How Transistor works - verifying

[Nikarus]

Hi All,
I'm trying to understand how transistor works. I have drew some pictures for NPN transistor to show my current understanding. Can you please verify this for me.

On the picture:
"-" and "+" are movable charges;
circled "-" and circled "+" are immovable ions.

1. No voltage applied. Two depletion layers are formed.

2. We apply voltage that we want to control - "-" to Emitter and "+" to Collector. Electrons from Emitter fill the Base forming negative ions. Electrons from Collector are gone; positive ions are formed in there. Because Base is filled with negative ions, electrons from Emitter can't go through - channel jammed.
Here I'm not sure about Base-Emitter depletion layer - will there be any interaction between Base and Collector..?

3. We apply controlling voltage to Base. If we have BJT than we will have current between Base and Emitter, if we have FET, electrons inside of Base will be just pulled up, leaving positive charges behind. Electrons from Emitter will flow thought Base to Collector and will not be jammed again because Base is very thin.

So, is my understanding correct? Do I do any mistakes..?
Thank you!

 

[Nikarus]

Guys please, I spent a lot of time drawing this picture!

 

[Tryp]

you can look at transistor as two diodes facing each other

--l>l-*-l<l---


this also forms NPN junction and will let thru if voltage is applied to the between.

Your assumption is basically correct ~ Applying + to emmiter will draw free electrons from there leaving + ions, applying - to collector will result in moving electrons to base and neutralizing some free 'holes'.
And when applying + to base most of electrons will go from collector to emmiter, but some will go to + in base and coming back to collector, so the smaller is base the better.
In ideal transistor Electron flow input in collector should = electron output in emmiter.

 

[rbj]

you can look at transistor as two diodes facing each other

--l>l-*-l<l---

that's hardly the case. it doesn't account at all for the mechanism of current amplification due to the thin base.

indeed...

And when applying + to base most of electrons will go from collector to emmiter, but some will go to + in base and coming back to collector, so the smaller is base the better.

you won't get a thin base layer by soldering two diodes facing each other.

 

[Tryp]

that's hardly the case. it doesn't account at all for the mechanism of current amplification due to the thin base.

it will work out transistor mechanism just fine.

you won't get a thin base layer by soldering two diodes facing each other.

never said so.

 

[valeriy2222]

It's difficult to find a good explanation about transistors, but here a link

<< link deleted by berkeman >>

It's a big article, but it leaves most questions behind once you read it.

 

[dlgoff]

I've posted this here several times. I hope it might help your understanding.

http://hyperphysics.phy-astr.gsu.edu/hbase/solids/trans2.html"

 

[cabraham]

It's difficult to find a good explanation about transistors, but here a link

<< link deleted by berkeman >>

It's a big article, but it leaves most questions behind once you read it.

This article is heresy. Peer-reviewed texts have the best info. I've covered this issue many times as well as others. A search for my post history will show these threads. Then I can address any questions. Stay away from the above web site. It contains just enough truth to look credible, but makes assumptions contrary to known science.

Claude

 

[Nikarus]

Hi All,
Sorry for the delay - I had a lot of thing to do these days. But I'm still full of enthusiasm :)
I tried to read this explanation before (<< link deleted by berkeman >>) And I agree with author that a lot of other authors does not really explain how transistors work. But since cabraham says that this is heresy, I stopped reading it :)

Actually I already read a lot of explanations about how transistor works, and I have my understanding. And I want to make sure of it. So this is why I posted this thread.

Guys, please, can you not just give me another explanations, but read my own and tell me what is wrong there on my picture. So I could use my own understanding as a reference point to move from.

 

[valeriy2222]

This article is heresy. Peer-reviewed texts have the best info. I've covered this issue many times as well as others. A search for my post history will show these threads. Then I can address any questions. Stay away from the above web site. It contains just enough truth to look credible, but makes assumptions contrary to known science.

Claude

Among other explanations this one gives you an understanding of principle of work. Having read this, you can correct mistakes made in the article by reading other more scientific articles and thus come up with the whole and right picture.

 

[valeriy2222]

.

Here is a video presentation that describes the principle of work of transistor.
http://www.youtube.com/watch?v=FODFowmDfvY&feature=related

 

[cabraham]

Among other explanations this one gives you an understanding of principle of work. Having read this, you can correct mistakes made in the article by reading other more scientific articles and thus come up with the whole and right picture.

Sure, but why not just read more scientific articles in the 1st place that don't have mistakes. Then there is no need to correct them. Peer-reviewed books from semiconductor producers & universities are the best source of info. These contrarian web sites just confuse the issue to the point where one does not know who to believe.

Claude

 

[cabraham]

Hi All,
Sorry for the delay - I had a lot of thing to do these days. But I'm still full of enthusiasm :)
I tried to read this explanation before (<< link deleted by berkeman >> ). And I agree with author that a lot of other authors does not really explain how transistors work. But since cabraham says that this is heresy, I stopped reading it :)

Actually I already read a lot of explanations about how transistor works, and I have my understanding. And I want to make sure of it. So this is why I posted this thread.

Guys, please, can you not just give me another explanations, but read my own and tell me what is wrong there on my picture. So I could use my own understanding as a reference point to move from.

In your pic, 3rd view at bottom, the base region is shown w/ an excess of electrons, herein called "e-", & an absence of holes herein "h+". Also, the collector has many e-, but few h+.

For the 3rd pic, where Vce > 0, & Vbe > 0, the device is in its active region. We will not cover the saturated region yet. In the collector, there are a small number of mobile h+, as they are minority carriers, & e- are majority carriers w/ high mobility. W/o Vbe biased > 0, the e- are crowded near the collector lead, & the h+ are near the base-collector boundary. The c-b jcn is rev biased & little current exists from base to collector.

But the b-e jcn is now fwd biased from an external source, & Ib/Ie > 0, & Vbe > 0. The emitter emits many e- toward the base, & the base emits not as many h+ towards the emitter. The h+ from the base enter the emitter & recombine w/ e- near the e-b boundary. The e- from the emitter transit into the base region. Some recombine w/ a h+ in the base, some transit through the base region out the base lead, & nearly all transit through the base into the collector region, where they again are majority carriers.

The number of e- which do not reach the collector are around 1 in 2000 to 1 in 5000. The number of h+ emitted from the base are about 100 to 200 times less than the number of e- emitted from the emitter. Thus the base current consists chiefly of holes emitted from base to emitter. This is the injection component of base current. The e- emitted from the emitter which recombine in base or exit base lead constitute a very small fraction of the base current & is called the transport component.

Another component of base current is the charging or displacement component. As frequency increases more h+ & e- get displaced but do not contribute to collector current. So the current gain decreases as freq increases. Eventually there is a frequency where the beta value equals one. This is the transition frequency of the bjt, called "ft" on data sheets.

The injection & transport components of base current determine the dc & low freq current gain, "hFE" on spec sheets. Adding the displacement component of base current decreases the current gain (beta value) at high freq. The ac & high freq current gain is called "hfe" on spec sheets.

It's all about the c-b jcn being reverse biased so that very little collector current exists. By forward biasing the b-e jcn, many e- flow from emitter to base. Nearly all are yanked into the collector by the electric field in the b-c boundary region. The base is made thin so as to minimize the number of e- recombining in base region. The base is doped lightly & the emitter heavy so as to minimize the injection component of base current.

The current gain, bets, is determined by Ic/Ib, where Ib consists of the 3 parts, injection, transport, & displacement.

Regarding heresy, I've addressed Mr. Beaty on another forum. We had a debate. I'll provide a link if you wish to see the points he & I presented.

Best regards.

Claude

 

[niwin]

Hi Claude,

I'm very interested in the debate you had with Mr. Beaty. Can you post the link? :)

 

[cabraham]

Hi Claude,

I'm very interested in the debate you had with Mr. Beaty. Can you post the link? :)

http://www.electro-tech-online.com/...oes-transistor-amplify-current-voltage-4.html

It's long, but worth the read. I address every issue brought up. Other posters made good contributions as well.

Claude

 

[niwin]

Man, I can't believe you guys had the patience to duke it out like that! And it ended with some level of agreement? Truly epic.

I actually enjoyed reading Mr. Beaty's site, but there should be a rebuttal to it, or a summary of the epic thread published on the web. I think a lot of confused people would appreciate it. Thanks for sharing.

 

[cabraham]

Man, I can't believe you guys had the patience to duke it out like that! And it ended with some level of agreement? Truly epic.

I actually enjoyed reading Mr. Beaty's site, but there should be a rebuttal to it, or a summary of the epic thread published on the web. I think a lot of confused people would appreciate it. Thanks for sharing.

One of these days I should copy & paste the whole thread into a text document. Then I can post the text. It gives all the different viewpoints from many different parties. That way, nobody is quoted out of context. I'll do that soon.

Claude

 

[wbeaty]

Hi Claude,

I'm very interested in the debate you had with Mr. Beaty. Can you post the link? :)

I'm interested too, since I've never had any such debate.

 

[cabraham]

I'm interested too, since I've never had any such debate.

Refer to the link I gave a few posts back. The poster called himself Bill Beaty. Once again, here it is:

http://www.electro-tech-online.com/g...voltage-4.html

If that isn't you, then someone used your name. Maybe the moderators on that forum can help you. Best regards.

Claude

 

[wbeaty]

Refer to the link I gave a few posts back.

I did.

The poster called himself Bill Beaty. Once again, here it is:
http://www.electro-tech-online.com/...es-transistor-amplify-current-voltage-4.html"

Where? No "beaty" at all on that page linked above, or on that thread as far as I've searched. On that site I see you debating with a "Ratchit"

 

[Studiot]

There have been other web debates about Beaty's site.

In general you cannot divorce voltage and current. So to claim that a transistor is wholely current controlled or wholely voltage controlled is pointless to say the least.

In this thread I pointed out that there are other ways to affect (control) the collector current than either changing the voltage or injecting base current.

One way is by injecting photons which increase the base current without increasing the base voltage and thereby increase the collector current.

http://forum.allaboutcircuits.com/showthread.php?t=12378&highlight=transistor+action&page=13

 

[Bassalisk]

Since you gave this link Studiot, I checked it out and one thing caught my eye. Discussion between Ebers-Moll and Gummel-Poon. I learned Gummel-Poon model of bipolar transistor at my college, and I know how to derive them from scratch. I see something about frequencies being mentioned.

Is this only limitation of this GP model? And which model generally is more accurate?

 

[cabraham]

I did.



Where? No "beaty" at all on that page linked above, or on that thread as far as I've searched. On that site I see you debating with a "Ratchit"

Sorry, I gave the wrong link. Here is the discussion you & I had, along w/ others, back in 2010:

https://www.physicsforums.com/showthread.php?t=390291&highlight=wbeaty

Again, my apologies. Best regards.

Claude

 

[wbeaty]

In this thread I pointed out that there are other ways to affect (control) the collector current than either changing the voltage or injecting base current.

One way is by injecting photons which increase the base current without increasing the base voltage and thereby increase the collector current.

Wow, interesting! Have some evidence that it happens as you say? Opto engineering refs about this issue?

Or forget talking, "Let the experiment be made." Instead just take a phototransistor and go see what actually happens. Try using a variable DC supply to hold the Vbe constant while flashing a light source and measuring Ie or Ic pulse. Then do it again with a floating Base lead. The usual BJT models predict that any expected large gain in current during the light pulse cannot occur if the height of the potential barrier at the BE junction isn't allowed to change. The Emitter won't be commanded to emit. Obviously you'll still get a photocurrent. But if the BE junction barrier is fixed, there shouldn't be any transistor action or any large emitter current.

Full disclosure: I spent some years as an EE at Eaton/Cutler-Hammer photosensors division. But PTs don't have well controlled characteristics, so all our products employed Photodiodes (PDs,) so I wouldn't have encountered this weird "no Vbe change" phototransistor phenomenon.

http://forum.allaboutcircuits.com/showthread.php?t=12378&highlight=transistor+action&page=13

Nothing about phototransistors that I find. Page 13 wrong?

 

[Studiot]

@wbeaty

Posts 55/66/80/84 in the link contain the information you asked for. You can get a photon induced variation to collector current without the base even being connected, let alone arguing whether it is current or voltage controlled.
I do not recommend this but there used to be a demonstration involving sawing the top off a 2N3055 and shining a light on it. Scratching an old glass-cased germanium transistor is a safer bet.

@bassalisk

Gemmel- Poon is some way beyond Ebbers Moll and is a good model over a wider current range and frequency range and certainly well beyond anything you have discussed here at PF before.
Both are circuit theory models, not models of the physics of transistors. You should always distinguish which you are needing or considering.

 

[wbeaty]

@wbeaty
Posts 55/66/80/84 in the link contain the information you asked for.

Thanks, I'll check it out

You can get a photon induced variation to collector current without the base even being connected

Certainly, that's how 2-lead phototransistors are supposed to work. The base must remain disconnected in order to let Vbe vary freely. On a 3-lead PT you can use a voltmeter to watch the changes. But if we use a power supply to hold Vbe constant, will the large transistor-effect current vanish?

demonstration involving sawing the top off a 2N3055

Nah, the ambitious or curious could just buy reliable devices designed for the task:

http://search.digikey.com/scripts/DkSearch/dksus.dll?vendor=0&keywords=phototransistors+ssg+through+hole&stock=1"J

http://www.mouser.com/Optoelectronics/Optical-Detectors-and-Sensors/Photodetector-Transistors/_/N-6jjuf?Keyword=osram&FS=True"

 

[cabraham]

Wow, interesting! Have some evidence that it happens as you say? Opto engineering refs about this issue?

Or forget talking, "Let the experiment be made." Instead just take a phototransistor and go see what actually happens. Try using a variable DC supply to hold the Vbe constant while flashing a light source and measuring Ie or Ic pulse. Then do it again with a floating Base lead. The usual BJT models predict that any expected large gain in current during the light pulse cannot occur if the height of the potential barrier at the BE junction isn't allowed to change. The Emitter won't be commanded to emit. Obviously you'll still get a photocurrent. But if the BE junction barrier is fixed, there shouldn't be any transistor action or any large emitter current.

Full disclosure: I spent some years as an EE at Eaton/Cutler-Hammer photosensors division. But PTs don't have well controlled characteristics, so all our products employed Photodiodes (PDs,) so I wouldn't have encountered this weird "no Vbe change" phototransistor phenomenon.

But we DO NOT ever hold Vbe constant. That is the point of controlling current, not voltage. When a bjt is *current driven*, the source providing the current also provides the voltage. In the case of current controlling a b-e jcn of a bjt, it is understood, no need to explicitly state, that a change in Ib/Ie inevitably produces a change in Vbe as well.

Nothing about phototransistors that I find. Page 13 wrong?

I'm not sure what you're trying to prove by holding Vbe steady with a dc lab constant voltage supply clamped right acrss the b-e jcn. That is not how to drive it. I believe that what you are demonstrating is that when photons are incident upon the jcn, & current is set up in response to the photon stimulus, that there will inevitably be a change in Vbe. Is that your point?

But I would have told you the same, as would many others here skilled in semicon physics. Once carriers are emitted from the emitter & base regions, they cross the barrier & recombination/ionization takes place. There is a change in the local E field due to this increased density of carriers. Since Vbe is just the line integral of the E field along the b-e path, a change in Ib/Ie results in a change in the net E field in the depletion zone, which results in a change in Vbe.

Whenever bjt action takes place, all quantities participate, namely Ib, Vbe, Ie, & Vbc. "Transistor action" requires all of them. The "current control" description we use, refers to the quantity that is directly controlled, & it is understood that the others are incidental, but still participate. We control Ie, or in limited applications we control Ib. But Vbe is indirect & consequential, although it is still an important parameter.

If we wish to use a bjt to run a motor requiring 1.0 amp, with a beta of 100, Si bjt material, then we need 10 mA & 0.65 volts to operate the bjt. Again, we DO NOT apply a 0.65V voltage source to the jcn, but rather we apply 10 mA to the jcn, & let Vbe fall where it may. At different temperatures Vbe will vary. But Vbe is important, very much so. If the source providing the 10 mA of base drive, has a maximum voltage compliance of just 0.40 volt, then we are out of luck. Ib & Vbe are both needed, but only 1 can be the directly controlled quantity.

Have I explained myself well? Further questions/comments are welcome.

Claude

 

[wbeaty]

Once carriers are emitted from the emitter & base regions, they cross the barrier & recombination/ionization takes place. There is a change in the local E field due to this increased density of carriers. Since Vbe is just the line integral of the E field along the b-e path, a change in Ib/Ie results in a change in the net E field in the depletion zone, which results in a change in Vbe.

Will Studiot agree? I was responding to his earlier message asserting that, when illumination is applied to a phototransistor, Vbe doesn't change.

BTW, my training background is BSEE, the usual 4-year degree. What's yours? (If you're remaining anonymous, no need to give personal info.)

 

[cabraham]

Will Studiot agree? I was responding to his earlier message asserting that, when illumination is applied to a phototransistor, Vbe doesn't change.

BTW, my training background is BSEE, the usual 4-year degree. What's yours? (If you're remaining anonymous, no need to give personal info.)

I am in my final year of the Ph.D. program in EE. I've been practicing EE for 33 yrs. I intended to get the Ph.D. at some point in time, since my mid-20's. But after the MSEE program, I was burned out from school & wanted to make some real money.

My MSEE graduate student pay was pretty meager, & I got an offer to make some real money for a large firm out of state, which I accepted in 1980. I figured then that I would "eventually" go for the Ph.D. It wasn't until 2007 that I returned to grad school to commence work on the Ph.D. The gap between MS & Ph.D was 27 yrs.

I am now making the meager pay of a grad student. The 1st 3 yrs. of my program were done when I was still working full time. But now I take contract jobs on the side, & do the Ph.D. research. I need to get this degree finished & this is the best way.

Maybe I didn't read the whole thread through carefully enough, but my position is that with a phototransistor, photons impart energy to valence band electrons, which transition into the conduction band. After the carriers cross the b-e jcn, recombination & ionization takes place, resulting in a change in the local E field, & consequently, a change in Vbe. So yes, Vbe eventually does change in response to incident light & current change.

Any further questions or comments are welcome.

Claude

 

[wbeaty]

I am in my final year of the Ph.D. program in EE. I've been practicing EE for 33 yrs. I intended to get the Ph.D. at some point in time, since my mid-20's. But after the MSEE program, I was burned out from school & wanted to make some real money.

Same here: 31yrs in EE, but half the time in embedded sw.

Area of specialization/dissertation?

At times I've been tempted by PhD when I can eventually afford it, but in physics of high-Q plasmon effects in non-contact friction and micromachines. There's all of three papers on that topic.

PS sorry for the delay, I've been out of town and checking laptop rarely.

 

[cabraham]

Same here: 31yrs in EE, but half the time in embedded sw.

Area of specialization/dissertation?

At times I've been tempted by PhD when I can eventually afford it, but in physics of high-Q plasmon effects in non-contact friction and micromachines. There's all of three papers on that topic.

PS sorry for the delay, I've been out of town and checking laptop rarely.

My dissertation is tentatively solar enrgy panels & associated power conversion electronics. I'm awaiting the funding to be finalized. As far as affording it goes, I applied in fall 2006 when I was employed by a company that agreed to finance my tuition & book expenses. Then, two thirds of the way through, the company encountered weak sales, & had to make cutbacks. The educational assistance program was hit. I was placed on "temporary furlough" along w/ dozens of others, mostly middle aged engineers, told I would be recalled, but no set date was given. They still plan to recall me. I've worked temp jobs since.

So my employer paid the 1st two thirds of my Ph.D. program, & I paid the final one third. It
s a private school, so the cost is really high. But, it should be worth it. I'd advise you to not put it off. The sooner you get it, the more years in life you will have to use it. BR.

Claude

 

[wbeaty]

Guys, please, can you not just give me another explanations, but read my own and tell me what is wrong there on my picture. So I could use my own understanding as a reference point to move from.

Diagram 1 looks OK, showing the polarity of mobile carriers in the three regions.

Those circled charges... I assume that they're immobile lattice ions of the depletion zones? There must be some better way to show that these ions cannot move and can't participate in currents in the DC case.

Diagram 2 isn't correct. Since the transistor is still turned off at that point, the Base region will still be full of holes as before, and since the EB depletion width doesn't change, the entire device should look almost like #1. Differences: the CB depletion zone is strongly reverse biased and becomes significantly wider, which shrinks the Base region a bit (look up "Early effect.") Also, net surface charge from the power suppl;y should be placed on the left and right ends of the device adjacent to the connecting wires for E and C. These surface charges always are associated with the power supply voltage, and they're the cause of the currents in circuit conductors. Remember that the applied voltage coming from power supplies has little effect upon the charge density inside a wire (little effect upon its resistance.)

For some excellent (and extensive!) info about these surface charges associated with power supply voltage, see this preprint below. It's very worthwhile going through it and understanding every detail.

A unified treatment of electrostatics and circuits, Chabay and Sherwood (1999)
http://www.matterandinteractions.org/Content/Articles/circuit.pdf
​

Diagram 3 isn't correct. The base is p-type and should still contain holes, even though a flood of negative carriers is pouring in from the Emitter. The CB depletion region should be wide because of reverse bias. The EB depletion zone should be narrow, showing the lowered potential barrier because of forward bias.

Do you have "Art of Electronics" by Horowitz and Hill? They go into the voltage controlled BJT, Shockley equation etc., where their transistor section is titled "Ebers-Moll Model Applied to Basic Transistor Circuits, Improved transistor model: transconductance amp." For the normal operation in active region, Ebers-Moll reduces to a transconductance equation: voltage in, current out. That's basically where the name "trans-istor" came from.


PS
It's probably better to start by analyzing interior diagrams of single diodes under forward bias and under reverse bias. Once you work out those concepts completely and draw accurate diagrams, then you can build your transistor diagram.

It's harder to teach yourself about diodes and transistors at the same time. Also, if transistor diagrams aren't GIF animations, they're almost impossible to understand! That's one reason we traditionally hold off on transistor explanations until the students acquire the skills to understand the math models. But one situation is different: science museum exhibits on semiconductor physics. Explaining it ...to my grandmother.

 

[wbeaty]

So my employer paid the 1st two thirds of my Ph.D. program, & I paid the final one third. Its a private school, so the cost is really high. But, it should be worth it. I'd advise you to not put it off. The sooner you get it, the more years in life you will have to use it. BR.

Claude

Whew, that sounds a bit familiar. Our entire eng staff in industrial laser scanners was laid off in 2004 during the first internet bubble. I'm a staff engineer in academia since then, at half the pay of the outside world. Interesting instrument design work, but the majority of time is spent in debugging failed systems and keeping ancient equipment alive.

PhDs aren't required for the work that's on the horizon: three private R&D orgs looking to develop some really radical solar stuff. Start-ups are searching for scientists; for idea-generators who've developed their "serendipity of mind" as in the quote below. The trouble with advanced degrees is that the usual employee pipelines are already choked with PhDs. So don't show them the piece of paper, instead provide copies of your unpublished research, and give them an off the cuff Feynman-esque physics rant. Deliver insights which send them reeling with delusions of ...patentable embodiments! :) Well, actually it's all just fantasy at this point since their funding is still "approaching."

"Many very serious-minded, solid and knowledgeable people work hard in science
all their lives and produce nothing of the smallest importance, while others, few by
comparison and not highly erudite, exhibit a serendipity of mind that enables them to
have valuable ideas in any subject they may choose to take up." - R.A. Lyttleton​

 

[wbeaty]

I actually enjoyed reading Mr. Beaty's site, but there should be a rebuttal to it, or a summary of the epic thread published on the web. I think a lot of confused people would appreciate it. Thanks for sharing.

If speaking for all the confused people ...which parts of http://amasci.com/ele-edu.html are giving the problems?

URL, and snip of specific paragraphs would be best.

 

[cabraham]

If speaking for all the confused people ...which parts of http://amasci.com/ele-edu.html are giving the problems?

URL, and snip of specific paragraphs would be best.

Honestly Bill, the rebuttals to your site have been thoroughly presented. A more fitting question is "which points in the rebuttals need clarification?" Everything I've presented is backed up with known proven science. You need to address what, if any, points I may have overlooked.

Once again, here is the thread where the matter was well discussed:

https://www.physicsforums.com/showthr...ghlight=wbeaty

BR, Claude