltage* that controls the behaviour in the transistor. Since the BE junction is a forward biased diode, the only way of increasing the voltage is to increase the current.

Actually, it is the current that changes greatly as Vbe changes slightly. Vbe is the cause, current is the effect. The physics of diffusion of the PN junction prove it.

It is possible to place a *voltage* on the base of the transistor and control the transistor directly (and ignoring the current flow). However, this is *very* sensitive, and the resulting current can easily destroy the transistor, so it's never done (except in current mirrors).

Correct, we put lots of resistance in the emitter so as to lower the voltage on the B-E junction. However, the voltage on the B-E junction is what is controlling the Ic.

What we *say* is that an increasing base current controls a larger current in the collector. What actually happens is that a voltage change on the base allows an exponentially increasing current in the collector. Because the relationship between base voltage and base current is similar to the relationship between the base voltage and the collector current, we get a far more linear relationship between base current and collector current. For many uses this is easier to understand and calculate.

Yes, understanding that the Ib is an indicator of Ic makes things a lot easier. But it is physically wrong to say that Ib controls Ic.

Ratch
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
No, we both aver that a BJT is voltage controlled.

WRONG.

The explanation that Arouse1973 gave for "voltage control" was in fact current control.

Read it again:

The fact that an NPN base clamps itself to what ever it's Vbe is +Ve doesn't really matter. I am still able to control the current through the device by increasing the voltage to the base resistor. Yes it might be extremely sensitive to this but is still the voltage that is controlling the current.

He is saying that

(1) Vbe is fixed
(2) current is controlled by changing the voltage on the other end of a resistor tied to the base

This is a current control explanation.

A voltage control explanation would have read:

"The fact that an NPN base HAS A CHANGING Vbe is +Ve IS ALL THAT MATTERS. I am still able to control the Vbe OF the device by increasing the CURRENT THROUGH THE base. Yes it might be extremely sensitive to this but is still the voltage that is controlling the DEVICE."​


The difference is that you cannot explain the control of diffusion with a current.

I'm not arguing that the transistor is current controlled, remember!

Actually, it is the current that changes greatly as Vbe changes slightly. Vbe is the cause, current is the effect. The physics of diffusion of the PN junction prove it.

Whether current controls voltage or voltage controls current is immaterial.

The issue is the electric field in the semiconductor, not the rate at which charges are flowing through it. And that is essentially the voltage control argument.

Yes, understanding that the Ib is an indicator of Ic makes things a lot easier. But it is physically wrong to say that Ib controls Ic.

Gee whiz! If you had been paying attention you would have noticed I agree with you.

However there are distinct practical reasons to use a current control model for almost all BJT uses.

One reason is to stop pointless arguments like this one.
 

Ratch

Mar 10, 2013
1,099
Joined
Mar 10, 2013
Messages
1,099
He is saying that
(1) Vbe is fixed

I disagree. I interpret him as saying that whatever the sum of the emitter and Vbe are, the Ic is controlled by Vbe. He never says that Vbe is constant.

(2) current is controlled by changing the voltage on the other end of a resistor tied to the base
This is a current control explanation.

Voltage change on the base causing change in Ic sounds like voltage control to me.

I'm not arguing that the transistor is current controlled, remember!

Whether current controls voltage or voltage controls current is immaterial.

I don't agree. The physics say that voltage is controlling the current in this case. It is confusing to use a current to apply a voltage. Because then you have to know what the components are and their values to know what the voltage is.

Gee whiz! If you had been paying attention you would have noticed I agree with you.
However there are distinct practical reasons to use a current control model for almost all BJT uses.

I have been paying attention, but I get confused when you equate applying current with applying voltage. The results might eventually be the same, the application method can be convoluted. Yes, using Ib as an indicator of Ic can be helpful. An argument is not senseless if it promotes understanding.

Ratch
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
Are you being deliberately thick?

He is saying that you have a voltage source --- resistor --- base and that changing that voltage means voltage control of the transistor.

Forgive me if that is not what he is saying, but it sure sounds like it.

And "clamps it's base to whatever it's Vbe is" says that it clamps it to a fixed voltage, not one determined by (or determining) anything.
 

Ratch

Mar 10, 2013
1,099
Joined
Mar 10, 2013
Messages
1,099
Are you being deliberately thick?

He is saying that you have a voltage source --- resistor --- base and that changing that voltage means voltage control of the transistor.

Forgive me if that is not what he is saying, but it sure sounds like it.

And "clamps it's base to whatever it's Vbe is" says that it clamps it to a fixed voltage, not one determined by (or determining) anything.

He said that changing the voltage source to the base through a resistor will change the Ic. It will, but he never said that it proves the transistor is voltage controlled. I first made that statement about voltage control, and used a physics argument to substantiate my assertion. He said clamps its Vbe to Ve. That would be the emitter voltage.

Ratch
 
Top