You've done something in little over 12 minutes what my profs (and the books) couldn't do for a lifetime - made me understand the Miller effect! Heartfelt thanks! Could you please do a video on - 1. why we introduce an emitter resistance in a common emitter amplifier and 2. what is "emitter degeneration" (a term you use in this video). I look forward to the circuit demo (if you actually do this video). Thanks once again, Anand
Anand Paralkar Adding a resistor into the emitter of a CE amplifier is a form of negative feedback. It reduces the gain, but provided benefits of bias stability and vastly improved linearity. This is also called emitter degeneration. You may find these videos helpful: ua-cam.com/video/NizrzRKQqII/v-deo.html ua-cam.com/video/c6cmkm3UPUI/v-deo.html ua-cam.com/video/zXh5gMc6kyU/v-deo.html ua-cam.com/video/YQlbPGNB-ys/v-deo.html
This is the most elegant description of Miller effect I've ever seen. The precise reason for the multiplication of Ccb is normally glossed-over in text books as a solution of equations. Very nice and thank you.
davet11 sometimes, IMHO, the book´s author want 1) that the reader won´t get the whole idea at first glance, so he needs to go and solve the equations him/hisself....(BORING) 2) the author didn´t get the whole toppic (this happens so much more than you ever expect) but he´s stil trying to teach you about it..... 3) the autor is simply arrogant and do not want you to understand what he´s talking about, thus leaving you thinking you´re a "jackass" and "not worth it...... I had some kind of books like these, (and they had been recommended by our university´s professor - DANMIT) but i sold them very quickly and searched myself for book covering that toppic a "little" more better;)
davet11 I was thinking the same! At my university it was taught like that, as the interpretation of a solution of equations. I always had the uncomfortable feeling of not understanding it thoroughly. Thank you ***** for this great and insightful video.
I found a cascode circuit in one of my Kenwood audio amps (KA-300) It did not ring a bell, but now it does. Thank you Alan, for this very clear explanation and demo.
l love cascodes, they have a million and one uses.. well 3 that I can think of right now: -Raising a pole to increase bandwidth, especially useful in a feedback loop. Although you still have to finish compensating the loop elsewhere, at least it can shift pesky poles above the 0dB crossover frequency. -Increasing the voltage rating of a transistor amplifier. BJTs only go up to around 400-500V for linear use (Vceo), ones with higher voltage rating that that are specified at the higher voltage only when base and emitter are shorted together (Vces) i.e. a well driven switch. -Reducing switching loss, since its during the miller plateau that drain current and voltage coexist (in a FET) and so reducing the duration of that plateau reduces the energy lost in switching. I like to think of it as the upper NPN's Vbe is setting its emitter at a fixed voltage as Vb is fixed, therefore the bottom NPN's collector is a fixed voltage regardless of current, so since that voltage doesn't change the voltage across Ccb doesn't change and the miller effect is reduced. As well as this the upper NPN doesn't have a significant miller capacitance because it's common base, the base is grounded to AC. I also like to think of the bottom NPN as a current source driving the emitter of the upper NPN. As you point out it's easier to see in a circuit with no emitter degeneration.
That's an excellent demo. I studied cascode amplifiers in an EE lab many years ago. There is an improvement in the bandwidth when a common collector input stage is used. Another way of looking at the high frequency roll off is Cμ is shunted by the collector resistor + input impedance of the CE amplifier. The inverse of this time constant is the main contributor to the high frequency roll off of the cascode amplifier.. With the cascode, the Miller Effect produces a low impedance looking into the emitter of the common base transistor., so the open circuit time constant is much lower. When a CC input stage is used, the input impedance is lowered, so the open circuit time constant is reduced further. The approximation of summing the inverse time constants to calculate the bandwidth is accurate to about 13%, sometimes better.
Thanks again, Alan. It is the clearest, most intuitive explanation of a cascode amplifier I have ever come across (and that includes Horowitz and Hill (2nd Ed)). 73, Marcus
Joe White just because, there is nothing "fancy-pancy" about it, it´s just that simple :) But the teachers at the universities keep on trying to make this looking like a "second order" problem only to be fully solved by using integration....
Thanks ! This circuit has puzzeled me for many years. I first saw it in a circuit that amplifies a tacho signal coming from a photo diode. I was told long ago that the additional transistor helped reduce the v_bc (AC) and therefore the effect of the parasitic capacitance C_cb, and the high impedance of the photo diode is then better matched with the input of the amplifier. Now that you explained the miller effect it made more sense to me :)
Your videos are some of the finest on the Internet! It never ceases to amaze my how effortlessly you seem to make the most nebulous, theoretical concepts as clear and obvious as the nose on your face! Every one of your posts should come with 1000 "Like" buttons!
Thank you for sharing your knowledge, I never miss one of your videos, and always learn something, even if I already knew the subject, and that is amazing. Please keep up the good work and greetings from Venezuela.
And i thought Miller effect was what people have after having a Miller's beer. Just kidding, it was a great video. Clear and direct to the point. Thumbs up!
The Miller effect (in vacuum tubes): Edwin Armstrong discovered in 1915 that oscillation occurs in triode amplifiers because of grid-plate capacitance, when the anode load is inductive (and below resonance) The actual (small) grid-plate capacitance is multiplied by (µ +1). In 1919 John Miller of the US bureau of standards presented a mathematical analysis of the phenomina and its effect, which now bears his name!
Nice video. The circuit gets its effectiveness by the low impedance of stage 2 on top. The common base amplifier also has a Miller cap but it’s working against an AC short on it’s base. Just a note -Your base cap, 68p, seems pretty low on the CB stage.
Nice video. I ve find this cascode configuration several years in an exam.... But only it's polarization. I've always said why anyone wants this complicated config? Now i know :) . And for Thevenin discussion similar thing happens with Kirchoff (i know this is a literalization of what it sounds)
Great video, thanks. Quick question: If we assume that the g of the CE amplifier transistor is 20, you would expect a BW improvement of ~10. Why do you think you got only 2-3? Do you know what is limiting the BW of the cascode amp?
I didn't design this circuit, so I'd have to look at it more carefully. There is a cap around the b-e junction of the emitter follower, so that is likely part of it.
Wow, they say there's no such thing as a free lunch but the cascode amplifier seems to get pretty close. For one extra transistor you get a nice bandwidth boost!
Gain of a mosfet based Cascode is not the same as common source amplifier as stated at 1:30ish. The gain can be as high as (gm × ro)^2. Largely it depends on what the values of Rd and RL are. If Rd is implemented as two stage cascode load using pmos or the circuit is biased using a current source with very high output impedance, the gain can be (gm × ro)^2. The big deal with cascode circuit is that one is able to boost the gain while not having to sacrifice the bandwidth. This happens because the voltage at drain of common source transistor is close to Vin since that voltage is set by gm1 times [~1/gm2 in parallel to ro1] which works out to be close to 1. Including gmbs in the circuit does not make much difference as that effect too cancels out. So, in effect, the voltage gain between gate and drain is almost 1 which ensures that the Miller effect on Cgd is non existent. The "only" thing you lose, and this really does matter when supply voltages are tiny, is the voltage headroom. For a BJT based amplifier, the gain can be as high as gm1 × (gm2× ro2) (ro1 parallel to rpi2) which is still much higher than common emitter version.
If your signal generator started at 1 mhz; why did the output peak at 2 or 3 mhz? Shouldn't it have started out high and decreased as the frequency increased? Is that gain boost at 2 or 3 mhz caused by that parallel RC network in the emitter circuit? I tried to see the component values off the sheet in the video but at 77 years old my eyes are getting pretty bad... I would also like to know how he determined the component values to provide the proper bias on the upper transistor. Thank you for the interesting video...
Thank you for this explanation. Ali Hajimiri's CalTech lectures mentioned cascode amp configurations, so it was nice to hear you explain it here. Question: Is there a good professional oscilloscope that you might recommend from your experience, that will be useful for precision analog, comparator, fast precise rise times, gyrator, and work involving transients?
You've got to be more precise about "fast rise times", as that will determine the bandwidth needed by the scopes and probes. Also, an idea regarding the available budget.
@@w2aew I've generally been looking at 4 channel scopes with 200-350Mhz. $1500-$6000, definitely under $10k. A EE mentor/colleague told me that 500mhz minimum for work which may involving noise and transient analysis but those 4ch 500mhz analog seem to get pricier. There's also some 10x frequency rule of thumb I heard. In terms of tasks, I will be constructing zero crossing detector circuits on sine waves, going up to 30MHz (that 10x rules would suggest a 300MHz scope). I am interested in coupling oscillators together, hooking them to gyrators and looking at the oscillatory current outputs. Trying to build synthetic large value inductors from op-amps in order to get precise inductance for oscillators. And constructing a few analog computers to compute some results.
Great video! Makes the cascode configuration easy to understand. The one issue that you didn't mention though is that it requires more headroom than a CE amp as I learned the hard way! The CB amp biasing forces the output swing to be more constrained than with a plain CE amp. One question: Is there anything that can be done to make the amplifier broadband and flat across its passband? I've done a fair bit of simulation of cascodes and can't seem to get much better than about 50-60 MHz BW. I realize that selecting a BJT with a low C_ cb improves freq response, but is there anything more that can be done to improve BW? Thanks
Eric Everton You will lose a little headroom, but you can usually limit the loss to a few hundred mV by biasing the CB stage so that the CE's collector-emitter voltage is a few hundred mV above saturation. Of course, you'll sacrifice a little BW this way because Ccb will be larger with less Vcb. BW is generally improved with larger Vce in the CE stage, as well as with some degeneration in the CE stage (will do a lot for passband flatness). Proper selection of transistors with low capacitance and high ft is required, as well as operating at generally larger Ic with lower value RC, will all help to increase BW.
Alan, Would you please update the links for the PDF's of your notes on this video and video 188 as well as any others that may be linked to your old site. Do you have a complete repository or zip of the notes somewhere? Thanks Sam W3OHM
how about a video on increasing miller capacitance with an additional capacitor on a common emitter amplifier to limit bandwidth, done commonly in the voltage amplification stage of op amps to improve stability
@@w2aew My goal is to improve the linearity by increasing the bandwidth. Ideally, I would like a relatively flat response in the FM broadcast band. I've tried several biasing configurations which I'm not sure qualify as "cascode" or not, but using two transistors instead of a single device does seem to help overall. see sim at : www.falstad.com/circuit/circuitjs.html?ctz=CQAgjCDMB0YEwFYCmBaMAGEdYA5IYTABYwcA2OdHIskBWhOzZNMAKABMQUdMiic3MLX6C44AHJF06ImwBO4Ms3hK+AkJgxsA7lEp0AnOMgGyAdnHo2AJW68Qo7oZEbMfLVc2boCXY40LcScEY002AHMAwSDHMiIQWPdIoREXVLpVZL14un5EhKT-YWZzWnSEMvC9EszxCqz-U2Y4QWa6ZXCAYxAK-IqqvmhzVDAINHRoC2V0MENIeaoiOATJmXYehvqGfKGR1nBJ+jH8ZaJzSBxCQTXZth64SzoqykFK2i1UQQnoWUgySAIaRwQwISD8CC3DZQahxBI0eG5T4ocyOaDLQFgSwkOatYQ+dYKLBPd7E8QUVaHGTWRSPYIaOl1bwYalsAA2aiZtVJWl81P5ApkcFQqJQt3QCDgFgQ6AuOClJGkNzYABcyc9aO0eeAfAJZOYiJVIODjoZVlNTP8yGCAUDDPKqSAOEgAGYAQwArmyVcVOqTXhrquqnO0nNYooyKer4u4g7kgQk+qsie1Yk4klSmmYnrkiopak40K4bpmAPJwisIxwybzWPSh9Khtz3DKF-FEdLI1FjaCPcE0RBLeVgRhQltasIhMJdnwg0wuBDy9BWsiGVFj-OdJyMsPgXjWNWMwgmAzH5lCaDGY2WM3W5dwKwWuYAyiIczzCi0FlO12e71Eo9VCPQYqXDGEEjPS4INaWtVX0dQ2lhXdxnMaAcGxIEwRwbDDQsHw7zXQxSAWMhyFCR1nXdL0fT0ItolbGs6wYzBIFkatsnA9jOKIUxwkUKCKwEs9MAfcVBX5SAiSraT4yyUCUirM9FJg5J+LY6QWKQxjHHMNgAHssC-bwSHQBgfEhcA2EucBxAAMQssBRggfYxiEEAAGE3QABzdLoAEsVTdAA7LokCswRXPs5knJAFzxggTyfP8wKQrCoA
Excellent video. One question: it looks like the collector of the input device is clamped at Vb-Vbe. If this is so, why is the Miller effect menationed at 2x ? Won't it be 1x ? Regards, from NZ, Mark
It's because the impedance looking into the emitter of the CB transistor is not zero. It is 1/gm. Thus, the CE stage has a gain of -gm*(1/gm), or -1. This makes the miller magnification equal to two.
This is a little off the subject, but you are the one that would know. I am trying to convert the 19.2khz scan rate from the crt output on my HP 8921a to feed into a modern LCD monitor. I was wondering if you have a schematic for a converter to achieve this? Great video by the way, keep up the good work!!!!
Thank you for the excellent explanation. But I'm still confused as to why there's still a multiplication of C_cb of CE stage by 2. CE stage gain is v_out/v_in, isn't it? The v_in is the change in base voltage, and v_out is the change in collector voltage. But, the collector voltage of CE stage should be constant at (V_b - V_be) of CB stage. So, the v_out of CE stage is always 0, and the CE stage gain is 0. Therefore, there should be no multiplication of C_cb of CE stage, and no Miller effect at all.
The gain of the CE stage is -1, which means that the collector voltage moves in the opposite direction of the base voltage. Thus, the change in voltage across the c-e capacitance is double what it would be if the capacitance was simply from base to ground or collector to ground - thus making it appear twice as big.
@@w2aew I'm sorry. I didn't google it myself. That's a great link and a fascinating read. I'll simulate some cascodes on an amp I'm working on later. Thanks again...
jlfqam Not sure why you'd want to do that - all it would would is add a DC offset to the CE's collector voltage, which would limit the maximum signal swing because the transistor will be that much closer to saturation.
***** Thanks, I just wondered about be the effect on the frequency response of the system compared to the jumper and the CB transistor sets. Should it fall in between?
jlfqam I would think it would be very close the the jumper result, in fact maybe even slightly worse because the voltage on the b-e junction will be lower, which will increase the capacitance.
Could the CB capacitance cause the a circuit to self oscillate under some circumstances? If so, then maybe that would be an interesting idea, to try to force a transistor to self oscillate at, I guess, its own resonance frequency. And furthermore, what's that that device with two parallel coaxes with inductors connected to ground? (The on with copper-clad, not the angle bracket.) A choke?
Collector-Base capacitance is usually not a source for oscillation (in fact it's often used to help make a circuit more stable). The device you are referring to is a directional coupler from a video I did a few months ago.
Gameboygenius if you want a self oscillating transistor, look at MOSFETs with no daming resistor on the gate. the lead inductance and the gate capacitance can cause a lot of ringing. You need some feedback gain to maintain oscillation though. In a bigger circuit with gain elsewhere and poor layout they can oscillate seemingly by themselves.
+Dr.Lecter The pronunciation of Thevenin might be region specific. W2AEW is American and I am Canadian, and I was taught in university, that that's how you say it (as the video said it). Perhaps different regions might pronounce it differently.
Thank you, it bugs me to no end when someone corrects another like that. It's not called for and I think it's not polite. On a side note; English is a living language and changes constantly.
Given that all of the gain here is coming from the common base stage, why keep the troublesome common emitter stage? Is it purely to ensure that we have high input impedance (or control over it, maybe)?
Thanks. A bit of googling shows that, in the 60s, there were 4 terminal "cascode transistors" (two transistors on the die) which I guess made sense, since that would allow better matching of the g_m of the transistors, which you are cancelling in your derivation of the gain.
Dennis Hill This was designed as an RF amplifier, so it has very little gain at audio frequencies. Of course, component (capacitor) values can be adjusted to improve the audio performance. The Miller effect is generally not an issue for audio amplifiers unless the source impedance is very high, so the cascode isn't used much for audio.
More current flow through Ccb, due to a large voltage swing at some frequency across it, has the "effect" of as if it were the gain times larger capacitance. Same impedance (fixed or single frequency) with a larger voltage equals more current (okay). So more current . . . so how does it work? It must somehow make the input impedance smaller to the source, so less input to be amplified? The Ccb doesn't actually change value. Or is it just more negative feedback due to large voltage swing, yielding less input vs frequency? Where does (what path) it shunt the signal to, like a low pass filter to ground, via this capacitance (or is that the "effect" or analogy)? Sure, at the same frequency, a larger capacitance would have lower Xc and draw more current. That is the effect of the higher voltage, but how does it work in the circuit . . . neg feedback, shunting, lower input impedance? Nice demo. Nice to have lots of test equipment. Thanks. Still learning.
ive looked this up because they mention it in some amps , but this is a small effect that will only roll off a bit of highs , but guess what in HI FI some roll off of highs is basically what ppl like . I guess in highly technical measurement stuff this would have to be fixed but in Audio , i doubt its noticeable . Did anyone do these tests , and knows how much is it exactly ?
When using higher gain transistors why do technicians add miller capacitors pf across the collector to base which they say it helps reduced the gain noise hiss? The miller capacitance will reduce or eliminate the gain noise any reasons why it does this?
@@w2aew yes you said that in your video but other techs have told me that it also reduces gain hiss noise not sure if that is true or not that miller caps help with reducing the noise floor?
It's a pity that time didn't permit because the more interesting effect of Miller feedback is when the collector load is not a resistance. It turns out that if the load is capacitive then the feedback capacitor appears as a resistance across the input. If the input came from a tuned circuit then it would be loaded down by this resistance. If the load is inductive then the feedback capacitance appears as a NEGATIVE resistance which could cause oscillations at the input - especially if the input came from a tuned circuit. It is no wonder that neutralization was such a big issue in the receivers of the 1920s.
And I was complaining today about the performance of my NanoVNA that22 has a drop in amplitude of about 2.5 dbm between 50k and 300M... Now it looks much better.
@@w2aew Have you tried the modern Swiss Micros HP15C copy? Beautifully made and a faithful repro of the original. They also do a lovely copy of the HP42 with a larger LCD screen and 4 stack levels displayed. Viva RPN!
A simple RC low pass filter rolls voltage off at 6db per octave. Is the roll off different for a Miller effect, voltage gain limited CE amplifier. It seems that it must be a lower db per octave because the capacitor is effectively variable and falling in effective value, as frequency rises, until the voltage gain approaches -1. What do you call such a reduced roll off response? A half pole roll off?
jpopelish I think it's about the same as a simple RC. The capacitance isn't effectively falling, since the *gain* isn't changing, just the signal amplitude. I'd have to examine it more closely to be sure though.
But an amplifier operating in the frequency range where Miller effect is in control of its gain, certainly shows changing gain, versus frequency, and that changing gain must have a changing multiplier effect on the capacitive (Cbc) feedback. So, changing RC time constant across input signal, versus frequency. Will your scope show a log plot of amplitude versus frequency?
jpopelish I think of it as the input being attenuated by the RC filter, not that the gain is being reduced. Again, I'll have to look at it again when I get back to my lab next week...
***** Never mind. It just occurred to me that the same reduction in voltage, as frequency rises, across the capacitor occurs with a fixed capacitor in an RC, low pass filter. I just had a mental block, because of the gain produced voltage in the CE amplifier. If, somehow, the voltage across the capacitor were forced to be independent of frequency, the result would not be a low pass, but a notch filter.
A cascode takes away from headroom. To increase headroom, you want to setup the bias conditions so that the collector voltage (with no signal applied) is halfway between the positive supply and the emitter voltage.
David Pilling it certainly comes from thermionic valves / vacuum tubes. You get various -odes. Triode, tetrode, pentode etc. which you can cascade to make a casc-ode. There's nothing to say you can't cascode tetrodes or pentodes though. But yes, pentodes have a reduced miller effect because their control grid is screened from the anode by the screen grid IIRC, so 2 triodes in cascode configuration would sort of be pentode-ish in that one single regard.
I'm sure the French telegraph engineer would have pronounced his own name 'tevenin' -- and that should probably be persuasive -- but not many people are aware of the subtleties of french pronunciation.
Apa itu Efek Millar ? Efek Millar sering kali menjadi faktor pembatas bandwidth utama dan desain penguat common-emitor, alasannya karena perbesaran efek kapasitansi parasit. Jadi, jika ada transistor yang memiliki kapasitansi di antara basis-emitor, basis-kolektor, dan kolektor-emitor maka setiap kapasitansi ini pada dasarnya mencuri arus dari transistor. Dan oleh karena itu, akan mempengaruhi bandwidth dan kinerja transistor.
You've done something in little over 12 minutes what my profs (and the books) couldn't do for a lifetime - made me understand the Miller effect! Heartfelt thanks!
Could you please do a video on - 1. why we introduce an emitter resistance in a common emitter amplifier and 2. what is "emitter degeneration" (a term you use in this video).
I look forward to the circuit demo (if you actually do this video).
Thanks once again,
Anand
Anand Paralkar Adding a resistor into the emitter of a CE amplifier is a form of negative feedback. It reduces the gain, but provided benefits of bias stability and vastly improved linearity. This is also called emitter degeneration. You may find these videos helpful:
ua-cam.com/video/NizrzRKQqII/v-deo.html
ua-cam.com/video/c6cmkm3UPUI/v-deo.html
ua-cam.com/video/zXh5gMc6kyU/v-deo.html
ua-cam.com/video/YQlbPGNB-ys/v-deo.html
@@w2aew Could also clarify why we decoupling circuit for any LNA or RFIC circuit design??
This is the most elegant description of Miller effect I've ever seen. The precise reason for the multiplication of Ccb is normally glossed-over in text books as a solution of equations. Very nice and thank you.
Indeed.
davet11 sometimes, IMHO, the book´s author want
1) that the reader won´t get the whole idea at first glance, so he needs to go and solve the equations him/hisself....(BORING)
2) the author didn´t get the whole toppic (this happens so much more than you ever expect) but he´s stil trying to teach you about it.....
3) the autor is simply arrogant and do not want you to understand what he´s talking about, thus leaving you thinking you´re a "jackass" and "not worth it......
I had some kind of books like these, (and they had been recommended by our university´s professor - DANMIT) but i sold them very quickly and searched myself for book covering that toppic a "little" more better;)
davet11 I was thinking the same! At my university it was taught like that, as the interpretation of a solution of equations. I always had the uncomfortable feeling of not understanding it thoroughly. Thank you ***** for this great and insightful video.
64,800 knowledges transferred.. and counting. Every single one of these videos is pure gold! Thanks
I found a cascode circuit in one of my Kenwood audio amps (KA-300) It did not ring a bell, but now it does. Thank you Alan, for this very clear explanation and demo.
l love cascodes, they have a million and one uses.. well 3 that I can think of right now:
-Raising a pole to increase bandwidth, especially useful in a feedback loop. Although you still have to finish compensating the loop elsewhere, at least it can shift pesky poles above the 0dB crossover frequency.
-Increasing the voltage rating of a transistor amplifier. BJTs only go up to around 400-500V for linear use (Vceo), ones with higher voltage rating that that are specified at the higher voltage only when base and emitter are shorted together (Vces) i.e. a well driven switch.
-Reducing switching loss, since its during the miller plateau that drain current and voltage coexist (in a FET) and so reducing the duration of that plateau reduces the energy lost in switching.
I like to think of it as the upper NPN's Vbe is setting its emitter at a fixed voltage as Vb is fixed, therefore the bottom NPN's collector is a fixed voltage regardless of current, so since that voltage doesn't change the voltage across Ccb doesn't change and the miller effect is reduced. As well as this the upper NPN doesn't have a significant miller capacitance because it's common base, the base is grounded to AC. I also like to think of the bottom NPN as a current source driving the emitter of the upper NPN. As you point out it's easier to see in a circuit with no emitter degeneration.
That's an excellent demo. I studied cascode amplifiers in an EE lab many years ago. There is an improvement in the bandwidth when a common collector input stage is used. Another way of looking at the high frequency roll off is Cμ is shunted by the collector resistor + input impedance of the CE amplifier. The inverse of this time constant is the main contributor to the high frequency roll off of the cascode amplifier.. With the cascode, the Miller Effect produces a low impedance looking into the emitter of the common base transistor., so the open circuit time constant is much lower. When a CC input stage is used, the input impedance is lowered, so the open circuit time constant is reduced further. The approximation of summing the inverse time constants to calculate the bandwidth is accurate to about 13%, sometimes better.
Good stuff!
Thanks again, Alan. It is the clearest, most intuitive explanation of a cascode amplifier I have ever come across (and that includes Horowitz and Hill (2nd Ed)). 73, Marcus
Great video. You make understanding circuits very intuitive and straight forward.
Joe White just because, there is nothing "fancy-pancy" about it, it´s just that simple :) But the teachers at the universities keep on trying to make this looking like a "second order" problem only to be fully solved by using integration....
Thanks ! This circuit has puzzeled me for many years. I first saw it in a circuit that amplifies a tacho signal coming from a photo diode. I was told long ago that the additional transistor helped reduce the v_bc (AC) and therefore the effect of the parasitic capacitance C_cb, and the high impedance of the photo diode is then better matched with the input of the amplifier. Now that you explained the miller effect it made more sense to me :)
Your videos are some of the finest on the Internet! It never ceases to amaze my how effortlessly you seem to make the most nebulous, theoretical concepts as clear and obvious as the nose on your face! Every one of your posts should come with 1000 "Like" buttons!
well said!
Best explanation of the cascode stage!
Merci, a great video. Very well explain!
This is just the video I needed right now, I was just researching the cascode and was really struggling to "get" it. Thanks so much!
Thank you for sharing your knowledge, I never miss one of your videos, and always learn something, even if I already knew the subject, and that is amazing. Please keep up the good work and greetings from Venezuela.
As always thank you. You make the subject so clear to understand.
Thank you for the instructive video making it easier to understand cascode amplifiers and the Miller effect.
And i thought Miller effect was what people have after having a Miller's beer.
Just kidding, it was a great video. Clear and direct to the point. Thumbs up!
great video. this stuff is really great supplementary resources for engineering students.
The Miller effect (in vacuum tubes):
Edwin Armstrong discovered in 1915 that oscillation occurs in triode amplifiers because of
grid-plate capacitance, when the anode load is inductive (and below resonance)
The actual (small) grid-plate capacitance is multiplied by (µ +1).
In 1919 John Miller of the US bureau of standards presented a mathematical analysis
of the phenomina and its effect, which now bears his name!
Thanks....
Excellent video, as always!
Your tutorials are always very well organized.
p.s. I dig your handwriting and your grid-ruled paper.
Thanks Alan for the very nice demo
Great video & explanation. Thank you!
Excellent material as always! Clear presentation.
Nice video. The circuit gets its effectiveness by the low impedance of stage 2 on top. The common base amplifier also has a Miller cap but it’s working against an AC short on it’s base. Just a note -Your base cap, 68p, seems pretty low on the CB stage.
Thank you for another excellent and very understandable video.
Nice video. I ve find this cascode configuration several years in an exam.... But only it's polarization. I've always said why anyone wants this complicated config? Now i know :) . And for Thevenin discussion similar thing happens with Kirchoff (i know this is a literalization of what it sounds)
Great video as always. One request, if it isn't much trouble, it would be nice to include the scope images in the notes. Thank you
I'll consider that in the future.
Great video and explanation as always! Thanks.
I have seen many cascode amplifiers in the schematics of the 7000 series Tektronix scopes!
Thanks Alan,I'm really enjoy your video's...
Kostas
Thank you! Very useful!
Very informative. Thanks for sharing.
Are identical devices used in this configuration, or should one have a higher current rating than the other?
Both transistors, being in series, carry the same current - so using identical devices is the way to go.
Great video, thanks.
Quick question: If we assume that the g of the CE amplifier transistor is 20, you would expect a BW improvement of ~10. Why do you think you got only 2-3? Do you know what is limiting the BW of the cascode amp?
I didn't design this circuit, so I'd have to look at it more carefully. There is a cap around the b-e junction of the emitter follower, so that is likely part of it.
Your the best. I always learn so much. Thanks
Now I know why I add a 100 pF capacitor between the base and the collector to drive LEDs around RF circuits!
You save done of my modules in 3rd year Mechatronics :D thank you
Many thanks for your notes and video.
Wow, they say there's no such thing as a free lunch but the cascode amplifier seems to get pretty close. For one extra transistor you get a nice bandwidth boost!
Yes, but only if the C-B capacitance is the source of the bandwidth limiting.
Thanks as usual for a great overview. 73 - Dino KL0S
Gain of a mosfet based Cascode is not the same as common source amplifier as stated at 1:30ish. The gain can be as high as (gm × ro)^2. Largely it depends on what the values of Rd and RL are. If Rd is implemented as two stage cascode load using pmos or the circuit is biased using a current source with very high output impedance, the gain can be (gm × ro)^2. The big deal with cascode circuit is that one is able to boost the gain while not having to sacrifice the bandwidth. This happens because the voltage at drain of common source transistor is close to Vin since that voltage is set by gm1 times [~1/gm2 in parallel to ro1] which works out to be close to 1. Including gmbs in the circuit does not make much difference as that effect too cancels out. So, in effect, the voltage gain between gate and drain is almost 1 which ensures that the Miller effect on Cgd is non existent.
The "only" thing you lose, and this really does matter when supply voltages are tiny, is the voltage headroom.
For a BJT based amplifier, the gain can be as high as gm1 × (gm2× ro2) (ro1 parallel to rpi2) which is still much higher than common emitter version.
If your signal generator started at 1 mhz; why did the output peak at 2 or 3 mhz?
Shouldn't it have started out high and decreased as the frequency increased?
Is that gain boost at 2 or 3 mhz caused by that parallel RC network in the emitter circuit?
I tried to see the component values off the sheet in the video but at 77 years old my eyes are getting pretty bad...
I would also like to know how he determined the component values to provide the proper bias on the upper transistor.
Thank you for the interesting video...
Thank you for this explanation. Ali Hajimiri's CalTech lectures mentioned cascode amp configurations, so it was nice to hear you explain it here. Question: Is there a good professional oscilloscope that you might recommend from your experience, that will be useful for precision analog, comparator, fast precise rise times, gyrator, and work involving transients?
You've got to be more precise about "fast rise times", as that will determine the bandwidth needed by the scopes and probes. Also, an idea regarding the available budget.
@@w2aew I've generally been looking at 4 channel scopes with 200-350Mhz. $1500-$6000, definitely under $10k. A EE mentor/colleague told me that 500mhz minimum for work which may involving noise and transient analysis but those 4ch 500mhz analog seem to get pricier. There's also some 10x frequency rule of thumb I heard.
In terms of tasks, I will be constructing zero crossing detector circuits on sine waves, going up to 30MHz (that 10x rules would suggest a 300MHz scope). I am interested in coupling oscillators together, hooking them to gyrators and looking at the oscillatory current outputs. Trying to build synthetic large value inductors from op-amps in order to get precise inductance for oscillators. And constructing a few analog computers to compute some results.
Great video! Makes the cascode configuration easy to understand. The one issue that you didn't mention though is that it requires more headroom than a CE amp as I learned the hard way! The CB amp biasing forces the output swing to be more constrained than with a plain CE amp.
One question: Is there anything that can be done to make the amplifier broadband and flat across its passband? I've done a fair bit of simulation of cascodes and can't seem to get much better than about 50-60 MHz BW. I realize that selecting a BJT with a low C_ cb improves freq response, but is there anything more that can be done to improve BW?
Thanks
Eric Everton You will lose a little headroom, but you can usually limit the loss to a few hundred mV by biasing the CB stage so that the CE's collector-emitter voltage is a few hundred mV above saturation. Of course, you'll sacrifice a little BW this way because Ccb will be larger with less Vcb. BW is generally improved with larger Vce in the CE stage, as well as with some degeneration in the CE stage (will do a lot for passband flatness). Proper selection of transistors with low capacitance and high ft is required, as well as operating at generally larger Ic with lower value RC, will all help to increase BW.
This Video is great! Especially The comparison! Thank you :)
Thanks a lot for your videos. Please do a video on phase sifter using variable capacitors.
Alan,
Would you please update the links for the PDF's of your notes on this video and video 188 as well as any others that may be linked to your old site.
Do you have a complete repository or zip of the notes somewhere?
Thanks
Sam
W3OHM
Very well explained, thanks!
Thank you for this great tutorial!!
Another great video
how about a video on increasing miller capacitance with an additional capacitor on a common emitter amplifier to limit bandwidth, done commonly in the voltage amplification stage of op amps to improve stability
Learn something new today ! Thanks
Brilliant explanation - thank you
Question: Can this or a similar technique be used in constructing a Colpitts Osc. to improve its bandwidth?
If it was the C-B capacitance of the amp that was limiting the bandwidth, yes.
@@w2aew My goal is to improve the linearity by increasing the bandwidth. Ideally, I would like a relatively flat response in the FM broadcast band.
I've tried several biasing configurations which I'm not sure qualify as "cascode" or not, but using two transistors instead of a single device does seem to help overall.
see sim at :
www.falstad.com/circuit/circuitjs.html?ctz=CQAgjCDMB0YEwFYCmBaMAGEdYA5IYTABYwcA2OdHIskBWhOzZNMAKABMQUdMiic3MLX6C44AHJF06ImwBO4Ms3hK+AkJgxsA7lEp0AnOMgGyAdnHo2AJW68Qo7oZEbMfLVc2boCXY40LcScEY002AHMAwSDHMiIQWPdIoREXVLpVZL14un5EhKT-YWZzWnSEMvC9EszxCqz-U2Y4QWa6ZXCAYxAK-IqqvmhzVDAINHRoC2V0MENIeaoiOATJmXYehvqGfKGR1nBJ+jH8ZaJzSBxCQTXZth64SzoqykFK2i1UQQnoWUgySAIaRwQwISD8CC3DZQahxBI0eG5T4ocyOaDLQFgSwkOatYQ+dYKLBPd7E8QUVaHGTWRSPYIaOl1bwYalsAA2aiZtVJWl81P5ApkcFQqJQt3QCDgFgQ6AuOClJGkNzYABcyc9aO0eeAfAJZOYiJVIODjoZVlNTP8yGCAUDDPKqSAOEgAGYAQwArmyVcVOqTXhrquqnO0nNYooyKer4u4g7kgQk+qsie1Yk4klSmmYnrkiopak40K4bpmAPJwisIxwybzWPSh9Khtz3DKF-FEdLI1FjaCPcE0RBLeVgRhQltasIhMJdnwg0wuBDy9BWsiGVFj-OdJyMsPgXjWNWMwgmAzH5lCaDGY2WM3W5dwKwWuYAyiIczzCi0FlO12e71Eo9VCPQYqXDGEEjPS4INaWtVX0dQ2lhXdxnMaAcGxIEwRwbDDQsHw7zXQxSAWMhyFCR1nXdL0fT0ItolbGs6wYzBIFkatsnA9jOKIUxwkUKCKwEs9MAfcVBX5SAiSraT4yyUCUirM9FJg5J+LY6QWKQxjHHMNgAHssC-bwSHQBgfEhcA2EucBxAAMQssBRggfYxiEEAAGE3QABzdLoAEsVTdAA7LokCswRXPs5knJAFzxggTyfP8wKQrCoA
Excellent video. One question: it looks like the collector of the input device is clamped at Vb-Vbe. If this is so, why is the Miller effect menationed at 2x ? Won't it be 1x ? Regards, from NZ, Mark
It's because the impedance looking into the emitter of the CB transistor is not zero. It is 1/gm. Thus, the CE stage has a gain of -gm*(1/gm), or -1. This makes the miller magnification equal to two.
This is a little off the subject, but you are the one that would know. I am trying to convert the 19.2khz scan rate from the crt output on my HP 8921a to feed into a modern LCD monitor. I was wondering if you have a schematic for a converter to achieve this? Great video by the way, keep up the good work!!!!
James skippinhopper I'm sorry - I don't have a circuit for that on hand.
Great, thanks. I learned something!
Thank you for the excellent explanation. But I'm still confused as to why there's still a multiplication of C_cb of CE stage by 2. CE stage gain is v_out/v_in, isn't it? The v_in is the change in base voltage, and v_out is the change in collector voltage. But, the collector voltage of CE stage should be constant at (V_b - V_be) of CB stage. So, the v_out of CE stage is always 0, and the CE stage gain is 0. Therefore, there should be no multiplication of C_cb of CE stage, and no Miller effect at all.
The gain of the CE stage is -1, which means that the collector voltage moves in the opposite direction of the base voltage. Thus, the change in voltage across the c-e capacitance is double what it would be if the capacitance was simply from base to ground or collector to ground - thus making it appear twice as big.
Fascinating, thankyou. Would there be any application in an audio-amplifier for cascodes?
There can be a benefit in lower distortion: www.passlabs.com/technical_article/cascode-amp-design/
@@w2aew I'm sorry. I didn't google it myself. That's a great link and a fascinating read. I'll simulate some cascodes on an amp I'm working on later. Thanks again...
thank you this is so easy to understand!
Great video Alan. 73s.
Great video!
Thanks again for the video. What would be the effect of using a diode instead of the metal jumper to replace the common base amplifier?
jlfqam Not sure why you'd want to do that - all it would would is add a DC offset to the CE's collector voltage, which would limit the maximum signal swing because the transistor will be that much closer to saturation.
***** Thanks, I just wondered about be the effect on the frequency response of the system compared to the jumper and the CB transistor sets.
Should it fall in between?
jlfqam I would think it would be very close the the jumper result, in fact maybe even slightly worse because the voltage on the b-e junction will be lower, which will increase the capacitance.
Really great vid. Thank you!
Could the CB capacitance cause the a circuit to self oscillate under some circumstances? If so, then maybe that would be an interesting idea, to try to force a transistor to self oscillate at, I guess, its own resonance frequency.
And furthermore, what's that that device with two parallel coaxes with inductors connected to ground? (The on with copper-clad, not the angle bracket.) A choke?
Collector-Base capacitance is usually not a source for oscillation (in fact it's often used to help make a circuit more stable). The device you are referring to is a directional coupler from a video I did a few months ago.
Gameboygenius if you want a self oscillating transistor, look at MOSFETs with no daming resistor on the gate. the lead inductance and the gate capacitance can cause a lot of ringing. You need some feedback gain to maintain oscillation though. In a bigger circuit with gain elsewhere and poor layout they can oscillate seemingly by themselves.
+Dr.Lecter The pronunciation of Thevenin might be region specific. W2AEW is American and I am Canadian, and I was taught in university, that that's how you say it (as the video said it). Perhaps different regions might pronounce it differently.
efox29 You can pronounce it the way you like but there is only one correct pronunciation.
efox29 W2aew is not far off btw ;)
Thank you, it bugs me to no end when someone corrects another like that. It's not called for and I think it's not polite. On a side note; English is a living language and changes constantly.
do you have a technique for measuring phase noise of A 500 Mhz carrier with a spectrum analyzer?
Given that all of the gain here is coming from the common base stage, why keep the troublesome common emitter stage? Is it purely to ensure that we have high input impedance (or control over it, maybe)?
Yes, mainly an input impedance issue (CB amp has a much lower input impedance).
Thanks. A bit of googling shows that, in the 60s, there were 4 terminal "cascode transistors" (two transistors on the die) which I guess made sense, since that would allow better matching of the g_m of the transistors, which you are cancelling in your derivation of the gain.
What does the slope of the signal look like down in the audio range?
Dennis Hill This was designed as an RF amplifier, so it has very little gain at audio frequencies. Of course, component (capacitor) values can be adjusted to improve the audio performance. The Miller effect is generally not an issue for audio amplifiers unless the source impedance is very high, so the cascode isn't used much for audio.
Nice video - don't remember any of this stuff from my EE classes way back when...
More current flow through Ccb, due to a large voltage swing at some frequency across it, has the "effect" of as if it were the gain times larger capacitance. Same impedance (fixed or single frequency) with a larger voltage equals more current (okay). So more current . . . so how does it work? It must somehow make the input impedance smaller to the source, so less input to be amplified? The Ccb doesn't actually change value. Or is it just more negative feedback due to large voltage swing, yielding less input vs frequency? Where does (what path) it shunt the signal to, like a low pass filter to ground, via this capacitance (or is that the "effect" or analogy)? Sure, at the same frequency, a larger capacitance would have lower Xc and draw more current. That is the effect of the higher voltage, but how does it work in the circuit . . . neg feedback, shunting, lower input impedance? Nice demo. Nice to have lots of test equipment. Thanks. Still learning.
ive looked this up because they mention it in some amps , but this is a small effect that will only roll off a bit of highs , but guess what in HI FI some roll off of highs is basically what ppl like .
I guess in highly technical measurement stuff this would have to be fixed but in Audio , i doubt its noticeable . Did anyone do these tests , and knows how much is it exactly ?
is this a capacitance multiplier? it is right. man i been meaning to make one of these
When using higher gain transistors why do technicians add miller capacitors pf across the collector to base which they say it helps reduced the gain noise hiss? The miller capacitance will reduce or eliminate the gain noise any reasons why it does this?
The miller capacitor reduces the circuit's bandwidth.
@@w2aew yes you said that in your video but other techs have told me that it also reduces gain hiss noise not sure if that is true or not that miller caps help with reducing the noise floor?
How do you deal with frequencies in the GHz range in an amplifier? Like in 2.4 GHz or 5 GHz WiFi signal boosters?
Microwave amplifiers certainly require suitable transistors and especially proper physical layout techniques.
Great video - thankyou
Do FET's suffer from Miller effect, and is there a cascode circuit for them ?
yes and yes.
Thanks for the video!
It's a pity that time didn't permit because the more interesting effect of Miller feedback is when the collector load is not a resistance. It turns out that if the load is capacitive then the feedback capacitor appears as a resistance across the input. If the input came from a tuned circuit then it would be loaded down by this resistance. If the load is inductive then the feedback capacitance appears as a NEGATIVE resistance which could cause oscillations at the input - especially if the input came from a tuned circuit. It is no wonder that neutralization was such a big issue in the receivers of the 1920s.
Love your videos......
And I was complaining today about the performance of my NanoVNA that22 has a drop in amplitude of about 2.5 dbm between 50k and 300M... Now it looks much better.
Nice explanation. HP Voyager user as well I see :-)
Yes, my favorite calculator.
@@w2aew Have you tried the modern Swiss Micros HP15C copy? Beautifully made and a faithful repro of the original. They also do a lovely copy of the HP42 with a larger LCD screen and 4 stack levels displayed. Viva RPN!
@@timcoates1925 I've seen them, but never pulled the trigger on one.
Gr8 work!
Thank you so much from.. India
A simple RC low pass filter rolls voltage off at 6db per octave. Is the roll off different for a Miller effect, voltage gain limited CE amplifier. It seems that it must be a lower db per octave because the capacitor is effectively variable and falling in effective value, as frequency rises, until the voltage gain approaches -1. What do you call such a reduced roll off response? A half pole roll off?
jpopelish I think it's about the same as a simple RC. The capacitance isn't effectively falling, since the *gain* isn't changing, just the signal amplitude. I'd have to examine it more closely to be sure though.
But an amplifier operating in the frequency range where Miller effect is in control of its gain, certainly shows changing gain, versus frequency, and that changing gain must have a changing multiplier effect on the capacitive (Cbc) feedback. So, changing RC time constant across input signal, versus frequency. Will your scope show a log plot of amplitude versus frequency?
jpopelish I think of it as the input being attenuated by the RC filter, not that the gain is being reduced. Again, I'll have to look at it again when I get back to my lab next week...
***** Never mind. It just occurred to me that the same reduction in voltage, as frequency rises, across the capacitor occurs with a fixed capacitor in an RC, low pass filter. I just had a mental block, because of the gain produced voltage in the CE amplifier. If, somehow, the voltage across the capacitor were forced to be independent of frequency, the result would not be a low pass, but a notch filter.
Why do FETS and Transistors have "Miller Plateau"? if you look at the turn on/rise time the middle region you will see the Miller Plateau.
Cascode also eliminates the Early effect making the gain independent of power supply voltage.
GREAT video, thanx!! Nothing more to say about it! EEVBLOG - µRuler spotted ;) ;) ;) ;)
Please give me a design of a Cascade amplifier whose input is 54 v and output is 540v.
A vital component is the ferrite bead on Q1 emitter. Without this the circuit will most likely oscillate at maybe 250 Mc/s
for not so worry about the miller effect but look for way to increase head room.
A cascode takes away from headroom. To increase headroom, you want to setup the bias conditions so that the collector voltage (with no signal applied) is halfway between the positive supply and the emitter voltage.
The name has always puzzled me, according to wikipedia it comes from "cascaded triodes having similar characteristics to a pentode"
David Pilling it certainly comes from thermionic valves / vacuum tubes. You get various -odes. Triode, tetrode, pentode etc. which you can cascade to make a casc-ode. There's nothing to say you can't cascode tetrodes or pentodes though. But yes, pentodes have a reduced miller effect because their control grid is screened from the anode by the screen grid IIRC, so 2 triodes in cascode configuration would sort of be pentode-ish in that one single regard.
Thévenin is actually pronounced "tevenin", rather than how you said it.
I'm sure the French telegraph engineer would have pronounced his own name 'tevenin' -- and that should probably be persuasive -- but not many people are aware of the subtleties of french pronunciation.
Sir is CB-CE amplifier possible?
Yes.
@@w2aew ok sir
Nicely done! de KQ2E
Apa itu Efek Millar ?
Efek Millar sering kali menjadi faktor pembatas bandwidth utama dan desain penguat common-emitor, alasannya karena perbesaran efek kapasitansi parasit. Jadi, jika ada transistor yang memiliki kapasitansi di antara basis-emitor, basis-kolektor, dan kolektor-emitor maka setiap kapasitansi ini pada dasarnya mencuri arus dari transistor. Dan oleh karena itu, akan mempengaruhi bandwidth dan kinerja transistor.
Sorry found not find. Damn small virtual keyboards
nice beefy vise you have there.
I love that old vise. Very stable for building and testing circuits. It's tough to find a nice heavy small machinists vise like that anymore.
nice thumbs up
god bless