BJT - a current-controlled device?

[Winterstone]

To all forum members who are convinced that the BJT is a current-controlled device.

I am aware that the problem as mentioned in the title was discussed already earlier in this forum (for my opinion, without a commonly agreed final result). And also in three currently running threads this question plays a major role:

https://www.electro-tech-online.com...rent-of-an-npn-common-emitter-circuit.134752/

https://www.electro-tech-online.com...tor-not-affect-the-amplifiers-working.134687/

https://www.electro-tech-online.com/threads/voltage-biasing-for-bjt.134806/

In some cases, the difference between „to control something“ and „to determine something numerically“ was not understood. Thus, we had to face some unfair and unobjective contributions leading to misunderstandings, accusations and reproaches.

With this post, it is my intention and hope to enable a fair and openhearted discussion about this - surprisingly - still disputed question. Therefore, I have listed my arguments in the following. To simplify the discussion I have neglected some minor effects that do not contribute too much the principle answer.
To make your replies and critics easier I have numbered the following

Chain of claims and statements:

1.) Due to charged carrier diffusion there is a depletion region (insulation layer) at the pn junction of a diode. As a consequence, there is a so called „diffusion voltage“ across the pn junction with an associated electrical field that establishes an equilibrium between drift and diffusion effects.

2.) A DC voltage applied across the pn junction disturbs this balance - the depletion area becomes smaller and the potential barrier is reduced. The mathematical description of the voltage-to-current relation is given with the famous Shockley equation, which contains as the most important part the expression exp(Vd/Vt).
(Vd being the voltage across the pn junction and Vt=kT/q is the temperature voltage).

3.) A bipolar junction transistor (BJT) is a three-terminal device (E,B,C) which consists - in principle - of two pn junctions connected back-to-back. The middle region (base region) intentionally is very thin. The emitter region is heavily (n) and the base region is lightly (p) doped. Thus, in the following an npn type transistor is assumed.

4.) In analogy to the pn diode there exists, of course, an insulation layer across both pn junctions.

5.) Now we assume normal operating conditions: B-C junction reverse biased and B-E junction forward biased with a rising voltage starting at Vbe=0. Then, the B-E junction will start to conduct as for classical diodes.

6.) There is no reason, why the voltage-to-current relationship across the B-E junction should not follow the above mentioned Shockley equation (exponential law). It is the momentary VOLTAGE across the junction that determines the width of the depletion area and, thus, the amount of moving carriers (current).
And - as everybody knows - all measurements confirm this exponential law.

7.) Because (a) the p-doped base region is very thin and (b) a rising voltage Vbe even decreases the width of the depletion barrier of this central region, most of the the n-charged carriers released from the emitter (equivalent to the emitter current Ie) do not enter the base node but move to the collector region: Ic=alpha*Ie (alpha<1, but close to unity).
Explanation: The E-field accelerates the mobile charges; they are attracted by the collector potential and they become majority carriers in the n-doped collector region.

8.) It is obvious that the corresponding current Ic (which is slightly smaller than Ie because of alpha<1) follows the exponential law as mentioned under 6.).

9.) Thus, we have a collector current that depends on the applied voltage Vbe - following Shockley`s equation: That gives the classical transfer characteristic Ic=f(Vbe).

10.) Note that - up to now - it was not necessary to quantify the current through the base node. This will happen now.
Because - as mentioned under 7.) - not all of the emitted n-carriers reach the collector, the rest forms the base current Ib1. For completeness, it is to be mentioned that there is a second part Ib2 that is caused by p-carriers of the slightly doped base region which recombine with the emitted n-carriers. Thus the resulting base current is Ib=Ib1+Ib2.

11.) It turns out that because of Ie=Ib+Ic the ratio Ic/Ib=(1-alpha)/alpha=beta is a factor, that is relatively constant for different operating conditions (in particular for different Ic values). This leads to a relation that has a clear practical significance: Ic=beta*Ib.
However, from this we cannot derive that Ic would be physically controlled by Ib.
As mentioned, the factor beta is relatively constant but its actual VALUE is connected with a very large tolerance.
Thus, Ib is an unwanted byproduct, that - however - should not be made arbritarily small because of other impacts and restrictions (breakdown voltage, leakage, etc).

12.) Result/Summary : According to Shockley`s equation the collector current Ic is determined by three physical quantities (Vbe, Vt, Is) and can be electrically controlled by an external voltage Vbe.
That means: The BJT is to be treated as voltage-controlled current source.The most important parameter that determines the amplification properties of the BJT is the transconductance g (slope of the Ic=f(Vbe) characteristic.

____________________________________________________________________________

To clarify the main question (current vs. voltage control) it would be helpful if the „defenders“ of current-control could tell me which of the above listed points are wrong (of course, with corresponding verification/explanation). I am hopeful this will lead to a „common agreement“.
I am sorry for the length of this contribution.
Thank you. and regards to all.

Winterstone

 

[crutschow]

I have no particular disagreement with your statements. They seem to rather nicely sum up the basics about the operation of the BJT. My only observation is to note that, although the physics of a BJT show it is a voltage controlled device, for some practical considerations, such as bias and large signal calculations and using them as switches, a current gain model is generally easier to use. This is related to the fact that the BJT has a low input impedance, so low that some circuit calculations are rather awkward if only the voltage-controlled model is used in those cases. For small signal AC calculations, where it can be assumed that the transconductance value is constant, then the voltage-controlled model is likely better and easier to use.

 

[MikeMl]

None of it "wrong", but I have not needed the material discussed above for 48+years that I have been working in electronics. I studied that material in SemiConductor Physics. That dusty textbook is still somewhere in my bookshelves, but I haven't had to refer to it for probably 30 years...

 

[Claude Abraham]

Once again the Ib straw man is played. Drs. Ebers and Moll published in 1954 that Ie is what controls Ic, not Ib. The equation Ic=alpha*Ie, is the law of transistor action. The beta equation is derived from the alpha equation plus Kirchoff's current law.

The fatal mistake that critics make is to assume that the Shockley diode equation implies V in charge of I. The equation

SDE1) Id = Is*exp((Vd/Vt)-1), could just as well be written as follows:

SDE2) Vd = Vt*ln((Id/Is)+1), w/o loss of accuracy. In fact this form of the equation, SDE2 (Shockley diode equation form 2), is generally more useful for computations with the diode in the forward bias mode, whereas form 1 is usually more applicable in the reverse bias mode of operation. Regarding the depletion region, I am at a loss to explain this any clearer. The width of the depletion region "controls" the current? The potential barrier at the junction is an obstacle for sure. It represents energy LOST per unit charge crossing the p-n junction with its associated barrier potential. But this hardly controls the forward current.

If we bias a diode in the forward direction, let it settle to steady state, then get out our volt-meters (VM), and ammeters (AM), and measure, what can we ascertain? Adjust the variable bench top supply and/or series resistor so that we bias the diode at a new level of I/V. Suppose we had Id=1.0 mA, Vd=0.625 V. We then adjust Vsupply or Radjust so that the new operating point is Id=2.0 mA, Vd=0.660 V. Here is the question that lies at the heart of all this arguing. Which is true?

a) The increase in Vd caused an increase in Id?

b) The increase in Id caused an increase in Vd?

c) It's chickens and eggs?

I firmly believe that critics who insist that bjt is "really" VC, will choose answer a). I choose c) for the most part.

We cannot observe this phenomena and understand which is correct if we only look at steady state values. To answer this question we need to perturb the I-V operating point quickly and observe where each quantity goes. Let's say we have initial conditions above, Id=1.0mA, Vd=0.625V, then the bench top supply outputs a brief pulse for a very short time. We will see a short spike in both Id and Vd. But what is the chronological sequence? Does the voltage Vd rise first, then Id follows as an effect to Vd? If that be the case, Vd might be what controls Id, but not for sure. But if Vd leads Id in time, Id cannot be controlling Vd.

Suppose Id leads Vd, i.e. the current responds quicker and voltage trails current eventually settling after current settled. What does this imply? Well, if Id leads Vd, Id might be what controls Vd, but not for sure. However, we can take 1 thing to the bank, that is that Vd does NOT CONTROL Id. A cause can never follow its own effect. If Id is determined by Vd, potential barrier width, etc., then a change in Id can never happen w/o Vd first changing. Id would then follow as a result of Vd changing.

Every diode model in every semiconductor physics text shows the depletion region equivalent circuit as a diffusion capacitance. The barrier region consists of charges separated by a void or "depletion zone". Charged carriers separated by space or dielectric material is indeed a capacitance. This capacitance, of course, is not linear over large signal excursions, but nonetheless it behaves as would a capacitance.

As frequency increases, this diffusion capacitance, Cd, presents a lower reactance/impedance. Also, this Cd exhibits a well known capacitive property, known to the electrical world as "Eli the ice man". We know Eli. He has been telling us since the 19th century that in an inductor "E Leads I" (Eli), and in a capacitor, that "I Leads E) (ice). Before Vd can change, charges must be transported into and out of the depletion region. This charge motion is by definition current, diode current Id. Any person such as me, who has done extensive design work with diodes, bjt, FETs, LEDs, etc., knows about diode characteristics and that Id will always change ahead of Vd.

That being the case, it is ludicrous to even consider the theory that Vd is what controls Id. On a scale of 10, that theory is a zero. Just as Vd can eventually settle to its equilibrium value, Id settles first.

In fact, we can illustrate the concept without using transients. Look at a buck converter SMPS with an inductor and catch diode, ref attached pdf schematic. When FET Q1 is on, input power supply voltage appears across catch diode D1. Here, Vin is forced across D1. Vd is forced, Id is determined by temperature, and by Vd. So we use form 1 of SDE to compute Id. If temp changes, Vd stays at Vin, but Id increases.

When FET Q2 turns off, L2 current continues into catch diode D1. Here Id is forced to be Iinductor, so that temperature and Id determine Vd. The diode voltage will be controlled by temp and Id per form 2 of SDE. D1 is in the forward direction, FET Q1 is off, and Id dictates Vd. When Q1 FET turns on again, D1 is reverse biased, and Vd dictates Id.

SDE (Shockley's diode equation) simply tells us the relation between Idand Vd, it does not imply which one controls the other. The environment determines that. Vd can control Id, often the case in reverse mode, but in forward mode, Id usually controls Vd.

The whole "voltage controls current" case is built on a flimsy assumption that SDE, a canon relation, supports a voltage controlling current viewpoint because of the way it's written. In semicon phy text references, form 1) is usually how SDE is introduced. Later in the chapter, form 2) may appear, but form 1) is more prominent. Do not misunderstand. Just as I is an exponential function of V, so it is also true that V is a log function of I. Either can "control" the other.

A bjt uses forward bias on the bjt, not reverse. We never force Vbe with a CVS directly at the b-e junction. We control the bjt emitter and/or base current, usually Ie. By forcing a specific value of Ie, we know that Ic will Ie times alpha. Vbe will be there as well, since carriers crossing the junction will exhibit finite lifetime then recombine. During said carrier lifetime, carriers crossing the junction will feel these carriers fields and lose energy since the depletion zone carriers tend to exert repulsive force onto the new incoming carriers, being the same polarity.

Thus the source providing Ie and/or Ib must provide enough potential to overcome the barrier, roughly 0.65V for Si. A source which outputs current to the junction must have at least 0.65V of compliance plus that needed for resistive V drops. One cannot ignore Vbe, it is there and should be considered. But just as a good circuit developer does not rely on Ib to control Ic (beta is not a fixed parameter), nor does one rely on Vbe (more uncontrollable than beta).

Shockley's diode equation seems to be at the heart of this whole contrarian position. Folks, it works both ways, form 1) as well as form 2) Study my schematic and feel free to ask. BR.

 

[Winterstone]

Hi Claude Abraham,

Thank you for your long answer. For the moment just a short reply.

"It's chickens and eggs?"
"Vd can control Id, often the case in reverse mode, but in forward mode, Id usually controls Vd."

We speak about physics, OK? Not about technical solutions or rewriting a formula, which never can tell us the physical truth.
Current is movement of charges. Why do they move? Because there is an electrical field.
Is the reverse mode possible? That means: Can charges move without such a force? Why should they?

Thus, my philosophy of understanding is that there can be no current without a driving voltage. Voltage is first - no "chicken & egg problem".

W.

 

[kubeek]

I am not quite sure what you are trying to accomplish with this yet another thread. Another lengthy thread that solves nothing?

Anyway, the way I see the difference between something being controlled by voltage and by current:
Current controlled: you need an appreciable amount of current to control it from zero to max, and the voltage remains more or less the same.
Voltage controlled: you need an appreciable amount of voltage to control it from zero to max, and the current remains more or less the same.

 

[Ratchit]

Claude Abraham,

Drs. Ebers and Moll published in 1954 that Ie is what controls Ic, not Ib.

No they did not. Look at the link and the E & M equations half way through the article. I see Vbe and Vbc prominently in the equations for Ic, Ie, and Ib. https://en.wikipedia.org/wiki/Bipolar_junction_transistor

The fatal mistake that critics make is to assume that the Shockley diode equation implies V in charge of I. The equation

SDE1) Id = Is*exp((Vd/Vt)-1), could just as well be written as follows:

SDE2) Vd = Vt*ln((Id/Is)+1), w/o loss of accuracy.

It is not a fatal mistake, it is a physical fact. Calculating Vd from Id (SDE2) just tells you what Vd has to be if you want a value of Id. But the physical fact is that you will never get that Id without putting a Vd across the diode. As explained by Winterstone and myself many times, the diode junction needs a forward voltage to counter the back voltage caused by the uncovered charges left behind when the mobile charge carriers diffuse across the junction.

a) The increase in Vd caused an increase in Id?

b) The increase in Id caused an increase in Vd?

c) It's chickens and eggs?

I firmly believe that critics who insist that bjt is "really" VC, will choose answer a). I choose c) for the most part.

The first answer is the correct one. You can drive it with a current, but the physics of the junction requires a voltage to counter the voltage caused by the uncovered charges. Another point, when you drive it with a current, you are in effect using a voltage with an external resistance. In other words, you are looking at the parameters of a diode circuit, not just the diode itself.

SDE (Shockley's diode equation) simply tells us the relation between Idand Vd, it does not imply which one controls the other. The environment determines that. Vd can control Id, often the case in reverse mode, but in forward mode, Id usually controls Vd.

The physics of the diode junction determine what controls what.

Every diode model in every semiconductor physics text shows the depletion region equivalent circuit as a diffusion capacitance. The barrier region consists of charges separated by a void or "depletion zone". Charged carriers separated by space or dielectric material is indeed a capacitance. This capacitance, of course, is not linear over large signal excursions, but nonetheless it behaves as would a capacitance.

As frequency increases, this diffusion capacitance, Cd, presents a lower reactance/impedance. Also, this Cd exhibits a well known capacitive property, known to the electrical world as "Eli the ice man". We know Eli. He has been telling us since the 19th century that in an inductor "E Leads I" (Eli), and in a capacitor, that "I Leads E) (ice). Before Vd can change, charges must be transported into and out of the depletion region. This charge motion is by definition current, diode current Id. Any person such as me, who has done extensive design work with diodes, bjt, FETs, LEDs, etc., knows about diode characteristics and that Id will always change ahead of Vd.

Now you are pettifogging the issue. We are talking about the active region at DC or low frequency where capacitance has a negligible effect.

We never force Vbe with a CVS directly at the b-e junction. We control the bjt emitter and/or base current, usually Ie. By forcing a specific value of Ie,

Yes, but Ie is controlled be Vbe, isn't it?

But just as a good circuit developer does not rely on Ib to control Ic (beta is not a fixed parameter), nor does one rely on Vbe (more uncontrollable than beta).

This guy does, and he is one of the best. Read the following thread, especially post #10. https://cr4.globalspec.com/thread/68055/voltage-vs-current

Shockley's diode equation seems to be at the heart of this whole contrarian position. Folks, it works both ways, form 1) as well as form 2) Study my schematic and feel free to ask.

What does that complicated schematic prove or have to offer?

Ratch

 

[Claude Abraham]

Hi Claude Abraham,

Thank you for your long answer. For the moment just a short reply.

"It's chickens and eggs?"
"Vd can control Id, often the case in reverse mode, but in forward mode, Id usually controls Vd."

We speak about physics, OK? Not about technical solutions or rewriting a formula, which never can tell us the physical truth.
Current is movement of charges. Why do they move? Because there is an electrical field.
Is the reverse mode possible? That means: Can charges move without such a force? Why should they?

Thus, my philosophy of understanding is that there can be no current without a driving voltage. Voltage is first - no "chicken & egg problem".

W.

Well just 1 problem with your analogy. Current can exist only with a "driving voltage". Take a battery and a lamp. You claim that charges move because of an E field, right? But that presents a problem. Outside the battery I have no problem with your position. Outside, the e- are moving "down hill", so to speak. It's inside the battery where your position has problems.

Inside the battery, e- are moving AWAY from the positive terminal, which is counter to the E field, and TOWARDS the negative terminal, another problem. When charges move in presence of an E field, they gain energy, don't they? But that energy is accounted for, since the E field by imparting said energy to charges, gave up some of its own energy.

Take the lamp and battery. As e- move through the wire, outside battery, going down hill, they are weakening said E field. As e- approach positive terminal their own field partially negates that E field at the positive terminal, likewise at opposite terminal. So the chemical reduction/oxidation reaction inside the battery does the following. It separates positive and negative ions, transporting positive ions towards positive terminal. This goes AGAINST the E field. Likewise for negative ions. THe redox process propels charges against the E field, which literally strengthens the E field.

But forcing ions to move against the E field means that redox process produces a current that counters the E field. The result of this current is to increase the E field, and increase the voltage as well. So the current outside the battery is simply charges being moved by the E field, and taking E field energy away. The current IN the battery produces E field as well as voltage. But this current is produced by redox process.

So it takes current to produce voltage, and the redox reaction process produces this current. Ultimately current flows, and voltage develops due to energy conversion, be it battery redox, electromagnetic generator, etc. A photodiode receives incident photon energy as light and a current results. No voltage "driving" it.

Voltage is NOT "first". Current moving as a result of an E field only works going "down hill". Not "up hill" as the case inside the battery. A roller coaster is a good analogy. I've been told that the "gravity field" imparts motion to the car. Sure it does, but ONLY DOWN HILL. An independent power source running on utility grid elevates the car at the start of the ride. Then gravity takes over.

Can charges move w/o a force? Yes. E force is one way to move charges, but ion redox, photonic light, induction, etc. involves energy conversion which propels charges AGAINST the E field which raises E field energy. If E fields alone are what propel charges, no current could sustain for a length of time. A capacitor has an E field when charged. Open circuited it holds its E field value. But connect a resistor across the cap and the charges move in response to E force. But as they move the e- move towards the positive plate neutralizing charge and energy on said plate. In a short time the E field will have decreased to zero.

This notion that voltage is first, then when it is across an impedance current is a result is pure nonsense. This theory cannot withstand scrutiny at all. Where did you learn this? I attended 2 different unis and taught at another. Your statements are at odds with every faculty member at all 3 unis, plus every EE practitioner at several companies I worked for, as well as EE people I've worked with from other companies (clients, customers), contractors brought in to assist, etc.

Please tell us where you obtained this idea that voltage is "first" or is it just plain intuitive to you? Nothing personal you sound intelligent but you seem to be determined to hold your ground to the point of not giving an opposing viewpoint its due consideration. Everybody who says voltage comes first gives your line of reasoning "charges move because there is an E field". But they never acknowledge this: "E field exist because charges are moved". It truly is chickens and eggs.

The key to understanding is realizing that current being motion of charges and said charges carry their own E field. Thus the E field in the moving charges counters external E field, weakening it. An independent power conversion source replenishes this field. I will elaborate if desired.

 

[Winterstone]

Wow! What have I done! A catastrophic failure - I forgot to look inside the battery.
Claude - don`t be disappointed but I am really not motivated to open a new "battle field" about chicken and eggs.
To me, it is more important to get a clear answer to my first post.
And - if I understood you right - your position is "Folks, it works both ways, form 1) as well as form 2)".
Just because you were able to rewrite Shockley`s equation?

Doesn`t your technical or physical "feeling" say that only one single answer is possible ? (Sorry for my bad english, I hope you can understand what I mean).

Or perhaps there is a misunderstanding between us?
Regarding "chicken and eggs": I do NOT speak about a current that can be injected into the base node, which could develop a corresponding voltage.
In that case you are right - it does not matter if I apply a voltage Vbe that causes a corresponding current Ib or if I inject a current Ib causing a voltage Vbe across the junction.

The question is: Which physical quantity causes the depletion layer to change its dimension? Current or voltage?
Please read again point 10 of my list of arguments: I did not need to mention the current at all! Why do you think (at which place of my list) I should introduce any current?
This is the key question!

W.

 

[Claude Abraham]

Wow! What have I done! A catastrophic failure - I forgot to look inside the battery.
Claude - don`t be disappointed but I am really not motivated to open a new "battle field" about chicken and eggs.
To me, it is more important to get a clear answer to my first post.
And - if I understood you right - your position is "Folks, it works both ways, form 1) as well as form 2)".
Just because you were able to rewrite Shockley`s equation?

Doesn`t your technical or physical "feeling" say that only one single answer is possible ? (Sorry for my bad english, I hope you can understand what I mean).

Or perhaps there is a misunderstanding between us?
Regarding "chicken and eggs": I do NOT speak about a current that can be injected into the base node, which could develop a corresponding voltage.
In that case you are right - it does not matter if I apply a voltage Vbe that causes a corresponding current Ib or if I inject a current Ib causing a voltage Vbe across the junction.

The question is: Which physical quantity causes the depletion layer to change its dimension? Current or voltage?
Please read again point 10 of my list of arguments: I did not need to mention the current at all! Why do you think (at which place of my list) I should introduce any current?
This is the key question!

W.

Which physical quantity current or voltage changes the depletion region? The answer is charge, but it takes time to change charge, so current is needed. It takes work to transport charge against an E field so voltage does come in. You treat I & V as if they are divorced, which is ludicrous. Besides you can change a diode's current before changing depletion region. In fact you don't have to counter the barrier to start the charge flow. To start charges moving through a junction, you need to counter inductance L and capacitance C. You need voltage for L and current for C. When the charges cross the junction the finite lifetime results in a depletion zone and barrier potential. If the driving source does not adjust, the barrier voltage will counter the forward from the source and current drops. If the source is CCS, the V will increase to overcome the new barrier.

So the barrier voltage is determined by current. The barrier voltage does not drive current. The battery source or generator energy conversion drives current.

A bjt is current controlled because Ie controls Ic. I have Ebers-Moll 1954 paper. They show a schematic model where Ic = alpha*Ie. They model the collector current as a current source controlled by Ie with alpha the scale factor. Tomorrow maybe I can post it. Ebers & Moll always modeled bjt as CCCS, never VCCS. Nobody has refuted them since 1954. Eli the ice man refuted the V controls I theory. Internal charge flow in a battery refutes "voltage drives current" theory, etc. Every claim made opposing the CCCS model cannot bear scrutiny. Regards.

 

[Claude Abraham]

Claude Abraham,



No they did not. Look at the link and the E & M equations half way through the article. I see Vbe and Vbc prominently in the equations for Ic, Ie, and Ib. https://en.wikipedia.org/wiki/Bipolar_junction_transistor



It is not a fatal mistake, it is a physical fact. Calculating Vd from Id (SDE2) just tells you what Vd has to be if you want a value of Id. But the physical fact is that you will never get that Id without putting a Vd across the diode. As explained by Winterstone and myself many times, the diode junction needs a forward voltage to counter the back voltage caused by the uncovered charges left behind when the mobile charge carriers diffuse across the junction.



The first answer is the correct one. You can drive it with a current, but the physics of the junction requires a voltage to counter the voltage caused by the uncovered charges. Another point, when you drive it with a current, you are in effect using a voltage with an external resistance. In other words, you are looking at the parameters of a diode circuit, not just the diode itself.

Id changes before Vd, so how can Vd still unchanged be what is changing Id? After Id settles, Vd continues to increase then settles later. Impossible for Vd to be what controls Id. Diode current is "controlled" by external source, and in some cases slightly by Vd. If we drive diode w/ battery abd resistor, the current is (Vbatt-Vd)/R. A temp change could affect Vd which can affect Id. If Vbatt >> Vd, it is minimal.

The physics of the diode junction determine what controls what.



Now you are pettifogging the issue. We are talking about the active region at DC or low frequency where capacitance has a negligible effect.

Capacitance has an effect even at low frequency, albeit displacement current is small compared to conduction. Not petfogging. Sequence of events is all important. You insist Vd causes or controls Id, and timing of events negated you. Id responds before Vd changes. Your Vd controlling Id is w/o merit.

Yes, but Ie is controlled be Vbe, isn't it?



This guy does, and he is one of the best. Read the following thread, especially post #10. https://cr4.globalspec.com/thread/68055/voltage-vs-current

Winfield attacks the base current straw man again then invokes Ebers-Moll. E-M uses emitter current and alpha as control parameter, nit Ib/beta. His arguments are just like yours. No OEM, Fairchild (started by Shockley), On Semi, TI, Linear Tech, Intl Rectif, even remotely suggests voltage control modeling of bjt, let alone driving them. The fact that you can find a web site, or forum member who shares your contrarian viewpoint does not make it credible. Winterstone admitted what I've been suggesting. Critics are not only at odds w/ current control, but they feel that in general that V is the driver/controller of I. Not just for bjt but in general.

Any electrical device, light bulb, LED, speaker, needs V & I to work. Since both are present, some will insist that V ultimately controls I so the device is VC. That is just a prejudice.

What does that complicated schematic prove or have to offer?

Ratch

------

 

[Mr RB]

Mods normally shut these threads down when people start nitpicking about the exact behavior of subatomic particles instead of keeping in mind what a transistor is and how we use it...

What actually steers the car? The guy turning the steering wheel? Or the angle of the front wheels where they contact the road? You can argue both sides, and both be totally right, and it just goes on forever.

 

[Winterstone]

Mods normally shut these threads down when people start nitpicking about the exact behavior of subatomic particles instead of keeping in mind what a transistor is and how we use it...

What actually steers the car? The guy turning the steering wheel? Or the angle of the front wheels where they contact the road? You can argue both sides, and both be totally right, and it just goes on forever.

Hi MrRB - I agree that`s one possible way to look at the "problem".
However - are you happy to read in some textbooks the BJT would be current-controlled and in some others it would be voltage-controlled (example: Horowitz & Hill) ?
For my opinion, this is a good reason to raise the question about the physical truth. And I don`t consider a discussion like this as "nitpicking".
Didn`t you ever think about this question? Don`t you have an answer?
Are we all just using parts without knowing why and how they work? I am afraid, in this case we wouldn`t get any new electronic parts, applications and circuits.

 

[Winterstone]

Answer to post#10 (Claude Abraham)

Hello Claude Abraham,

I would suggest to discuss the matter with a calm and relaxed attitude. Please find below some comments.

1.)

Which physical quantity current or voltage changes the depletion region? The answer is charge, but it takes time to change charge, so current is needed. It takes work to transport charge against an E field so voltage does come in. You treat I & V as if they are divorced, which is ludicrous.

„The answer is charge“. Yes - I agree.
And more detailed: Charges with different polarities (charge separation), OK?
And this is equivalent to voltage, is it not?
More than that, I didn`t „divorce“ V & I. Of course, I have mentioned that there is an associated current. But not as a primary (dominating) effect.
I am aware that voltage and current are somewhat connected with each other.

2.)

A bjt is current controlled because Ie controls Ic. I have Ebers-Moll 1954 paper. They show a schematic model where Ic = alpha*Ie. They model the collector current as a current source controlled by Ie with alpha the scale factor. Tomorrow maybe I can post it. Ebers & Moll always modeled bjt as CCCS, never VCCS. Nobody has refuted them since 1954.

Ie controls Ic ? Thus, the question arises: How do you define „to control“?
To me, alpha is a factor that cannot be controlled externally. And Ie? How do you determine/control Ie ?
I suppose you know that, of course, even the basic Ebers-Moll model contains the expression exp(Vbe/Vt)?
Why don`t you mention this fact?
But - it is a model! And you know - there are other models like Gummel & Poon, which is used in most, if not all, simulation programs.
Models do not always reflect true physical laws. The only goal is to behave as the modeled device (as good as possible/necessary)
Simple example: It is common practice to model opamps using either (ideal) controlled current or voltage sources.

3.) My last attempt to describe/explain my position:
At the start of this thread I have mentioned that I will concentrate on the major effects only - for the sake of clarity and simplicity. Therefore I didn`t consider any switching effects and time periods in which the amount of charge changes (and the associated current). To make my position clear I propose to concentrate on two initial states only:

case 1: Ic1=Is[exp(Vbe1/Vt)-1]
case 2: Ic2=Is[exp(Vbe2/Vt)-1]

And this leads to my question: To describe both states as above, do I need the base current Ib ? Or can both Ic values externally fixed by Vbe only?
I kindly ask you to answer, in particular, this question.

To avoid misunderstandings: Of course, there is a current Ib1 and Ib2, respectively. However, according to Shockley`s equation they do not determine the Current Ic.
They are really nothing else than a byproduct that can be calculated Ib=Ic/B (B~beta=hfe).

Regards
W.

Finally, a short side note: It is interesting that one of the best-known researchers in the field of electronics - Barrie Gilbert - needs only „Four Basic Truths“ to start a new monolithic design project.
And one of them is: Ic=Is*exp(Vbe/Vt). No mentioning of Ib, beta, B, hfe,...

 

[Claude Abraham]

Answer to post#10 (Claude Abraham)

Hello Claude Abraham,

I would suggest to discuss the matter with a calm and relaxed attitude. Please find below some comments.

1.)

„The answer is charge“. Yes - I agree.
And more detailed: Charges with different polarities (charge separation), OK?
And this is equivalent to voltage, is it not?
More than that, I didn`t „divorce“ V & I. Of course, I have mentioned that there is an associated current. But not as a primary (dominating) effect.
I am aware that voltage and current are somewhat connected with each other.

2.)

Ie controls Ic ? Thus, the question arises: How do you define „to control“?
To me, alpha is a factor that cannot be controlled externally. And Ie? How do you determine/control Ie ?
I suppose you know that, of course, even the basic Ebers-Moll model contains the expression exp(Vbe/Vt)?
Why don`t you mention this fact?
But - it is a model! And you know - there are other models like Gummel & Poon, which is used in most, if not all, simulation programs.
Models do not always reflect true physical laws. The only goal is to behave as the modeled device (as good as possible/necessary)
Simple example: It is common practice to model opamps using either (ideal) controlled current or voltage sources.

3.) My last attempt to describe/explain my position:
At the start of this thread I have mentioned that I will concentrate on the major effects only - for the sake of clarity and simplicity. Therefore I didn`t consider any switching effects and time periods in which the amount of charge changes (and the associated current). To make my position clear I propose to concentrate on two initial states only:

case 1: Ic1=Is[exp(Vbe1/Vt)-1]
case 2: Ic2=Is[exp(Vbe2/Vt)-1]

And this leads to my question: To describe both states as above, do I need the base current Ib ? Or can both Ic values externally fixed by Vbe only?
I kindly ask you to answer, in particular, this question.

To avoid misunderstandings: Of course, there is a current Ib1 and Ib2, respectively. However, according to Shockley`s equation they do not determine the Current Ic.
They are really nothing else than a byproduct that can be calculated Ib=Ic/B (B~beta=hfe).

Regards
W.

Finally, a short side note: It is interesting that one of the best-known researchers in the field of electronics - Barrie Gilbert - needs only „Four Basic Truths“ to start a new monolithic design project.
And one of them is: Ic=Is*exp(Vbe/Vt). No mentioning of Ib, beta, B, hfe,...

Once again you edited Ebers-Moll equation. Ic = alpha*Ies*exp((Vbe/Vt)-1). Why is alpha always omitted? Vbe is related to Ib and Ie. THe alpha in the E-M equation equals beta/(beta+1). So the E-M equation you invoke contains beta implicitly. But these equations do not contain Vbe:

Ic = alpha*Ie

Ic = beta*Ib.

Again, the fact that Vbe does not appear above is not proof that it is insignificant. An electric heater can have its power expressed 3 ways:

1) P = I^2*R

2) P = V^2/R

3) P = V*I

In eq 1), V does not appear. So is V irrelevant? In eq 2), is I irrelevant? In eq 3), is R irrelevant.

A bjt is a 2 port device. At the input there must be an I and a V, as it is neither superconducting nor superinsulating. The output is a current source with a shunt resistance due to Early effect. This output current is mathematically related to all 3 input variables, Ib, Vbe, and Ie. We can express Ic as a function of either of the 3.

We "control" Ic by setting Ie, or sometimes Ib, never Vbe. A bjt is uncontrollable when Vbe is the input directly driven. Hence "current control" is the phrase used to describe bjt. But one can never input Ie or Ib unless Vbe is included. Ib/Vbe/Ie are mutually inclusive, inseparable, inter-related, joined at the hip, etc. But bjt behavior is markedly different depending on which of the 3 parameters are directly driven as the input.

Barrie Gilbert has little formal education beyond high school. Most self taught people consider V to be the primary quantity with I as secondary. I attached the computation sheet yesterday. It details what is needed to design an amp stage. The op amps in production were designed by many EEs, all who use the approach like the one I attached yesterday. Please review my calculations and point out where you disagree.

 

[Claude Abraham]

Just curious Winterstone, what is your academic background, physics, EE, chemistry? How far did you go in uni? I only ask because I want to know if a uni taught you what you state, or if you learned it on your own through books, forums, web sites, etc. What do you do for a living? R&D, production, FAE, etc.

 

[MrAl]

What actually steers the car? The guy turning the steering wheel? Or the angle of the front wheels where they contact the road? You can argue both sides, and both be totally right, and it just goes on forever.

Hi MrRB - I agree that`s one possible way to look at the "problem".
However - are you happy to read in some textbooks the BJT would be current-controlled and in some others it would be voltage-controlled (example: Horowitz & Hill) ?
For my opinion, this is a good reason to raise the question about the physical truth. And I don`t consider a discussion like this as "nitpicking".
Didn`t you ever think about this question? Don`t you have an answer?
Are we all just using parts without knowing why and how they work? I am afraid, in this case we wouldn`t get any new electronic parts, applications and circuits.

Hi guys,

The only answer can be "Energy". Energy causes things to move, and without things moving we dont get to cause anything no matter how big or small.
And as we all know it takes BOTH current AND voltage to produce energy (ie two dimensions).

That means that every other view is just a single dimension view of that is really happening, and we as humans like to come up with single dimensional formulas so that we can more easily explain what is happening. So if we chose one or the other that means we've chosen to follow a given methodology, not a physical truth.

For the diode or transistor, to set up a field internal to the device, we have to have a way to move charge otherwise we can not get the field to change. We like to simplify this by saying that the voltage causes the field and so the voltage controls the device...but that's only when we like to explain it that way. We never find a voltage that suddenly appears from out of nowhere without pushing some charge around, and that takes energy, and that movement represents current.

So it we look at the physical basis for what is happening, it must be both current and voltage, which is energy.

For an example of using energy in a different way to control a transistor, consider shining a laser onto the die of the transistor near the base emitter pn junction. The photoelectric effect causes charge to move and the transistor turns 'on' harder. Here we never even apply a voltage directly. So again we've controlled the transistor with energy.

I'd like to say more but i'll keep it short for now

Quick note:
We have a 100v battery located three feet from a transistor not connected to the transistor. The transistor is 'OFF'. Why is the transistor off when we have a 100v source nearby? Clearly we have enough voltage right?
It doesnt turn on because something ELSE has to happen in addition to having that voltage available. the voltage can not get to the transistor without inducing some other kind of change as well. It takes some amount of energy to get the voltage to the transistor.
So the deepest physical view of the transistor control is that it is caused by energy, which requires both current and voltage.
To get from the system State1 to State2 requires the addition of energy.

 

[Claude Abraham]

Hi guys,

The only answer can be "Energy". Energy causes things to move, and without things moving we dont get to cause anything no matter how big or small.
And as we all know it takes BOTH current AND voltage to produce energy (ie two dimensions).

That means that every other view is just a single dimension view of that is really happening, and we as humans like to come up with single dimensional formulas so that we can more easily explain what is happening. So if we chose one or the other that means we've chosen to follow a given methodology, not a physical truth.

For the diode or transistor, to set up a field internal to the device, we have to have a way to move charge otherwise we can not get the field to change. We like to simplify this by saying that the voltage causes the field and so the voltage controls the device...but that's only when we like to explain it that way. We never find a voltage that suddenly appears from out of nowhere without pushing some charge around, and that takes energy, and that movement represents current.

So it we look at the physical basis for what is happening, it must be both current and voltage, which is energy.

For an example of using energy in a different way to control a transistor, consider shining a laser onto the die of the transistor near the base emitter pn junction. The photoelectric effect causes charge to move and the transistor turns 'on' harder. Here we never even apply a voltage directly. So again we've controlled the transistor with energy.

I'd like to say more but i'll keep it short for now

Quick note:
We have a 100v battery located three feet from a transistor not connected to the transistor. The transistor is 'OFF'. Why is the transistor off when we have a 100v source nearby? Clearly we have enough voltage right?
It doesnt turn on because something ELSE has to happen in addition to having that voltage available. the voltage can not get to the transistor without inducing some other kind of change as well. It takes some amount of energy to get the voltage to the transistor.
So the deepest physical view of the transistor control is that it is caused by energy, which requires both current and voltage.
To get from the system State1 to State2 requires the addition of energy.

I've been saying that for decades. Energy is the "cause" of all. When A causes B, it always involves energy transfer from A to B. A cue ball is struck by the cue stick, hitting the 1 ball, combo-ing into the 2 ball into the pocket. One can say that the 1 ball caused the 2 ball to move. The cue ball was the "cause" of the 1 ball's motion. The cue stick caused the cue ball to move. The pool player caused the cue stick to move.

It all gets back to an independent source providing energy. I said this re a vocalist singing into a condenser mic. The element has a charge, the vocalist air pressure vibrates the mic element resulting in both V & I. Mr. Al we seem to agree that charge and energy are at the heart.

So why are some devices like FETs called VC, whereas LEDs, bjt are called CC? It has to do with which of the 2 is directly controlled which is indirect, i.e. determined by device and network conditions. A bjt driven by connecting a CVS right at the b-e junction will be incinerated. A FET has a very high g-s impedance. A CCS into the g-s terminals will charge up the g-s capacitance indefinitely. Since charges cannot punch through the insulation at low voltages, the capacitance will acquire charge and increasing voltage. Once dielectric breakdown is exceeded, g-s punches through and FET is now useless. So a FET must be voltage driven, never current driven.

But if I wish to drive a FET but all I have is a 0.10 amp current source, I can place a 100 ohm resistor across the g-s terminals. The 0.10A with 100 ohm limits Vgs to 10 volts, a safe value for standard FET parts. For a bjt we can use a CVS to drive it. We just insert a resistor in the emitter or base side of the b-e junction and all is well.

I believe your summary pretty much puts a period on this issue. I and V are both needed for every electric device in the world. One of them is directly driven, the other is incidental but needed nonetheless. The p-n junction due to its "Is" value which increases with temp, has thermal runaway when voltage driven, stability when current driven, so we control I and let V acquire its value via Shockley and network values.

Thanks for your input.

 

[Winterstone]

Just curious Winterstone, what is your academic background, physics, EE, chemistry? How far did you go in uni? I only ask because I want to know if a uni taught you what you state, or if you learned it on your own through books, forums, web sites, etc. What do you do for a living? R&D, production, FAE, etc.

Fine - now we have reached the point I have expected from you.
I am not sure if this nice (serious and objective) reply from you can convince beginners in this forum that your arguments meet the physical truth.
By the way: You didn`t answer one of my questions. Good enough. No problem.
Finally, you and MrAl have arrived at the term "energy". Right - this is always true and you cannot make any errors.

Barrie Gilbert has little formal education beyond high school. Most self taught people consider V to be the primary quantity with I as secondary.

I will not further comment these two sentences, which show your real attitude.
However, perhaps there are some young members of this forum who didn`t hear about Barrie Gilbert up to now.
Therefore, here some excerpts from his vitae (to be found at the end of most of his articles):

Barrie Gilbert (Life Fellow, IEEE) consulted to Analog Devices Inc. (1972–77). He manages the Northwest Labs, ADI’s first remote design center, in Beaverton, developing a variety of IC products for the communications industry, and holds 65 patents. He has authored papers in JSSC and other journals, is a contributor to several texts, and a co-editor of a recent book.
For work on merged logic he received the IEEE Outstanding Achievement Award (1970) and for contributions to nonlinear signal processing the IEEE Solid-State Circuits Council Out-standing Development Award (1986).
He was Oregon Researcher of the Year in 1990, and received the Solid-State Circuits Award in 1992, the ISSCC Outstanding Paper Award on five occasions, the Best Paper Award at ESSCIRC twice, and several awards for Best Product of the Year.
He received an Honorary Doctorate from Oregon State University in 1997.

See also: https://en.wikipedia.org/wiki/Barrie_Gilbert

I have opened this thread hoping it would be possible to exchange some different - even conflicting - views on a technical subject in a calm, relaxed, objective manner.
But it did not work. All the readers may judge why this was not possible.
Thank you and bye bye.

W.

Ohh - Claude A., I forgot to point to an interesting information for you. The man who "considers V to be the primary quantity with I as secondary" (Barrie Gilbert) has published an article (2004, Analog Integrated Circuits...) with the title: Current Mode, Voltage Mode, or Free Mode? A Few Sage Suggestions. Isn`t funny? Didn`t you know that?

 

[Claude Abraham]

Fine - now we have reached the point I have expected from you.
I am not sure if this nice (serious and objective) reply from you can convince beginners in this forum that your arguments meet the physical truth.
By the way: You didn`t answer one of my questions. Good enough. No problem.
Finally, you and MrAl have arrived at the term "energy". Right - this is always true and you cannot make any errors.



I will not further comment these two sentences, which show your real attitude.
However, perhaps there are some young members of this forum who didn`t hear about Barrie Gilbert up to now.
Therefore, here some excerpts from his vitae (to be found at the end of most of his articles):

Barrie Gilbert (Life Fellow, IEEE) consulted to Analog Devices Inc. (1972–77). He manages the Northwest Labs, ADI’s first remote design center, in Beaverton, developing a variety of IC products for the communications industry, and holds 65 patents. He has authored papers in JSSC and other journals, is a contributor to several texts, and a co-editor of a recent book.
For work on merged logic he received the IEEE Outstanding Achievement Award (1970) and for contributions to nonlinear signal processing the IEEE Solid-State Circuits Council Out-standing Development Award (1986).
He was Oregon Researcher of the Year in 1990, and received the Solid-State Circuits Award in 1992, the ISSCC Outstanding Paper Award on five occasions, the Best Paper Award at ESSCIRC twice, and several awards for Best Product of the Year.
He received an Honorary Doctorate from Oregon State University in 1997.

See also: https://en.wikipedia.org/wiki/Barrie_Gilbert

I have opened this thread hoping it would be possible to exchange some different - even conflicting - views on a technical subject in a calm, relaxed, objective manner.
But it did not work. All the readers may judge why this was not possible.
Thank you and bye bye.

W.

Ohh - Claude A., I forgot to point to an interesting information for you. The man who "considers V to be the primary quantity with I as secondary" (Barrie Gilbert) has published an article (2004, Analog Integrated Circuits...) with the title: Current Mode, Voltage Mode, or Free Mode? A Few Sage Suggestions. Isn`t funny? Didn`t you know that?

I've read those papers. People w/o degrees invent things. Here in USA, we have a retired heavyweight boxing champ named George Foreman, big time boxer in the 70's. On my kitchen counter top is a grill he invented. He has no engineering degree. I'm sure he hired consultants to help with the design and pass UL agency requirements. But he is creative.

I've used the Gilbert multiplier circuit and I really like it. BG has creatively authored some good circuit designs, like George Foreman did with his grill. I would never go to George Foreman for thermodynamics help, and BG in my opinion has little in the way of theory. His statements are all based on subjectivism with some selective math thrown in only if it supports his point.

I've assisted lay people with patent filings. I only have 5 patents myself, but I don't believe that having a degree should be mandatory for invention disclosure. If my buddy Paul, a high school graduate, George Foreman and Barrie Gilbert, have an idea for an invention, their lack of a formal engineering degree should not stop them.

BG has produced some clever IP, but his papers from the last 15 years or so are pretty much fluff and nothing innovative or informative. His voltage mode current mode free mode paper touched on a few thoughts that are interesting to ponder, but nothing in that paper was really useful. The problem with some of these authors is that they take themselves way too seriously. EE is a great field, but sometimes the egos are just more than I can handle.

You asked many questions which I did answer, but I may have missed one. Ask me that question which I have not yet answered and I will PROMPTLY answer it, scout's honor. BR.