Yay, AOA! They're about the only one who gets transistors right. Horowitz, Hill, and also Lanterman. (Also go find yourself a copy of their Lab Manual. The authors really let "current control" have it, pulling out all the stops. If you deeply believe in holy hfe, then your analog designs will all fail ...because slight temperature changes ruin the bias calculations, and because no two BJT transistors have similar hfe, even if the part numbers are the same.) These same issues came up to bite Win Hill in his first EE job. Exactly the same thing happened to me! I had to abandon hfe and current-control, and instead re-teach myself the BJT physics. I ended up using longtail-pairs and cascode level-shifters, both which are voltage-input circuits, and cannot be explained in terms of current input. Also Wikipedia had all the current-control garbage on their BJT transistor page. I and others set them straight, but then it all got reverted. Then, WIN HILL HIMSELF came storming in, and carved those "experts" a new a-hole. The BJT page on Wikipedia then had all the voltage-control material. (I haven't checked it recently, so I wouldn't be surprised if the ignorant masses had deleted it all again, years later.) And finally ...just tell students that there's no mechanism where one current can directly affect another. In the BJT, we can **pretend** that Ib affects Ic ...but in reality, Ib can be used to adjust Vbe, and then Vbe directly controls Ie. The physics says that BJTs are voltage-based devices, "transimpedance-istors." Same as FETs. But with BJTs, the depletion-zone extends across the current path, and varies the current by altering the DZ thickness. Fets are different in that the DZ invades from the side. If FETs are like an optical iris with variable aperture, then BJTs are like sunglasses with variable tint! PS We can prove that MOSFETS are controlled by current. After all, if we vary the nanoamps on the gate, it controls the drain current! (This is actually true.) Well sure, but then everyone complains that the MOSFET physics shows that it's controlled by voltage. GOTCHA ...because the BJT physics shows that Ie is controlled by Vbe, and that base-current plays no role in this, except as a leakage-current, same thing as FET gate-leakage.
Go read Streetman if you want to understand how base current can control collector current. Mind you, it's not a general electronic circuits textbook, but it's a semiconductor devices textbook.
That's one of the questions on the German amateur radio license test. Unfortunately I had to choose the BJTs are current-controlled devices answer to pass. That's how they break ya 😂
Doug Self, one of the "gurus" of analog audio amplifiers, also really emphasized that the BJT is fundamentally a voltage controlled current source, and made almost the exact same argument as you, and the Art of Electronics do---pretty good company overall.
I am SO glad somebody has finally addressed the current control myth that is in almost every electronics book! I taught my students the facts about the BJT's in terms of the governing (voltage) equation and explained that the need for base current is just a side effect of the operation, not the "transistor law". The fact that HFE varies so much should give a clue to the falsity of current control argument
Having spent a lot of time characterising exponential-voltage-to-current designs for music VCOs recently, I'm always quite amazed at how accurately a real life circuit follows the Ebers-Moll equation if you compensate the circuit for the bulk emitter resistance of the bjt. Another nice video Aaron!
Thanks for addressing this subject. Another great resource, one that focuses on radio frequency circuits, is Communication Circuits Analysis and Design by Clark & Hess 2nd ed. 1978. The book begins with a great explanation of the BJT as a voltage driven device. This simplifies the design of RF oscillators, mixers and much more.
You have probably heard it a million times already professor Lantermann but you're sense of humour is really high class. Everytime after i watched one of your videos i have to sit in silence for a minute to take in your "ironic/sarcastic/comical/silly/superintelligent" one-liners and jokes you weave in. I appreciate and respect that a lot.
when i saw the thumbnail i had to watch this. My "electronic devices" teacher drilled this into our heads of how and external electric field across base-emitter modulates the internal "built in" voltage fighting the current diffusion across the depletion region, hence controlling he emitter-collector current .
I studied BJTs a million years ago using Millman and Halkias as the text. We studied both the current model (Ic = Beta*Ib) and the voltage model (Ic = gm*Vpi).
The irony is the famous "Transistor Man" drawing early in Chapter 2 of the Art of Electronics showing a man controlling collector current by looking at a meter measuring base current! Of course it's not a horrible approximation, but it IS an approximation, and the Ebers-Moll model is introduced later in the chapter. But the "Transistor Man" model is likely what most people remember.
You should think of the base current as an unwanted side effect of the transistor. When designing a circuit around Ib and hfe it will be very difficult to reproduce it expecting it to have the same behavior even with the same type of transistor. Also because hfe is depend on the collector current, called beta droop, this means driving driving an amplifier with current will result in a lot of distortion bucause of the non linear behavior.
Nice video, I'm a hobbyist myself. I have the book, the art of electronics, but also, learn the art of electronics. And that doesn't give a bad explanation on this subject either and keeps it nice and practical. It's not that your videos aren't practical, it's just another explanation on the subject
I think the best example of the fact is how one can hugely speed up closing a bjt switch by forcing the base-emitter voltage to zero or slightly negative. (it otherwise takes several microseconds for the collector current to cease flowing after stopping the base current)
This can as well be the best example of the need for a base current: you need to remove charge from the base and the best way is to... Make it flow out (with due sign convention).
Great video! Although I think we have to be a bit careful between differentiating modelling for engineering and physics for design. Although several models allow us to behaviourally accelerate how to make a useful system, a tradeoff has to be made between hiding detail for ease.
I have two "art of electronics" books and I see these books the best for actual design work after you have learned the fundamentals. I'm using Donald A Newman's book and Sedra/Smith's books for a more detailed dive into the operation of the devices.
I recently talked with the guy from SSI over the SSI2164 and he said it requires a voltage, not a current, but i only saw transistors in the simplified schematic. A while back i played with the TB-303 vco, where 2 transistors act like a comparator. it clearly shows voltage play an important role (if the voltage on base is higher than the emitter).
@@Lantertronics Ya, but it's weird when you look at the schematic in the datasheet and only see transistors. to remove it's exponential property you need to make a anti log circuit as the SSI man told me. so the SSI21664 can be used as VCO with REAL linear input on the control pin put aside the normal voltage. i did a research on Anti Log circuits, but for now it's a bit unclear for me, i tried simulating these circuits found and found nothing special or the simulator isn't up to it.. Would you give it a shot on the Anti Log circuit to be used on the SSI2164 control pin? i have made a cool VCO circuit and it works, accept it's useless to build if it doesn't follow 1v/oct.
Personally I've always felt like what people are really asking when they talk about current controlled or voltage controlled devices is which parameter has the more linear relationship with power through the device, not necessarily which parameter can be used to most neatly model a devices characteristics. Generally this is because that's how they've learned to think of two terminal devices and they are trying to extend that understanding to transistors. For example a diode is often described as a current controlled device because if you were to linearly vary the power through a diode you would see that the current through it would seem to change almost linearly while the voltage would seem fairly constant. Is it true that the power through a diode is controlled with it's current and not it's voltage though? Of course not. Your thumbnail for this video is the Shockley diode equation demonstrating that after all, but as the old saying goes, "All models are wrong, some are useful." The idea that a component is either voltage controlled or current controlled is a useful simplification for someone just learning about the basic principles of circuits. That simplified model which sees every component as having a linear relationship with either voltage or current can quickly outlive its usefulness though.
Indeed, I usually see people support the current-controlled viewpoint by saying because it's linear it's easier to think about. I think that just winds up just transferring the complexity elsewhere...
@@Lantertronics Yep, because some black-box models are more convenient. But if we actually believe a black-box model is true, that's called Reification. Looking at the transistor physics, we find no mechanism where the base current can somehow alter the collector current. One current doesn't affect another. Transistors literally cannot be current-controlled. Then, to explain the APPARENT current-control, we just introduce Vbe and e-fields of the junction. In AOA, the "little transistor man" who produces the "current gain," he's the EB junction voltage. He measures the base current, then proportionally adjusts the emitter current. (But AOA authors forgot to mention that "transistor man" is just the Vbe potential! Base current and collector current are communicating via changes in Vbe. If we force Vbe value to be fixed, then current-gain disappears.)
It reminds me of, "all models are wrong, but some models are useful." I started considering BJT's as voltage controlled shortly after I started working with differential pairs. Am I right in saying that, if you apply a current from a voltage to a BJT's base (i.e. through a resistor, either on the base or the emitter) the voltage on the BE junction will be the log of the input voltage. However the output current (and therefore output voltage across the collector resistor) will be proportional to the anti-log of the BE voltage. I.e. the log/anti-log functions cancel each other out (with other influences such as temperature, etc) ?
So... is the main significance of the base current that it needs to have a DC path to ground provided for it, and so must be considered but shouldn't figure in the gain calculations? I've been doing some transistor matching, and it seems from reading the experience of others that it's more worthwhile matching the Vbe rather than hfe. I've been measuring it by connecting the DUT as a diode in a circuit that pulls a constant(ish) 100uA through it while I measure the voltage drop across it. Do you agree about the Vbe, and can you suggest the right way to measure it please?
I"m not sure I'm following what you're trying to say in the first paragraph. For BJT or enhancement mode MOSFET you'll generally have connections to the DC supply rails somewhere to provide bias. Regarding matching, google: Ian Fritz transistor matching. Ian is a very clever guy.
@@Lantertronics I too was reading AoE, and it was something in the differential amplifier section of the BJT chapter.... hang on, thanks to having a digital as well as a print version I can cut & paste it: "Be sure to remember that this amplifier, like all transistor amplifiers, must have a dc bias path to the bases. If the input is capacitively coupled, for instance, you must have base resistors to ground."
I think the current-controlled description is useful for comparing from FETs. FETs only need a voltage and draw almost no current (very high input impedance). You can put a voltage on the base of a transistor with a very limited current and it will not work. Any current requires a voltage but you can have voltage without current.
I'm not sure I'm following what you mean by "put a voltage on the base of a transistor with a very limited current?" If the voltage on the base is so small that you don't see any collector current, then the Ebers-Moll equation predicts that.
@@Lantertronics For example, if you try to drive a transistor from a source that can supply the voltage but not at the turn on current the transistor will not turn on. The voltage makes current but you need current for it to turn on vs. an FET.
Consider driving a bunch of transistors with a small microcontroller. Maybe it can produce the VBE but not the current and in that case it won't work. You can say it can't supply the voltage at that current I guess but it's all interconnected and knowing what physically turns the device on is useful.
With tubes you typically want a high voltage to control the flow of electrons but the electrons are not supplied by the high voltage parts of the tube. They're supplied by the low voltage but high current of the heaters. The effect of voltage on the grids is what makes things happen. Not current. Same with FETs.
Go read the original Ebers and Moll paper, and read it till the end. You will see that they also give their celebrated equations in the form of a current controlled model. You can see the device as either a voltage controlled or a current controlled device. If anything, these devices are charge controlled.
For a list of references, see "Working of a bipolar transistor with electron flow" and "Why is a BJT considered current controlled" on Electrical Engineering Stack Etc. I had to put this in a separate comment because yt deletes anything pointing to external referenced. It interferes with their gaslighting.
Is it just for me, that Prof. Lanterman's voice is solely in the right stereo channel in this video? The content is great though. Mostly in my right ear. :-)
Man, this is bit of sophistry. The electric potential (base emitter voltage) induces current, i.e. charge carriers into the base. Whether you consider the BJT as a voltage or a current controlled device is simply a matter of viewpoint. Either approach is valid. If you wanted to spilt hairs even more you could say that the BJT, the FET and even the vacuum tube are all charge controlled devices. Yes, I do this for a living.
It has practical implications. I contend the voltage-controlled viewpoint leads to much cleaner solutions that are ultimately easier to teach. The current-controlled viewpoint just puts off the complexity but it eventually works its way back in.
@@Lantertronics That's where the rage and flamewars arise: from people who have used current-control all their lives, and consider themselves experts, but suddenly are shown to be wrong on an absolute scale. Nothing will shift their prior beliefs, because for that to occur, first they'd have to see themselves as little children, still learning. Children, not educated experts with decades of experience. If instead we become "experts," learning mostly ceases. I think this was Einstein's great secret. He considered himself a little child. That way he never got trapped by the dogma of the time, the way that any "top physics experts" get trapped.
@@Lantertronics > Professor Lanterman puppet - - That's the way to do it! I always say, don't try to become everybody's friend, instead make it our goal to outrage certain kinds of people. My favorite trick: attract people who love to read, while repelling everyone else. (Just type many-pages responses to every comment.)
I'd argue the opposite -- the voltage-controlled viewpoint allows for much cleaner explanations of many circuits than the current-controlled viewpoint does.
@@jachfeng6201 I think we must be misunderstanding each other. I'd argue it's very difficult to precisely control Ib. It's much easier to think about voltage control. Let VBE be the DC bias point, and vbe be the small-signal perturbation around that bias point. Let vBE = VBE + vbe be the total voltage. Looking at the slope of ic, the small signal collector current, vs vbe (evaluated at the DC bias point) gives you gm, the small-signal transconductance. The total current is iC = IC + ic, where IC is the DC bias current. If you have a common collector amplifier with grounded emitter, the gain is just gm times RC, the collector resistor. Very elegant and clean. Derivations of the voltage gain that try to work through beta are much more awkward. And the point is you're generally specifying the input signal as a voltage.
Actually, current-control is the "for-dummies" description of transistors. It's so over-simplified that it's wrong. When we go to engineering school and take semiconductor courses, we learn the real stuff: the actual physics of the BJT, and the math which describes it, called "Ebers-Moll equations." Current-control doesn't explain transistors, because base-current has no effect on collector current. Also, current-control cannot be used in professional designs, except in the most simplified examples. The pros use voltage-control, because it always works. Current control causes designs to fail (as Horowitz and Hill describe in great detail, in the lab-manual which comes with the AOA textbook.)
I'm always perplexed when people think the current-control view is easier to understand, because when you actually try to use it, it winds up being really awkward compared with using the actual voltage-control view.
I'm not sure what you're trying to say when you say Vbe/Vt is unitless? Yes, of course it is. The saturation current is in units of amps, so you get a collector current in units of amps.
Sorry, but no mathematical model changes physical principles. What you call fundamental... I don’t consider it such at all. It's really good, but hey - just a model, unnecessary in too many practical cases. FET is a voltage controlled device, meaning it does not require current to control it. BJT is a current controlled device which means that current is required, and this is because BJT uses charge carriers injected by current rather than voltage. This explains, without going into detail, how and why a transistor works. I would call this knowledge more fundamental. You would probably call this basic. And just one more thing about resistors-sources. Even if you can replace A with B, that doesn't mean A is B.
By that reasoning, FETs require current. There is no such thing as infinite resistance, and also, the gate input is a capacitor, and requires significant currents to charge and discharge. But if we deny this, and insist that FETs are voltage-controlled only, then we also deny that BJTs require current, since the physics revelas that base current is just an unwanted leakage, same as in FETs. The controlling parameter in both is an e-field across a depletion layer. FETs differ from BJTs in that the variable depletion layer in the BJT is stretched across the current path, while in FETs, it invades from the sides. Gate currents are never eliminated, but can be drastically reduced. Same thing with BJTs, the "super-beta" components with hfe up in the tens of thousands. They still work fine, because the e-field in the junction is controlling the emitter current, regardless of any leakage through the base. Apparently this was not recognized back in the 1950s, when BJTs had single-digit hfe. If they'd had hfe=500 devices right at the start, instead William Shockley's name might be on the Ebers-Moll equations.
Yay, AOA! They're about the only one who gets transistors right. Horowitz, Hill, and also Lanterman. (Also go find yourself a copy of their Lab Manual. The authors really let "current control" have it, pulling out all the stops. If you deeply believe in holy hfe, then your analog designs will all fail ...because slight temperature changes ruin the bias calculations, and because no two BJT transistors have similar hfe, even if the part numbers are the same.)
These same issues came up to bite Win Hill in his first EE job. Exactly the same thing happened to me! I had to abandon hfe and current-control, and instead re-teach myself the BJT physics. I ended up using longtail-pairs and cascode level-shifters, both which are voltage-input circuits, and cannot be explained in terms of current input.
Also Wikipedia had all the current-control garbage on their BJT transistor page. I and others set them straight, but then it all got reverted. Then, WIN HILL HIMSELF came storming in, and carved those "experts" a new a-hole. The BJT page on Wikipedia then had all the voltage-control material. (I haven't checked it recently, so I wouldn't be surprised if the ignorant masses had deleted it all again, years later.)
And finally ...just tell students that there's no mechanism where one current can directly affect another. In the BJT, we can **pretend** that Ib affects Ic ...but in reality, Ib can be used to adjust Vbe, and then Vbe directly controls Ie. The physics says that BJTs are voltage-based devices, "transimpedance-istors." Same as FETs. But with BJTs, the depletion-zone extends across the current path, and varies the current by altering the DZ thickness. Fets are different in that the DZ invades from the side. If FETs are like an optical iris with variable aperture, then BJTs are like sunglasses with variable tint!
PS
We can prove that MOSFETS are controlled by current. After all, if we vary the nanoamps on the gate, it controls the drain current! (This is actually true.) Well sure, but then everyone complains that the MOSFET physics shows that it's controlled by voltage. GOTCHA ...because the BJT physics shows that Ie is controlled by Vbe, and that base-current plays no role in this, except as a leakage-current, same thing as FET gate-leakage.
I have the lab manual and love it.
I'm pinning this comment because it is epic.
Go read Streetman if you want to understand how base current can control collector current.
Mind you, it's not a general electronic circuits textbook, but it's a semiconductor devices textbook.
"but in reality, Ib can be used to adjust Vbe, and then Vbe directly controls Ie". That's current control.
I have no idea the wrath around this. I personally found the distinction useful. Ohms law makes everything interconnected.
That's one of the questions on the German amateur radio license test. Unfortunately I had to choose the BJTs are current-controlled devices answer to pass. That's how they break ya 😂
Doug Self, one of the "gurus" of analog audio amplifiers, also really emphasized that the BJT is fundamentally a voltage controlled current source, and made almost the exact same argument as you, and the Art of Electronics do---pretty good company overall.
I am SO glad somebody has finally addressed the current control myth that is in almost every electronics book! I taught my students the facts about the BJT's in terms of the governing (voltage) equation and explained that the need for base current is just a side effect of the operation, not the "transistor law". The fact that HFE varies so much should give a clue to the falsity of current control argument
Having spent a lot of time characterising exponential-voltage-to-current designs for music VCOs recently, I'm always quite amazed at how accurately a real life circuit follows the Ebers-Moll equation if you compensate the circuit for the bulk emitter resistance of the bjt. Another nice video Aaron!
Thanks for addressing this subject. Another great resource, one that focuses on radio frequency circuits, is Communication Circuits Analysis and Design by Clark & Hess 2nd ed. 1978. The book begins with a great explanation of the BJT as a voltage driven device. This simplifies the design of RF oscillators, mixers and much more.
I don't know what changed, but I like the style and tone of this video much more than the one about mosfets. Great content as always
I guess I should do more "Story Time with Professor Lanterman." :)
You have probably heard it a million times already professor Lantermann but you're sense of humour is really high class.
Everytime after i watched one of your videos i have to sit in silence for a minute to take in your "ironic/sarcastic/comical/silly/superintelligent" one-liners and jokes you weave in.
I appreciate and respect that a lot.
Thank you for your kind words! I'm glad someone appreciates them. :)
when i saw the thumbnail i had to watch this. My "electronic devices" teacher drilled this into our heads of how and external electric field across base-emitter modulates the internal "built in" voltage fighting the current diffusion across the depletion region, hence controlling he emitter-collector current .
I studied BJTs a million years ago using Millman and Halkias as the text. We studied both the current model (Ic = Beta*Ib) and the voltage model (Ic = gm*Vpi).
The irony is the famous "Transistor Man" drawing early in Chapter 2 of the Art of Electronics showing a man controlling collector current by looking at a meter measuring base current! Of course it's not a horrible approximation, but it IS an approximation, and the Ebers-Moll model is introduced later in the chapter.
But the "Transistor Man" model is likely what most people remember.
They should have given Ebers-Moll a catchy mascot. ;)
You should think of the base current as an unwanted side effect of the transistor. When designing a circuit around Ib and hfe it will be very difficult to reproduce it expecting it to have the same behavior even with the same type of transistor. Also because hfe is depend on the collector current, called beta droop, this means driving driving an amplifier with current will result in a lot of distortion bucause of the non linear behavior.
Nice video, I'm a hobbyist myself. I have the book, the art of electronics, but also, learn the art of electronics. And that doesn't give a bad explanation on this subject either and keeps it nice and practical. It's not that your videos aren't practical, it's just another explanation on the subject
Oh, I didn't know about the "Learning the Art of Electronics" book. I will need to check it out.
教授, 您的影片內容都很有趣!
I think the best example of the fact is how one can hugely speed up closing a bjt switch by forcing the base-emitter voltage to zero or slightly negative. (it otherwise takes several microseconds for the collector current to cease flowing after stopping the base current)
This can as well be the best example of the need for a base current: you need to remove charge from the base and the best way is to... Make it flow out (with due sign convention).
Great video! Although I think we have to be a bit careful between differentiating modelling for engineering and physics for design. Although several models allow us to behaviourally accelerate how to make a useful system, a tradeoff has to be made between hiding detail for ease.
I have two "art of electronics" books and I see these books the best for actual design work after you have learned the fundamentals. I'm using Donald A Newman's book and Sedra/Smith's books for a more detailed dive into the operation of the devices.
gotta get them gainz!
Woot!
Barry Gilbert called this way of understanding BJTs, Translinear circuits.
I met Barrie Gilbert once. He was absolutely brilliant.
WOW!
I recently talked with the guy from SSI over the SSI2164 and he said it requires a voltage, not a current, but i only saw transistors in the simplified schematic.
A while back i played with the TB-303 vco, where 2 transistors act like a comparator. it clearly shows voltage play an important role (if the voltage on base is higher than the emitter).
Yeah, the SSI2164 control pin takes a voltage input.
@@Lantertronics Ya, but it's weird when you look at the schematic in the datasheet and only see transistors. to remove it's exponential property you need to make a anti log circuit as the SSI man told me. so the SSI21664 can be used as VCO with REAL linear input on the control pin put aside the normal voltage. i did a research on Anti Log circuits, but for now it's a bit unclear for me, i tried simulating these circuits found and found nothing special or the simulator isn't up to it.. Would you give it a shot on the Anti Log circuit to be used on the SSI2164 control pin? i have made a cool VCO circuit and it works, accept it's useless to build if it doesn't follow 1v/oct.
Look again.... the SSI2164 datasheet does have resistors in the simplified circuit.
Personally I've always felt like what people are really asking when they talk about current controlled or voltage controlled devices is which parameter has the more linear relationship with power through the device, not necessarily which parameter can be used to most neatly model a devices characteristics. Generally this is because that's how they've learned to think of two terminal devices and they are trying to extend that understanding to transistors. For example a diode is often described as a current controlled device because if you were to linearly vary the power through a diode you would see that the current through it would seem to change almost linearly while the voltage would seem fairly constant.
Is it true that the power through a diode is controlled with it's current and not it's voltage though? Of course not. Your thumbnail for this video is the Shockley diode equation demonstrating that after all, but as the old saying goes, "All models are wrong, some are useful." The idea that a component is either voltage controlled or current controlled is a useful simplification for someone just learning about the basic principles of circuits. That simplified model which sees every component as having a linear relationship with either voltage or current can quickly outlive its usefulness though.
Indeed, I usually see people support the current-controlled viewpoint by saying because it's linear it's easier to think about. I think that just winds up just transferring the complexity elsewhere...
@@Lantertronics Yep, because some black-box models are more convenient. But if we actually believe a black-box model is true, that's called Reification.
Looking at the transistor physics, we find no mechanism where the base current can somehow alter the collector current. One current doesn't affect another. Transistors literally cannot be current-controlled. Then, to explain the APPARENT current-control, we just introduce Vbe and e-fields of the junction. In AOA, the "little transistor man" who produces the "current gain," he's the EB junction voltage. He measures the base current, then proportionally adjusts the emitter current. (But AOA authors forgot to mention that "transistor man" is just the Vbe potential! Base current and collector current are communicating via changes in Vbe. If we force Vbe value to be fixed, then current-gain disappears.)
I LOVE your content, wish i had you as a professor
🙏🙏 great video
Thanks!
It reminds me of, "all models are wrong, but some models are useful."
I started considering BJT's as voltage controlled shortly after I started working with differential pairs. Am I right in saying that, if you apply a current from a voltage to a BJT's base (i.e. through a resistor, either on the base or the emitter) the voltage on the BE junction will be the log of the input voltage. However the output current (and therefore output voltage across the collector resistor) will be proportional to the anti-log of the BE voltage. I.e. the log/anti-log functions cancel each other out (with other influences such as temperature, etc) ?
I'd rephrase it as "current from a current source to a BJTs" base -- and then yeah, your explanation does make sense.
So... is the main significance of the base current that it needs to have a DC path to ground provided for it, and so must be considered but shouldn't figure in the gain calculations?
I've been doing some transistor matching, and it seems from reading the experience of others that it's more worthwhile matching the Vbe rather than hfe. I've been measuring it by connecting the DUT as a diode in a circuit that pulls a constant(ish) 100uA through it while I measure the voltage drop across it. Do you agree about the Vbe, and can you suggest the right way to measure it please?
I"m not sure I'm following what you're trying to say in the first paragraph. For BJT or enhancement mode MOSFET you'll generally have connections to the DC supply rails somewhere to provide bias.
Regarding matching, google: Ian Fritz transistor matching. Ian is a very clever guy.
@@Lantertronics I too was reading AoE, and it was something in the differential amplifier section of the BJT chapter.... hang on, thanks to having a digital as well as a print version I can cut & paste it:
"Be sure to remember that this amplifier, like all transistor amplifiers, must have a dc bias path to the bases. If the
input is capacitively coupled, for instance, you must have base resistors to ground."
@@gerryjamesedwards1227 Yup, need biasing. But that's just something you need in general, not particularly because current flows through the base.
@@Lantertronics thank you!
@@gerryjamesedwards1227 You are welcome!
What do you recommend as a deep-dive to revisit "looking into terminals" and deriving expressions?
Definitely check out Marshall Leach's ECE3050 webpage. Read his notes on "The BJT" and then the notes on various amplifier models after that.
I think the current-controlled description is useful for comparing from FETs. FETs only need a voltage and draw almost no current (very high input impedance). You can put a voltage on the base of a transistor with a very limited current and it will not work. Any current requires a voltage but you can have voltage without current.
I'm not sure I'm following what you mean by "put a voltage on the base of a transistor with a very limited current?" If the voltage on the base is so small that you don't see any collector current, then the Ebers-Moll equation predicts that.
@@Lantertronics For example, if you try to drive a transistor from a source that can supply the voltage but not at the turn on current the transistor will not turn on. The voltage makes current but you need current for it to turn on vs. an FET.
This seems to be a thing with Art of Electronics where they throw the baby out with the bathwater for simplification maybe.
Consider driving a bunch of transistors with a small microcontroller. Maybe it can produce the VBE but not the current and in that case it won't work. You can say it can't supply the voltage at that current I guess but it's all interconnected and knowing what physically turns the device on is useful.
With tubes you typically want a high voltage to control the flow of electrons but the electrons are not supplied by the high voltage parts of the tube. They're supplied by the low voltage but high current of the heaters. The effect of voltage on the grids is what makes things happen. Not current. Same with FETs.
Go read the original Ebers and Moll paper, and read it till the end. You will see that they also give their celebrated equations in the form of a current controlled model.
You can see the device as either a voltage controlled or a current controlled device. If anything, these devices are charge controlled.
For a list of references, see "Working of a bipolar transistor with electron flow" and "Why is a BJT considered current controlled" on Electrical Engineering Stack Etc.
I had to put this in a separate comment because yt deletes anything pointing to external referenced. It interferes with their gaslighting.
Is it just for me, that Prof. Lanterman's voice is solely in the right stereo channel in this video? The content is great though. Mostly in my right ear. :-)
I can't stop reading BJT as Brazilian Jiujitsu
Man, this is bit of sophistry. The electric potential (base emitter voltage) induces current, i.e. charge carriers into the base.
Whether you consider the BJT as a voltage or a current controlled device is simply a matter of viewpoint. Either approach is valid.
If you wanted to spilt hairs even more you could say that the BJT, the FET and even the vacuum tube are all charge controlled devices.
Yes, I do this for a living.
It has practical implications. I contend the voltage-controlled viewpoint leads to much cleaner solutions that are ultimately easier to teach. The current-controlled viewpoint just puts off the complexity but it eventually works its way back in.
@@Lantertronics Hmm, yeah, I guess. Whatever works for you.
@@Lantertronics That's where the rage and flamewars arise: from people who have used current-control all their lives, and consider themselves experts, but suddenly are shown to be wrong on an absolute scale. Nothing will shift their prior beliefs, because for that to occur, first they'd have to see themselves as little children, still learning. Children, not educated experts with decades of experience. If instead we become "experts," learning mostly ceases.
I think this was Einstein's great secret. He considered himself a little child. That way he never got trapped by the dogma of the time, the way that any "top physics experts" get trapped.
@@wbeaty I just recorded another "Story Time with Professor Lanterman," this time with a Professor Lanterman puppet my wife made. ;)
@@Lantertronics > Professor Lanterman puppet - -
That's the way to do it!
I always say, don't try to become everybody's friend, instead make it our goal to outrage certain kinds of people.
My favorite trick: attract people who love to read, while repelling everyone else. (Just type many-pages responses to every comment.)
This is nothing new. We got taught this in uni. I guess a lot of people were sleeping in the lecture hall that day.
A lot of people get uptight at me when I contend that it's really voltage-controlled...
@@Lantertronics Can't teach the willfully ignorant.
The title is literarily correct, but not in practical usage.
I'd argue the opposite -- the voltage-controlled viewpoint allows for much cleaner explanations of many circuits than the current-controlled viewpoint does.
@@Lantertronics Yes, you are right. But in practice, it's almost impossible to use Vbe to control Ic
@@jachfeng6201 I think we must be misunderstanding each other. I'd argue it's very difficult to precisely control Ib. It's much easier to think about voltage control.
Let VBE be the DC bias point, and vbe be the small-signal perturbation around that bias point. Let vBE = VBE + vbe be the total voltage. Looking at the slope of ic, the small signal collector current, vs vbe (evaluated at the DC bias point) gives you gm, the small-signal transconductance. The total current is iC = IC + ic, where IC is the DC bias current.
If you have a common collector amplifier with grounded emitter, the gain is just gm times RC, the collector resistor. Very elegant and clean. Derivations of the voltage gain that try to work through beta are much more awkward. And the point is you're generally specifying the input signal as a voltage.
And they all lived happily ever after.
The End.
No, they are current controlled
In the equation Vbe/Vt is unitless
Actually, current-control is the "for-dummies" description of transistors. It's so over-simplified that it's wrong.
When we go to engineering school and take semiconductor courses, we learn the real stuff: the actual physics of the BJT, and the math which describes it, called "Ebers-Moll equations."
Current-control doesn't explain transistors, because base-current has no effect on collector current. Also, current-control cannot be used in professional designs, except in the most simplified examples. The pros use voltage-control, because it always works. Current control causes designs to fail (as Horowitz and Hill describe in great detail, in the lab-manual which comes with the AOA textbook.)
I'm always perplexed when people think the current-control view is easier to understand, because when you actually try to use it, it winds up being really awkward compared with using the actual voltage-control view.
I'm not sure what you're trying to say when you say Vbe/Vt is unitless? Yes, of course it is. The saturation current is in units of amps, so you get a collector current in units of amps.
@@wbeatyin the original paper by Ebers and Moll the equations are given also in the current control form. Has anyone ever read that paper?
Sorry, but no mathematical model changes physical principles. What you call fundamental... I don’t consider it such at all. It's really good, but hey - just a model, unnecessary in too many practical cases.
FET is a voltage controlled device, meaning it does not require current to control it. BJT is a current controlled device which means that current is required, and this is because BJT uses charge carriers injected by current rather than voltage. This explains, without going into detail, how and why a transistor works. I would call this knowledge more fundamental. You would probably call this basic.
And just one more thing about resistors-sources. Even if you can replace A with B, that doesn't mean A is B.
What controls the amount of current flowing through the base?
By that reasoning, FETs require current. There is no such thing as infinite resistance, and also, the gate input is a capacitor, and requires significant currents to charge and discharge. But if we deny this, and insist that FETs are voltage-controlled only, then we also deny that BJTs require current, since the physics revelas that base current is just an unwanted leakage, same as in FETs. The controlling parameter in both is an e-field across a depletion layer. FETs differ from BJTs in that the variable depletion layer in the BJT is stretched across the current path, while in FETs, it invades from the sides.
Gate currents are never eliminated, but can be drastically reduced. Same thing with BJTs, the "super-beta" components with hfe up in the tens of thousands. They still work fine, because the e-field in the junction is controlling the emitter current, regardless of any leakage through the base. Apparently this was not recognized back in the 1950s, when BJTs had single-digit hfe. If they'd had hfe=500 devices right at the start, instead William Shockley's name might be on the Ebers-Moll equations.