EEVblog Electronics Community Forum
EEVblog => EEVblog Specific => Topic started by: EEVblog on May 29, 2015, 11:00:07 am
-
Dave explains how BJT and MOSFET transistors work at the silicon chip level.
How does a BJT transistor actually amplify current?
P and N type doping, charge carriers, conduction channel, field effect, holes and electrons, all the other good stuff.
https://www.youtube.com/watch?v=qUeK7pHe0rI (https://www.youtube.com/watch?v=qUeK7pHe0rI)
-
Great video. There's one point that I was confused about, where you just say that the C-B depletion region shrinks as if it's obvious that would happen. But now the N and P are the other way around, wouldn't the same current flow now expand the depletion region?? But I think I figured it out while writing this post. So I'll leave it as my question and then my own answer, in case it's useful (and to get confirmation it's correct).
Question: You say that the charge carriers flooding from the emitter into the base of the BJT flow through the C-B depletion region and out the collector. But that depletion junction is P on the bottom, N on the top, that's a reverse biased diode! If I replaced the emitter N region with a metal contact, clearly that'd be a reverse biased diode and no current would flow. What's different here?
I think the answer is: A metal contact would flood the base with holes, which (according to how a diode works), would just expand the depletion region even more, and no current would flow. But in the BJT, the emitter is flooding the base with electrons. Perhaps this point could have been highlighted a bit; the P material in the base, which is normally supposed to be full of holes, has completely changed and now has an excess of electrons. And so it's temporarily a NNN device, and in turn that's why the depletion regions are gone.
Is that right?
-
In the mid-60's when I was studying electrical engineering at university, we started with the Ebers-Moll model, which relates emitter current to Vbe and to think of a BJT as a voltage controlled device.
While in many cases - perhaps even most - practical design can consider a BJT as a current controlled device, the variation in HFE from device to device and with temperature can and will bite you in the backside on occasion. If I wish to use a BJT to switch, for example, a relay on and off, it's entirely reasonable to use the worst case HFE, throw in some temperature margin and a safety factor.
-
There will be another voltage-vs-current argument here, and I will be watching for entertainment.
*passes around popcorn* :popcorn:
-
I will just ignore that :-)
I look forward to Dave digs again into his project box and shows us some other of his projects he designed and built. I consider that much more interesting than PN-junction V/I war :-)
-
There will be another voltage-vs-current argument here, and I will be watching for entertainment.
For the folks interested in creating a flame war, just read this entire thread (https://www.eevblog.com/forum/beginners/transitor-the-base-pin/) as many times necessary until your itch is calmed... :scared:
c4757p, sorry for my attempt to burst your bubble... :)
-
Dave, you just hit a troll nerve with saying " current driven" :)
-
Well I never went to college, but I have learned a lot from these videos. Thanks Dave
-
No argument intended about voltage vs. current.
A forward-biased PN junction has a monotonic curve of current vs. voltage, and therefore has a monotonic curve of voltage vs. current.
It is often easier to consider the voltage as a function of current, since the voltage is approximately a logarithmic function that changes much less rapidly with respect to current and the parameters.
-
Thanks for the great video. Makes much more sense that other explains I have seen. I realize thinking back to old project a MOSFET would of been much better suited. I had a microcontroller pin connected to a TIP120 via a resistor of course to turn a 9V RF Camera on and off. Never worked very well.
On a side note, I don't know why anyone would give a thumbs down... and more than one at that.
-
I did go to college, but I only studied Maths and various CS subjects there. So I learn a lot from these videos too. Thanks!
I agree that some parts of the explanation about the depletion region sounded kind of vague. But I can see how it would have needed a 10-fold increase in complexity to get a 1% increase in understanding, so I won't fault Dave for staying below that threshold. ^-^
-
Great video. There's one point that I was confused about, where you just say that the C-B depletion region shrinks as if it's obvious that would happen. But now the N and P are the other way around, wouldn't the same current flow now expand the depletion region?? But I think I figured it out while writing this post. So I'll leave it as my question and then my own answer, in case it's useful (and to get confirmation it's correct).
Question: You say that the charge carriers flooding from the emitter into the base of the BJT flow through the C-B depletion region and out the collector. But that depletion junction is P on the bottom, N on the top, that's a reverse biased diode! If I replaced the emitter N region with a metal contact, clearly that'd be a reverse biased diode and no current would flow. What's different here?
I think the answer is: A metal contact would flood the base with holes, which (according to how a diode works), would just expand the depletion region even more, and no current would flow. But in the BJT, the emitter is flooding the base with electrons. Perhaps this point could have been highlighted a bit; the P material in the base, which is normally supposed to be full of holes, has completely changed and now has an excess of electrons. And so it's temporarily a NNN device, and in turn that's why the depletion regions are gone.
Is that right?
I'm not surprised that you are confused because Dave's explanation is not only wrong, but doesn't even make sense.
Dave says as you forward bias the base-emitter junction the depletion region shrinks and then flips around. This is wrong. The depletion region does shrink, but never shrinks to zero. There is always an emitter-base depletion region and a collector-base depletion region.
Dave implies carriers can't flow through a depletion region. This is wrong as well.
As you pointed out, his explanation of why current flows in the collector doesn't make any sense.
The short explanation is the collector-base is always reverse biased. There is always a depletion region there. As the base is forward biased, electrons flow from the emitter into the base. Because there is a positive voltage on the collector, there is an electric field that sweeps the excess electrons right through the base-collector depletion region and out of the collector. A very few electrons in the base re-combine with holes in the base and this results in a small base current flow.
I love Dave's videos, but if you really want to understand how transistors work, I suggest you take this one for entertainment value only and look at some of the many other sources of information on the internet on how transistors work.
-
I still consider the concept of "hole flow" to be a fraud. Just made up to try and justify the original explorers of electronics incorrectly guessing that current flowed from positive to negative. After science started to actual discover and understand the structure of atoms and how it allows current to flow via electrons in it's valence layer, it become clear to me that the only thing moving were electrons. So holes were invented to support hole flow so they could continue to say that current (in the conventional definition) flows from positive to negative. But sense they still hang on to the fraud of 'conventional flow direction' in most all EE programs, we are still stuck with it.
So when a newcomer to electronics asks what direction does current flow, why must we honestly answer that electrons flow - to +, but then go on to explain why all the semiconductor arrow symbols point against this very flow because of holes? Sounds like pulling out of the burning bush or virgin birth to explain what they didn't originally understand well enough to state what actually moves (flows) and in what direction.
There are just too many holes in the concept of hole flow. ;)
I'll probably be trolled to shut my hole, but there it is. :box:
-
I still consider the concept of "hole flow" to be a fraud. Just made up to try and justify the original explorers of electronics incorrectly guessing that current flowed from positive to negative.
Consider a bubble rising from the bottom of a glass of fizzy drink. Note that it's far easier to think of the bubble moving up, rather than thinking about the water molecules all moving down slightly at just the right time to leave a water void that appears to be moving up. I grant you, not a perfect analogy since the bubble actually contains air, but the point remains valid I think. Similarly, in semiconductors, a "hole" is a "bubble" in the sea of electrons. If you see two snapshots, one with a hole on the left of the page, and the next with a hole on the right of the page, it's far more natural to say that the hole has moved to the right, rather than saying that a whole bunch of electrons have moved left by one atom.
In any case, your claim is completely flawed because semiconductor descriptions rely on simultaneously tracking holes and electrons -- holes and electrons have equal importance in semiconductor theory, they're both invoked together in the description of a BJT, which is hardly what you'd expect if you thought holes were just a historical mistake. If they weren't truly distinct manifestations, why would a single explanation conflate them. If you show me a (hypothetical) picture of a piece of silicon, I can point out the free electrons, and I can point out the holes. Holes are real; distinct entities from electrons.
Conventional current vs electron flow in metals is a completely separate issue, you seem to be conflating the two. If someone were claiming that holes carry current in metals, then I'd agree with you. But no-one does that.
-
I'm not buying it.
in semiconductors, a "hole" is a "bubble" in the sea of electrons.
A hole, as a construct and justification to define flow direction, is simply a single atom that is carrying a positive charge, and this 'hole' will soon be occupied, just as the atom that will fill that 'hole' is simply a negatively charged atom. Bubble in a sea of electrons, indeed weird.
Every hole flow advocate I've run across has to hold their tongue just right, as if tweeking a muti-turn trimmer. ;)
-
Dave, you just hit a troll nerve with saying " current driven" :)
Too bad for them :P
Undeniable fact that BJT's have base current at the application level, and is the basic operational difference between BJT's and FET's.
-
On a side note, I don't know why anyone would give a thumbs down... and more than one at that.
I have serial haters. Thumbs down come in the first few minutes after upload before they even have time to watch it. Happens on every video.
-
That being the basic operational difference is not undeniable. I'd call the basic operational difference the very-close-to-pure exponential characteristic between both VBE and IC at low currents and VCE and IC at high currents (the reason why translinear circuits work so well with BJTs, and the reason IGBTs are used at very high currents). The base current is incidental. MOSFETs are much more complicated in terms of transfer function.
Damn, I told myself I wouldn't get into this.
-
The short explanation is the collector-base is always reverse biased. There is always a depletion region there. As the base is forward biased, electrons flow from the emitter into the base. Because there is a positive voltage on the collector, there is an electric field that sweeps the excess electrons right through the base-collector depletion region and out of the collector.
Yes, correct, my goof. Have clarified this in annotation.
The "flipping" terminology was a poor choice, "overcome" would have been much better.
-
Methinks math-hating, religion-loving Dave has only ever used BJTs as switches, and therefore believes them purely current-operated. :box:
(Fuck, now there's some flamebait... if y'all don't see me again, let it be known why... :-DD :-DD )
Tim
-
Jerry! Jerry! Jerry! Jerry! Hahahahaha.. Or if you are older and remember the Jeraldo show, whatch out for that chair! Mere Griffin and Dick Cavett never did such things. Sorry to all the non yanks who might not get these references.
-
Great video, Something you didn't touch on for MOSFETs is why they are not 'symmetric'. For the BJT it is clear that the collector has the lightly doped region. But from simplified MOSFET drawing you used (excellent 3d davecad modelling there), it appears that source and drain are identical which can be confusing if you take the drawing literally. And maybe bring up why J-FETs can be made symmetric.
-
Great video, Something you didn't touch on for MOSFETs is why they are not 'symmetric'. For the BJT it is clear that the collector has the lightly doped region. But from simplified MOSFET drawing you used (excellent 3d davecad modelling there), it appears that source and drain are identical which can be confusing if you take the drawing literally. And maybe bring up why J-FETs can be made symmetric.
Yes, deliberately left out. It was already getting way longer than I wanted.
-
Yes, deliberately left out. It was already getting way longer than I wanted.
Hey, I'd watch the extended remix where you add the humble vacuum triode at the beginning and insert the JFET in between BJT and MOSFET. :)
-
Great video, Something you didn't touch on for MOSFETs is why they are not 'symmetric'. For the BJT it is clear that the collector has the lightly doped region. But from simplified MOSFET drawing you used (excellent 3d davecad modelling there), it appears that source and drain are identical which can be confusing if you take the drawing literally. And maybe bring up why J-FETs can be made symmetric.
Keep in mind that MOSFETs can be symmetric (http://www.nxp.com/documents/data_sheet/BSS83_N.pdf).
-
Most common example is a CMOS analogue switch, which uses this property to act as a switch. If you use a CD4007 you can connect the one device as an amplifier, and if you connect it with the drain and source reversed it acts exactly the same with identical characteristics, as the substrate connection is a separate connection to a supply rail pin.
-
Yes, deliberately left out. It was already getting way longer than I wanted.
Hey, I'd watch the extended remix where you add the humble vacuum triode at the beginning and insert the JFET in between BJT and MOSFET. :)
Don't forget IGBTs!
-
IGBT is dead easy, you take a very big PNP transistor ( acts pretty much like a NPN one, just substitute holes for electrons in conduction) and use a power MOSFET between base and collector, with the collector being common. Mosfet is a converter from gate voltage to current, and the PNP device is a non saturated switch, so turn on and turn off can be pretty fast. Drawback is the non saturated switching, you always will have 2V across the device when on. Advatage is high current capability and high voltage ability, though not both at the same time, power dissipation ability in linear mode is very poor, and you really want to switch only with it.
-
In the mid-60's when I was studying electrical engineering at university, we started with the Ebers-Moll model, which relates emitter current to Vbe and to think of a BJT as a voltage controlled device.
I'm a retired lawyer and I specialized in patent litigation. I love the intersection of science and law. Most of my cases involved semiconductor design and fabrication. (You can tell by my forum username.) I met John Moll. He came to court to serve as an expert witness. He actually showed up wearing farmer jeans and a flannel shirt. He was the nicest, most unassuming guy. He passed away in 2011.
Explaining semiconductor device physics to a jury and judge with no EE background was the biggest challenge. I would try to use analogies to things in everyday life. One case involved the ultimate shape of a P-N junction in a power MOSFET after implant and diffusion of dopants through a mask into the epi region (specifically, the shape of the base region in a planar power MOSFET when viewed from above the substrate). The longer the time and higher the temperature of diffusion, the more rounded the junction. I used an analogy of spraying ink on a stencil and seeing that the outline of the resulting image wasn't exactly like the stencil opening, but was more rounded at the corners.
I really enjoy Dave's tutorials. I could have used him as an expert. He would have gotten a free trip to the states.
-
Taking a seemingly difficult subject and explaining it as simply as possible is the start of understanding it. Great idea this Fundamental Friday.
-
Most common example is a CMOS analogue switch, which uses this property to act as a switch. If you use a CD4007 you can connect the one device as an amplifier, and if you connect it with the drain and source reversed it acts exactly the same with identical characteristics, as the substrate connection is a separate connection to a supply rail pin.
Really interesting, I have used 4000 series switches before but I guess I never thought about it much and assumed they had a way to integrate JFETs. symmetric CMOS makes a lot more sense from a process standpoint.
-
I'm a retired lawyer and [...]
Great!
I would like to sue semiconductors for the shameless infringement of Ohms law. :scared:
Can you help me with that?
-
On a side note, I don't know why anyone would give a thumbs down... and more than one at that.
I have serial haters. Thumbs down come in the first few minutes after upload before they even have time to watch it. Happens on every video.
apply kirchoff and convert them to parallel haters ...
-
IGBT is dead easy, you take a very big PNP transistor ( acts pretty much like a NPN one, just substitute holes for electrons in conduction) and use a power MOSFET between base and collector, with the collector being common. Mosfet is a converter from gate voltage to current, and the PNP device is a non saturated switch, so turn on and turn off can be pretty fast. Drawback is the non saturated switching, you always will have 2V across the device when on. Advatage is high current capability and high voltage ability, though not both at the same time, power dissipation ability in linear mode is very poor, and you really want to switch only with it.
What about the upcoming high voltage HEMTs? From what I read, they can be thought of as very fast and very low loss MOSFETs, but exactly what makes them work so well?
-
I'm a retired lawyer and [...]
Great!
I would like to sue semiconductors for the shameless infringement of Ohms law. :scared:
Can you help me with that?
You need to re-read that third word: RETIRED!! No more arguing for me. But I might reconsider if your name was Georg Ohm. :box:
-
IGBT is dead easy, you take a very big PNP transistor ( acts pretty much like a NPN one, just substitute holes for electrons in conduction) and use a power MOSFET between base and collector, with the collector being common. Mosfet is a converter from gate voltage to current, and the PNP device is a non saturated switch, so turn on and turn off can be pretty fast. Drawback is the non saturated switching, you always will have 2V across the device when on. Advatage is high current capability and high voltage ability, though not both at the same time, power dissipation ability in linear mode is very poor, and you really want to switch only with it.
What about the upcoming high voltage HEMTs? From what I read, they can be thought of as very fast and very low loss MOSFETs, but exactly what makes them work so well?
An exercise for the student..........
Not familiar with them, must do some research some time on them and see.
-
Thank you very much for the lesson. I'm a mathematician, not a physicist, and I'm starting with semiconductor physics for fun (going through this magnificent online reference here (http://ecee.colorado.edu/~bart/book/book/contents.htm)).
Your explanation has made clear for me the workings of a BJT (I didn't reach that part in the course yet), though there seem to be some holes :) . If I understood well, the key to amplification is as follows. The BE current modulates the width of the BE junction: the greater the current, the narrower the depletion region. The narrower that region, the more electrons diffuse through it. Now, due to the thinness of the base, many of these electrons will also cross the BC depletion region and be captured by the strong field in the weakly n-doped layer, therefore being sent to the collector. Really interesting.
I understand why there may be flame wars about current or voltage driving the transistor. It is the base current which determines the BE depletion width, but that width is a function of the BE voltage. So the casual relationship is a bit fuzzy.
Again, thank you very much for the enlightening lesson.
-
What about the upcoming high voltage HEMTs? From what I read, they can be thought of as very fast and very low loss MOSFETs, but exactly what makes them work so well?
HEMTs and HBTs work by manipulating the band gap (the energy levels, and populations, of conduction electrons and holes) of a semiconductor by varying the composition of the semiconductor, into materials which are compatible (chemically and in crystal structure), but have different band gaps.
The basic idea is something like: the average level of the band gaps must meet (for continuity reasons: their electrons all have to be at the same thermal energy, defined by the Fermi level). The top (conduction band) and bottom (valence band) of the band gap must vary, because the material's band gap is varying with position. This can lead to some interesting consequences. By stacking layers appropriately, the conduction band can be made to dip down below the Fermi level. This causes the region to become metallic, i.e., permanently occupied by a 'gas' of free conduction electrons.
The capper is: by varying the electric field imposed on that region (by placing an insulator or depletion region right on top of the 'gas' layer), the band potentials can be "bent" up or down, enhancing or completely eliminating the 'gas' layer. The result: low resistivity (when on), high transconductance (it doesn't take much voltage to affect a large change in current), and low capacitance (it doesn't take much area to achieve a required conductance). This is a HEMT. The "High electron mobility" refers to the electron gas (which, because it is confined to a layer between different semiconductor compositions, acts planar, i.e., the electrons are excluded from occupying the volume above or below the plane, and therefore is called a 2-dimensional electron gas (2DEG)), which has significantly higher mobility than the ordinary diffusion mobility in the material. Which goes directly proportionally with Rds(on), for a given size of device.
The HBT (heterojunction BJT) similarly plays a band-structure trick. In short, rather than merely letting charge carriers wander randomly through the base region, eventually finding their way into the collector: the base is formed with a gradient of semiconductor materials, which creates a built-in potential (independent of, and additional to, the E/C depletion regions), which practically vacuums them up. Result: much higher hFE and current density. Typical materials are SiGe:C (i.e., silicon substrate, germanium gradient, carbon doping). One practical consequence is very high Early effect voltage (practically infinite), which makes these high-performance transistors excellent for instrumentation amplifiers (high compliance current sources, bootstrap followers, etc.) as well as RF amplifiers.
Tim
-
Bipolar junctions, unijunctions, junction gate FETs, MOSFET junctions, heterojunctions...
Roll on the "perfect transistor (http://physicsworld.com/cws/article/news/2010/mar/01/junctionless-transistor-makes-its-debut)" which doesn't have any junctions.
-
I'm a retired lawyer and [...]
Great!
I would like to sue semiconductors for the shameless infringement of Ohms law. :scared:
Can you help me with that?
go after the esaki diode ! ( called tunnel diode) that thing blatantly showes the laws of physics up nature's nose ... it does have negative resistance...
-
Consider a bubble rising from the bottom of a glass of fizzy drink.
I might, but they don't always rise
http://arxiv.org/pdf/1205.5233v1.pdf (http://arxiv.org/pdf/1205.5233v1.pdf)
-
go after the esaki diode ! ( called tunnel diode) that thing blatantly showes the laws of physics up nature's nose ... it does have negative differential resistance...
FTFY ;)
-
What about the upcoming high voltage HEMTs? From what I read, they can be thought of as very fast and very low loss MOSFETs, but exactly what makes them work so well?
I assume you are referring to GaN (Gallium Nitride) HEMTs. They are in production now, mostly for RF applications, but also for power management devices. CREE is a big GaN manufacturer. GaN HEMTs have high electron mobility which means lower resistance for the same size, so relatively smaller capacitance which means higher frequency operation. GaN can have a high breakdown voltage and has excellent thermal conductivity.
By the way, HEMTs are nothing new. It was 30 years ago in 1985 that Gould Electronics announced the first production HEMT (I worked there at the time).
These so called hetero-junction devices are typically made with exotic material structures from Gallium-Arsenide or Gallium-Nitride among others.
-
I thought that Dave’s explanation was fine up to the moment where he skipped over the important bit ie. how the gain is actually generated. So here’s the explanation I was given as an apprentoid back in the sixties.
Starting from the basis that we have an NPN sandwich of doped silicon with a depletion layer at each junction this explanation assumes what we all really know ie that electrons behave like tiny negatively charged billiard balls. If you think that electrons behave like waves or probability fields then you might as well stop reading now.
Applying a positive voltage to the collector reverse biases its PN junction and widens its depletion layer. Hence the base-collector junction drops all the applied voltage and consequently has a high voltage field across it. The base-emitter junction doesn’t see the collector voltage at all.
Applying a positive voltage to the base forward biases the base-emitter junction, reduces the width of its depletion region and causes a current to flow ie electrons flow from the emitter into the base region. The idea that it starts to conduct at 0.6V or any other voltage is a purely practical approximation. Just like a diode, current flows at all forward voltages right down to very small values but as the current is exponentially related to the voltage it only become significant at a few tenths of a volt.
This base-emitter current then tries to flow through the base region to the base terminal. It doesn’t matter whether you consider it to be holes flowing towards the emitter or electrons flowing towards the base, the actual charge carriers are electrons moving between silicon atoms.
Here’s the important bit, the base region is extremely thin, you wouldn’t believe how incredibly thin it is (oops sorry, slipped into THHGTTG mode there) in fact it’s only a few tens of atoms across (Edit - This is WRONG, see next post). At this point you might imagine an orderly stream of electrons marching in single file through the base region but it’s actually nothing like that. The silicon atoms are held in a crystal matrix but are violently vibrating due to thermal energy and in the process the electrons are being bounced around all over the shop and hence tend to spread out sideways. This is called diffusion. At room temperatures the diffusion path length is larger than the width of the base region so the electrons tend to stray into the two depletion layers. Any electrons (or rice crispies) that enter the base-collector depletion layer are whipped away by the strong electric field (remember that?) and become collector current. Because the base region is so thin compared to its length very few of the electrons make it to the base terminal and so a small base current supports a large collector current. QED – gain.
This behaviour explains why the gain is positively dependent on temperature and why gain falls off at low collector voltage when the voltage field is insufficient to prevent some of the electrons from being bounced back into the base region. The thinner the base region the higher the gain but it cannot ever be infinite because there must be some base conduction to maintain a voltage across the base-emitter junction.
Perhaps I should mention that in a forty year career as a circuit designer I have never actually needed to know any of this and it might be completely wrong, it’s just what I was told.
-
Bipolar junctions, unijunctions, junction gate FETs, MOSFET junctions, heterojunctions...
Roll on the "perfect transistor (http://physicsworld.com/cws/article/news/2010/mar/01/junctionless-transistor-makes-its-debut)" which doesn't have any junctions.
I think they are making those now, as FinFETs and GAAFETs. Cool devices, by the way, especially the multi-gate implementations.
-
Here’s the important bit, the base region is extremely thin, you wouldn’t believe how incredibly thin it is (oops sorry, slipped into HHGTTG mode there) in fact it’s only a few tens of atoms across.
No, not nearly so! You're off by several orders of magnitude.
The critical measurement is the diffusion length in silicon. This varies by temperature (for obvious reasons?) and more strongly by material.
The distance a charge carrier can travel is limited by Brownian motion and recombination time. In silicon, it's a few microns. In germanium, it's tens of microns -- which is why BJT were first developed with germanium. (Ge is also less sensitive to impurities, because the intrinsic carrier density at room temperature is equivalent to something like a ~0.001% doping of random P/N junk. Or something like that. So it was doubly easier to work with!)
A base region thicker/wider than the diffusion length obviously won't have many charge carriers left to diffuse into the collector junction, even if it starts out saturated by lots of forward bias on the emitter.
So, practical (usually diffused, then later, epitaxy as well) junctions are in the single to fractional micron range.
I don't think BJTs are ever much thinner than that; the doping at least would have to be much higher, leading to lower voltage ratings, and Early effect would be pretty significant (i.e., base junction thinning due to the size of the collector depletion region).
It's worth noting that the channel conduction region in modern CMOS is about that level -- atoms (single nm) scale, that is. The doping levels are very high, which makes high conductivity, high shielding effect (i.e., the channel region that appears beneath the gate is very thin), and high leakage (relative to the size) and low voltage tolerance (maybe 2V breakdown!).
Also for related info: early (60s-70s) IC processes had resolution on the order of one or a few microns. So, one could draw some closely spaced lines, and apply P-N-P doping along them, to produce a lateral (current flow is sideways, not depthwise) transistor. These were symmetrical and had the voltage rating of most collectors (i.e., 30V), hence the high differential input voltage range of pretty much every classic analog circuit (uA741, LM339, etc.). hFE was pitiful (because of the wide base), maybe 5 at the most. But that was still good enough for differential input pairs (which is why these devices almost always have negative input bias current) and current mirrors (the accuracy was poor, but just to get any bias was good enough).
This behaviour explains why the gain is positively dependent on temperature and why gain falls off at low collector voltage when the voltage field is insufficient to prevent some of the electrons from being bounced back into the base region. The thinner the base region the higher the gain but it cannot ever be infinite because there must be some base conduction to maintain a voltage across the base-emitter junction.
The rest is correct, as far as I know. So for not having to use it, you remembered it awfully well, I guess :)
Some consequences:
- Superbeta transistors (hFE > 1000) require thin base layers
- So they should have terrible Early effect, and may even achieve punch-through rather than avalanche breakdown (i.e., the base thins so much that hFE effectively becomes infinite, amplifying its own leakage current into what looks like avalanche current; alternately, the base thins so much that it ceases to exist (punch-through), and collector and emitter join together, effectively shorting out the device at that terminal voltage). So one should also expect low Vceo ratings.
Which as far as I know, are true, so it's good physics to know. :)
Tim
-
Here’s the important bit, the base region is extremely thin, you wouldn’t believe how incredibly thin it is (oops sorry, slipped into HHGTTG mode there) in fact it’s only a few tens of atoms across.
No, not nearly so! You're off by several orders of magnitude.
Thanks for the correction T3sl4co1l (catchy name by the way) my memory has obviously substituted "tens" for "thousands" or maybe "tens of thousands". What's a few orders of magnitude between friends eh? I'm impressed that I could remember any of the explanation at all considering that I can barely remember what I had for breakfast today.
-
Thousands or tens of thousands would do (an atom is on the order of 0.2nm, so that would be ~um). Simple omission. :)
Cheers,
TIm
-
Taking a seemingly difficult subject and explaining it as simply as possible is the start of understanding it. Great idea this Fundamental Friday.
You must be joking, this is a simple explanation (fast forward to 1:35 http://youtu.be/R_VlWQa0lpc?t=1m35s (http://youtu.be/R_VlWQa0lpc?t=1m35s)):
https://www.youtube.com/watch?v=R_VlWQa0lpc&t=1m35s (https://www.youtube.com/watch?v=R_VlWQa0lpc&t=1m35s)
-
Taking a seemingly difficult subject and explaining it as simply as possible is the start of understanding it. Great idea this Fundamental Friday.
You must be joking, this is a simple explanation (fast forward to 1:35 http://youtu.be/R_VlWQa0lpc?t=1m35s (http://youtu.be/R_VlWQa0lpc?t=1m35s)):
https://www.youtube.com/watch?v=R_VlWQa0lpc&t=1m35s (https://www.youtube.com/watch?v=R_VlWQa0lpc&t=1m35s)
Remember how I said the resistor gives you the same amount of voltage on the other side, just lower current? It just lowers and limits the amperage that can go through it?
Stopped there.
-
Remember how I said the resistor gives you the same amount of voltage on the other side, just lower current? It just lowers and limits the amperage that can go through it?
Stopped there.
Good move, you saved yourself from hearing "A transistor is just a resistor. A transistor is just a resistor. A transistor is just a resistor." At this point I pinchoff-ed the video :)
-
I thought that Dave’s explanation was fine up to the moment where he skipped over the important bit ie. how the gain is actually generated.
Agreed.
It all went a bit wishy washy as in 'bloke from the pub' once the BJT operation was described in detail with a few errors and (as you say) he managed to miss the most important bit of all.
I only watched it up to the end of the section on the BJT.
What I found surprising was how some people commented that it was the best video/tutorial on this subject they had ever seen. However, I suspect that Dave could show a video of himself describing a turd on a string and still get high ratings. i.e. the best turd on a string video they had ever seen :)
-
Remember how I said the resistor gives you the same amount of voltage on the other side, just lower current? It just lowers and limits the amperage that can go through it?
Stopped there.
Good move, you saved yourself from hearing "A transistor is just a resistor. A transistor is just a resistor. A transistor is just a resistor." At this point I pinchoff-ed the video :)
Im sure both of you love hearing about moving holes and doping, but reality is novices do not understand any of it. Resistor is easy to understand, comparing transistor to resistor you can control is brilliant.
-
I think there is one thing to note regarding FETs, that is misleadingly (especially for newbies) taught a lot in literature:
FET's in most publications are described as voltage controlled, but this is only correct at the physical level of the conductive channel itself.
At the application level keep in mind you have to charge/discharge a capacitor (the gate capacitance) to get the needed voltage to build up an electric field to change the conductivity of the channel. Keep in mind you can not simply switch off an enhancement FET by switching off the control voltage, you have to actively discharge the gate capacitance to turn it off.
In low frequency applications the term of a "voltage controlled" transistor may be approximately correct (I nevertheless think this is misleading), but in high-frequency switching-applications with several 10s ore 100s of kHz a lot of driving current may be needed to change the charge of the gate capacitance.
I know about a former fellow student who burned his power-MOSFET circuit several times because he relied on the definition of a FET being a voltage controlled device.
So in fact FETs are not voltage controlled, they are charge controlled.
Best regards,
Andreas
-
Remember how I said the resistor gives you the same amount of voltage on the other side, just lower current? It just lowers and limits the amperage that can go through it?
Stopped there.
Good move, you saved yourself from hearing "A transistor is just a resistor. A transistor is just a resistor. A transistor is just a resistor." At this point I pinchoff-ed the video :)
Im sure both of you love hearing about moving holes and doping, but reality is novices do not understand any of it. Resistor is easy to understand, comparing transistor to resistor you can control is brilliant.
To be fair to the original inventors of the transistor, it's very name gets to a simple construct that even beginners should understand without having to snow them or make up strange analogies or inventing 'bubbles'.
The transistor name is a contraction of 'transfer of resistance'. That is by controlling the resistance of the BE junction one can control the resistance of the CE current path. Building from that start should make for a pretty simple to grasp explanation of how a transistor works. No need to get into the chemistry of doping material or junction construction.
-
but in high-frequency switching-applications with several 10s ore 100s of kHz a lot of driving current may be needed to change the charge of the gate capacitance.
So in fact FETs are not voltage controlled, they are charge controlled.
Andiz is absolutely correct. For designers of switch mode power supplies, a major consideration is getting enough drive current into the main switching MOSFET in order to turn it on and off sufficiently fast. This is why MOSFET drivers exist and why switch mode supply ICs have output current capabilities in the Amps range when it only takes microAmps at dc. Gate charge is a parameter on the data sheet for a good reason.
-
What's more, BJTs are also charge controlled, at least if you want to use them at the peak of speed. (Calling out "charge" versus "voltage" is a fairly useless tweak, because to put charge on the terminal, you must change the terminal voltage -- driving it from a resistor divider, for example, doesn't count, for the reason that, you aren't changing the voltage on the input node itself quickly, due to its capacitance.)
The most easily forgotten quirk of BJTs is that they store charge, exactly like a battery stores charge. Just extremely faster (microseconds, not hours), and faster self-discharge (tens of microseconds). The discharge curve even looks like a battery: if you fully charge the B-E junction of a BJT, then let go, it'll quickly drop down from the 0.7-0.9V it had in forward bias (or forward charge, if you will), then take a long time passing the 0.5-0.7V range (where most of the charge is stored -- exponentials being what they are), then collapse in the <0.5V range. At the same time, the collector voltage doesn't budge until the 'collapse' phase. A B-E resistor, or anything to sink current from the junction ("clearing stored charge"), quickens turn-off substantially.
Tim
-
Remember how I said the resistor gives you the same amount of voltage on the other side, just lower current? It just lowers and limits the amperage that can go through it?
Stopped there.
Good move, you saved yourself from hearing "A transistor is just a resistor. A transistor is just a resistor. A transistor is just a resistor." At this point I pinchoff-ed the video :)
Im sure both of you love hearing about moving holes and doping, but reality is novices do not understand any of it. Resistor is easy to understand, comparing transistor to resistor you can control is brilliant.
The sentence I quoted is wrong, not just oversimplified.
-
I very much agree with the term charge controlled for both MOSFETs and BJTs. The difference is that in a BJT the charge is intrinsically decaying (one the µs time scale) and leaking through the forward biased BE junction.
I thing there were quite a lot of mistakes in Daves explaination of the BJT. Teh One from number33 is much closer to reality.
To me the simple picture of the NPN BJT is, that the emitter is injecting electrons to the base. Due to lower doping of the base, only few holes are injected to the emitter. The base to collector junction can be very much seen like a photodiode. Electrons in the base, close to the junction are collected by the collector and allow current flow. So the electron injected by the emitter act very much like electrons generated by light. Just like the quatum efficiency of photodiodes can be rather close to 1, most electrons end up at the collector and not the base. This picture also explains why the collector - Emitter saturation voltage can be smaller than 0.6 V. The low doped layer at the collector is more of a detail to get a resonable high voltage rating - so the simple picture may work without it.
The one thing that this simplified model does not explain is the function of a photo-transistor however.
-
I agree number33's explanation is very good. Especially because he mentions that the carriers in the base move into the collector junction because of diffusion and are swept through because of the electric field (drift). It might have been easier for Dave to start out with a video about "How do diodes work", and talk about drift and diffusion and recombination. Not very difficult concepts. After covering PN junctions, the bipolar transistor would have been easy.
The collector-emitter saturation voltage is just Vbe - Vbc, where both junctions are forward biased when the transistor saturates.
-
I agree number33's explanation is very good. Especially because he mentions that the carriers in the base move into the collector junction because of diffusion and are swept through because of the electric field (drift). It might have been easier for Dave to start out with a video about "How do diodes work", and talk about drift and diffusion and recombination. Not very difficult concepts. After covering PN junctions, the bipolar transistor would have been easy.
I agree. As a student I was first taught how a diode works, how to model and predict its behaviour and then how to apply this to the BJT when configured in common base with the base terminal grounded. The similarity in behaviour to a diode here is easy to show and to model in terms of input impedance and emitter current vs Vbe etc. Also the transistor 'alpha' in common base mode is easier to explain and understand in terms of how nearly all of the electrons flow from emitter through the base region and then get swept into the collector. The potential benefits this offers in terms of signal amplification are easy to demonstrate at this point. The transconductance of the BJT is easier to explain in common base. But that's just my opinion of course :)
-
Thank you very much for the lesson. I'm a mathematician, not a physicist, and I'm starting with semiconductor physics for fun (going through this magnificent online reference here).
http://ecee.colorado.edu/~bart/book/book/chapter1/ch1_2.htm#1_2_1 (http://ecee.colorado.edu/~bart/book/book/chapter1/ch1_2.htm#1_2_1)
Great now I can start understanding the formulas in the Big Bang Theory tv show.
Now for a serious question - does anyone get a headache reading this? (I do after about 20 minutes) I do not think there are any pills that can help (seriously). I was got a stack of Science magazines from the library and I tried really really hard to read them but always ended up with bad headaches. So I was wondering if it is just me or does someone else have this problem?
-
Now for a serious question - does anyone get a headache reading this? (I do after about 20 minutes) I do not think there are any pills that can help (seriously). I was got a stack of Science magazines from the library and I tried really really hard to read them but always ended up with bad headaches. So I was wondering if it is just me or does someone else have this problem?
Doesn't happen to me normally. Maybe you try to cover too much material too fast. Unless I'm under pressure, I tend to digest physics material like that quite slowly, repeating calculations and derivations in a notebook, and referencing results I'm not sure about to textbooks. If I'm interested, I can keep doing that for hours without pause. That way, I eventually form a picture of the subject that makes sense to me. For example, the section you linked is a very quick review of basic quantum mechanics (first course in a physics degree); you can go through that very quickly if you already know it. Otherwise, trying to assimilate these contents is better done with a good textbook and liberal doses of free time.
-
I still consider the concept of "hole flow" to be a fraud. Just made up to try and justify the original explorers of electronics incorrectly guessing that current flowed from positive to negative. After science started to actual discover and understand the structure of atoms and how it allows current to flow via electrons in it's valence layer, it become clear to me that the only thing moving were electrons. So holes were invented to support hole flow so they could continue to say that current (in the conventional definition) flows from positive to negative. But sense they still hang on to the fraud of 'conventional flow direction' in most all EE programs, we are still stuck with it.
So when a newcomer to electronics asks what direction does current flow, why must we honestly answer that electrons flow - to +, but then go on to explain why all the semiconductor arrow symbols point against this very flow because of holes? Sounds like pulling out of the burning bush or virgin birth to explain what they didn't originally understand well enough to state what actually moves (flows) and in what direction.
There are just too many holes in the concept of hole flow. ;)
I'll probably be trolled to shut my hole, but there it is. :box:
The "holes" are like quasiparticles. They exist in the sense that they behave as though there really was a positive particle there with the same properties as a hole. If you don't talk about holes flowing, then instead you're forced to talk about all of the other electrons surrounding the hole moving the other way, so that the absence of an electron in effect moves. They are 100% equivalent viewpoints, but it's much simpler and more intuitive to talk about the "hole".
It's like when you talk about the group velocity of a wave. Well, there's only one real velocity, and that's the velocity of some individual particle, but it makes sense to treat a waveform as it's own thing, even though it only exists as a logical construct.
But really, it's not any different than how we use the word "hole" in everyday life. For example, if you need to bolt a few things to each other, it's frustrating, isn't it? Somehow, you have to get this bolt to line up with all of these different holes, all at the same time. What do you think to yourself? "I have to get all of these holes lined up", and you move the holes around to line them up. Well, you could also think to yourself "I need to move all of this metal out of the way to find a place where there's no metal", but no one would ever think that. It's equivalent, but it's much simpler to think in terms of moving the hole. It's exactly the same thing with semiconductors. If the hole moves, the surrounding electrons move...if the surrounding electrons move, the holes move. You pick whichever model is easiest for the task at hand.