Maker Pro
Maker Pro

Assistance with Transistor Theory

Status
Not open for further replies.

Laywah

Dec 19, 2014
9
Joined
Dec 19, 2014
Messages
9
Hey All,

I am sure theory is the last thing many people want to do and understand. I certainly hated theory years ago!
I am looking to better understand the transistor device and how it works including the understanding of the electrical symbol.
I have already done some study and know things such as the base, emitter and collector. I have also used them in a very basic scenario

some questions I have are:
  1. Why is the emitter on an NPN transistor conned to ground? The arrow points in the wrong direction for what I understand electricity travels from negative to Positive.
  2. Is their primary function just an electronic switch?
Sorry if these questions sound stupid or have been discussed somewhere before. I haven't been able to find these answers myself.
 

mofy

Dec 19, 2014
46
Joined
Dec 19, 2014
Messages
46
1. There is a tradition in electronics that dates back to before electrons were known to be negative charge carriers, that current flows from positive to negative. That is why schematics have the positive supplies above the negative, to show a downward flow like gravity. In semiconductors both holes(positive) and electrons(negative) are charge carriers and mobile.
2. No the primary function is a current amplifier. The current injected into the base of the transistor gets multiplied by the Beta of the transistor and pulled down through the collector. Assuming a well biased transistor. This is a very simplified explanation!
 

Laywah

Dec 19, 2014
9
Joined
Dec 19, 2014
Messages
9
Thanks mofy,
So my understanding of transistors being like electric switch is wrong, and that the current comes from base not from ground and goes out the collector.

This makes things a little more clear.
I will keep experimenting.
 

mofy

Dec 19, 2014
46
Joined
Dec 19, 2014
Messages
46
Technically it flows into the base, into the collector and out the emitter for NPN.
 

davenn

Moderator
Sep 5, 2009
14,254
Joined
Sep 5, 2009
Messages
14,254
So my understanding of transistors being like electric switch is wrong,

no, not completely, a transistor makes a VERY GOOD switch
its just one of its uses :)
 

KJ6EAD

Aug 13, 2011
1,114
Joined
Aug 13, 2011
Messages
1,114
The Wikipedia article on transistors is pretty good.
 

hevans1944

Hop - AC8NS
Jun 21, 2012
4,878
Joined
Jun 21, 2012
Messages
4,878
Actually, many of us here love theory. Sure beats trial-and-error, most of the time, if you know what is supposed to happen.
 

LvW

Apr 12, 2014
604
Joined
Apr 12, 2014
Messages
604
No the primary function is a current amplifier. The current injected into the base of the transistor gets multiplied by the Beta of the transistor and pulled down through the collector. Assuming a well biased transistor. This is a very simplified explanation!

Laywah - I must confess that I hesitate a little to write the following (because in the past I had some - bad and disappointing - experiences with similar discussions, even personal attacks! ), nevertheless:

If you really want to understand the working principle of a bipolar junction transistor (BJT), you must know that the BJT is a voltage controlled device.
That means: The current Ic is solely determined by the applied base-emitter voltage Vbe.
It is true that, of course, there is also a current Ib (through the base terminal) - however, this current has absolutely no controlling fuinction.
In contrary - it is a current that is unwanted and is kept as small as possible, but it cannot be avoided.

One reason for this false understanding of the BJT working principle (even contained in some textbooks) may be the fact that - unfortunately - the ratio Ic/Ib from the beginning was called "current gain" - and some people therfore believe that the BJT is a current amplifier. But it`s simply false. A mathematical formula like Ic=beta*Ib never can tell you anything about cause and effect.

When you know the working principle of a pn diode - and, in particular, the fact that the current-voltage relation across the pn junction is desribed by the famous exponential equation from Shockley, it is just a tiny and logical step to accept that, of course, the pn junction in the BJT follows the same rules.
It is really funny, even the defenders of the current-control party use in their calculations the voltage Vbe for determining different operational points (app. 0.65...0.7 volts for A-operation and Vbe=0.1...0.3 volts for AB or B-operations), but they still think that Ib controls Ic - although in their calculatiuons they use Vbe only.

To me, the whole story is really a phenomenon: There is not a single proof or indication that Ib would control Ic, in contrary: There are many proofs that it is the voltage Vbe that controls Ic.
But the fairy tale lives on.....
 

Merlin3189

Aug 4, 2011
250
Joined
Aug 4, 2011
Messages
250
When I saw LvW's intro, I felt a bit sympathetic. People do seem a bit intolerant of dissident views. But having read the rest, my sympathy waned!
Laywah can decide what to think, but I'd like to offer a PoV.
"you must know that the BJT is a voltage controlled device. " I disagree: A BJT is a lump of semiconductor with wires attached. It does what it does.
What I think you should say is, "a BJT can be modelled as a voltage controlled device" or "a BJT can be modelled as a current controlled device."
You can use whichever model helps you most in any particular situation, but the result will only be an approximation to the reality.
Generally people choose to use the model which is simplest and easiest to use, because good enough is generally good enough.

I really don't know whether a voltage model or the current model is generally better. I was brought up on the current model and have never used a voltage model.
As I am old enough to have learnt with valves and had to convert to transistors, I would be very surprised to learn that there was a simple voltage model. Valves used a voltage model and we had to be dragged (many kicking and screaming) to the current model when transistors arrived. Had there been an easy voltage model for the BJT, why did no one mention it back then? They'd have been VERY popular and the current modellers would have never gotten a hearing.

If LvW understands a voltage model, I at least wish him well to use it. If he uses it to explain his answers to questions here, I will be happy to try to understand them and will be very pleased if it helps me understand a circuit better than the current model.
 

LvW

Apr 12, 2014
604
Joined
Apr 12, 2014
Messages
604
Merlin - thank you for your reply. I understand what you mean (and, please, excuse my bad english while explaining my view), but there is one point I cannot agree with you:
For my opinion, we should strictly discriminate between a MODEL (which can be good or bad) and a desription of the PHYSICAL working principle. I agree with you that in many cases the simple model of current-control will work. However, there are many counter examples - and I have mentioned one of them (class-A and clas-B operation).
I do not intend to go down to the level of charge carriers - that`s not necessary.

Here is another example: From feedback theory we know that the input resistance goes up for voltage feedback and goes down for current feedback.
Now -what does the input resitance of a common-emitter amplifier with Re-feedback? It goes up because - internally ! - it is the base-emitter voltage that causes the feedback.
You cannot explain this fact based on the model of current-control. There are many other examples!

Quote: Valves used a voltage model and we had to be dragged (many kicking and screaming) to the current model when transistors arrived.
May be - however the physical reality is different.

Quote: Had there been an easy voltage model for the BJT, why did no one mention it back then.
Did you never hear about the Gummel-Poon model (incorporated in simulation programs) ?

May I ask you one single question (and forgetting for the moment everything you have learned) :
Is there any single reason not to assume that the base-emitter pn junction will have the same properties as the junction of the classical diode?
Up to now, nobody (from the "current-control party") has made an attempt to prove that the BJT would be current controlled. They have learned that, they did believe that ... and that`s all.
And - finally - in a discussion like this, they claim to be right.
But it is not a big problem to show why the BJT is voltage-controlled - just from circuit behaviuour. (Sometimes I have the feeling to speak against a wall).

I could give you many, many references supporting my position - here is one:
http://www.kevinaylward.co.uk/ee/voltagecontrolledbipolar/voltagecontrolledbipolar.xht

Other positive references are from Universities of Berkeley and Stanford.

I would be very happy if you would answer my (underlined) question above .
 
Last edited:

hevans1944

Hop - AC8NS
Jun 21, 2012
4,878
Joined
Jun 21, 2012
Messages
4,878
The Wikipedia article (almost cited by @KJ6EAD) provides a "beginners" explanation of bipolar junction transistors, or BJTs. The starting equation for BJT design is the Shockley Diode Equation which led to the Ebers-Moll model. As @Merlin3189 implies in his post #9 above, this is only an approximation to reality. A better approximation is the Gummel-Poon model mentioned by @LvW.

As computers become ever more powerful, we will eventually reach the point where an almost exact quantum-mechanical model of any semiconductor device can be produced. We aren't quite there yet, but any model (however accurate it may be) is only a model, not reality. The "map is not the territory" is an old saying that is appropriate here.

I think most of the "controversy" of current-control versus voltage-control revolves around cause-and-effect arguments, which may be non-productive to getting things accomplished at the beginner level.

It is certainly true that NO collector current (except reverse leakage current) flows through the reverse-biased base-collector junction if the base-emitter junction of a BJT is shorted. That is, zero base-emitter voltage and zero base-emitter current results in no collector current.

And it is certainly true that if there is an electrical field applied to forward-bias the base-emitter junction, with a reverse-bias field applied to the base-collector junction, then current will flow in the collector that is proportional to the base-emitter junction voltage. This is cause and effect: the electrical field in the base-emitter junction causes the collector current effect in the base-collector junction.

Perhaps it is not so obvious that it is base-emitter voltage that produces an electrical field across the junction between the base area and emitter area. The junction between base and emitter is very thin, so a small voltage can produce a very large electrical field. It is this electrical field that cause carriers to migrate from the heavily doped emitter region into the base region. Without the field, some do wander into the base region by thermal motion, where they neutralize oppositely charged carriers, creating a depletion region whose thickness is then diminished by a forward biasing electric field.

What happens as a result of this migration of carriers from the emitter is complicated, but the end result is collector current that now flows through the reverse-biased base-collector junction. The fact that there is a base current is almost incidental. It is the base-emitter voltage causing an electrical field to be produced across the base-emitter junction that is responsible for the collector current.

Externally, the base-emitter voltage is almost always a consequence of a voltage applied to a resistor in series with the forward-biased base-emitter junction, which voltage then causes a base current, Ib, to flow as approximated by the Shockley Diode Equation. Experimentally, this base current is related to the collector current, Ic, by a "gain" factor, beta, so Ic is approximately equal to Ib times beta. This is a "good enuf" explanation for simple circuit design, but it doesn't accurately describe everything that is going on. No model, so far, accurately describes everything that goes on in a BJT.

Consider for example a Darlington-connected pair of transistors, or a Sziklai Darlington-connected pair of complementary NPN and PNP transistors. It appears that current in the base of the first transistor is multiplied by its beta to cause a much larger base current in the second transistor, which then multiplies that base current by its own beta to create an overall collector current of Ib x beta1 x beta2. It is circuits like these that perpetuate the "myth" that transistors are current-controlled devices.

My advice? Go buy a few hundred 2N3904 (NPN) and 2N3906 (PNP) small-signal transistors. These are really inexpensive, a few cents each, so you can afford to "blow up" or let the "magic smoke" out of few while learning how they work. Also buy a bag of 1/4-watt resistors in assorted values, a solderless prototyping board, two 9 V "transistor radio" batteries, and snap-on leads for the batteries. Buy a decent 3-1/2 digit multimeter so you can measure electrical things: voltage, current, resistance. Now sit down and build stuff from simple circuits you can find on-line.

Oh, most people also throw in a few cheap LEDs (Light Emitting Diodes) to the basic experimenter's kit since these give immediate visual feedback when you do something right. You can add stuff, like mechanical switches or relays, to the kit as you progress in your learning. A few low-power LM555-type timers are fun to play with. Eventually you will want your very own line-operated bench power supply. Come back here and ask questions, but do some web surfing yourself to answer most questions. Google is your friend.
 

chopnhack

Apr 28, 2014
1,576
Joined
Apr 28, 2014
Messages
1,576
My advice? Go buy a few hundred 2N3904 (NPN) and 2N3906 (PNP) small-signal transistors. These are really inexpensive, a few cents each, so you can afford to "blow up" or let the "magic smoke" out of few while learning how they work. Also buy a bag of 1/4-watt resistors in assorted values, a solderless prototyping board, two 9 V "transistor radio" batteries, and snap-on leads for the batteries. Buy a decent 3-1/2 digit multimeter so you can measure electrical things: voltage, current, resistance. Now sit down and build stuff from simple circuits you can find on-line.

Oh, most people also throw in a few cheap LEDs (Light Emitting Diodes) to the basic experimenter's kit since these give immediate visual feedback when you do something right. You can add stuff, like mechanical switches or relays, to the kit as you progress in your learning. A few low-power LM555-type timers are fun to play with. Eventually you will want your very own line-operated bench power supply. Come back here and ask questions, but do some web surfing yourself to answer most questions. Google is your friend.

While theory is great, and hevans certainly knows his stuff and has the background to understand it - the best piece of advice and the most fun has just been given IMHO :D

That should keep me busy for awhile.
Thanks!
 

Merlin3189

Aug 4, 2011
250
Joined
Aug 4, 2011
Messages
250
I am happy to agree that the base-emitter junction will behave like any other diode.

I did not know of the Gummel-Poon model and have only a vague recollection of the Ebers-Moll equation. No doubt if I understood these, I could do some complicated calculations and might gain some useful insights into transistors. But I don't, so I just have to look at manufacturers' data sheets and use the value of beta they give.

It is interesting that you say these were around from the start of transistors. Presumably even though they are voltage models, they must be different from the transconductance model we used for valves or we would not have had a problem. I guess the difference is that with valves we generally assumed grid current (DC) was zero, but with BJTs there is always some DC base current.

Perhaps I should apologise for my scepticism about your voltage model. I really can't disagree with your "description of the PHYSICAL working principle." since I don't understand that. I have no idea how transistors work as a physical principal.
People talk about holes. I know there are places where there could be an electron, but isn't. But to say a hole moves, seems nonsensical to me. If there were a row of empty seats and one occupied by a person, then the person got up and moved to sit in an empty seat, I would not say the seat had moved! You could say the seat doesn't move, it is the emptiness that moves. But emptiness is just a concept not a physical object. I could give more reasons why holes are silly, but that is irrelevant.

So I just have to work with the numerical models that other people have derived from physical principles that they understand. In the words attributed to Richard Feynman, "Don't speculate, calculate." If their formulae work, it doesn't matter to me if their physical reality is little pixies kicking footballs around inside the transistor.

Now if I take the voltage model, then I will get very accurate and detailed results (so WikiP tells me) but that is only if I can solve their complex equations and if I can control the BE voltage (which I don't think I generally can.)

It is much easier for me to accept the current model and use Ic = beta x Ib with Vbe about half a Volt. I know it doesn't give me the accurate results you get, but it's dead easy and good enough for most simple circuits. Since beta is reasonably large and many circuits use significant negative feedback, the exact transistor gain (however you calculate it) does not matter very much.

As I said, I can't disagree with your voltage model, because I don't understand it. But if you use it to work out what happens in a circuit or to design a circuit, I'll try to follow as best I can and be as pleased as anyone if it works.
The only reason I did not like your original post was that you insist people must use the voltage model. People use the current model because it helps them design & understand circuits. Does the voltage model do that? I cannot see how it helps the average person at all. What you would need to do would be to show how your model (or reality) can be used to get everyday results.

PS. just seen Heavans fine post and must agree with his recommendation: play with some circuits.
 
Last edited:

Laplace

Apr 4, 2010
1,252
Joined
Apr 4, 2010
Messages
1,252
I would just ask LvW to use the 'voltage model' to show why the collector current is proportional to the base current. Then I would use the 'current model' for doing circuit design.
 

LvW

Apr 12, 2014
604
Joined
Apr 12, 2014
Messages
604
I would just ask LvW to use the 'voltage model' to show why the collector current is proportional to the base current. Then I would use the 'current model' for doing circuit design.

Quote Merlin:"The only reason I did not like your original post was that you insist people must use the voltage model. People use the current model because it helps them design & understand circuits. Does the voltage model do that? I cannot see how it helps the average person at all. What you would need to do would be to show how your model (or reality) can be used to get everyday results."

Gentlemen, I am afraid there is a misunderstanding between us - and, certainly, I am responsible due to my limited knowledge of the english language.
Merlin - I never have said that "people must use the voltage model" . I didn`t speak about usage, I did speak about theoretical explanation of the working principle.
And that is an important difference (see below).
Laplace - the same applies to your contribution. I didn`t speak about "doing circuit design".
_______________________

Now let me explain:
* I think, we strictly must discriminate between the two activities: (a) Designing a BJT stage and (b) Understanding the working principle of the BJT.

For designing an amplifier stage (i.e. common emitter) we all use the same design principles and the same equations.
There is absolutely no difference - whether we rely on current-control or on voltage control.
And - of course - for designing the resistive voltage divider for base biasing I take the base current into account (Ib=Ic/beta).
But this has nothing to do with the principle how the BJT`s collector current is controlled.
And - as far as the "current-control party" is concerned - don`t they use the quantity Vbe=0.65...0.7 volts in their calculation?
And, of course, they again are using other Vbe values for clas-A/B or class-B operation.
The same applies to the Temp-Co of the base-emitter voltage (-2mV/K) to keep the Ic value constant.
Is there any similar figure which describes how beta=Ic/Ib depends on temperature?

In short: Everybody is using the same equations - the whole design process is completely independent on the question "current or voltage-controlled".

But remember: The original question was NOT how to design an amplifier stage but "Assistance with transistor theory"!

I hope, I could somewhat clarify things.
And, finally, thank you to Hevans1944 and his valuable contribution in which he explains things from the physical point of view.
 

Laplace

Apr 4, 2010
1,252
Joined
Apr 4, 2010
Messages
1,252
And - as far as the "current-control party" is concerned - don`t they use the quantity Vbe=0.65...0.7 volts in their calculation?
Of what use is the quantity Vbe? We force a current to flow through the base-emitter junction and the semiconductor junction generates a voltage drop. It is a voltage that must be accounted for but Vbe is not otherwise used for any particular purpose. Vbe can be related to collector current through semiconductor physics theory, but we do not control Vbe directly. Instead we control the base current that generates Vbe across the PN junction. The BJT has the useful property that its collector current is fairly well proportional to its base current. I would ask again for you to use the voltage-model to show why this proportionality exists. But the BJT is not a voltage-controlled device for the simple reason that we do not control the base voltage -- we control the base current in order to control the collector current, and that makes the BJT a current-controlled device. This is not a discussion about semiconductor physics, it is about how the English language dictionary defines 'control'.
 

LvW

Apr 12, 2014
604
Joined
Apr 12, 2014
Messages
604
This is not a discussion about semiconductor physics, it is about how the English language dictionary defines 'control'.

Ahh - I see, may be that both of use have a different understanding of the term "control". I am using this term as it is used and defined in classical closed-loop control systems.
That means (my explanation):
Of course, I can vary the base current (using a so-called "current source", in fact a large series resistor and a voltage source) - and, as a result, the collector current will vary according to Ic=beta*Ib.
But, to my understanding, this does not mean that Ib controls Ic. Instead, it is still the voltage develloped across the BE junction that controls Ic.
But I agree with you that it is absolutely necessary to come to a common understanding about terms like "control".

As I have mentioned, there are many effects to be observed in BJT based circuits which support my claim of voltage control (my understanding of this term).
Thus, it is not necessary to go down to the charged carrier domain to find a justification.
1.) Common-emitter stage: Signal input resistance goes up for Re-feedback. According to feedback theory this happens only if the feedback quantity is a voltage.
2.) Early effect: This effect is solely caused by the rising electrical field (or the corresponding voltage) across the pn junction.

More than that, there are many other electronic circuits which can be explained (working principle) using Vbe-control only.
(And many sources with high reputation - e.g. Stanford and Berkeley Universities - support this view.
 

BobK

Jan 5, 2010
7,682
Joined
Jan 5, 2010
Messages
7,682
Laplace: Look at this nifty device called the lever. If I push on the long side of the lever, I can control a much larger weight on the other side of the lever.

LvW: No, your pushing on the long end of the lever is not controlling the weight, it is the multiplied force on the other side of the lever that is controlling the weight.

Bob

Edit: Bob: No, you are both wrong, it is the weight on the short end of the lever that is controlling how much you have to push on the long end.
 

LvW

Apr 12, 2014
604
Joined
Apr 12, 2014
Messages
604
LvW: No, your pushing on the long end of the lever is not controlling the weight, it is the multiplied force on the other side of the lever that is controlling the weight.
.
Bob - I must confess that I do not understand your reply (due to my limited knowledge of your language). Is it an anlogy?
 

KJ6EAD

Aug 13, 2011
1,114
Joined
Aug 13, 2011
Messages
1,114
If I interpret the analogy correctly, Bob's making the point that both ways of modeling transistors have enough historical validity to work for construction of circuits; that it's mostly a difference in perspective (point of view) and semantics that results in people choosing sides and trying to convince the others to see it their way.
 
Status
Not open for further replies.
Top