How does a transistor amplify current or voltage?

[Sceadwian]

During a quantum moment in time current can NOT flow. Zero time as you said. in those moments the currents you're referring to are vectors, which are in fact the voltage fields. Time is not always flowing so current can't always flow, even time has a quanta where things become fuzzy =)

 

[BrownOut]

That's incorrect. Current is always flowing. Making a trivial boundary of zero time does not change current. Current isn't a vector, but it does have velocity and direction, even in instantaneous time. A "voltage field" does not arise from current vectors. In fact, there is no such thing as a voltage field.

Really, this is all a flight of fantasy, and not grounded in any scientific principle. You're only making a trivial condition that doesn’t change anything. Any physicist or engineer will rip this right apart.

 

[Sceadwian]

Poor chose of words BrownOut, my pardon. Electric field.

 

[BrownOut]

Sceadwian, I understand the point you're trying to make. You've seen pictures of P-N junctions with all its E fields, energy bands and Fermi levels. It appears when looking at a picture that current isn't flowing and only E fields exist. The reason for that is because there are no "flip" books that explain device physics. Too bad, because then people would get a much better idea of how it actually works. If current actually stopped flowing, the E fields would quickly vanish and nothing would happen. Thus, current is essential and not just some side effect. The other thing you must consider is, even though you see voltages connected to diodes in the books, the V-I relationships are reciprocal, and neither dominates the other. The reason that I and others promote the CC model is, aside from the fact that nobody has yet to factually invalidate it, is because it just works better in BJT's. That in spite of the fact that the voltage model is most often shown in pictures. To really understand BJT's one must get past the pretty pictures and dig deep into the theory ( alternatively, one can just turn voltage and current around and realize they are equally valid and reciprocal results )

 

[Sceadwian]

That made too much sense not to agree with Brownout.
I hate being wrong, but I'll eat shoe leather any day rather than muddy the water more beyond this point. I like the viewpoint of reciprocity with voltage/current relationship.


<edit>After reading this again I realize how horrible a medium the Internet is for sarcasim =\ <end edit>

 

[Ratchit]

Claude,

Once again, Ratchit, we agree that CC does not describe internal/atomic phenomena in a bjt. But you still insist that CV describes how & why internally. Every OEM & semicon physics reference uses QC, then for a really close look, QM. VC is never used at the atomic level. The late Dr. William Shockley stated in the early bjt days, that only QM can fully explain bjt inner action. You earlier asked what Dr. S would say about bjt innards, & if you search, you can find his early papers, where he employs QM. I don't have any of those papers.

Is the the diode equation at the quantum level? It uses voltage to define current.

So you acknowledge that singing Sue imparts the needed energy. But you state that once the energy arrives at the bjt device, then Vbe "determines" Ic. But how can that be? In a SMPS, input power is taken from the mains, then stored in an inductor when the power switch is on. Then it turns off, so the inductor transfers its energy to the output cap & load. Although the inductor needs constant replenishment from the mains supply, it is an intermediate storage device.

How does that apply to the inner control of a BJT?

But Vbe is not an intermediate variable. When Sue sings, her energy is imparted to the mic element & converted to electrical energy, I & V. The current in the mic cable enters the bjt in the mic preamp. The electrons have energy already. When they enter the base & emitter terminals, Vbe is still unchanged. Thus Ib/Ie have increased before Vbe. When the number of electrons entering the emitter lead increase, then the number collected at the collector also increases as a result. Hence Ic increases due to an increase in Ie. Ib increased as well, but Ib does not contribute to Ic.

Ib and Vbe change simultaneously. The important thing is that a voltage get established across the base-emitter.

In the meantime, these extra electrons entering the emitter, as well as the extra holes injected from base to emitter, cross the depletion zone & recombine. The statistical average time for this to occur is called "mean carrier lifetime". A result of this charge increase is change in the depletion width, an increase in Vbe, & an increase in the stored excess minority charge carries. Thus the change in depletion width, change in Vbe, are merely responses consequential to Sue cranking up her volume.

Yes, and in my analogy, the rack and pinion movement is the result of the driver turning the wheel. As I said before, those charges need voltage to move.

Ie & eventually Ic have already begun to increase before the depletion width has even had time to change. But let me make this clear. Just as Ib/Ie/Ic are not consequences of Vbe, it is equally true that neither Ib nor Ie is the sole driver of bjt action. Like I stated earlier, the E field due to the reverse biased c-b jcn with its associated voltage Vcb, & displacement current, is what "collects" the electrons emitted from the emitter.

Ic will be able to sustain its changed value unless the depletion width changes. I still believe that Ib and Ic are consequences of Vbe.

So a bjt relies on Vcb, Icdisp, Ib, Vbe, & Ie, as well as an external power source as the stimulus. The people telling you that Vbe controls depletion width, which then controls Ie/Ic, w/ Ib as a residue, cannot explain in solid state physics terms the sequential order of events. If you do not believe me, just set up a bjt, or even a FET. Use a rather high freq, say 1 to 10 MHz. Carefully place voltage & current probes at Ib, Ie, Vbe, & Ic. Measure the timing. You will be surprised as the bjt does not behave as the contrarians describe.

Yes, it takes time to sweep the charges out of the semiconductor when the voltage reverses. That is why I did not want to get involved with diffusion capacitance and its effect on frequency.

This constant lecture about "CC is only valid externally, not internally" is preaching to the choir. Everybody knows the CC model is only external, & not internal. Your beef is between VC vs. QC & QM. I have studied this for 3 decades at the BS, MS & now Ph.D. level. I've developed electronic hardware in industry for 32 yrs. I've designed hundreds of bjt & FET networks, operating them as a linear amplifier, as well as a cutoff/saturated switch. The OEM data is amazingly reliable.

No doubt it is.

For 32 yrs., they've been telling me that the bjt is CC device at low speeds, & a QC device when used as a high-speed switch. Their CC & QC equations have been "dead-right-spot-on" for as long as I remember. My 1965 reference text "Pulse, Digital,& Switching" covers CC & QC very thoroughly. Nowhere in any of my uni-approved peer-reviewed reference texts in EE or semicon phy will we find that Vbe controls depletion width which controls Ic, with Ib a residue. I've never seen it.

I am sure that different explanations have to be made for high speeds due to the diffusion capacitance. But would you not agree that in the active region, a higher Vbe will increase the Ic or vice versa, regardless of the reason why?

The contrarians simply cannot accept anything being controlled by a current. They firmly insist that every current has a voltage driving it, making voltage the ultimate controlling quantity. Attempts to illustrate the reason for CC modeling are met with hostility & dogma. Every contrarian has firmly stated that Vbe is dog wagging the tail, but conservation of energy, charge, & close examination of sequence of events refutes their position.

I sure would have liked to be present during some of those arguments. Especially with the professor.

If I'm a fool for accepting the OEM/university/engineering profession's teachings, so be it. That is the type of fool I prefer to be. There is, however, an even bigger type of fool. One who trusts his own judgement, intuition, ability to reason, to the point of rejecting the whole science community because his intuition cannot be wrong. Believe me, that is the biggest type of fool one can be. Such a person is both the leader & the follower of a 1 person cult. The notion that current is not driven by voltage, that neither is the dog nor the tail, that they are mutually inclusive, & are both driven by the energy conversion process in the power source, is a tough pill for some to swallow. They have to define everything in terms of voltage. They acknowledge the role of current, but only as secondary, & consequential, with voltage as the main driver.

Or a minority cult. Do you think the prof is a fool?

I've argued this issue since I was an undergrad in the mid-70's. A small minority always says I'm wrong. But every OEM still publishes the same result since the bjt was introduced.

As I said before, they are concerned with design results. VC is not a good design practice in a BJT.

Ratch

 

[Ratchit]

Brownout,

No they can't be considered the same. There is a depletion layer or space-charge layer. There is no boundary layer. You have to care enough to at least get the vocabulary correct. I know how much you know by reading your misinformation. Better educated members have tried in vain to educate you with factual, correct information. By contrast, you've written meaningless tripe ( like boundary layer or waste current ) If you had correct information, it doesn't make any sense that you wouldn't want to share it. As others have already pointed out, your information has zero usefulness.

I stand corrected on the vocabulary, but it does not change anything. Information is in the eye of the beholder.

I was actually prepared to discuss the physics of CC and VC, but when I read you repeating the same wrong statements each time someone wrote anything useful, I though it just wouldn't be worth the effort. Why should I take the time to get it right, just so you can summarily and categorically contradict, never with any useful alternative data or information, but with cheesy, uniformed lazy comments. I wish the mods watched for trolling more closely and cleaned house of those who troll. Had I not been involved in this thread, I would have reported you for trolling. Next time, I will. Not out of retribution, but because this site is better without trolls.

It is still not too late to do so.

What I would have discussed is pretty much what Claude A. has already said. Every knowledgeable person has this same position.

You are in the majority, all right.

Basically, the argument of every non-knowledgeable person I've read involves VBE being in the terminal or internal equations. To the uninformed, uneducated or just intellectually lazy, this looks like voltage control. Those who know better know that VBE isn't required as a controlling variable and can be a result of current. And for many, very good, well thought out reasons already given, BJT's, though can be controlled by voltage or current, are considered current controlled devices. Nothing in this discussion has reasonably confirmed otherwise.

The discussion was not about the many ways or the best way to control a BJT. It was about whether a BJT was a CC or CV.

Ratch

 

[BrownOut]

The discussion was not about the many ways or the best way to control a BJT. It was about whether a BJT was a CC or CV.

I said nothing about the many or best ways to control a BJT. I detailed why BJT's are CC devices. Look, read and understnad, and you just might learn something.

 

[Ratchit]

Claude,

One more thought. All 3 eqns are useful & needed. When using a bjt as a linear amplifier, we have 3 optional topologies, common emitter, common base, & common collector aka emitter follower. With EF we have voltage gain just under unity, current gain of beta+1. With CB we voltage gain much greater than unity, or Av = Rc/Re, where Rc/Re are the collwctor/emitter resistors resp., & with CE we get current gain of beta, w/ voltage gain of -Rc/Re.

The reason that all 3 are important is that no amp stage can ever have a gain larger than the raw bjt device, current or voltage. Eqn 1:

Ic = beta*Ib, conveys the effectiveness of the bjt device at amplifying current. The bias resistors on the bjt input side draw some current, so that the current gain of the stage, Ai, is always less than that of the device, which is beta.

The 2nd eqn, 2) Ic = alpha*Ies*exp((Vbe/Vt)-1), conveys the effectiveness of the bjt device at amplifying voltage. THe transconductance of the overall amp stage can never exceed that of the raw bjt device. Some voltage is dropped across elements in the biasing network, including emitter degenerating resistance, as well as bjt internal resistances. If we examine the ic vs. vbe ratio by differentiating eqn 2), we get a small signal transconductance parameter "gm" as gm = Ic/Vt, so that ic = gm*vbe, where the lower case "i & v" indicate ac small signal quantities.

In the CB mode, the current gain is limited to alpha, per eqn 3) Ic = alpha*Ie.

All 3 eqns provide useful info as to what a bjt can offer in terms of signal gain, both current & voltage. Neither is the dog or the tail. Both are needed. What makes an active device useful is its ability to amplify current & voltage BOTH with gain values >> 1. To output a specific value of current & voltage requires minimum input values of current & voltage. How much is needed? The 3 eqns provide the answer.

None of the 3 eqns involve "causality", consequential, etc. They convey useful info to the user detailing device limitations. The QC equations detail the speed limitations of the device.

I agree with all you said above. That is basic Transistors 101.

I have millions of products whose hardware I designed out in the field worldwide. My products don't come back for repair. They work first time & indefinitely after that.

You appear to be a skilled and sure engineer.

I worked for an aerospace/defense firm in Baltimore with thousands of engrs, tech, Ph.D. etc. I never once heard of bjt being VC. I've never met anyone who actually develops hardware saying otherwise.

As I said before, I sure would not use Vbe to design something.

I swear they get a rush just putting down actual practitioners. In industry, I've met people who were hostile to me on my 1st day on the job for no good reason. Being an EE seems to annoy a handful of people for reasons I wish I knew.

EE is an honorable profession. I don't know why anyone would denigrate it.

Ratch

 

[Claude Abraham]

Claude,

Is the the diode equation at the quantum level? It uses voltage to define current.

How does that apply to the inner control of a BJT?

Ib and Vbe change simultaneously. The important thing is that a voltage get established across the base-emitter.

Yes, and in my analogy, the rack and pinion movement is the result of the driver turning the wheel. As I said before, those charges need voltage to move.

Ic will be able to sustain its changed value unless the depletion width changes. I still believe that Ib and Ic are consequences of Vbe.

Yes, it takes time to sweep the charges out of the semiconductor when the voltage reverses. That is why I did not want to get involved with diffusion capacitance and its effect on frequency.

No doubt it is.

I am sure that different explanations have to be made for high speeds due to the diffusion capacitance. But would you not agree that in the active region, a higher Vbe will increase the Ic or vice versa, regardless of the reason why?

I sure would have liked to be present during some of those arguments. Especially with the professor.

Or a minority cult. Do you think the prof is a fool?

As I said before, they are concerned with design results. VC is not a good design practice in a BJT.

Ratch

When you say that the diode equation uses voltage to define current, are you referring to Id = Is*exp((Vd/Vt)-1)? Are you aware that the same equation can be written as Vd = Vt*ln((Id/Is)+1)? Just as current is an exponential function of voltage, voltage is also a logarithmic function of current. The equation is usually introduced in I vs. V form to emphasize how a small change in voltage accompanies a large change in current. There is no pecking order here. In Ohm's law, V = I*R, I = V?R< & R = V/I, all say the same thing.

Since Ib & Ic change before Vbe, how can Ib/Ic/Ie be consequences of the change in Vbe. The consequence cannot occur before the controlling quantity. As far as a higher Vbe increasing Ic, again, the external stimulus increases Ie & Ib first, then Ic & Vbe increase as consequences afterward. Vbe is not controlling Ic, rather Ie is controlling Ic. The sequence of events affirm this. The notion that Vbe controls Ic is one you cannot demonstrate w/ semicon physics, you just dogmatically insist upon it. I'm at a loss to make that clearer.

Which prof are you referring to? The debates in my undergrad years were not with any prof. An undergrad ME student once told me that V comes before I. He also told me that electric cars should use generators to recycle energy. When the battery is driving the motors, he felt that adding a generator would replace some of the energy, allowing for extended battery life. Attempts to show him why that won't work were futile. My profs always emphasized that I & V are circular in relation to each other. Neither controls the other.

Do you develop hardware? What field if so? Have you studies semicon physics? You cannot address the energy issues. You ignore the fact that a change in energy requires both Ib & Vbe. It's pointless to continue. People reading this thread will hopefully know that OEMs have a reason for their positiions. They can research this matter themselves, & learn why the CC & QC models are employed for the bjt.

Claude

 

[BrownOut]

When you say that the diode equation uses voltage to define current, are you referring to Id = Is*exp((Vd/Vt)-1)? Are you aware that the same equation can be written as Vd = Vt*ln((Id/Is)+1)? Just as current is an exponential function of voltage, voltage is also a logarithmic function of current. The equation is usually introduced in I vs. V form to emphasize how a small change in voltage accompanies a large change in current. There is no pecking order here. In Ohm's law, V = I*R, I = V?R< & R = V/I, all say the same thing.


Claude

These are the details behind my point about reciprocity.

 

[Ratchit]

Brownout,

I said nothing about the many or best ways to control a BJT. I detailed why BJT's are CC devices. Look, read and understnad, and you just might learn something.

You said " And for many, very good, well thought out reasons already given, BJT's, though can be controlled by voltage or current, are considered current controlled devices."

I took that to mean you were talking about ways to control a BJT.

Ratch

 

[BrownOut]

Brownout,

You said " And for many, very good, well thought out reasons already given, BJT's, though can be controlled by voltage or current, are considered current controlled devices."

I took that to mean you were talking about ways to control a BJT.

Ratch

Then you don't understand simple english. "Are considered current controlled devices" doesn't sound anything like "ways to control BJT's"

The only mention of control was a subordinate clause. Look it up in a freshman English book.

 

[Ratchit]

Claude,

When you say that the diode equation uses voltage to define current, are you referring to Id = Is*exp((Vd/Vt)-1)? Are you aware that the same equation can be written as Vd = Vt*ln((Id/Is)+1)? Just as current is an exponential function of voltage, voltage is also a logarithmic function of current. The equation is usually introduced in I vs. V form to emphasize how a small change in voltage accompanies a large change in current. There is no pecking order here. In Ohm's law, V = I*R, I = V?R< & R = V/I, all say the same thing.

Yes, I am aware of inverse relationships. My point is that Vbe and Ic are tied together by a relatively simple equation that does not involve QM to express. I iterate that I would not design a BJT using that equation, however.

Since Ib & Ic change before Vbe, how can Ib/Ic/Ie be consequences of the change in Vbe. The consequence cannot occur before the controlling quantity. As far as a higher Vbe increasing Ic, again, the external stimulus increases Ie & Ib first, then Ic & Vbe increase as consequences afterward. Vbe is not controlling Ic, rather Ie is controlling Ic. The sequence of events affirm this. The notion that Vbe controls Ic is one you cannot demonstrate w/ semicon physics, you just dogmatically insist upon it. I'm at a loss to make that clearer.

I don't agree that Ib & Ic change before Vce. I would think they changed together. You are right, I cannot imagine one of those variables changing before the other.

Which prof are you referring to? The debates in my undergrad years were not with any prof. An undergrad ME student once told me that V comes before I. He also told me that electric cars should use generators to recycle energy. When the battery is driving the motors, he felt that adding a generator would replace some of the energy, allowing for extended battery life. Attempts to show him why that won't work were futile. My profs always emphasized that I & V are circular in relation to each other. Neither controls the other.

The one I referred to earlier. Here is his website. **broken link removed** . He responded to me with a email when I asked him if a BJT was CC.

Ratch,
I hate to be the one telling you this but the BJT is indeed a voltage
controlled device. The voltage applied to the base emitter junction controls
the collector current and the base current is a result of the additional
hole injection (for an npn BJT) into the emitter as well as the
recombination in the base-emitter depletion region and the quasi-neutral
base region. It is tempting to claim that the BJT is controlled by the base
current, since that is how a BJT is typically biased; the exponential
variation of the current with the base-emitter voltage makes a voltage bias
impractical. Any circuit designer will also tell you that any voltage bias
can be replaced by its Thevenin equivalent current source. Hopefully this
provides you some ammunition to claim that either one can be claimed when
treating the device as a black box. Finally, you'll find that a MOSFET
biased in the subthreshold region has characteristics that are very similar
to that of a BJT.
Bart Van Zeghbroeck
Professor
University of Colorado
Department of Electrical and Computer Engineering
Campus Box 425
Boulder, CO 80309-0425
Office ECEE1B41
Tel: 303-492-2809
Fax: 303-492-2758
Email: b...@colorado.edu

The ME you referred to was off base, of course.

Do you develop hardware? What field if so?

No, I do not.

Have you studies semicon physics?

Yes.

You cannot address the energy issues. You ignore the fact that a change in energy requires both Ib & Vbe.

I agree it does. And if the BJT has a higher β, then it requires less energy, less current, but the same Vbe.

It's pointless to continue. People reading this thread will hopefully know that OEMs have a reason for their positiions. They can research this matter themselves, & learn why the CC & QC models are employed for the bjt.

Certainly they can.

Ratch

 

[Ratchit]

Brownout,

Then you don't understand simple english. "Are considered current controlled devices" doesn't sound anything like "ways to control BJT's"

The only mention of control was a subordinate clause. Look it up in a freshman English book.

Sorry for the misinterpretation.

Ratch

 

[Mikebits]

Okay, I tried to bite my lip through this thread but, somebody has to speak up.
Brownout, surely there are other ways to settle an argument without resorting to ever so slight insulting innuendos. Despite your constant demeaning tone towards Rachit, he has not responded to you in the same manner, and he has reciprocated in a polite and cordial tone. He has not questioned your engineering skills, nor has he questioned your grasp of the English language. Agree with him or not, I think he deserves the same respect that he has shown to you. So how about it?

 

[BrownOut]

Okay, I tried to bite my lip through this thread but, somebody has to speak up.
Brownout, surely there are other ways to settle an argument without resorting to ever so slight insulting innuendos. Despite your constant demeaning tone towards Rachit, he has not responded to you in the same manner, and he has reciprocated in a polite and cordial tone. He has not questioned your engineering skills, nor has he questioned your grasp of the English language. Agree with him or not, I think he deserves the same respect that he has shown to you. So how about it?

Oh really? Before you ride in on your high horse again, consider the experienced people who took the time to try to help him understand, and the way he repeatedly threw it back in their faces. Do you think those on here who go to the trouble to help someone should have to put up with impetulence?

ericgibbs: I think you will realise that there are a lot learned and experienced members who will try to convince you otherwise.


Ratchit: Bring them on. But I hope they can show me their reasoning, and not just throw out some links that only parrot the false viewpoint that has been published before.

Ratch

This goes on for another 10 pages or so. Maybe you're OK when our senior people who, overwhelmingly provide the most assistance on here, getting kicked in the teeth for their efforts. I'm not.

 

[Claude Abraham]

Claude,

Yes, I am aware of inverse relationships. My point is that Vbe and Ic are tied together by a relatively simple equation that does not involve QM to express. I iterate that I would not design a BJT using that equation, however.

I don't agree that Ib & Ic change before Vce. I would think they changed together. You are right, I cannot imagine one of those variables changing before the other.

The one I referred to earlier. Here is his website. **broken link removed** . He responded to me with a email when I asked him if a BJT was CC.

Ratch,
I hate to be the one telling you this but the BJT is indeed a voltage
controlled device. The voltage applied to the base emitter junction controls
the collector current and the base current is a result of the additional
hole injection (for an npn BJT) into the emitter as well as the
recombination in the base-emitter depletion region and the quasi-neutral
base region. It is tempting to claim that the BJT is controlled by the base
current, since that is how a BJT is typically biased; the exponential
variation of the current with the base-emitter voltage makes a voltage bias
impractical. Any circuit designer will also tell you that any voltage bias
can be replaced by its Thevenin equivalent current source. Hopefully this
provides you some ammunition to claim that either one can be claimed when
treating the device as a black box. Finally, you'll find that a MOSFET
biased in the subthreshold region has characteristics that are very similar
to that of a BJT.
Bart Van Zeghbroeck
Professor
University of Colorado
Department of Electrical and Computer Engineering
Campus Box 425
Boulder, CO 80309-0425
Office ECEE1B41
Tel: 303-492-2809
Fax: 303-492-2758
Email: b...@colorado.edu

The ME you referred to was off base, of course.

No, I do not.

Yes.

I agree it does. And if the BJT has a higher β, then it requires less energy, less current, but the same Vbe.

Certainly they can.

Ratch

Ib/Ie change before Vbe. Look it up. You mentioned diffusion capacitance already. In a cap, the current change leads the change in voltage. Not disputable.

Now regarding the energy required to change the E fiels, I've shown it is proportional to Ib*Vbe. You replied that for a higher beta, less Ib is needed, but the same Vbe. This is WRONG! The Ib I'm referring to is displacement cuurent. The beta you refer to is conduction current. Not the same. Let's say we have 2 identical bjt devices, equal in all respects except 1, part A, has a base doping density lower than the other, part B. Part A will have a higher beta than part B.

The energy required to change the b-e region E field is the SAME! The areas are equal,as well as the geometry. Although part A requires less base conduction current than part B, it still needs the same value of displacement current as part B. Which semicon phy courses have you taken? What univ? Which text(s). Don't take it personally, I just want to know your sources for your info. Are you intimately familiar with the difference between conduction vs. displacement current?

A bjt with very high beta is still a CC device. Take for example, the supergain bjt devices used in the front end of op amps. They have beta values in the 4000 to 5000 range. They are still CC devices. They need current to operate (as well as voltage), but less current than an ordinary bjt with a beta value of 100-400. Likewise, some FETs require more gate drive current than others. A small signal FET operating at 10 Hz, may require only NANOamps of gate drive. A large power FET switching at 2 MHz requires AMPS of gate drive current. It's still a VC device.

Also, just as bjt parts have varying beta, so do FET parts have varying gm values. A FET may have a gm of 5 mS for given current, temp, die size, etc. Another FET may have 10 times that value, or 100 times. As gm increases, less voltage at the g-s terminals is needed. But as gm increases, the FET does not act like a CC device. It is still VC, but less "V" is needed. This business about "as beta increases a bjt approaches VC" does not hold water. A forward biased p-n junction cannot have zero current with non-zero voltage. A power bjt operating at very high current levels, at low temp, may have a beta value of 3 or 4. With beta = 4, 50A of Ic requires 12.5A of Ib. A supergain bjt at the op amp input can have beta of 5000, so 100 uA of Ic can be had with just 20 nA of Ib.

They are both CC devices. Likewise 2 FETs w/ grossly differing gm values & gate currents & g-s voltage swings are both VC.

 

[Mikebits]

This goes on for another 10 pages or so. Maybe you're OK when our senior people who, overwhelmingly provide the most assistance on here, getting kicked in the teeth for their efforts. I'm not.

Okay, I think your right, I was having a preachy moment.

 

[Ratchit]

Claude,

Sorry for the delay in answering, but it was a busy day.

Ib/Ie change before Vbe. Look it up. You mentioned diffusion capacitance already. In a cap, the current change leads the change in voltage. Not disputable.

In sinusoidal operation, yes. But as I said before, I wanted to just consider the DC characteristics.

Now regarding the enrgy required to change the E fiels, I've shown it is proportional to Ib*Vbe. You replied that for a higher beta, less Ib is needed, but the same Vbe. This is WRONG! The Ib I'm referring to is displacement cuurent. The beta you refer to is conduction current. Not the same. Let's say we have 2 identical bjt devices, equal in all respects except 1, part A, has a base doping density lower than the other, part B. Part A will have a higher beta than part B.

As I said before, I am only considering DC characteristics.

The energy required to change the b-e region E field is the SAME! The areas are equal,as well as the geometry. Although part A requires less base conduction[/b[ current than part B, it still needs the same value of displacement current as part B. Which semicon phy courses have you taken? What univ? Which text(s). Don't take it personally, I just want to know your sources for your info. Are you intimately familiar with the difference between conduction vs. displacement current.

I took the regular undergraduate courses that are standard for just about any student of EE. I went to a big 10 land grant college. The texts were the modular series on solid state physics from Purdue University by Robert F. Pierret. and Gerold W. Neudeck. It has been a long time since I studied semiconductors intensively, so I am somewhat rusty on the details. I never worked with semiconductors for a living.

A bjt with very high beta is still a CC device. Take for example, the supergain bjt devices used in the front end of op amps. They have beta values in the 4000 to 5000 range. They are still CC devices. They need current to operate (as well as voltage), but less current than an ordinary bjt with a beta value of 100-400. Likewise, some FETs require more gate drive current than others. A small signal FET operating at 10 Hz, may require only NANOamps of gate drive. A large power FET switching at 2 MHz reqires AMPS of gate drive current. It's still a VC device.

I already granted that current and power is inevitable. My focus is where the control circuit the control lies. It is the driver, or the rack and pinion steering.

Also, just as bjt parts have varying beta, so do FET parts have varying gm values. A FET may have a gm of 5 mS for given current, temp, die size, etc. Another FET may have 10 times that value, or 100 times. As gm increases, less voltage at the g-s terminals is needed. But as gm increases, the FET does not act like a CC device. It is still VC, but less "V" is needed. This business about "as beta increases a bjt approaches VC" does not hold water. A forward biased p-n junction cannot have zero current with non-zero voltage. A power bjt operating at very igh current levels, at low temp, may have a beta value of 3 or 4. With beta = 4, 50A of Ic requires 12.5A of Ib. A supergain bjt at the op amp input can have beta of 5000, so 100 uA of Ic can be had with just 20 nA of Ib.

Let's just stick to BJT'S. I already said that power and current are necessary to drive a BJT. Again, my focus is where the control is.

Ratch