Maker Pro
Maker Pro

How Transistors Work

Status
Not open for further replies.

chopnhack

Apr 28, 2014
1,576
Joined
Apr 28, 2014
Messages
1,576
As I try to learn more about these components, I stumbled onto this site. Can anyone review the information and let me know if this is a valid viewpoint or if its bunk. The author did state that it is his opinion as I recall, but so far reading it, it was making sense...
http://amasci.com/amateur/transis.html
 
Last edited by a moderator:

Arouse1973

Adam
Dec 18, 2013
5,178
Joined
Dec 18, 2013
Messages
5,178
He has copied a lot of this from another website I know very well so yes he would be accurate. But there is quite a lot to go through. I will look over this later for you and give my opinion.
Adam
 

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
That's an interesting site and I would be interested to hear opinions from Adam and Steve, both of whom have mentioned that it's more accurate to consider a transistor a voltage-controlled device than a current-controlled device.

But to be able to make practical use of transistors, the current-controlled model is adequate, and more straightforward; I guess this is why it is taught. And going deeply into the physics of one (increasingly less-used) type of semiconductor device is only an unnecessary diversion from learning more general, important, and useful things about circuits in general, and other types of components. I think it would be a waste of time, unless you're specifically interested in the topic and want to learn about it in great detail.
 

chopnhack

Apr 28, 2014
1,576
Joined
Apr 28, 2014
Messages
1,576
That's an interesting site and I would be interested to hear opinions from Adam and Steve, both of whom have mentioned that it's more accurate to consider a transistor a voltage-controlled device than a current-controlled device.

But to be able to make practical use of transistors, the current-controlled model is adequate, and more straightforward; I guess this is why it is taught. And going deeply into the physics of one (increasingly less-used) type of semiconductor device is only an unnecessary diversion from learning more general, important, and useful things about circuits in general, and other types of components. I think it would be a waste of time, unless you're specifically interested in the topic and want to learn about it in great detail.
To be honest, that becomes a problem for me.... I can get bogged down in the minutia and loose the big picture! I do keenly appreciate the details and how things work. I don't like to treat things like black boxes. I try to understand it so that I can better know when to apply the information. It seems I am spiraling out of control from my one projec!!! LOL. I started and still haven't finished that PCB - from there I branched off to learn C, which brought me to Code Blocks and its issues, PIC's, pic programmers, PCB design which is an art unto itself, a breadboard that currently has a half dozen small circuits crammed unto it... :D It's an eclectic grouping!

Thanks Adam, I appreciate your feedback! If you get a chance could you link the other site?
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
I have issues with a few things he says.

He mangles the idea when he says that wires are already "full of charge". They have no net charge and they have mobile charge carriers. But they're not cancelled charge, or even worse, "uncharged charge".

Wires contain "uncharged" charge. Better call it "cancelled charge." Yet even though the electrons are cancelled by the nearby protons, the electrons can still flow among the protons.

Electrons don't flow among the protons...

He bangs on and on about it's not current that flows. And I'll agree with him up to a point. It's like talking about an "ATM machine" or a "PIN number" in some respects. However the phrase "current starts to flow" can be readily interpreted as "that which moves starts moving".

In effect he's simply stating the physics of the device, that is that the bipolar junction transistor is voltage controlled.

This may give you a valuable insight into how the device works, but it's far less useful in practice. As a rough approximation, base current controlling collector current almost always allows you to perform the required calculations to a reasonable degree of accuracy with far less effort.

I also think he skips over why there are "extra electrons" and "holes" in N and P doped silicon respectively. This is actually important stuff if his explanations are not going to throw up a whole lot more questions or statements that need to be taken on faith. Why don't the holes just fill up? Why don't the extra electrons fall out? Why don't all the extra electrons just fill up all of the holes?

Then later he bangs on about holy scripture Vs dark heresy. It's as if he has never conceived of looking at a problem at different levels of abstraction. Imagine if we taught children to cross the road by explaining Newton's laws of motion. We would never be able to cross the road unaided until high school. And then there would be people who would insist that unless we force children to consider relativity they are in some significantly more danger. No, we tell children to stop, look both ways (we generally ignore the possibility of one-way roads), we tell them to listen, and we tell them to cross "when it's safe".

Teach your child how to cross roads safely. Children must first stop at the kerb. Then they need to look and listen for traffic, and then decide whether it is safe to cross. (“STOP, LOOK, LISTEN and THINK”)

How can that possibly work without a deep an abiding knowledge of Newton's laws? And it ignores the possibility of supersonic vehicles (not to mention those traveling at appreciable fractions of the speed of light)!

No, we make it simple. We use rules that work in most simple cases. We're aware that they may not work in certain special cases. But as heuristics, they work.

As we get a little better at crossing roads, we may walk onto a road with oncoming traffic, we may rely on the fact that some of the traffic will pass before we get to that point on the road, or that there is an island between oncoming lanes. Or (shock) we might decide to run!
 

chopnhack

Apr 28, 2014
1,576
Joined
Apr 28, 2014
Messages
1,576
I have issues with a few things he says.

I found it interesting, but I too took note of his very opinionated style. I found his statement of being able to sweep the charge out of a conductor with sufficient potential odd... how would the material keep its properties??

<<In silicon, you can sweep all the charges out of the material by using a few volts of potential, while in a metal it would take billions of volts to accomplish the same thing.>>
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
Oh, and another thing, he decries Ohms law because it (in his mind):

But Ohm's law is "wrong" in that it incorrectly implies forward and reverse causation between e-fields and carrier drift.

But this is because he has already failed to grasp what Ohms law is. He says:

Yet isn't all of this crazy talk, because Ohm's law works fine when we assume that current causes voltage. Yes, that's right, and it's because Ohm's law ignores the internal physics. Ohm's law is a simplified abstract model; a mental tool, and a very useful one. It treats a resistor as a black box.

Perhaps he needs to refresh his mind on what a scientific law is:

While scientific theories and laws are both based on hypotheses, a scientific theory is an explanation of the observed phenomenon, while a scientific law is a description of an observed phenomenon.

Kepler's Laws of Planetary Motion, for example, describe the motions of planets but do not provide an explanation for their movements.

So, a law is a description of observed phenomena. It is not an explanation, perhaps the mathematical statement of correlation.

And what does Ohm's law say?

Ohm's law states that the current through a conductor between two points is directly proportional to the potential difference across the two points

And we'll ignore that it only applies to material to which it applies. And one of those materials is NOT a PN junction.

Well, Ohm's law, as it is stated seems to satisfy his concern. The law itself says that current is dependent on voltage and some other constant of proportionality.

So I = kV, where k is the constant of proportionality.

If we call the inverse this constant of proportionality "resistance" we get I = V/R, which is a more literal statement of Ohms law.

He essentially criticizes the rearrangement of this into V = IR because (in his mind) it says that voltage is dependent on current. This is as stupid as saying that rearranging F = mg to g = F/m is stating that for a given force, gravity is dependent on the mass it is acting on, thus rendering the equation wrong. It's not wrong, neither are, they simply state a relationship which allows us to calculate a missing value given the other ones.
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
I found it interesting, but I too took note of his very opinionated style. I found his statement of being able to sweep the charge out of a conductor with sufficient potential odd... how would the material keep its properties??

<<In silicon, you can sweep all the charges out of the material by using a few volts of potential, while in a metal it would take billions of volts to accomplish the same thing.>>

Yeah, that's funny.

He may mean "all the free charge carriers", but even then that would be stretching the point.
 

chopnhack

Apr 28, 2014
1,576
Joined
Apr 28, 2014
Messages
1,576
Yeah, that's funny.

He may mean "all the free charge carriers", but even then that would be stretching the point.

It's kinda where I stopped reading and looked for help! Law of conservation of energy seemed to come to mind as well as transmutation, LOL
 

LvW

Apr 12, 2014
604
Joined
Apr 12, 2014
Messages
604
In effect he's simply stating the physics of the device, that is that the bipolar junction transistor is voltage controlled.
This may give you a valuable insight into how the device works, but it's far less useful in practice. As a rough approximation, base current controlling collector current almost always allows you to perform the required calculations to a reasonable degree of accuracy with far less effort.
Hi Steve,
Every time I read a statement like this (..far less useful in practice..) I ask myself: How can you design a simple transistor stage without using the voltage-control property of the BJT?
It is simply not true that "base current controlling collector current almost always allows you to perform the required calculations to a reasonable degree of accuracy with far less effort."

Let´s take a very basic simple example: Common emitter BJT stage (without emitter resistor) and with "base current injection" using a resistor RB between voltage supply Vcc and the base node. (Intentionally, I am using the term "current injection" because it is commonly used - but it is wrong!).
After selecting a suitable collector current IC and based on the known (or estimated) value of B=IC/IB the next step is to find a value for RB.
What are you doing? You start with VBE=(0.65...0.7)V and the you calculate RB=(Vcc-VBE)/IB.
That means (in words): You use a current-limiting resistor RB between Vcc and the base node with the aim to for produce the desired voltage VBE .
Or with other words: We want to "open" the transistor with a desired collector current Ic - and we want to use the only available DC source Vcc. Hence, we have no other choice than to use a current-limiting resistor RB for producing the necessary voltage VBE wwhich allows the current Ic.
Remember: Such a current limiting resistor is used also in other simple circuits to protect some parts which are sensitive against overvoltage (pn diode or zener).

Question: At which point of the design process did you apply current-control (with far less effort ) ?

Summary: It is really a "fairy-tail" that (a) the BJT would be current-controlled and (b) that this approach would be "useful in practice".
More than that, it is really funny but all people believing they would apply the "current-control approach" (for what ever reason) do NOT apply this principle in reality.
Instead, they apply the voltage-control feature (however, without knowing this). Nevertheless, they are voting for current-control.
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
Question: At which point of the design process did you apply current-control (with far less effort ) ?

The method you used above is what I would call the current method of control.

If you were to calculate the change in Vbe required to get a particular collector current, *THAT* would be voltage control.

Summary: It is really a "fairy-tail" that (a) the BJT would be current-controlled and (b) that this approach would be "useful in practice".
More than that, it is really funny but all people believing they would apply the "current-control approach" (for what ever reason) do NOT apply this principle in reality.
Instead, they apply the voltage-control feature (however, without knowing this). Nevertheless, they are voting for current-control.

I apologize to you if your method was actually using the change in Vbe as a controlling condition. But I didn't see that because of B=IC/IB and RB=(Vcc-VBE)/IB. They seem to be using the base current to me.

We know that is is actually the change in Vbe which is important, but do we ever calculate the change in Vbe required? I'm pretty sure the answer is "rarely".
 

LvW

Apr 12, 2014
604
Joined
Apr 12, 2014
Messages
604
The method you used above is what I would call the current method of control.
.
Steve - I cannot see where I have used a current to "control" something. Of course, there is a current IB and, therefore, we must consider its influence; and, thus, we consider and calculate the voltage "loss" at the resistor RB caused by this current IB. But all this only to allow the required/designed voltage VBE. What about my example with the zener diode? Here, we are using also a limiting resistor in series with the pn junction. Why? Just to allow operation in the zener voltage range. And exactly the same applies to the BE junction.
Question: Steve - even if it is appropriate to call the calculation steps in my example "current control" (as you propose) - please, can you tell me at which point this method requires "far less effort" than the voltage control approach ? (If - at the same time - you state that it´s just the same procedure!).


I apologize to you if your method was actually using the change in Vbe as a controlling condition. But I didn't see that because of B=IC/IB and RB=(Vcc-VBE)/IB. They seem to be using the base current to me.
We know that is is actually the change in Vbe which is important, but do we ever calculate the change in Vbe required? I'm pretty sure the answer is "rarely".

1.) Yes - of course, we must not neglect the base current because it exists! But this is no prove that it controls something. However, for a good design with sufficient Ic stabilization (RE feedback) we can design a base biasing voltage divider even WITHOUT considering the base current. We can forget it - and the corresponding error is comparable with other uncertainties/tolerances.
2.) You are asking "do we ever calculate the change in Vbe required?" No - why should we?
We don´t calculate it but we are using it always because it belongs to the most important parameter for gain calculation: The derivative dIC/dVBE=gm (transconductance) is the slope of the transfer characteristic which connects input and output signals of the transistor. Consequently, this derivative determines the voltage gain of a transistor stage.
 

Arouse1973

Adam
Dec 18, 2013
5,178
Joined
Dec 18, 2013
Messages
5,178
I think we can all agree that a BJT is a transconductance device. This is the collector current is a function of applied base emitter voltage. This means that a 57.6 mV change in base emitter voltage with produce an order of magnitude change in collector current.

But how do we calculate all of this on a day to day basis? Well if we were to take the Vbe voltage method we would have to do the following.

New Picture (5).png

This is called the diode equation and tells us everything we need to know about a PN junction and its electrical characteristics. This approach is valid for all Vbe voltages if the temperature is also known and the reverse saturation current of the transistor. Also at very low Vbe levels the gain rises, so this method is also accurate at low levels of Vbe bellow 0.6 V. This is because most of the data sheets show gain graph at say 100 uA and greater.

The other method involves using the base current and transistors hFE. This is only valid for common emitter configuration, after all that’s what it means, hybrid, forward, common emitter. This is also only valid when the base emitter junction is forward biased by 0.6 V anything less than this the hFE can’t be used accurately as mentioned above.

New Picture (6).png

So they all seem to match pretty well. Which one do you use? Well most, no actually all engineers I have worked with use the later. This is because the data sheets will have the gain curves readily available and the above workings show it’s much easier to remember (less effort) this way rather than the diode equation method.

The fact that the above is actually modifying the Vbe to achieve the same result as the diode equation might not seem apparent but at the end of the day who cares as long as we have something to work with. After all transistor hFE are so variable we almost always end up putting a POT in there somewhere if accuracy is needed.

So if you are starting out in electronics then I would recommend using the later method, but knowing that a transistor is a voltage controlled device is off course of paramount importance but something that is rarely thought of today and for most transistor switching /amplification applications it is of no real importance as other simpler method exists that will give you the same results. You choose your own method, but all I would say is trying to find a text book with lots of examples that use the first method might be tricky.

Diagram below shows some examples. But be aware that applying too high a voltage to the base without current limiting is not a good idea, BANG!
Adam
New Picture (7).png
 

Attachments

  • BJT.JPG
    BJT.JPG
    28.4 KB · Views: 124

chopnhack

Apr 28, 2014
1,576
Joined
Apr 28, 2014
Messages
1,576
Hope it makes sense to you?
Adam
On a conceptual level I can follow along. Knowing when to use certain equations and how to find the information from the data sheet seems to be the problem for me. It will require more experience and exposure for it to make sense and I certainly do appreciate your help!
 

LvW

Apr 12, 2014
604
Joined
Apr 12, 2014
Messages
604
Hope it makes sense to you?
Adam
Hi Adam - for my opinion you didn´t really met "my point". Let me explain:
As the result of your contribution we have two formulas which - for my opinion - cannot be compared to each other.
That means: They cannot reveal any differences between both approaches.
The first equation describes the physical behaviour of the pn junction, that means: How the collector current depends on the applied base-emitter voltage.
Nobody would use this equation for designing an amplifier stage. Nevertheless - it reflects the physical truth.
The second equation shows how in practice the current limiting resistor RB is calculated. In words: The formula shows how a base-emitter voltage of VBE=0.6 V can be created if a DC source of 5V is available. For this purpose, we are using a resistor RB in series with the B-E path. The value for RB is calculated using the current through it (IB). You will notice that in my last contribution I have mentioned exactly the same formula.
Question: Is this method to produve VBE=0.6 V something that can be called "current- control?".

That is my main point: Everybody is using the same set of formulas - however, some people think that they exploit a current-control feature of the BJT. But this is not true!
Even the last example presented by you is based on the requirement to create the necessary voltage VBE.

Last comment: Of course, you can argue that - if we all use the same formulas - it does not matter if we call this procedure "current" or "voltage" control.
Let me explain why I think it is important to use the correct description and the correct wording:

Up to now I have seen no physical explantion WHY the base current IB should be able to control IC. This is no surprise because it cannot be justified.
More than that, also from the energy point of view it is impossible that a certain quantity (in our case IB) can directly control a larger quantity of the same kind (in our case: IC).
The only relation we have is IB=IC/B - however, this simple relation does only state that IB is a fraction of IC and does not tell us anything about Ic control.
In general, it is my opinion that students should start learning to be critical against simple equations and simple explanations.
I think, they shouldn´t simply believe something and - as often as possible - they should ask „why?“. This is the first step to really understand how an electronic circuit works.
Therefore, I think it is important to realize the physical truth behind the formulas we are using.
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
In my opinion, if your calculations require in Ib with Vbe ignored or assumed to be constant then you are describing current control.

To talk about voltage control you need to describe the change in Vbe with current either ignored or described in terms of Vbe.

Essentially this is a restatement of Arouse1973's argument above.

If we teach the use Ib and hFE in those cases where we can, then regardless of the fact that we know it is not the base current controlling the collector current, that's a reasonable reading of the formula.

If we are answering a question about how to switch a given load then we can either use this formula and get to an answer quickly, offer we can get into a long discussion about transconductance before getting back to saying what is effectively "but we'll just use a formula that seems to ignore all of that". And then in practice we continue to ignore the arguably correct method as long as we're in the fairly broad operating region where the formula works well enough.

This is simply going back to the crossing the road analogy. Unless the person is a budding electronic engineer without an immediate problem to solve, the simple solution is probably good enough. Later when that person has a better under standing of the heuristic you can expose a little more of what is hiding behind that level of abstraction.

And its not as if the voltage control model actually describes 100% accurately what's going on. I'm sure we could be more accurate by looking at the transistor at an even more fundamental level. At some point quantum effects will provide some small iota of additional accuracy (excluding where we take things to such a small scale that they dominate).

I accept that if you're teaching EE then you start with a more deeper theoretical model, but you do that specifically in the knowledge that you're not solving an immediate technical problem, you're arming them for solving a broader class of problem. In this case it may be fine to teach the general before moving on to specific techniques, but then you also have a captive audience...
 

Arouse1973

Adam
Dec 18, 2013
5,178
Joined
Dec 18, 2013
Messages
5,178
Yes I agree with Steve on this one. If you are using base current in your calculation then you are using a form of current control. But without knowing it you are changing the Vbe which is what is controlling the collector current. In this respect do we really care what the Vbe is doing as long as we have a method of some form of control. We could say we are using the current in RB to control the voltage between the base and emitter which in turn controls the collector current.
Adam
 

LvW

Apr 12, 2014
604
Joined
Apr 12, 2014
Messages
604
Hi Steve, I think we are pretty close together - only some "minor" differences:
In my opinion, if your calculations require in Ib with Vbe ignored or assumed to be constant then you are describing current control.
To talk about voltage control you need to describe the change in Vbe with current either ignored or described in terms of Vbe.
All of us assume a VBE value while designing an amplifier stage, do we not? And what about IB? No tolerances/uncertainties for IB=IC/B ?
I don´t think I have to "describe" the change in VBE separately. I have mentioned in my former post that delta(VBE) is part of the transconductance.
I like to emphasize that I do not argue on a pure theoretical basis - just the opposite is true:
When calculating the gain of a stage I need a good estimate for the transconductance gm (for a given IC). We all know that we have gm=IC/Vt.
This equation is derived from the exponential I-V relationship. That means: For very practical tasks we exploit the voltage control feature - and each beginner should not only use this relation, he also should know where it comes from. Don´t you agree?

I accept that if you're teaching EE then you start with a more deeper theoretical model, but you do that specifically in the knowledge that you're not solving an immediate technical problem, you're arming them for solving a broader class of problem. In this case it may be fine to teach the general before moving on to specific techniques, but then you also have a captive audience...

I suppose, before learning transistor basics all EE students will learn how a pn diode works. And they will hear about Shockley´s exponential equation.
Therefore, my question: Why should it be a problem to tell the students that the pn junction within a BJT shows a similar behaviour? I cannot understand.
In contrary, according to my experience the opposite happens very often: Student´s hear or read something about "IB controls IC" and they ask: Why? Does Shockley´s equation not apply to this kind of pn junction? What do you propose to answer?
 
Status
Not open for further replies.
Top