# How does a transistor amplify current or voltage?

Discussion in 'General Electronics Chat' started by jac4b, Jul 23, 2010.

1. ### Claude AbrahamMember

Joined:
Mar 17, 2006
Messages:
197
Likes:
23
Location:
NE Ohio, USA
But how does one establish a voltage on the emitter-base junction? You keep re-iterating that Vbe changes the depletion width. In order for Vbe to change, an external power source must do work moving charges through the b-e jcn. Take a PA system as an example.

The dc bias point is established in the microphone 1 stage preamp. The singer forms words, & her voice is amplified by a bjt. What gave rise to a change in Ic, namely "ic" the small signal component? It's not "vbe", nor is it "ib", nor "ie". They are all related & ic cannot exist w/o all 3 being present beforehand. However, the work required to change the depletion region is done by the singer. By moving lips & forming sounds, the acoustic pressure impinges on the mic diaphragm. The diaphragm vibrates, which is energy, or work being done.

This work results in a generation of current & voltage. The diaphragm motion generates the current/voltage that is the input to the mic preamp stage. The input resistance of the stage, including rbb' the base resistance, any emitter resistance present, & the depletion barrier are in series. Energy from the mic gives charge carriers the energy needed to transition from valence to conduction band. Since a p-n jcn is not ideal, there is a forward voltage drop. A portion of the energy is lost driving the b-e jcn. Likewise, some ib is needed to forward bias the jcn.

Thus ib, vbe, & ie, are all consequences of the work done by the singer. It is the singer doing work with her lungs & vocal cords, lips, etc. creating acoustic energy, that gives rise to ib, vbe, ie, as well as ic. The 3 quantities ib/vbe/ie are mutually inclusive. Neither 1 controls another. I've taken 1 solid state physics course from the phy dept as an undergrad. I've taken 4 such courses from the EE dept as a grad. I used highly esteemed texts, peer-reviewed, such as Kittel, Muller/Kamens, & Sze. Nowhere is it mentioned that Ib/Vbe/Ie have a pecking order, i.e. 1 does not control the other 2. Only contrarian non-peer-reviewed web sites like the 1 you referred to take such a position.

FWIW, you acknowledge that a bjt must be externally driven with current, not voltage. You agree that driving Vbe directly is thermally unstable. But when I refer to a bjt as current controlled, that is all I'm inferring. The current control view is only an external view, looking at the device from a circuit perspective. Current control was never claimed to be a microscopic atomic viewpoint.

At the atomic level, only quantum mechanics, QM, can explain device behavior. The ultimate model is Kronig-Penney. The equation is that of Schroedinger. My current control model is just used to externally drive the device. I never said it was valid at atomic level physics. As long as we understand that, all is well. It is also important to state that QM is the only valid FET model when viewing atomic action. The FET voltage control model is a good external model for driving the FET device. It cannot explain semiconductor physics at atomic scale.

Also, FWIW, in a log amp, we do not control voltage, but current. We measure the Vbe with a specific value of current. I spent years working w/ log amps, & I have a patent. Look up US no. 5,670,775. It uses a diode as the log element. Don't be angry with me, but how much formal education have you in semiconductor physics? You keep saying that this controls that, yet you cite no verifiable laws, or peer-reviewed texts. You earlier stated that I & others present valid facts, but err when drawing conclusions. But that is what you are doing, except that not all of your info is factually right. BR.

Last edited: Jul 25, 2010
• Like x 1
2. ### RatchitWell-Known Member

Joined:
Mar 12, 2008
Messages:
1,935
Likes:
78

I already said that regarding a BJT as a current controlled device is useful for design and analysis, even though it does not operate that way under the hood. That is because it mimics a current controlled device.

Sure they require current, especially in the CE path. I never said they didn't. And leakage current in the base circuit is inevitable also. I think the OP got his question answered. If this were in a different thread, would he not have the same problem?

Ratch

3. ### BrownOutBanned

Joined:
Jun 10, 2009
Messages:
4,630
Likes:
60
Location:
Huntsville, Alabama USA
That voltage is required for current to flow in semiconductors is completely irrelevant to the way current is controlled in BJT's. Collector current is controlled by charge injected into the base, and it matters not what it is that injected it. In fact, that has already been treated in some of the previous posts. Some information is right and some is wrong. There is no point in confusing noobs with wrong information, like the false, bizarre "waste" current, and other weird things that are being claimed. Hopefully, the OP can wade through the tangle of information, and choose the correct answer. If he wants more clarification, he may feel free to PM me.

Last edited: Jul 25, 2010
4. ### DaveNew Member

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

Joined:
Oct 27, 2006
Messages:
14,047
Likes:
141
Location:
Rochester, US

No Brownout, collector current is controlled by the voltage field induced at the PN junctions by the charge injected into the base. That is the clinical reality.

I do think waste current is a bad way of looking at things though, and muddies the waters horrible for true understanding. While voltage fields are required for a BJT to function, current is required to flow for those fields to exist. It's Baron Munchhausen'ish to say that one or the other is the dominant effect, both are absolutely required by the basic physical topology and atomic structure of a simplified BJT. The topology and manner the materials are connected require currents to flow, the actual amplification ability is directly derived from the applied voltage field at the semi conductor junctions.

So the voltage field causes the semi conducton state change. The current flows in response to this. You can't have the current without the applied voltage, otherwise semi conductors would work at 0 volts without bias. The magical voltage threshold that the semi conductor effect starts to occur at is the voltage field atomicly required to be applied to modify the valence bands of the semi conductor material itself. Does Ratchit, Brownout, or Claude see the circular logic involved in this argument enough to stop arguing and let this go? Agree to disagree, and be done with it.

Last edited: Jul 25, 2010
6. ### BrownOutBanned

Joined:
Jun 10, 2009
Messages:
4,630
Likes:
60
Location:
Huntsville, Alabama USA
Nope. Collector current is independent of collector-base voltage. The control mechanism is change injected into the base ( current ) The collector junction is a reversed biased diode which is controlled by electrons injected at the base-emitter junction. That’s the reality.

Claude and I just want to get it right. The mechanism of collector current control is current injection, and we feel it’s important enough for people studying electronic to understand. That devices need voltage for current is not relevant to how the collector current is controlled. There is no circular argument. Though the terminals have current and voltage relationships, just as every device would have, the control mechanism is still injected current.

Last edited: Jul 25, 2010
7. ### RatchitWell-Known Member

Joined:
Mar 12, 2008
Messages:
1,935
Likes:
78
Claude Abraham,

There are a myriad of ways to do that. The Vbe is what is important, Ib is the consequence.

So far, so good.

It depends on what you mean by "control". Each of the variables will have a value based on the other two, but Vbe will determine what that value will be, because it controls the depletion width, which is the mechanism for bias placement and modulation. In other words, it controls the inner physical part of BJT. The Ib controls Ic through Vbe.

My guess is that they are mute on that question, and only espouse the current control viewpoint because it is useful and works for design and calculations. But a concensus of opinion does not make it correct unless it is backed up by facts. I believe my position is factual when looking at it from "under the hood".

And I acknowledged already that is a good way to look at a BJT to obtain results. Other and I have always said that voltage control occurs within the transistor mechanism at the depletion layer level.

As I said before, I agree that the current model is very useful as long as you don't get too far inside of the BJT. FETs have not come up for this type of discussion yet.

I am sure there are plenty of ways to make a log amp. But you will get a exponential current if you drive a BJT with a voltage source. I have a BSEE. Here is a reference that says the depletion region narrows and widens according to Vbe. Depletion region - Wikipedia, the free encyclopedia . Do I need to show a reference that says the BE depletion region controls Ic? I would like to be corrected on any faulty info I presented.

Ratch

8. ### RatchitWell-Known Member

Joined:
Mar 12, 2008
Messages:
1,935
Likes:
78
Brownout,

In the active region, BJT collector current is independent of collector-base voltage to a first approximation. Deep inside the BJT, the depletion layer controls the Ic current. The base current injected or ejected into/out of the BJT is an unavoidable consequence of putting a Vbe across the base-emitter junction. This Vbe controls the depletion layer and therefore Ic. The Ib is a useful indicator of Ic.

Ratch

9. ### BrownOutBanned

Joined:
Jun 10, 2009
Messages:
4,630
Likes:
60
Location:
Huntsville, Alabama USA
Here's an example of factually incorrect information. The depletion layer does not control collector current. As Claude has pointed out, you only repeat this over and over and have not proven anything concerning this statement. In fact, it is incorrect. Depletion layer width ( which I suspect you mean when you say "depletion layer"... ) is not the dominant effect of current control. It's all about the injected current at the emitter side of the base, not depletion layer width. Base current is not a consequence of current control, it is the controlling mechanism.

Last edited: Jul 25, 2010
10. ### RatchitWell-Known Member

Joined:
Mar 12, 2008
Messages:
1,935
Likes:
78
Brownout,

In the link I previously referenced in my answer to Claude, Depletion region - Wikipedia, the free encyclopedia , I will quote only the forward bias case.

"(2) Forward bias (P positive with respect to N) narrows the depletion region and lowers the barrier to carrier injection. The diffusion component of the current greatly increases and the drift component decreases. In this case the net current is rightward in the figure of the pn junction. The carrier density is large (it varies exponentially with the applied bias voltage), making the junction conductive and allowing a large forward current.[3] The mathematical description of the current is provided by the Shockley diode equation."

The large forward current is, of course, Ic.

Now, how would you interpret that statement?

Ratch

11. ### BrownOutBanned

Joined:
Jun 10, 2009
Messages:
4,630
Likes:
60
Location:
Huntsville, Alabama USA
I would interpret that statement as woefully misleading. You need to cease trying to learn device physics from Wikipedia, and open an actual book on the subject. Any when you do, don't rely on just P-N junctions. BJT's are more complex than simple junctions.

Joined:
Oct 27, 2006
Messages:
14,047
Likes:
141
Location:
Rochester, US
And for my final comment I will repeat again. Semi conductors are controlled by voltage fields, charge flow is always required to establish or change these fields.

13. ### Claude AbrahamMember

Joined:
Mar 17, 2006
Messages:
197
Likes:
23
Location:
NE Ohio, USA
Wikipedia is not peer-reviewed. In AP level physics at the high school level, I know students who cannot use Wiki as a reference. They must use peer-reviewed info. If they present Wiki w/o supporting peer-reviewed info - automatic F. Wiki is a joke, although they do get some facts correct. You examine Beatty & Wiki & draw the conclusions favorable to your pre-disposed view that voltage is primary, current is just a consequence.

Regarding your depletion width argument, here is my reply. To say that Vbe controls depletion width is circular reasoming. You keep accusing me of circular reasoning, but I have been stating forever that Ib/Vbe/Ie are circularly related, that neither is more primary, all are consequences of the singer imparting energy to the mic.

How does the depletion region vary in width. By the energy produced by the singer. In order to change Vbe, you must inject charge. This requires energy. The added charge results in an increased E field, hence the depletion width changes as does the Vbe value. Vbe & depletion width are related, & both change in unison as charge is extracted from/ injected into the region. Of course work is needed, & that is provided by the singer transferring energy to the mic element.

Vbe is purely incidental, or consequential in the process. A change in vocal energy produces current/voltage in the mic element, resulting in charge injection in the b-e region, changing depletion width & voltage Vbe, as well as current Ib. Vbe is not an active driving source, the mic is. Vbe is the energy per unit charge LOST in forward biasing the b-e junction. Vbe cannot "drive" anything. Ib is not a consequence of Vbe, rather Ib is a consequence of Sue the Singer. Vbe is likewise a consequence of Sue, as is the depletion width. It takes energy to change these quantities. That is where Sue comes in.

Sceadwan, there you go again with that nonsense term "voltage fields"! There is no such thing! There are 4 fields, B - magnetic flux density in volt*second/turn*meter^2, H - magnetic field intensity in amp*turn/meter, E - electric field intensity in volt/meter, & D - electric flux density or displacement in amp*second/meter^2.

The volt & amp units appear in both magnetic & electric field quantities. There is no "voltage field"! Where do you get this stuff? Voltage field??? Nothing personal. BR.

Last edited: Jul 25, 2010
• Like x 1
14. ### RatchitWell-Known Member

Joined:
Mar 12, 2008
Messages:
1,935
Likes:
78
BrownOut,

OK, let me get this straight. You don't think that Vbe widens or narrows the depletion region? Perhaps Wiki is not the best source to reference from, but do you think they got that wrong?

Ratch

15. ### RatchitWell-Known Member

Joined:
Mar 12, 2008
Messages:
1,935
Likes:
78
Claude Abraham,

True, I am predisposed because of their material. I make no apology of it if I think it is correct.

Yes, I know your position. And my position is that Vbe is the dog that wags the tail, even if they are all attached together.

I don't disagree with the above paragraph too much. A counter voltage is needed to beat down the barrier voltage that keeps the emitter-collector current from existing. Whatever makes it happen, a voltage is needed.

I would say that Vbe is essential. Let's just say that Singin' Sue is the cause of Vbe being where it is needed. Sure there is energy involved, and Vbe delivers it through charge carriers, which controls a lot more energy in the emitter-collector circuit. But I still believe that the charge carriers are the leakage current which unfortunately exists.

Ratch

Joined:
Mar 12, 2008
Messages:
1,935
Likes:
78
17. ### Claude AbrahamMember

Joined:
Mar 17, 2006
Messages:
197
Likes:
23
Location:
NE Ohio, USA
If you search hard enough, you will always be able to find people who view the world with a "voltage is primary, current is incidental" viewpoint. As far as Vbe being the dog that wags the tail, you keep saying that it is so, but you cannot show why. Of course, one can accuse me of not providing a definitive answer regarding which of the 3, Ib/Vbe?ie, is the primary controlling quantity.

The fallacy is assuming that 1 of these 3 has to be primary. Singing Sue is responsible for all action. Nothing happens until Sue emits sound. Sceadwan keeps reiterating that Ib/Vbe/Ie are all mutually inclusive, & all indespensible. He is correct. Then he claims that Vbe is ultimately in control, with nothing other than his word as backing.

Not everything can be proved. Some laws are axioms that are not derived from more fundamental laws. But they can be empirically verified through repeatable measurement. The notion that voltage is ultimately in control with current being a consequence albeit a necessary consequence, has been put forth by Ratchit, Sceadwan, & now KA at Anasoft is referenced. But no measurements can affirm that claim. Since Ib/Ie take place in time ahead of Vbe, the notion that Vbe is first, w/ Ib/Ie as consequences cannot withstand scrutiny. Again, many technically oriented people are so accustomed to independent power sources being constant voltage designed, that they see current as a consequence of connecting voltage across an impedance.

Batteries, generators, regulated bench top supplies, are constructed & designed for constant voltage operation. As the load impedance varies, so does the current. This makes it easy to view the voltage as the INdependent variable, & current as secondary or DEpendent.

But the power company spins their turbines at constant speed to produce constant voltage. If they employed constant torque, then the generators would output constant current. It isn't done because losses would be too great. CV works much better, as well as providing constant frequency so that synchronous motors can run at constant speed.

Batteries can be produced with constant current, CC, but the lifetime & performance is greatly degraded. Nuclear batteries, however, work much better in CC mode, & that is how they are designed. Someday, we may use nucells.

The view that voltage is primary, current a consequence, is nothing but a prejudice. Neither comes first. Both are inclusive. The fact the we always design power supplies for CV operation is due to the near ideal nature of insulators, & the lossy nature of conductors. In addition, the electrical world has defined many standards in terms of voltage instead of current. Logic levels in CMOS, TTL, etc. are defined as voltages. This makes sense because the power supply is constant voltage. Video/TV signals are defined as voltages, but they could be referenced to current.

If insulators were very lossy, & conductors were ideal at room temp, the power company would generate CC. We would grow accustomed to current being present always, with voltage being dependent on load impedance. Then this debate would be inverted. It would be universally agreed that bjt is current controlled. But the contrarians would insist that a FET is also current controlled. I would be arguing that it is voltage controlled.

Again, "V before I' is just a prejudice.

18. ### RatchitWell-Known Member

Joined:
Mar 12, 2008
Messages:
1,935
Likes:
78
Claude,

Did I not agree that in a magnetic amplifier, the current controls the output?

You are talking about external inputs to the transistor. I am talking about how that gets translated to the internal control of the transistor.

BJT transistors are a mature technology. There is very little about the operation of the BJT that engineering does not know about. At least on the level we are dealing with. I don't think that electrical energy supplies being mostly constant voltage has much to do with this question.

If a power company were to supply a constant current, then they would have to supply a high voltage to enforce the CC. So if I unplugged an appliance from the wall, I could expect a constant arc across the outlet contacts and perhaps 15 kv to enforce the constant current. What would the current be, by the way?

To get back to the question at hand. I finally found what I was looking for. An archived communication I had with a professor of electrical engineering at the University of Colorado back in Dec 2004. This is his website: Department of Electrical, Computer, and Energy Engineering (ECEE) | University of Colorado at Boulder . I once believed that BJTs were CC until I found out differently. I asked the professor about that, and this was his answer to me.

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
Department of Electrical and Computer Engineering
Campus Box 425
Boulder, CO 80309-0425
Office ECEE1B41
Tel: 303-492-2809
Fax: 303-492-2758

As you can see, I do have good sources for my beliefs.

Ratch

19. ### Claude AbrahamMember

Joined:
Mar 17, 2006
Messages:
197
Likes:
23
Location:
NE Ohio, USA
Every EE prof I've ever had says that the bjt is CC, every last one. This VC viewpoint is from a very vocal minority. My sources for my beliefs are every semiconductor OEM in the world. Your ref says that the voltage applied to the b-e jcn "controls" Ib & Ic. That is the crux of the problem. Apply voltage & get current. He simply states as a fact, w/o proof, the very thing being challenged. As far as his statement "either one can be claimed", I fully agree. The Ib & Vbe are inclusive & both participate in bjt action. However, he falls into the same mis-assumption that all critics make. When I say "current controlled", he immediately assumes that I'm referring to "base current control". I'm refering to emitter current. How many carriers are "collected" is precisely controlled by how many carriers are "emitted". If the emitter emits 417 e-/psec, the collector cannot collect 425, 639, 738, etc. Ic is precisely controlled by Ie. Even with specimam & temp variation, alpha varies over a very narrow range, typ 0.98 to 0.998. The Ie to Ic relationship is very precise & predictable.

Vbe on the other hand has a relation w/ Ic that varies widely w/ speciman & temp. The Ies parameter, reverse saturation current scale factor, varies by a factor of over a million over the industrial temp range of -40C to +85C. Also, 2 different bjt devices have differing Ies values. Likewise, w/ base current control of Ic, beta varies over temp, speciman, & current level, by as much as 10. It can vary from 50 to 500, so in applications where we need precise gain, we employ mathods to "beta-proof" the network.

Once again, every person opposed to CC viewpoint, states as a matter of plain obvious fact, that "Vbe controls Ib/Ic/Ie". They cannot show how or why, but they just "know" that is how it is. Refer to any bjt data sheet. Most have a graph showing Vbe vs. Ic. The Ic is the independent parameter, w/ Vbe dependent. Let me emphasize that Vbe is NOT a consequence of Ic. But Ic is established by the biasing network, & for a specific Ic value, the graph shows the expected Vbe, which varies w/ temp & speciman.

Again, my CC model is EMITTER, not base current. Iagree w/ the prof that "voltage drive" is impractical. But I wopuld say to him the following. It is not merely the exponential I-V relation that makes voltage drive unworkable. But rather, Ies has a strong positive temp coefficient. As the device heats up, Ies increases greatly. So since Ic = alpha*Ies*exp((Vbe/Vt)-1), & Ies has a positive temp coeff, connecting a CVS across b-e jcn results in a current per the above Ebers-Moll eqn. The power is non-zero, so the junction temperature rises. But Ies increases greatly as temp rises, resulting in higher value of Ic, more power, higher temp, higher Ies, higher Ic, higher power, higher temp. etc., etc., etc. We get thermal runaway.

A p-n junction should never be biased w/ a constant voltage source, CVS. But a CCS works fine. Since Vbe = Vt*ln((Ie/Ies)+1), if we apply a current Ie to the emitter, we get a voltage Vbe, & a power, w/ a temp rise. The temp rise increases Ies. But Ies is in the denominator, so thast the quantity Ie/Ies DEcreases, so that Vbe decreases. A rise in temp results in a reduction in power, assuring thermal stability.

If a CCS is not available, a CVS plus a resistor works fine. Your prof is correct that Thevenin & Norton circuits are interchangeable. Any increase in Ib/Ie results in an increase in voltage dropped across the resistor. This results in a decrease in Vbe, & a decrease in power. Again, thermal stability is assured. So b-e jcns MUST be driven w/ either a CCS, or a CVS plus a sufficient series resistor. That is all that "current control" refers to.

Inside the atomic lattice structure, neither CC nor VC can accurately describe the physics. Charge control is used when time delays & bandwidth need to be determioned. For doping, geometry, etc., the QM Kronig-Penney model & Schroedinger's wave eqn are used. VC & CC models are totally inadequate to describe atomic level phenomena.

Every OEM of bjt/FET/diode/LED/photodiode/SCR/triac, etc. concurs. A few contrarians & even a few profs challenge the convention. But they don't produce semiconductors. The OEMs are the horses mouths. FWIW, your prof re-iterates that CC is valid at the external level. But I've always said exactly that. He errs when he says that VC model is valid at the atomic level, because it is not.

Peace, cheers, & BR.

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

Joined:
Mar 12, 2008
Messages:
1,935
Likes:
78
Claude,

I never did understand why you wandered over to saying that Ie controls Ic when Ic is a major component of Ie.

Is that important when we consider what is the causal between Vbe and Ic?

Vbe is what lowers the barrier voltage of the depletion region. The professor explained very briefly why there is a current in the base. The way Vbe and Ic are plotted has nothing to do with the casual relationship.

I think we both agree that a voltage drive on a BJT is unthinkable.

I believe that the prof said that a Thevenin equivalent works for the black box representation of the BJT. He specifically said that a BJT is a voltage controlled device.

I don't think it is necessary to go down to that level. Especially for DC operation.

OEMs manufacture semi's, but they don't necessarily disclose all their information, especially if it is proprietary. They find it easier to say that BJTs are CC, and then show how design and calculations can be made from that assumption. Why should they get into a abstruse argument when all the want to do is sell BJTs as easily as possible?

I think have exhausted all my arguments with you. As you can see, my viewpoint is not a wild theory I embraced, as some folks have implied. I quoted two experts in the field, especially the professor, and I don't have any way of easily locating more who share the VC viewpoint. Looking at the professor's website, and noting all his publications and patents, one has to listen to and respect his opinion. Especially since condensed matter physics appears to be his specialty. I would like to try to get Bill Schockley to comment on this question, but he has been unavailable since 1989. Everyone who reads this thread will have to decide for themselves who is right.

Ratch

21. ### BrownOutBanned

Joined:
Jun 10, 2009
Messages:
4,630
Likes:
60
Location:
Huntsville, Alabama USA
In engineering, expert "opinion" without proof doesn’t mean much. We settle things with physics, math and rigorous proof. In their book, Analog Integrated Circuits, Paul Grey and Robert Mayer develop an expression for forward current gain that does not include any terminal voltages, only device and physical parameters. It does not matter that the input has a voltage/current relationship; every device does.

I trust the experts here, like Mr. Eric Gibbs, who has spent a half-century designing electronic products, over any arbitrary achedemic "expert". It actually sounds like he gave you an opinion that was counter to how you believed. You should have just done your own research.

Last edited: Jul 26, 2010