# BJT physics

Discussion in 'Feedback and Suggestions' started by LvW, Jul 31, 2013.

Not open for further replies.
1. ### LvW Thread Starter Active Member

Jun 13, 2013
674
100
Reference is made to:

(Volume III Semiconductors, Solid State Device Theory, Bipolar Junction Transistors)

The following two paragraphs need correction:

*Paragraph just above the second figure

Also, few electrons entering the base flow directly through the base to the positive battery terminal. Most of the emitter current of electrons diffuses through the thin base into the collector. Moreover, modulating the small base current produces a larger change in collector current.“

*End of paragraph below the 3rd figure:

Thus 99% of the emitter current flows into the collector. It is controlled by the base current, which is 1% of the emitter current.“

Correction: Both quoted sections claim - without any proof - that the base current Ib would „produce“ or „control“, respectively, the collector current Ic. These statements are not correct.
A small current can never directly control a larger current.
Instead, the current Ic is controlled by a modulation of the width of the base depletion layer. Thus, it is evident that the base-emitter voltage Vbe controls Ic.
The well-known Ebers-Moll equations show the exponential relationship between both quantities. As a result, the slope of the function Ic=f(Vbe) gives the transconductance gm of the device - a parameter that clearly determines the gain properties of the BJT.

Some references (Stanford University, Berkeley University, W. Hill/co-author „The Art of Electronics“) :

http://www.eecs.berkeley.edu/~hu/Chenming-Hu_ch8.pdf

2. ### Wendy Moderator

Mar 24, 2008
20,764
2,534
Fraid not.

IC = β IB

You are not the first one to bring this up. The above formula is pretty well universal. It also happens to work, except for saturation.

There is a voltage controlled model for BJTs, but it is not for beginners.

Not Approved.

3. ### MrChips Moderator

Oct 2, 2009
12,415
3,354
I think it is nitpicking on semantics. My vote is it is acceptable to say the base current controls the collector current.

4. ### Wendy Moderator

Mar 24, 2008
20,764
2,534
Which goes back to our basic premise Ic = β Ib. We have had other users get quite vehement on the subject. They didn't last. It keeps coming up now and again, so this ones answer was easy.

5. ### LvW Thread Starter Active Member

Jun 13, 2013
674
100
May I point to the fact that the quoted sentences were found under "Solid States Device Theory"?
I know that the simple formula "works". However, for my opinion, this is not reason enough to "forget" the correct relationship in a chapter that explains the physical background. Why not mention the exponential function even for beginners? They know this formula already from the diode operation.
According to my experience - no student (beginner) had any problems in accepting this exponential law.
And I like to repeat: The voltage control model (it is not only a "model") automatically leads to the transconductance gm.
Question: How do you arrive at the gain formula based on the term "gm*Rc" without the relationship Ic=f(Vbe)?
Don`t you think it is only logical that the transfer parameter gm=ic/vbe appears also in the gain formula?

More than that, the model of current-control does not "work" always!
I think, the function of the well-known current mirror and the principle of translinear circuits (B. Gilbert) can be explained only using the base-emitter voltage as input quantity.

Remark: I wrote this correction in response to the headline:

Is this request still valid?

Last edited: Jul 31, 2013
6. ### LvW Thread Starter Active Member

Jun 13, 2013
674
100
For my opinion, there is only one single physical truth, it`s not only "semantics".
For designing a working circuit it may be "acceptable" to use Ic=beta*Ib.
However - in a chapter dealing with device physics?

7. ### Georacer Moderator

Nov 25, 2009
5,142
1,266
First off, I 'd like to state that I hate semiconductor theory and probably I don't have enough insight to make comments. However, I 'm a mod and an e-book editor so I have to take a shot.

I got my hands at a semiconductor theory book. Specifically, this one.

In chapter 10, on the bipolar transistor section 10.1.2, the author first expresses the collector current as a function of the gradient of the (assumed linear) electron distribution in the base $n_B(0)-0$, which in turn expresses in terms of the voltage $v_{BE}$.

Then he makes a comparison with the emitter current, deriving the term a, and finally mentioning that the regeneration of holes in the base is proportional to the collector current and also to the base current. Thus he defines the term beta.

If we follow the logic of that book, yes, one could say that the primary control variable for the collector current is the voltage $v_{BE}$.
I quote the book:

8. ### MrChips Moderator

Oct 2, 2009
12,415
3,354
But the current Ib through the base-emitter junction is also a function of Vbe.
Hence is follows that Ic is a function of Ib.

9. ### LvW Thread Starter Active Member

Jun 13, 2013
674
100
To Bill_Marsden:

Perhaps I am allowed to justify and explain my view:

Everybody will agree that the voltage-current relationship across the pn junction of a diode follows the well-known exponential law as given by Shockley many decades ago.
And that`s what we teach in our basic semiconductor lessons.
And I cannot see why we would expect too much from the same audience if we describe a similar effect of thickness modulation for the pn junction of a BJT.
In contrary - there is no physical explanation for the claim that the base current would „produce or control“ the collector current.
For my opinion, it is surprising and funny that some people misuse the experimental observation Ic=constant*Ib (it holds only approximately!) as a physical control law.
Just because it is so simple? Is this the only justification?
Just a statement without explanation in an E-book chapter that is called „Solid State Device Theory“?

In addition to the two examples (mentioned in my former post), which cannot be analyzed using the current-control scheme, I like to ask:
How can we explain the effect of current-controlled voltage feedback caused by an emitter resistor Re using the current-control approach?
LvW

10. ### LvW Thread Starter Active Member

Jun 13, 2013
674
100
Yes - there is a mathematical relationship (function) between Ic and Ib.
However, I was speaking not about such a function, but about a physical control effect (cause and effect). And that is something different.
Remember: The quoted sentences contain the words "produce" and "control".

The formula Ic=B*Ib tells us only that Ib is a certain percentage of Ic - nothing else. It does not contain any information about physical effects.

11. ### MrChips Moderator

Oct 2, 2009
12,415
3,354
Your point about "Solid State Device Theory" is well taken.

However, in All About Circuits Forum and discussion of BJT circuit design it is simpler to use the relationship Ic = β Ib than one that uses Vbe.

12. ### Georacer Moderator

Nov 25, 2009
5,142
1,266
What MrChips means to say is that the e-book (even though it is not stated explicitly) is meant to be of assistance to the site.
Personally, I don't think that the e-book is neither detailed nor exhaustive of the topics it covers. It is a simple introduction for whomever wants to get into electronics.

You are right to say that the explanation of the BJT might be a bit sketchy, but if one wanted to expand on this, it would outweigh the rest of the book content. Notice that not even the diode formula is stated in the corresponding chapter (unless I'm missing it).

13. ### LDC3 Active Member

Apr 27, 2013
920
160
I think it is a mute point since Ib is proportional to Vb. If you want to increase Ib, you need to increase Vb and vise versa. Since the web page is not going into the details of the semiconductor, there is no need to be explicit in what is the actually reason how Ic is controlled by Ib.

14. ### LvW Thread Starter Active Member

Jun 13, 2013
674
100
To MrChips and Georacer:

OK - I see your points. And I agree with MrChips that for BJT circuit design the relation Ic = β Ib can be used. More than that - it even MUST be used for proper design of the base resistive divider. I never have claimed that it is false.
Thus, taking all your comments into account, I think it would be sufficient just to mention that - similar to the diode pn junction - the collector current depends on the base-emitter voltage following an exponential law.
By the way, in the chapter dealing with Re-feedback suddenly the base emitter voltage is used to explain the effect. And that is, of course, correct because it is the only possible explanation. To me, a sort of inconsistency.
Thank you
LvW

15. ### studiot AAC Fanatic!

Nov 9, 2007
5,005
513
As Bill said,

We've been into this before.

How does the voltage model account for the measurable collector current that flows with the base disconnected or held at zero volts?

How does the voltage model account for the direct injection of charge carriers by photons in photodevices?

The voltage model is just that.

A model.

As is the current model.

Neither fully reflect reality.

But to suggest that a small quantity, judiciously applied, cannot control a large one is ridiculous, there are many instances in many branches of engineering of this.

With respect I note your base is in Germany, perhaps your understanding of the word control is different from mine? It does not mean overwhelm.

Last edited: Jul 31, 2013
16. ### MrChips Moderator

Oct 2, 2009
12,415
3,354
Glad to know that we agreed to agree.

And someone kindly corrected me recently. The word is "moot" not "mute".

17. ### Georacer Moderator

Nov 25, 2009
5,142
1,266
@LvW,

Do not think that we do not understand that you have a point and that you have presented reasonable justification.

However, please understand that the e-book is quite hard to be updated drastically, and that it mostly reflects the needs of the userbase.

In all, we thank you for your remarks and the time you took to express them.

18. ### LDC3 Active Member

Apr 27, 2013
920
160
OK, so spelling is not my forte.

19. ### LvW Thread Starter Active Member

Jun 13, 2013
674
100
Ridiculous? I spoke about DIRECT control only. Without any unit conversion in between. It would be interesting for me to learn which examples you are referring to. In our case (BJT) - how does the current Ib influence the thickness of the depletion layer?

I feel familiar with control systems and - even in Germany - it is known what "to control something" means - otherwise I wouldn`t use this term.

May be - but I am not sure.
Nevertheless, if I have the choice between a model that has no relation to reality (as far as the control mechanism is concerned) and another model that
(1) can be physically explained (in analogy to the pn diode) and
(2) must be used for voltage gain calculations,
my preference is obvious.
(Remark to point (2): How can we explain the gain dependence on the quiescent current Ic without using the transconductance gm=Ic/Vt, which is derived from Shockleys equation?).
Regards
LvW

20. ### LvW Thread Starter Active Member

Jun 13, 2013
674
100
OK - no problem for me.
Thank you.
LvW