# Understanding Transistor

Discussion in 'General Electronics Chat' started by sdmuashr, Mar 24, 2012.

1. ### RatchitWell-Known Member

Joined:
Mar 12, 2008
Messages:
1,934
Likes:
78
Claude,

It sounds like a clunky, clumsy way to distribute power. So if a heating element in a electric oven or a light bulb broke, we would have an electric arc until the arc got shorted?

Sure it does, the equations given prove it. The page was from Gerold W. Neudeck's The Bipolar Jujnction Transistor, by the way.

The energy lost and current in the base is very low because most of the current is shunted to the collector because of the geometry of the BJT. This includes a very thin base, asymmetrical doping, and a Vcc. Vcc is what supplies most current and power in the active region. Vbe is what enables the charge carriers to keep entering the base region where they are swept onto the collector. If it were possible, Ib could be eliminated. But it is inevitable that a small amount of Ib wastage current will be present.

Whichever way you figure it out, the Vbe controls the diffusion which brings charge carriers into the base region where they are sent on to the collector. The base current is not needed, nor is it in the equation for Ic. The base current is a waste charge flow that happens to be proportional to Ic.

No, not after. Vbe is what enables diffusion in the first place. Current drift has nothing to do with this.

The physics also apply to DC and lower frequency. It is wrong to assert that cause and effect cannot be determined at DC or lower frequencies. Diffusion still exists under those conditions. Charge storage does affect the timing of the effects, but it does not affect the cause and effect. I did address that in page 5, posts 49 through 51 in this thread. Did you read it?

We are in the active region at DC or low frequencies. I do recognize that storage charge affects the timing of responses between voltage and current. I showed that earlier in this thread. Sedra and Neudeck use Vbe in their equations to show a causal, not functional relationship.

All the above paragraph has nothing to do with Vbe controlling Ic. Of course AC voltage can exist without current if no capacitance if present.

The only causal current activated device I can think of at the moment is a magnetic amplifier. A LED works well when driven by a current, but like all junction diodes, it is causal voltage activated device which activates the current present within.

Yes, current does not exist without voltage, either internal or external. Voltage represents energy per unit charge, and the charges just are not going to move anywhere without energy to do so.

Ratch

2. ### RatchitWell-Known Member

Joined:
Mar 12, 2008
Messages:
1,934
Likes:
78
Mr RB,

Just like teaching Schockley's equations is "silly"? Did you get your answer from the previous post?

It's OK to teach models, but don't try to use them to explain how devices work. Models only are useful in determining what devices do.

I guess we gave him his money's worth with respect to how BJTs work, didn't we? But his question was not specific enough and too broad and general to answer succinctly.

Why current control? You can make a black box be voltage controlled using only BJTs and nonactive components, or make a black box be current controlled using only FETs and nonactive components. Same with an opamp.

After the original question is answered, then further discussion can be in order.

Ratch

3. ### RatchitWell-Known Member

Joined:
Mar 12, 2008
Messages:
1,934
Likes:
78
skyhawk,

Yes, a LED is just a special type of junction diode which has the basic characteristics of a regular junction diode. That means diffusion made possible with bipolar materials, and diffusion control by voltage across its terminals.

Further response: I forgot to mention that the exponential relationship of voltage and current is due to the diffusion mechanism of the diode. Diffusion of one liquid into another is also a exponential relationship.

Ratch

Last edited: Aug 14, 2012

Joined:
Jan 12, 1997
Messages:
-
Likes:
0

5. ### Mr RBWell-Known Member

Joined:
Jul 22, 2008
Messages:
4,716
Likes:
194
Location:
Out there

This time I'll get "hopelessly pedantic" with you.

He said " the pratical use of transistors (BJT, FETs) and how they work in circuit about their biasing. The material should be down to earth."

What part of that do you find to be "not specific"?

You just WANT to argue the physics of the insides of the transistor because it's capable of fueling a long argument with educated people, where you get to be "hopelessly pedantic".

But in regards to the more important point of providing the OP with his requested info your argument is wrong, covering the inner workings of a transistor in the fashion you have has NOT been about "practical use" "how they work in circuit" and is far from being "down to earth".

So in regards to the OP's stated requirements, all your replies have been of a low level of competence.

If the OP asked you how to use some bricks to make a brick wall, and you started going on about the ceramic crystal lattice inside a brick, I would start to think you are not very good at this bricklaying stuff.

• Like x 1
6. ### RatchitWell-Known Member

Joined:
Mar 12, 2008
Messages:
1,934
Likes:
78
Mr RB,

OK, works for me.

Yes, he did.

The short answer is all the four things he mentioned. Do you realize how extensive the subject is about how transistors work in a circuit? And he wanted to learn about two different transistor families. In addition, he wanted to find out about biasing, which is a subject all by itself. Do you find anything "specific" about that question? The OP did not ask a question, he asked to attend a transistor course. Did he expect us to award him a certificate also?

True, I do enjoy a good discussion. But if their arguments are true, and mine are false, then it would not last long, would it?

True, it would have to be a transistor course to cover what he asked about.

What has competence got to do with anything? Did I say something wrong? If so, show me and prove it.

But that is not what happened is it? The OP asked a question that anyone would have to write a book to answer properly.

Ratch

7. ### WinterstoneBanned

Joined:
Mar 24, 2012
Messages:
536
Likes:
15
Location:
Europe
Regarding the OP's question in post#1:

(1) In my post#81 I have explained how to calculate the gain of a typical common emitter stage with Re feedback. I have used no ("theoretical, academic") 4-pole parameters but, of course, the well known and always used transconductance g=Ic/Vt. During this calculation it was not necessary at all to use the relationship Ic=B*Ib or ic=beta*ib. Instead I have used the definition of the transconductance which is based on the voltage control property of Ic.

Response: Zero. Nobody from the "current-control"-party did respond - in spite of my request to comment.

(2) In the same contribution I have described how such a stage is designed: Choice/calculation of Ic, Rc, Re, R1 and R2 (bias network).
For calculation of the bias network I have used, of course, the dc voltage Vbe=0,65 volts and the relation Ib=Ic/beta as a very rough information
to have an estimate for the maximum current Ib. I am sure everybody knows the reason (current through the bias network larger by a factor 0f 6...10).
An exact calculation is not possible because the value of beta is known typically with a tolerance of 200...300% (or even more).

I think, this is the well-known engineering approach without measurements of the 4-pole parameters prior to design. Or not?

Was any control scheme (indirectly) used during the calculation? Yes - because the Re-feedback relies on voltage-feedback (i. e. voltage control) only.
Everybody knows that this is the reason to select a bias network that - as good as possible - can mimic a dc voltage source.
The equation Ib=Ic/beta was used only for a rough estimate for Ib.

Response: Zero. Nobody from the "current-control"-party did respond - in spite of my request to comment.
_________________________________________________________________________________________

Finally, I think that both informations (1) and (2) are answering some of the OP's questions.
I would be happy (and perhaps also the OP) to see similar technical answers (circuit design and gain calculation) based o the current-control "model".
(Why "model"? Didn`'t we try to find out the physical truth?).
This would be a good proof for Claude's claim that the "current-model" works best from outside (black box).

W.

Last edited: Aug 16, 2012
8. ### WinterstoneBanned

Joined:
Mar 24, 2012
Messages:
536
Likes:
15
Location:
Europe
To all who are interested:

Please find as an pdf-attachement an excerpt from W. Shockley's patent.

on page 2 the lines 33-39 and 45-49 , and
on page 3 the lines 3-5 (Ib is neglected to concentrate on the main effect!) and 50-53.

Last edited: Aug 15, 2012
9. ### Mr RBWell-Known Member

Joined:
Jul 22, 2008
Messages:
4,716
Likes:
194
Location:
Out there
Still theoretical huh?

Is that really relevant to the OP's "down to earth" "how to bias the transistor in a circuit"?

Are you saying that th OP should learn that if he wants to get 100mA Ic he should bias the Vbe to 0.5973v? And watch out for temperature changes?

Or is it better the OP learns that he can plug the transistor into his multipmeter, read a beta of 200, and bias it with 0.5mA Ib to get 100mA IC? And not have to worry too much about temperature changes.

There are important reaons that it is taught in classrooms that bipolar trannies are current controlled and FETs are voltage controlled, and the OP is a good example of those reasons.

I'll stick with the bricklaying analogy, electronics design is a lot like bricklaying; knowing how to connect all the bricks and make it work. The actual processes occurring inside any brick are of a lower importance, especially with beginner bricklayers.

10. ### WinterstoneBanned

Joined:
Mar 24, 2012
Messages:
536
Likes:
15
Location:
Europe
You ask if it is relevant how to bias a BJT? Is this a serious question? I cannot believe.

I don't know how much transistor stages you have designed - however, a good design always utilizes Re-feedback (as mentioned in my referenced examples!) in order to cope for temperature changes and to stabilize the dc operating point. It is easy to show that - in this case - it really does not matter if your calculations are based on 0.65 or 0.7 volts. Thus, your "counter example" with 0.5973 volts is based on false assumptions. Didn't you read my example in post#81 ?

This statement causes three questions:
1.) Do you really think it is a good approach to measure for each single transistor the respective beta? That is exact the opposite of a good and robust engineering approach aiming at a design that is independent as much as possible on active parameters.
2.) In this case - how do you realize the biasing with Ib=0.5 mA? (To me, THIS sounds a bit to theoretical). A circuitry explaining your idea would be very helpful.
3.) Why do you think that - in this case - we "not have to worry too much about temperature changes" ? Is beta not dependent on temperature?
4.) What is your approach for calculating the voltage gain of a simple BJT amplifying stage ? During evaluation of the formula you will recognize the problem.

Again - as in all former contributions from the current-control defenders - just a claim without any proof and without any examples.
Please, list some of these "important" reasons.
______________________________________

To make my position clear: I am aware of the relation Ic=beta*Ib and I will use it - if it seems necessary or appropriate (seldom enough) - but why should I teach the physical untruth?
Just because this equation looks a bit simpler than an exponential function? Where are the advantages? Just to state that there are "important reasons" without any single example is not enough!

In contrary: Shall I list some severe disadvantages connected with this false understanding?
Just one example (in addition to my former common emitter example - calculation of parts values and gain):
As you know - the differential amplifier is a basic circuit block used in many, many applications. Did you ever try to verify/explain its tanh-transfer function based on the current control approach?
You will run in a similar dilemma trying to explain the function of current mirrors, Gilbert cell, translinear circuits, log-domain applications,...
And what about the classical opamp? To understand its operation one should realize - at least - that the input stage is a differential amplifier. Surprisingly, all beginners learn that it is voltage controlled (even for a BJT input stage).

Remark: Physicians solve the dilemma for explaining the nature of light by accepting a kind of "duality" - as long as the have no other choice.
But is there any reason for defining such a duality for BJT operation (sometimes current and sometimes voltage controlled) ???

Last edited: Aug 17, 2012
11. ### Mr RBWell-Known Member

Joined:
Jul 22, 2008
Messages:
4,716
Likes:
194
Location:
Out there
Sometimes I would swear that you and Ratchit are the same pedantic person.

So rather than listen to another 300 word essay from you guys, here's a simple A or B question;

For a beginner like the OP, asking for simple, real-world, down-to-earth biasing, who needs to set the collector current to 100 mA, is it better that he tries to;
A. Produce an Ic current of 100mA by applying X volts Vbe
B. Produce an Ic current of 100mA by applying Y amps Ib

Please answer A or B with no other words. Yes the two answers are not perfect, but we need to pick one and teach it to the beginner.

12. ### WinterstoneBanned

Joined:
Mar 24, 2012
Messages:
536
Likes:
15
Location:
Europe
Thank you very much for the compliment. Now I have learned what "pedantic" means: Asking for examples which can support and verify some of your statements (example: "important reasons" to teach the physical untruth of current-control).
Is it really necessary to start with personal attacks? I think I should stop the discussion with you at this point.

May I remind you on ericgibbs and his words in post'81 of the thread "forum dying?" :

Its unfortunate that such a small number of members seem to consider they are not subject to the rules and have some form of diplomatic immunity and are able to post insulting and offence remarks and not be disciplined, what their personal agenda's are, heaven knows.

By the way- do you remember your own words? Post13 (Mr RB): Give it >0.6v at the base, and it turns ON.

It is really interesting that you do not hesitate to ask questions in spite of the fact that - up to now - you did not answer ONE SINGLE technical question from my side (see my post#129).
By the way - your question has been answered already several times. It shouldn't be a surprise for you (my post#129).
I only can hope that other (passive) visitors of this thread - in particular beginners - don't consider the way and the style of this "discussion" as typical for engineers. It is not!

Last edited: Aug 18, 2012
13. ### Mr RBWell-Known Member

Joined:
Jul 22, 2008
Messages:
4,716
Likes:
194
Location:
Out there
My post was not meant as an attack, just an observation that (to me) your posts seemed to be getting as pedantic as the self-confessed "hopelessley pedantic" Ratchit.

I'm exiting the thread, as it is obvious some people are insistant on discussing theory of the transistor internals rather than addressing the OP's points of "down to earth" "how to bias a transistor in a circuit" etc and those people continuing to refuse to accept they are wrong for doing so.

14. ### RatchitWell-Known Member

Joined:
Mar 12, 2008
Messages:
1,934
Likes:
78
Mr RB,

Just can't seem to "get it", can you? Maybe you can ask the OP if he is interested in biasing the transistor for operating point or stability. If stability, does he want to stabilize for variations in Vcc, Icbo, β, or Vbe? Does he want to use one voltage source or two? He also wants to cover FETs, right. Oh, and do it all without knowing much theory about what he is doing. "Down to earth", I believe you called it. How long do you think it would take to teach him all those things? I countered with the same point previously, but you have not addressed it. How about doing so now?

Ratch

15. ### WinterstoneBanned

Joined:
Mar 24, 2012
Messages:
536
Likes:
15
Location:
Europe
Do you realize that YOU are one of those persons who were not able - up to now - to answer the question "how to bias a transistor in a circuit" - based on your current-control scheme?
May I remind you again that in post#81 it was outlined how the engineering approach for biasing a transistor looks like - based on the voltage-control principle ?
________________________________________________________________________________________________

Final remarks and conclusions: As already indicated in my former post I think we should close this thread.

The thread was started (post#1) with the following question
Therefore, I would be very thankful if anyone here can guide me to any on-line material or a book (preferably low cost that can be bought) which is about the practical use of transistors (BJT, FETs) and how they work in circuit about their biasing.

Then, starting with post#8 the thread merged into a long discussion about „how a transistor really works“: Controlled by a voltage or a current ? Unfortunately, later it has turned into into a dispute rather than an objective discussion based on technical/physical facts.

I was and still I am of the opinion that the working principle of the BJT relies on voltage control.
I am aware that some textbooks (and some forum members) state the contrary, however, many competent persons with superb technical (not only academic) reputation – W. Shockley, Barry Gilbert, Paul Brokav, George Wilson, Bob Widlar, Bob Pease,.. – are also supporting the principle of voltage control. More than that, their inventions are based on this principle.

In response to the original question (post#1) some details were outlined how to bias a common emitter stage and how to determine/calculate the voltage gain (my posts #81). It was shown that a reasonable and engineering approach is based on BJT voltage control only (concept of transconductance and voltage-feedback using Re).

At the same time the „current-control“-defenders have been asked several times
* to comment/criticize these explanations („what is wrong“?), and
* to present similar formulas/calculations which could support their meanings (see, for example, post 81, 91, 100, 126).

Unfortunately, no answer - instead just general claims/statements like "there are important reasons" for relying on current-control.
That’s the situation – and I am afraid, it will not improve.
So everybody can draw his own conclusions.
Perhaps we'll meet again in another thread.

Winterstone

Last edited: Aug 19, 2012
16. ### deanwesterburgNew Member

Joined:
Jun 28, 2013
Messages:
1
Likes:
0
Location:
Cincinnati, Ohio
You get it!! A person who understands the technology and can break it down to the masses is an invaluable resource! Thank you so much for this post. I have been trying to understand this with the plumbing analogy for some time now. I signed up for the forum just to say thanks!