BJT - a current-controlled device?

[Claude Abraham]

Drs. Ebers & Moll equivalent bjt circuit 1954 as cccs

This excerpt is from the Dec. 1954 issue of IRE journal, by Drs. Ebers and Moll. I refer you to "fig. 5, equivalent circuit of junction transistor valid in regions I and II, and the small signal equivalent circuit."

In the top figure, the emitter current Ie is depicted as related to "phi_e" (Drs. Ebers-Moll used this for "Vbe") per Shockley's exponential equation. This is expressed in the upper left. In the lower left the 2nd component of emitter current is expressed solely in terms of alpha_n, alpha_i, and Ieo, the emitter current representing the collector leakage current in the upside down bjt, where emitter and collector are swapped.

For Ic, please observe. Collector current Ic is expressed in 2 components, the lower is "Ico" which is collector leakage current due to c-b junction reverse bias, a small value. Most bjt users are familiar with Ico.

The upper expression details Ic as a current source whose control parameter is given as "alpha_n*Ie". Alpha_n is alpha in "normal mode" i.e. the right side up bjt. "Alpha_i" is the inverse alpha value for upside down bjt.

Refer to the lower figure, The 3 bjt regions are modeled as resistances re, rb, and rc. The collector has a CCCS across rc. This current controlled current source is controlled by Ie with alpha_n as a parameter. Let me emphasize that in the infancy of the bjt, the current which controls Ic is not Ib, but rather Ie.

This is a very old paper, and much new info regarding bjt operation exists now. But 1 fact that cannot be refuted is that the number of charge carriers "collected" in a time span by the collector, is dictated by the number of carriers "emitted" by the emitter in that time span.

Of course, Vbe, Ies, Ics, Vbc, Vbe, etc., are involved. I never attempted to deny the role that other factors play. Although the E-M equations express Ie/Ic as a function of Vbe and Vbc, they model the collector current as a current source controlled by Ie. The rc resistance across the current source accounts for Early effect. The Ic value does vary with Vce, so it is a current source with less than infinite shunt resistance. The rc inclusion accounts for Early effect.

I have to concur because Shockley's equation is all important but it describes the I-V relation in any p-n junction, be it LED, rectifier, JFET g-s terminals, etc. But bjt action occurs by emitter carriers emitted, transiting through base, and collected by collector. If base region is super wide, say 1.0 mm, a huge number of holes would transit from base to emitter, even more electrons from emitter towards base. Very few electrons would reach the collector. Ib would about equal Ie, and Ic is small, merely leakage across reverse biased c-b junction.

Shockley's equation says that the b-e junction exhibits an exponential I-V relation, or a logarithmic V-I relation, however you look at it. It certainly is correct. But with alpha near zero or at zero, there is zero transistor action. I believe that Drs. Ebers and Moll studied the device and ascertained that this relation:

Ic = alpha*Ie,

is the key to understanding bjt functioning. The Shockley is very important is it relates I to V. All amp stages possess both voltage and current gain. Alpha and beta give us the upper limit on current gain. The gm factor is derived from Shockley, and gm is the upper limit for stage transconductance. Transconductance times load resistance is overall stage voltage gain.

Shockley diode equation and gm provide useful info on the amp stage voltage gain ability. Beta and alpha convey current gain capability. They are both important since active devices have both greater than unity. The ability to increase both V & I is what makes active devices, such as the bjt, so attractive.

I see no reason to disagree with Drs. Ebers and Moll. BR.

 

[MrAl]

Hello again Claude,


Well i agree with some of what you said, but to get my point across i have to disagree with your statement that the FET is voltage controlled. This may come as a surprise to you because we often call an FET voltage controlled, and it does make sense to call it that. But there is always some current involved again because we cant get anything to happen without energy.
So in the strictest sense, nothing can be either current controlled or voltage controlled, but must have both. There are a lot of reasons for this, and i can not go into every single reason, but the most basic physical reason is because capacitance exists everywhere between any two electrodes, and it always takes at least two electrodes to be able to have a voltage difference. Therefore where ever there exists a voltage there is also capacitance there as well. If we wish to alter the state of the system, we must deal with that capacitance in order to change the voltage.

A better example instead of the FET though is the LED. That's because the FET is at the extreme end of the list where current is minimal whereas the LED is a little hard to understand based on voltage instead, so they are opposite but the LED is a little easier to work with i think, so i will use the LED for the example.

Now normally we think of the LED as being current controlled, because the current level is responsible for the amount of light output. And also it is hard to control this current with the voltage because of various imperfections like when the characteristic voltage changes with temperature. So rather than deal with that we deal instead with the control of the current.
But is it impossible to control with voltage? Certainly not. It's hard to do, but it can be done. We can monitor the LED and determine if we should apply more voltage or not. For example, we can monitor the light output (we dont have to monitor the current) and adjust the voltage so that we get a constant light output. We can also monitor the temperature and make it a high reliable LED light source. All while controlling voltage.
Is it harder to do than controlling the current? You bet it is But it is certainly doable and that takes us back to square 1 where we can say that it can either be voltage controlled or current controlled, although we prefer current control.

We find this same thing in transistor circuits so that's why i have to say the same thing about them, especially looking at the base emitter diode.

So i believe that any alteration in field must come about from a movement of charge at some time or another. In basic theory we try to remove the other dimensions in order to understand the underlying single dimensional mechanism, but in practice it never works out that way.

Thanks for listening

 

[Mr RB]

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) ?
...

I learned 30+ years ago that textbooks (and classrooms) are full of errors and subject to recent fashion. No it doesn;t bother me that some books are wrong, or focus on an aspect of the transistor internals that really doesn't matter in electronics.

...
Didn`t you ever think about this question? Don`t you have an answer?
...

I think about this question every time people start to argue about it.

My answer is the same. Of the two choices, a transistor is better described as "current controlled" as the relationship of base current to collector/emitter current is better approximated to the real world than base voltage to collector voltage.

And in electronics we don't open up transistors, what we do is connect stuff to a transistor to make to do things. And "current controlled" is definitely a better model when you need to make a transistor do something in the real world.

So in electronics, "current controlled" is both a better description, and a more useful description.

 

[Winterstone]

I learned 30+ years ago that textbooks (and classrooms) are full of errors and subject to recent fashion.

Fully agreed. And that`s the reason I like to clarify - in the interest of newcomers/beginners/students, who are confused.

And in electronics we don't open up transistors, what we do is connect stuff to a transistor to make to do things. And "current controlled" is definitely a better model when you need to make a transistor do something in the real world. So in electronics, "current controlled" is both a better description, and a more useful description.

Also agreed. For most applications it works and is "a better model", because it is simpler. But it is a model only! It is the primary task of a model to behave (as good as possible) as the real device. But this does not necessarily mean that it mirrors the physical truth. For example: Opamps can be modeled using (ideal) controlled current or voltage sources or a mixture of both.
And it was my intention with this thread to collect some different opinions/explanations targeting towards the "physical truth". Nothing else.
And - as you know, there are some applications resp. circuits that do not work based on the current-control approach.

Regards
W.

 

[Winterstone]

The above reply to MrRb was my last contribution in this thread.
I am a bit disappointed about the way some members try to disqualify other opinions - in one case down to personal attacks (post#16) .

To all who are interested to read how W. Shockley explains the working principle of the BJT, I attach an excerpt from his patent document.
For my opinion and in the context of this thread, most interesting are the following lines:

*Left columne: 13-15, 38-40, 64-70
*Right columne: 33-35, 45-63
(note that after the main working principle was explained, the description of the role of the base current starts with "next there may be taken into account..." )

Winterstone

 

[Claude Abraham]

The above reply to MrRb was my last contribution in this thread.
I am a bit disappointed about the way some members try to disqualify other opinions - in one case down to personal attacks (post#16) .

To all who are interested to read how W. Shockley explains the working principle of the BJT, I attach an excerpt from his patent document.
For my opinion and in the context of this thread, most interesting are the following lines:

*Left columne: 13-15, 38-40, 64-70
*Right columne: 33-35, 45-63
(note that after the main working principle was explained, the description of the role of the base current starts with "next there may be taken into account..." )

Winterstone

Still beating up that base current straw man. I attached Ebers-Moll priginal equivalent circuit showing bjt as EMITTER-current-controlled device. Emitter, not base current. The emitter current is the control signal, the base current is indirect. Emitter direct, base indirect. You keep presenting evidence refuting "BASE current control" a position nobody adopted.

Shockley wrote what we've already been saying. We intentionally dope the base region (ref npn) with a sparse population density of acceptor ions, while doping the emitter with a high density of donor ions. OEM intentionally fabricate the device in order to minimize base current. Base current Ib is necessary for reasons I already stated like Vce break down capability, low c-b leakage, etc.

Depending on the temperature, desired leakage current limit, and Vce limit, we can use enough acceptor doping in the base to meet requirements, but no more than that. If doping the base with 1e11 atoms/cc results in sufficient Vce break down ability, low leakage c to b, then we do not use 1e13 atoms/cc because the base current would increase a hundred fold.

Since Shockley's paper, the world has known that it is desirable to dope the base in a manner that meets all Vce and leakage limits, yet keeps Ib as low as feasible. I've been explaining for nearly a decade that Ic is controlled by Ie, that "current control" implies Ic=alpha*Ie, NOT Ic=beta*Ib. Yet every time a VC advocate gets cornered, refuted, etc., the attack the base current straw man and present publications from experts which prove that it is desirable to minimize Ib. I've already said that.

An op amp, traditional voltage feedback, is a VCVS. The "control signal" is the input differential voltage "Vd". This control signal is intentionally made as small as possible. An op amp with the highest gain will force Vd to a smaller value, which is desirable. But too high an open loop gain is difficult to stabilize with low closed loop gain.

Any amp stage has I and V gain. It is always desirable to minimize the input signal needed for a given output swing. A FET has transconductance gm which ideally should be large. Vgs should be as small as possible for high gain. Vgs is the control signal. It varies as much as the bjt beta, yet nobody challenges Vgs as the control input. The arguments presented that Ib is less significant because we always try to minimize it is trivial.

Any amp is more attractive if input voltage and input current is as small as possible. That is what makes amps so attractive. One which can output the same magnitude with a lower input w/o side effects is better.

Your Shockley paper is so well known, it is trivial to rely on it to prove your case. A bjt is intentionally injected with asymmetrical doping just so that Ib is minimized, but not to the point where Vce and Irevcb are compromised. BR.

 

[Ratchit]

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...

This discussion is not about what constitutes a transistor and how useful it is. It is about how it physically works at its lowest basic level. Since a transistor manipulates subatomic particles, they have to be discussed, too.

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.

At the lowest level, for the two choices you gave, it is the angle of the wheels. If you limit the argument to control at the lowest level, only one answer is possible.

MrAL,

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).

True, but what kind of energy? And how and where is it applied?

Claude Abraham,

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.

Since Ie contains Ic, I guess that would be so. But then the question of what controls Ie has to be asked. What would a model using current generators prove? Models do not show what controls a device. There are models that show voltage control also.

Eli the ice man refuted the V controls I theory.

How? Just because some stored energy or released energy causes the current and voltage to be out of phase with each other does not disprove that one is controlling the other.

Internal charge flow in a battery refutes "voltage drives current" theory, etc. Every claim made opposing the CCCS model cannot bear scrutiny.

How? There are still internal electrochemical voltages at the molecular level. If a model works to show what a device does, then it cannot be disputed. And EM does that very well. But that does not mean it explains how the device works.

Ratch

 

[Claude Abraham]

Mr. RB,



This discussion is not about what constitutes a transistor and how useful it is. It is about how it physically works at its lowest basic level. Since a transistor manipulates subatomic particles, they have to be discussed, too.



At the lowest level, for the two choices you gave, it is the angle of the wheels. If you limit the argument to control at the lowest level, only one answer is possible.

MrAL,



True, but what kind of energy? And how and where is it applied?

Claude Abraham,



Since Ie contains Ic, I guess that would be so. But then the question of what controls Ie has to be asked. What would a model using current generators prove? Models do not show what controls a device. There are models that show voltage control also.



How? Just because some stored energy or released energy causes the current and voltage to be out of phase with each other does not disprove that one is controlling the other.



How? There are still internal electrochemical voltages at the molecular level. If a model works to show what a device does, then it cannot be disputed. And EM does that very well. But that does not mean it explains how the device works.

Ratch

In signal, systems, and control theory, control and causality have stipulations that the control quantity must be chronologically ahead of the output or "controlled" signal. The unit step function, Dirac delta function, etc., were formulated so that input signal begins at time t = 0. Output is referenced to this time base. If the output starts its excursion before time 0 then the input signal can never be causal or the control signal.'

The unit step was invented to account for time base. How does "out of phase" matter? It matters plenty. Ie begins changing before Vbe has even budges, Ie is being controlled by a transducer, like a mic, signal generator, photodiode, optical radiation, etc. Vbe responds eventually and catches up with Ie. In steady state Vbe for a bjt, or Vd for a diode, reaches the steady state value per Shockley equation. This is all important. If Vd controls Id then Id initially responds to a "change in Vd" that is still in the future and has not happened yet! This is just too absurd to even consider. Id settles to its final value then a moment later Vd continues then settles. Id first changes w/o Vd changing. Id settles then Vd continues changing then settles. A controlling signal always precedes or "leads" the controlled signal.

With diode, LEDs, bjt, nobody is disputing that Vd/Vbe are quantities that should be considered, I never ignore them, nor do I ignore Id/Ib/Ie. Also FWIW, Ib, Vbe, and Ie cannot exist separately. It takes energy to change a bjt state and all 3 quantities must be non-zero for that to happen. But I cannot force all 3 to be a specific value. I control 1 and let the other 2 be determined by bjt characteristics and external network. Internally what controls Ic has to be Ie because carriers collected are directly related to carriers emitted.

You ask "what controls Ie?". I know you think it is Vbe but Eli demolishes that theory. Ie is controlled through the source driving the bjt. If I sing into a mic I change the element I & V, the charrges move through the cable arriving at the b-e junction. They cross the barrier. Ie is already determined w/o Vbe having changed yet. After carriers cross and depletion zone changes resulting in modified barrier, Vbe increases.

If the signal generator is constant voltage, followed by a resistor in the base or emitter side of b-e junction, then the additional vbe at the barrier can slightly reduce ie since (vsource-vbe)/re = ie. They are interactive but a well designed network minimizes vbe influence on ie. If the signal generator is constant current, vbe increase is compensated by the current source.

No current source has infinite shunt impedance so current control is never perfect, voltage and impedance do have influence. Likewise voltage sources have finite series impedance making true voltage control impossible. But it is safe to say that the value of Ie is not controlled by Vbe.

In a battery electrochemical voltages at molecular level does not change the fact that charges in the form of ions are separated by a redox process chemically propelling ions against the E field. This is quantum mechanical in nature and frankly saying that this involves "electrochemical voltages" is not helpful. Nobody understands QM. How energy is discretely quantized, transfers, emitting photons, etc., cannot be explained with arm-waving theories like "there is electrochemical voltage involved".

I cannot explain just how ions are separated and propelled against the E field because that takes me to the limit of my QM knowledge, and I admit that QM is over my head. The late great Richard Feynman, a world renowned physicist summed it well stating that he does not understand QM, and those who claim they do are fooling nobody but themselves. I agree.

 

[MrAl]

Hello Claude,


I have read that in quantum mechanics that the voltage comes first and the current second because the current involves the movement of mass, and to move mass there must be a force, even when that mass is very small. But when we extrapolate up to classical physics we see that the mass is so small that it takes little to get it moving, and also from the differential equation of the system (no matter how tiny because after all we are now working within the framework of classical physics where there is no quanta) we see that the start of the voltage and the start of the current can not be found to be at different times in the same frame of reference.

But again if that voltage is to do anything it has to get there somehow, within a reasonable proximity to the quanta so how did it suddenly just appear there. It cant, something must have initiated THAT too.

Just adding a few comments

 

[Claude Abraham]

Hello Claude,


I have read that in quantum mechanics that the voltage comes first and the current second because the current involves the movement of mass, and to move mass there must be a force, even when that mass is very small. But when we extrapolate up to classical physics we see that the mass is so small that it takes little to get it moving, and also from the differential equation of the system (no matter how tiny because after all we are now working within the framework of classical physics where there is no quanta) we see that the start of the voltage and the start of the current can not be found to be at different times in the same frame of reference.

But again if that voltage is to do anything it has to get there somehow, within a reasonable proximity to the quanta so how did it suddenly just appear there. It cant, something must have initiated THAT too.

Just adding a few comments

My background in QM is too limited to argue that point either way. Having limited academic study of a topic does not dissuade others on this thread from criticizing canon teaching, but I don't believe in doing that. If QM says that V comes before I please provide a link or reference. Is this established canon law, or a theory someone has been advancing? Personally I can't refute such a claim because QM is at the limit of my knowledge. Ponder this.

It's been suggested that at the sub-atomic level there are voltages electrochemical or so. No dispute there, it is well known. But we need to SEPARATE charges. Ions getting separated requires work. The internal fields between ions of opposite polarity are attractive in polarity. To separate them we cannot use the sub-atomic voltage or E field. Likewise ions are composed of positive nuclei and negative electrons. The sub-atomic voltages and E fields does not separate them but rather the opposite.

To separate ions, how can sub-atomic E fields/voltages be the motivating force. Fields cannot be the force that do long term sustained work. Every field gices up energy when it imparts motion to charge carriers. Charge motion by its very nature depletes the energy of said field.

Also, the energy of charge carriers certainly includes kinetic, as you mentioned they have mass, but it takes energy to transition an electron from valence to conduction, quantum energy. To say that a voltage at the sub-atomic level is motivating the current is pure arm waving. The micro-voltages are due to E fields whose force is directed towards attraction of + and - ions, + and - sub-atomic particles.

Even if we go inside a proton or neutron, we see up and down quarks. An up quark has a +2/3 charge, a down quark has -1/3. The voltage/E field inside a nuclear particle (proton or neutron) provides a force which would draw the up and down quarks together fusing them. That doesn't happen. Anyway, I don't think one can just say that V before I is valid at any level, and likewise I don't believe in I before V either. E force is the result of charge separation which requires charge motion. But what provided that motion? It truly is a vicious circle as futile as the chicken/egg riddle.

Who knows, we may learn someday more about quarks and such. But we will always have the smallest entity that can be measured, and be unable to completely explain it because it is influenced by something smaller beyond our detection capability.

 

[Ratchit]

Claude Abraham,

In signal, systems, and control theory, control and causality have stipulations that the control quantity must be chronologically ahead of the output or "controlled" signal. ...... A controlling signal always precedes or "leads" the controlled signal.

Make it simple. How does all the above relate to the operation of a BJT? We are not discussing signals and control blocks. We are discussing the fundamental operation of a BJT, specifically the physics of how it operates. Therefore, when you introduce applications of the BJT, and try to consider it as a control block within control theory, and try to extrapolate from those applications the physical principles of its operation, you are wandering off the reservation.

With diode, LEDs, bjt, nobody is disputing that Vd/Vbe are quantities that should be considered, I never ignore them, nor do I ignore Id/Ib/Ie. Also FWIW, Ib, Vbe, and Ie cannot exist separately. It takes energy to change a bjt state and all 3 quantities must be non-zero for that to happen. But I cannot force all 3 to be a specific value. I control 1 and let the other 2 be determined by bjt characteristics and external network. Internally what controls Ic has to be Ie because carriers collected are directly related to carriers emitted.

What is the point of the above paragraph? That Ic is controlled by Ie? If Ic is in lock step with Ic, then what controls Ie?

You ask "what controls Ie?". I know you think it is Vbe but Eli demolishes that theory.

How? Surely you are not going to throw out the phase relationship again.

Ie is controlled through the source driving the bjt.

Yes, Ic and Ie are all controlled by the source driving the BJT. Specifically by the way the source affects Vbe.

If I sing into a mic I change the element I & V, the charrges move through the cable arriving at the b-e junction. They cross the barrier. Ie is already determined w/o Vbe having changed yet. After carriers cross and depletion zone changes resulting in modified barrier, Vbe increases.

If you apply a voltage to the base of the BJT, you will change the Vbe, and increase or decrease the voltage across the depletion region, thereby increasing or decreasing the diffusion process in the emitter circuit, which changes Ie. That makes a BJT a voltage controlled current source, or a basic transconductance amplifier.

If the signal generator is constant voltage, followed by a resistor in the base or emitter side of b-e junction, then the additional vbe at the barrier can slightly reduce ie since (vsource-vbe)/re = ie. They are interactive but a well designed network minimizes vbe influence on ie. If the signal generator is constant current, vbe increase is compensated by the current source.

No current source has infinite shunt impedance so current control is never perfect, voltage and impedance do have influence. Likewise voltage sources have finite series impedance making true voltage control impossible. But it is safe to say that the value of Ie is not controlled by Vbe.

No matter what resistance is added or how you drive it, the Ie and Ic will be controlled by the Vbe. Other more convenient relationships may exist for design and calculation, but Vbe is what is controlling a BJT.

n a battery electrochemical voltages at molecular level does not change the fact that charges in the form of ions are separated by a redox process chemically propelling ions against the E field. This is quantum mechanical in nature and frankly saying that this involves "electrochemical voltages" is not helpful. Nobody understands QM. How energy is discretely quantized, transfers, emitting photons, etc., cannot be explained with arm-waving theories like "there is electrochemical voltage involved".

I cannot explain just how ions are separated and propelled against the E field because that takes me to the limit of my QM knowledge, and I admit that QM is over my head. The late great Richard Feynman, a world renowned physicist summed it well stating that he does not understand QM, and those who claim they do are fooling nobody but themselves. I agree.

You are wandering off the reservation by bringing up chemical energy to electrical energy conversion. A lot is known about this subject, but not everything. Feynman did not know everything about QM, but he knew a lot more than most people did.

Ratch

 

[MrAl]

Hello Claude,


Well my point was simply that if you want to move something you first have to apply a force. And note that the application of force does not mean that we have yet done any work. So i think this is how we might think of it, that the voltage is the start of the process. In classical physics this isnt so, but in the world of the quanta the smallest thing that can happen is limited to some larger value, and we know that mass can not move in a zero time, so it makes more sense now to think of the voltage as being the first thing to get there, then the particle starts to move in response. This would happen i think because there is a minimum force that is applied unlike classical physics where an infinitesimal force is allowed. When we look closer and closer in classical, we NEVER reach a point where the voltage appears and the charge does not move. That's because even at the lowest level of energy something ALWAYS happens. In the quantum world, there has to be some minimum energy present to get anything to happen, so it is conceivable that at some small time interval the voltage is not yet enough to start any charge movement. It is only after the required quanta have arrived that the charge in question starts to move. This should be obvious. But what is not obvious is that the charge in question is not the only charge to consider, because as you said we have to move charge to get the voltage 'force' to appear in the first place. But that's the external charge, and that is why i always say that it takes energy to get anything to happen.

But in theory we can have a voltage appear between two terminals and not have any charge move in the quantum world, while in the classical world we always have charge moving when there is voltage present so we can not distinguish between the external and the internal charge so well.

I am sure this is a very simple concept, but i am not sure i i have explained it as well as a theoretical particle physicist would.

 

[Claude Abraham]

Timing is everything. If A causes B, then a change in A must chronologically precede corresponding change in B. Causality stipulates that and every controls, signals and systems, and physics text I've seen stipulates the same.

I studied semiconductor physics, herein called "semicon phy", at undergrad and grad level, 5 courses, from EE and physics dept. Not once was I told that the junction potential barrier, Vd (or Vled, Vbe, Vgate-cathode, etc.) is what "controls" the current, Shockley equation is bi-directional, I is an exponential function of V, just as V is a logarithmic function of I. Ratchit you keep claiming that Vd (Vbe) controls diffusion process thus controlling current, but you always omit the drift current. Do you know diffusion from drift? Every semicon phy text covers both. When charges move under influence of external field, such as mic into mic preamp, that is drift, not diffusion.Drift is as important as diffusion, yet you completely ignore it. The "J" in Ohm's law "J = sigma*E", is drift, not diffusion.

Also, in the now closed thread I brought up that Vd/Vbe are voltage drops, not emf's so they are losses not energy gains by charge carriers. Drops and emf are both measured in volts since they are dimensionally equivalent. Drop is energy lost per unit charge, emf is energy gained per unit charge. Only the loss/gain aspect differs, both are energy/charge, and thus emf is measured in volts just as is drop.

"EMF", is not an actual "force". This term was coined before the Lorentz force concept was fully developed. Take a simple CRT used in oscilloscopes and older tv sets. If a beam of moving electrons, velocity = u, enters the space between 2 plates w/ charge/potential, or in the vicinity of a B field, the free body diagram of all forces acting on the electrons includes the 2 Lorentz forces, Fe and FM (electric and magnetic resp.). Fe = qE, while FM = q(uXB). The total force on the electron is F = q(E + uXB). There is no "electromotive", nor any "magnetomotive" force on the electron. The terms "emf" and "mmf", again, were coined before we knew all that is happening. They are not true forces. EMF is rightly expressed in volts because it is literally energy/charge, same as a drop except that emf is gained energy/charge, vs. loss for drops.

Mr. Al - in order to move a charge we certainly need a force, but is it an electric/magnetic force? Is a chemical redox eraction an electric force imparting motion to ions? I say not. To separate ions we work against E fields. We are separating charges of opposite polarity so voltages at the atomic level cannot be supoplying the energy for this, nor the local E fields, nor B fields as well. An electric generator must be mechanically moved to generate power. Steam, hydro force, internal combustion engine w/ belt drive, is needed to provide forces to move charges.

Sure a charge has mass and needs a force to move, but electric forces as well as the integral of E field. i.e. voltage, cannot generate it's own moving force. A bunch of atomic charges have their E fields and voltage. But they cannot propel themselves as a group into drift motion. The conservation of momentum says not.

An external power source is needed. A group of ions has a net momentum of zero, or some constant. In order to move these charges into a drift type current flow, we need an external force to come in. Local E fields cannot change system momentum.

It's tough for some to accept this, bit I & V are simply mutually inclusive, interactive, and one does not generally "control" let alone "cause" the other. Physics does not support any notion of V controlling I, nor vice-versa. To say that a force is needed is not at all contentious, universally accepted. But the notion that system internal E forces and voltage impart motion to its charges resulting in net system drift and momentum change is heresy. Conservation of energy and momentum dictates that an external agent must input work. Again, steam, wind, ICE, etc. is what drives the charges, not V, not I, not B nor E.

Fields cannot "drive" anything because a field loses whatever energy it imparts to charges. Unless the external source replenishes said field, it will be exhausted quickly.

 

[Ratchit]

Claude,

Timing is everything. If A causes B, then a change in A must chronologically precede corresponding change in B. Causality stipulates that and every controls, signals and systems, and physics text I've seen stipulates the same.

I will agree to that, as long as you don't try to aver that the initial conditions represent a response that occurred previously to the stimulus.

I studied semiconductor physics, herein called "semicon phy", at undergrad and grad level, 5 courses, from EE and physics dept. Not once was I told that the junction potential barrier, Vd (or Vled, Vbe, Vgate-cathode, etc.) is what "controls" the current, Shockley equation is bi-directional, I is an exponential function of V, just as V is a logarithmic function of I.

Both I and V have a relationship that can be inverted no matter what the equation. That does not prove or define what is controlling what. Only the physics of the device can determine what is in control.

Ratchit you keep claiming that Vd (Vbe) controls diffusion process thus controlling current, but you always omit the drift current. Do you know diffusion from drift? Every semicon phy text covers both. When charges move under influence of external field, such as mic into mic preamp, that is drift, not diffusion.

Yes, I know about drift and diffusion. You did not mention recombination-generation (R-G). Drift current in a BJT can only come from thermally generated currents. That is not like a full conductor where a sea of charge carriers are present. The net effect of forward bias in a diode is a large increase in the diffusion current components while the drift components remain fix near their thermal equilibrium values. In reverse bias, the drift current predominates because the diffusion current is miniscule or zero. Just look at the reverse voltage characteristics of a diode and see how small the drift current is compared to the forward current. That is because the drift current is restricted to the small amount of thermal generated current in a diode.

In the depletion region, R-G current predominates only at low levels of current. The bottom line is that in a BJT, drift and R-G current are insignificant in the active region. We are not discussing wires, mics, and preamps. Their conductors do not have restricted charge carriers as p-type and n=type material do. Their charge movement is by drift, but that is irrelevant to what happens in a BJT.

Drift is as important as diffusion, yet you completely ignore it.

I disregard it because it is not relevant in the active region in a diffusion based device, which a BJT is.

The "J" in Ohm's law "J = sigma*E", is drift, not diffusion.

In a BJT, the current density J is both, with diffusion predominating.

Also, in the now closed thread I brought up that Vd/Vbe are voltage drops, not emf's so they are losses not energy gains by charge carriers. Drops and emf are both measured in volts since they are dimensionally equivalent. Drop is energy lost per unit charge, emf is energy gained per unit charge. Only the loss/gain aspect differs, both are energy/charge, and thus emf is measured in volts just as is drop.

So how does that tie into anything?

"EMF", is not an actual "force". This term was coined before the Lorentz force concept was fully developed. Take a simple CRT used in oscilloscopes and older tv sets. If a beam of moving electrons, velocity = u, enters the space between 2 plates w/ charge/potential, or in the vicinity of a B field, the free body diagram of all forces acting on the electrons includes the 2 Lorentz forces, Fe and FM (electric and magnetic resp.). Fe = qE, while FM = q(uXB). The total force on the electron is F = q(E + uXB). There is no "electromotive", nor any "magnetomotive" force on the electron. The terms "emf" and "mmf", again, were coined before we knew all that is happening. They are not true forces. EMF is rightly expressed in volts because it is literally energy/charge, same as a drop except that emf is gained energy/charge, vs. loss for drops.

Interesting, but irrelevant with respect to what is in control of the BJT.

OK, in a previous post I said that I could not explain how a current could lower the diffusion barrier voltage caused by uncovered charges resulting from the diffusion process. So far, you have not done so either. I still contend that forward Vbe voltage is needed to lower that barrier voltage in order to increase the diode diffusion current, and I have seen nothing to contradict that view.

I will let MrAl answer the rest of your post you addressed to him.

Ratch

 

[MrAl]

Fields cannot "drive" anything because a field loses whatever energy it imparts to charges. Unless the external source replenishes said field, it will be exhausted quickly.

Hi,

I am not sure i understand what you mean by this. On the one hand you seem to be saying that a field can not drive anything, but on the other hand you seem to be saying that it can drive if the energy is replenished.

The question comes up, what comes first the field or the current? The classical answer is they both happen simultaneously. But in quantum theory i think the field comes first.
For example, a laser at point A and a solar cell at point B one mile away. The solar cell does not put out any current with the laser off. When the laser is turned on, it takes a finite measurable amount of time for any photons to reach the surface of the solar cell, therefore there is a delay between the time the photons are generated and the current flows in the solar cell. Therefore the photons came first and it took time for them to reach the cell. So the photons had been in existence long before any charges moved. When the distance is made larger, the delay is larger. When the distance is shorter, the delay is shorter. But it takes zero distance to get it to happen in no time at all, and that's not possible because it always takes some time for photons to reach different parts of an apparatus regardless how small it is.

 

[vlad777]

Vbe controls Ib , Ib controls Ic.

If you have a resistor in emitter circuit transistor is voltage controlled,
but if emitter is dead short to ground than transistor can be treated as current controlled.
Because Ib will not depend so much on Vbe but it will mostly depend on Vb/Rb.

BTW

If you manage to somehow move charge carrying particles, then electric field will follow.
So voltage causes current but also current can cause voltage.
Math equations do not describe any causality, it is a question of what you can physically do.
For example rubbing transfers electrons and electric field follows.

 

[Winterstone]

Vbe controls Ib , Ib controls Ic.

If you have a resistor in emitter circuit transistor is voltage controlled,
but if emitter is dead short to ground than transistor can be treated as current controlled.

Hi vlad777,

So your statement is

(a) with Re the BJT is voltage-controlled, and
(b) without Re the BJT can be treated as current controlled (my question: is it or is it not? The physics cannot depend on Re existence)

Why do you use such a rather vague description? Are you not sure?

Regards
W.

 

[vlad777]

Hi Winterstone,

It is not vague,it is math approximation because above 0.7V
Vbe-Ib curve is pretty much flat ie independent of Vbe.
So you can directly control Ib.
Also imagine putting a current source directly on base and ground when emitter is also grounded.
It can be argued that then current source causes both Vbe and Ic.

 

[Winterstone]

Hello agian,

yes - one can agree to everything you have mentioned. But - does it answer the main question of the topic (current or voltage-controlled) ?

 

[vlad777]

I would (and did) say, it depends on how you use it.