For you it was 2010. For me it's 2019 and very helpful for starting to understand and pieces together the knowledge and mysterious world of electronics. Explaining things primarily in the simplest forms really helps, thanks eh'.
Mecheng here. I had the opportunity to choose my subjects when I did the study abroad program. I decided to throw myself in the deep end and do a final year analogue electronics theory course. It was one of the most interesting subjects I have ever done. Learning to model these circuits in the frequency domain was super useful later on when I had to do modelling of mechanical systems!
This was exactly the video I needed to see after reading on Wikipedia and watching images of disassembled switching mode power supplies. I went from confused to relaxed in a single video! Thanks!
I think for the LDO with PNP as the pass device. The negative feedback should be connected to positive polarity of the amplifier. Because the PNP already shift the phase by 180 degree. The rest of explanation were very good. Thank you.
There's a huge glut of small buck converter modules based on the LM2596 on eBay. For around a pound (a couple of bucks) you can play with them, blow them up and learn all you need to know. Great fun. And for a bit more cash you can get them with LED voltmeters and adjustable current limiting.
I wish I had YOU as my teacher in engineering college. Infinitely much better than the idiots who where so obsessed with theoretical proofs, math exercises that were only tailored to their solution. Then looking up an actual device datasheet showed features that made half of the learned theory not applicable. 95% of what I'm capable of today I just learned from internet resources and a few books which I bought myself and a little critical thinking.
While I sympathize with the sentiment, those obsessed with the math and physics help push for funding of research that leads to these electronics technologies being manufactured.
Same here. Most of what I'm using today I learned by staying in the labs until 5 in the morning tinkering with electronics, watching videos from Dave and Louis Rossmann and probing them with O-scopes and DLAs. the theory is certainly important, but too much emphasis is put on cranking out 3 page proofs and solving complex circuits by hand. More effort should be given to encouraging actual design work and teaching students how to THINK like engineers.
LDO’s uses a PNP transistor to allow lower loop saturation during propagation in a switch mode also a great explanation of linear efficiency vs switch efficiency
You, i like you, you are enthusiastic, and extremely explanitory. You break things down into the smallest possible pieces, and in any good understanding of electronics, modular learning is an absolute must! so props to you, and maybe one day when i get out of binary absolution wiring, and into digital IRL wiring, i will learn much more from you lol
Thanks for adding that bit about efficiency at the end there. Other videos says that linear are awfully inefficient but obviously it depends on input and output voltages.
Thank you very much for the insightful tutorial, mate. I am working on a 12VDC to 240VAC inverter for my car and was not sure whether to use a switched mode or linear. But this has helped me make the decision. I am pretty sure I will go for the switched mode. Purely because the 1000VA iron core transformer I have weighs about 4 kilograms and will totally defeat all possibility of making the power inverter even remotely portable. Thanks again and please keep these videos coming! :)
Thanks Dave. From my very limited experience with switching regulators I'd say that the tricky part is finding the right value for the inductor. In one case when I wasn't getting the proper output voltage (it was too low) I had to make the inductor value *smaller* (10uH instead of 100uH, for example), which seems counter-intuitive. ...continued...
@Films4You Yes, there are many different techniques for SMPS control, this in only one of them. Fixed freq PWM, fixed on time types, and more exotic ones exist.
Thank you Dave! Learned alot from that one. Would love to see some actual design on boost type, powering a 5v circuit from say a lower voltage, how you calculate current needs and what it will take from the battery or power supply. Like your LED head torch, what it would take to power a high current LED from a small input power, and how to keep the current somewhat constant to the end of the battery life or something. Would certainly be interesting.... But time is limited I know :(
I made a very discrete DC/DC Buck Converter ONLY with BC547, going to up to 95% PW, so if you input 150V, you can still get 140V out stable. Needs 2 isolated 100mA supplies, which makes it a bit more complicated, but it can switch 1000WATTS! Just needs 1 external Power transistor with a gain of at least 40 withstanding the voltage and current, huge inductor and POOF. 150 to 250V to 140V Stable. Efficiency of over 90% in ANY case!
A zener diode is usually built into the silicon of the regulator. It can be made to give approximately any voltage you desire (5 or 10 or 3.3 volts) for reference.
I love your shows and its my biggest inspiration for taking up electronics project back again after being in IT for a while. But you gotta write a book or something regarding these, it would be a life saver for college students who are dying coz of those over complicated "piss poor dog turd" of books out there... keep on rocking!
Anirudh Vasudevan We are in 2015, this youtube Channel is one of the best books about electronics in the history. He doesn't need to write a real book. Interactive, very up-to-date, accessible from anywhere in the word, "free", extremely practical, with pages of comments in every chapter and written by someone who actually used the electronics in real-life (not some whole-life teacher without any practical experience). Which book could possible beat that?!
Last night I threw together a custom buck converter (a first) using a PIC, a MOSFET, an arbitrary inductor, a capacitor, a bulb for load and a Schottky diode. No detailed calculations, just 1 hr of messing around and I easily achieved 75% efficiency...switch mode regulators were perceived to be scary critters but after yesterday's experiment they seem a little less intimidating.
Dave, did you ever made a video on the principles of something simple as a fuse? Like why they blow and how their ratings are designed? I'm still unsure what really makes a fuse go blow. Just the amps? Then why there is also a voltage rating on most fuses? Grrrr. I don't get it together
+Axel Werner Power dissipated in a fuse will blow it. The fuse uses a really thin wires that will melt when it gets too hot. Because it has a finite resistance, when a current flows through it there will be a small voltage drop, that voltage drop is directly proportional to the current. Power dissipated by the fuse is V*I, I^2*R, or V^2/R. The voltage dropped across the fuse is small, it has to be small. We don't want to waste power. However, when it blows, it effectively becomes an open circuit. The full voltage will be present across it, and if that voltage is too high then it will arc over from one end all the way to the other end and that can cause serious damage to everything closeby. Because fuses do not blow instantly (the wires do have thermal mass) it can take some time to heat up enough to blow. Even a really large current of many orders of magnitude can flow through and it will still take a small amount of time to blow. If a huge enough surge of current flows through the fuse, it can actually explode. That would be the rupture rating of the fuse. HRC fuses are designed to be relatively safe in this regard and have really high rupture ratings, however, exceed that, and they can also explode.
Nice vid, i always understood LDO regulators to gain their efficiancy from the use of a FET rather than a BJT. LDO regulators have got NPN as well just in the form of a FET instead with a gate driver of course which makes an LDO a little more complex internal circuit.
If you want, I would recommend Linear Technology (www.linear.com). They offer simulation software so you can test transients and stability and all that jazz and power supplies are kind of their thing.
I've just been getting into electronics and in between going through various tutorials one video at a time, I also like to jump around on different topics and one topic I've become interested in is boost converters and various other switch mode power supplies. I've never heard of linear power supplies before but you can bet I'll be learning about them real soon. My point is, I went into this with no expectations so I wasn't afraid of tackling switch mode power supplies before anything else. I didn't find them terribly difficult to understand, especially after playing with a circuit simulator app. But Dave says most newbs are afraid of them. So I dunno. I just thought I'd share that.
Very good video for beginners, I hope you could add some more detailed explanations on switch-mode regulators, such as buck, boost and buck-boost regulators.
From min. 9:13 onward, a mistake has occured in your presentation! You've just changed the Darlington NPN(common collector- non phase inversion) with a PNP(common emitter-phase inversion). Note that PNP inverts the phase while the NPN-Darl. doesn't. As a result, the LDO type (with PNP) must have the inputs of OpAmp swapped, i.e. reference voltage tied to inverting input(-), and feedback tied to non-inverting input(+) in order to maintain the negative feedback loop, otherwise it will latch-up(positive feedback!).
I think he just wanted to keep it simple and straightforward, so he did not change much else, knowing not many people would probably try to rebuild the circuit. I did, and found that out myself the hard way lol! Once I realized that the PNP in that config. get more and more 'turned off' as the current into the base is increased (it is inverting) I tried inverting the op amp inputs and voila, I got that circuit to work too! I have also had success laking an LDO w/ MOSFETs, though they like to oscillate parasitically!
i've managed to make up my own adjustable regulator using a beefy 10amp power transistor and a 741 opamp and a couble of silicon diodes for reference, the low end voltage comes out about 2 volts and near supply voltage at upper, you get about a few hundred mV dropped depending on load but the transistor doesnt even break a sweat even with heavy current draw
...continued... Another gotcha with the high frequency (900kHz, for example) switch mode regulators is that they will require high capacity (~22uF) _CERAMIC_ capacitors on both the input and output. These aren't something most of us hobbyists have in our kits. You can get away with electrolytic caps if the regulator's frequency is ~100kHz or lower so you might stick with the lower frequency regulators if you're just experimenting. ...continued...
@Albinorama I agree that was great, in fact I wish it was even more in-depth !! industry chat and annecdotes are great to but they're more entertainement and trivia while this is pure unadulterated learning & experience !
pretty helpful. im working on a renewable energy source pulse which i have to smooth and step down with minimal power loss from 100v to a smooth 24v dc line
Dave when you said "So what's all this bullshit I was talking about that SMPS aren't that much different from Linear Supplies?" I nearly burst out laughing. Cheers man, always love watching your vids.
Thanks Dave. But in my opinion the reason new people are reluctant to using switching PSU's is not because they don't know how similar they are to linear ones, that doesn't matter that much. They are reluctant to use them because of the inductor. Resistors and capacities are easy to buy and use, but inductors aren't.
+Mac Cartier Transistors are not binary in nature. In simple terms, when used digitally they are either cutoff or saturated. When used linearly they are biased at some point between cutoff and saturation so they can be used as amplifiers for example.
The first question that came to my mind after seeing Dave's linear regulator diagram is: how is a fixed reference voltage produced when the input to the regulator can be variable. e.g. how does a 7805 regulator produce a fixed 5 volt reference for the op-amp if the regulator input is anything between 7 and 20 volts?
In this video the SMPS is not much difference and still easy to understand because it is just Buck Converter, one of the easiest SMPS. What is hard and intimidating is the real power supply that supply high current from 220VAC, such as what you see in PC ATX Power Supply. The ATX Power Supply especially above 500W is intimidating and very complex because it is not just about Buck Converter, it is more than that: Power Factor Correction, Transient Damage protection, 3.3V, 5V, + and -12V regulated power supply each with overcurrent, overvoltage protection including step down transformer using Half Bridge topology. The circuit diagram is immensely complex. So, the real high power SMPS that we encounter in real life is not that simple.
I noticed something strange: the basic schematic for an LDO linear regulator should have the + and - inputs switched, otherwise it creates a positive-feedback loop and is just a latch.
Thanks again Dave. Great stuff! Inductor selection is a big thing also - as is the size of some of them. Not that I'd want you to do a blog on inductors, just saying it's a 'Jedi' type skill of analog designers maybe.
Linear regs burn a lot of energy as heat and are very inefficient (7805 is only about 60% efficient w/ 12vin, for example), while buck step down switchers can be up to 95% (and some even higher) efficient due to energy storage capabilities.
It doesn't really matter what transistors they use, they still operate the same way. MOSFETs are typically used in SMPSes because they dissipate less power when fully on (milliohms D-S * I^2) than a bipolar does in saturation (about 0.3 volts C-E * I).
linear power supply often have a transformer in front,and they often are big and low on efficiency,that eat a lot of power and work at fix frequencies like 50 or 60 Hz Depends on where you live.
I'm gonna go out on a limb and say that most SMPMs today don't use BJTs, but MOSFETs, due to the fact that junction behaves like a resistor and not a diode (Constant resistance instead of constant drop-out voltage) and has a really low resistance when saturated. Maybe not really relevant to mention in this video though. It's 17 minutes already.
That was a grate tutorial! But I have a little question, that: Capacitors are also the energy storing devices, so can we replace that series inductor in switching mode buck converter with a capacitor in shunt to the output (by the principle of duality) or similar in other topologies? Will it work? Thanks in advance!
It will theoretically work into a consumer which has a high-impedance power input. But you can only have high-impedance inputs if they are not current-carrying, because you'd be losing a lot of power there. It may work with compromises, if the device you're powering needs little power, the voltage drop you're trying to achieve is small, and the device doesn't actually care about sitting on a rather noisy power rail. Consider, the output of your switching IC is basically a bunch of sharp current bursts. And you have designed a junction there where that current can go into your smoothing capacitor to ground, or it can go into your powered device. Naturally, you want your device to receive a smooth regulated voltage on the input, and you want the sharp edge of that burst to go into your smoothing capacitor. But the only way to accomplish that is if you can guarantee that the impedance of your capacitor to ground path is infinitely lower than the impedance of the power input of your powered device. So a choke is the right choice here, it will limit the sharp current bursts regardless of what's behind it.
...continued... Finally, keep in mind that a buck regulator's max. duty cycle is going to limit the maximum output voltage. If the regulator's max. duty cycle is 90% then the maximum output voltage is going to be (at best) 90% of the input voltage.
you explained this perfectly. although my only question is would the switch mode be more prone to wear and tear due to the constant switching on and off. both to the actual switch itself and the NPN?
Respected Sir, It was great experience when I watched the video. It definitely helped me a lot to understand the basics. I want to thank you for the same and want to wish you a very very long life so that you can give us your deep experience. But I hopped that you will show some oscilloscope live wave form of linear & switch mode which may be more exciting or not I think you may understand better. Best wishes and thanks.
I thought the hard part of SMPS was selection of external components; IE, the required inductor rating is dependent on both expected load and voltage (and must not be too big or too small), and that the capacitor must have a specific ESR and capacitance as well, and it must be able to tolerate current at the switching frequency, and then there's the parasitics from traces... and if anything is off even a slight amount, it will apparently cause a negative vacuum gradient that will destroy the entire universe. At least that's what I've been told.
Could you do a video on how to make a completele SMPS? Values on the output components and such? I've also read that SMPS might need some carefull board layout aswell
in 4:29, you explained that since Ref. Voltage was equal to Output Voltage, it will turn on NPN Switch to provide IN Voltage to OUT Voltage. But if we just apply input voltage ho to get FeedBack Voltage from OUT to open NPN?
Hi Dave. The output voltage always has to be Vo(Op-Amp output voltage) - 0.7V(transistor vbe drop). This dictates the maximum Vout, since the Op-Amp's output voltage (Vo) swing can be = supply - a few volts at max. But Vout = Vin - vce(transistor). So Vout is always the min of these two constraints?
Do LDO regulators use a Szikali pair pass transistor stage? This will not need an inverse connection for the op-amp as the driver transistor is still a NPN transistor. I do get confused about the forward gain of the Szikali pair. I think it's mainly the driver transistor in this calculation.
You put an op amp there, but almost every representation of a linear regulator I've seen use a zener diode and there's no mention of an op amp. Also I understand the inductor part of the switching regulator, but I want to know more on the oscillator/switching. Everybody just put a magical switching box there and don't explain how it works!
Hey Dave, Good day.. if your switch mode power supply is trying to turn ON again after the input voltage have been removed? what might have been causing it? on the other hand, on your LDO, if there's a fluctuation in your input voltage will it's output have same waveform as it's input?
Hi Dave, On top of the question that Ricardo has asked below, I'd also like to ask why a PNP has a lower dropout compared to an NPN. Would it be due to the fact that in the case of the PNP, the voltage drop across the transistor Vbe junction is much higher ?
![ROSÉ (로제) - number one girl [더 시즌즈-이영지의 레인보우] | KBS 241129 방송](http://i.ytimg.com/vi/A9oByH9Ci24/mqdefault.jpg)
everytime i look for a tutorial, and i see EEVblog has one, i get really excited because you just know it's gonna be very good :)
always lol
Yup
For you it was 2010. For me it's 2019 and very helpful for starting to understand and pieces together the knowledge and mysterious world of electronics. Explaining things primarily in the simplest forms really helps, thanks eh'.
Mecheng here.
I had the opportunity to choose my subjects when I did the study abroad program.
I decided to throw myself in the deep end and do a final year analogue electronics theory course.
It was one of the most interesting subjects I have ever done.
Learning to model these circuits in the frequency domain was super useful later on when I had to do modelling of mechanical systems!
This was exactly the video I needed to see after reading on Wikipedia and watching images of disassembled switching mode power supplies. I went from confused to relaxed in a single video! Thanks!
I think for the LDO with PNP as the pass device. The negative feedback should be connected to positive polarity of the amplifier. Because the PNP already shift the phase by 180 degree. The rest of explanation were very good. Thank you.
There's a huge glut of small buck converter modules based on the LM2596 on eBay. For around a pound (a couple of bucks) you can play with them, blow them up and learn all you need to know. Great fun. And for a bit more cash you can get them with LED voltmeters and adjustable current limiting.
ill check it out thanks
I wish I had YOU as my teacher in engineering college. Infinitely much better than the idiots who where so obsessed with theoretical proofs, math exercises that were only tailored to their solution. Then looking up an actual device datasheet showed features that made half of the learned theory not applicable. 95% of what I'm capable of today I just learned from internet resources and a few books which I bought myself and a little critical thinking.
I've had some pretty good professors. Although I started at a community college so I don't know if they're part of educational Illuminati.
same here
While I sympathize with the sentiment, those obsessed with the math and physics help push for funding of research that leads to these electronics technologies being manufactured.
like everywhere learning is your personal responsibility.
Same here.
Most of what I'm using today I learned by staying in the labs until 5 in the morning tinkering with electronics, watching videos from Dave and Louis Rossmann and probing them with O-scopes and DLAs.
the theory is certainly important, but too much emphasis is put on cranking out 3 page proofs and solving complex circuits by hand. More effort should be given to encouraging actual design work and teaching students how to THINK like engineers.
My Thesis project would hadn't been possible without this blog
I wIsh I could visit you in Australia and thank you personally for all the knowledge you have given me. You are a great instructor. Thank you !!!
LDO’s uses a PNP transistor to allow lower loop saturation during propagation in a switch mode also a great explanation of linear efficiency vs switch efficiency
You, i like you, you are enthusiastic, and extremely explanitory. You break things down into the smallest possible pieces, and in any good understanding of electronics, modular learning is an absolute must! so props to you, and maybe one day when i get out of binary absolution wiring, and into digital IRL wiring, i will learn much more from you lol
That makes the whole subject of switch mode much less worrying and difficult. Thanks, Dave.
Very impressed with the way things have been explained in such a simple manner. Thank you!
Thanks for adding that bit about efficiency at the end there. Other videos says that linear are awfully inefficient but obviously it depends on input and output voltages.
Thank you very much for the insightful tutorial, mate.
I am working on a 12VDC to 240VAC inverter for my car and was not sure whether to use a switched mode or linear. But this has helped me make the decision. I am pretty sure I will go for the switched mode. Purely because the 1000VA iron core transformer I have weighs about 4 kilograms and will totally defeat all possibility of making the power inverter even remotely portable.
Thanks again and please keep these videos coming! :)
Hi Dave,
please could you do a more in depth video on switch mode power
supplies/chargers with some scope measurements so we can understand each step?
Another very good video from Dave comparing linear versus switch mode regulators. Thank you !
This guy is the most awesome teacher on the planet
Thanks Dave.
From my very limited experience with switching regulators I'd say that the tricky part is finding the right value for the inductor. In one case when I wasn't getting the proper output voltage (it was too low) I had to make the inductor value *smaller* (10uH instead of 100uH, for example), which seems counter-intuitive.
...continued...
This is the best tutorial I have ever seen. Thanks Dave
*Woah old school! He has grown changed so much and he lets his real voice come through in new videos*
Technical college. Dang, I finished that over 18 months ago! I'm up to designing basic circuits now. It's insane how much I have learnt since.
@chriscchrisc
Look up "Bandgap Voltage Reference". It's easy to get a 1.25V reference using the inherent properties of the silicon substrate.
@Films4You
Yes, there are many different techniques for SMPS control, this in only one of them. Fixed freq PWM, fixed on time types, and more exotic ones exist.
Thank you Dave! Learned alot from that one. Would love to see some actual design on boost type, powering a 5v circuit from say a lower voltage, how you calculate current needs and what it will take from the battery or power supply. Like your LED head torch, what it would take to power a high current LED from a small input power, and how to keep the current somewhat constant to the end of the battery life or something. Would certainly be interesting.... But time is limited I know :(
What a great educational video. I'm about to learn about SMPS for my next assignment, and this was perfect to get some prior knowledge.
I made a very discrete DC/DC Buck Converter ONLY with BC547, going to up to 95% PW, so if you input 150V, you can still get 140V out stable. Needs 2 isolated 100mA supplies, which makes it a bit more complicated, but it can switch 1000WATTS! Just needs 1 external Power transistor with a gain of at least 40 withstanding the voltage and current, huge inductor and POOF. 150 to 250V to 140V Stable. Efficiency of over 90% in ANY case!
A zener diode is usually built into the silicon of the regulator. It can be made to give approximately any voltage you desire (5 or 10 or 3.3 volts) for reference.
with your manner of speaking and the way you simplify things, you can probably teach this on sesame street..
I love your shows and its my biggest inspiration for taking up electronics project back again after being in IT for a while. But you gotta write a book or something regarding these, it would be a life saver for college students who are dying coz of those over complicated "piss poor dog turd" of books out there... keep on rocking!
Anirudh Vasudevan We are in 2015, this youtube Channel is one of the best books about electronics in the history. He doesn't need to write a real book.
Interactive, very up-to-date, accessible from anywhere in the word, "free", extremely practical, with pages of comments in every chapter and written by someone who actually used the electronics in real-life (not some whole-life teacher without any practical experience). Which book could possible beat that?!
I just finished TAFE for the year (including of SMPS, which was fairly easy). I CAN FINALLY UNDESTAND THIS!
Last night I threw together a custom buck converter (a first) using a PIC, a MOSFET, an arbitrary inductor, a capacitor, a bulb for load and a Schottky diode. No detailed calculations, just 1 hr of messing around and I easily achieved 75% efficiency...switch mode regulators were perceived to be scary critters but after yesterday's experiment they seem a little less intimidating.
thand for the vid.i really appreciate the ldo part i understand them better and wont be afraid to buy the in the future.God bless from Oklahoma city
Dave, did you ever made a video on the principles of something simple as a fuse? Like why they blow and how their ratings are designed? I'm still unsure what really makes a fuse go blow. Just the amps? Then why there is also a voltage rating on most fuses? Grrrr. I don't get it together
+Axel Werner Power dissipated in a fuse will blow it. The fuse uses a really thin wires that will melt when it gets too hot. Because it has a finite resistance, when a current flows through it there will be a small voltage drop, that voltage drop is directly proportional to the current. Power dissipated by the fuse is V*I, I^2*R, or V^2/R.
The voltage dropped across the fuse is small, it has to be small. We don't want to waste power. However, when it blows, it effectively becomes an open circuit. The full voltage will be present across it, and if that voltage is too high then it will arc over from one end all the way to the other end and that can cause serious damage to everything closeby.
Because fuses do not blow instantly (the wires do have thermal mass) it can take some time to heat up enough to blow. Even a really large current of many orders of magnitude can flow through and it will still take a small amount of time to blow. If a huge enough surge of current flows through the fuse, it can actually explode. That would be the rupture rating of the fuse. HRC fuses are designed to be relatively safe in this regard and have really high rupture ratings, however, exceed that, and they can also explode.
Nice vid, i always understood LDO regulators to gain their efficiancy from the use of a FET rather than a BJT. LDO regulators have got NPN as well just in the form of a FET instead with a gate driver of course which makes an LDO a little more complex internal circuit.
You have them swapped. Switching Mode Regulators use gate drivers which drive high-side and (hopefully) a low-side FET.
funkyguy4000
Aaa ok, sounds about right. I am looking at these new Micrel Super Beta LDO regulators at moment. Very low voltage dropout.
If you want, I would recommend Linear Technology (www.linear.com). They offer simulation software so you can test transients and stability and all that jazz and power supplies are kind of their thing.
funkyguy4000
Thank you, i shall take a look see what they offer as Micrel seem to be very expensive regarding their LDOs.
I've just been getting into electronics and in between going through various tutorials one video at a time, I also like to jump around on different topics and one topic I've become interested in is boost converters and various other switch mode power supplies. I've never heard of linear power supplies before but you can bet I'll be learning about them real soon. My point is, I went into this with no expectations so I wasn't afraid of tackling switch mode power supplies before anything else. I didn't find them terribly difficult to understand, especially after playing with a circuit simulator app. But Dave says most newbs are afraid of them. So I dunno. I just thought I'd share that.
Thank you for this. Very helpful. Keep up your great tutorials Dave.
Good video, and a good explanation, but I always feel like I need a nap after watching your videos :)
Great stuff. Dave is a champ!
Very good video for beginners, I hope you could add some more detailed explanations on switch-mode regulators, such as buck, boost and buck-boost regulators.
This is a great video. I learned fast with your explanation. Really appreciate your effort to put this up.
I think this is an excellent way to explain things.
From min. 9:13 onward, a mistake has occured in your presentation! You've just changed the Darlington NPN(common collector- non phase inversion) with a PNP(common emitter-phase inversion). Note that PNP inverts the phase while the NPN-Darl. doesn't. As a result, the LDO type (with PNP) must have the inputs of OpAmp swapped, i.e. reference voltage tied to inverting input(-), and feedback tied to non-inverting input(+) in order to maintain the negative feedback loop, otherwise it will latch-up(positive feedback!).
I think he just wanted to keep it simple and straightforward, so he did not change much else, knowing not many people would probably try to rebuild the circuit. I did, and found that out myself the hard way lol! Once I realized that the PNP in that config. get more and more 'turned off' as the current into the base is increased (it is inverting) I tried inverting the op amp inputs and voila, I got that circuit to work too!
I have also had success laking an LDO w/ MOSFETs, though they like to oscillate parasitically!
I noticed that in first instance too... but well, not important in dis explanation all that little details ;-)
So precisely explained !!!!!!!!!! Love you Dave :)
i've managed to make up my own adjustable regulator using a beefy 10amp power transistor and a 741 opamp and a couble of silicon diodes for reference, the low end voltage comes out about 2 volts and near supply voltage at upper, you get about a few hundred mV dropped depending on load but the transistor doesnt even break a sweat even with heavy current draw
In the case of an LDO, you must invert the input pins of your opamp so the loop stays a negative feedback
We're doing switch-mode power supplies next semester. You need to come give a guest lecture one day :P
...continued...
Another gotcha with the high frequency (900kHz, for example) switch mode regulators is that they will require high capacity (~22uF) _CERAMIC_ capacitors on both the input and output. These aren't something most of us hobbyists have in our kits. You can get away with electrolytic caps if the regulator's frequency is ~100kHz or lower so you might stick with the lower frequency regulators if you're just experimenting.
...continued...
@Albinorama
I agree that was great, in fact I wish it was even more in-depth !!
industry chat and annecdotes are great to but they're more entertainement and trivia while this is pure unadulterated learning & experience !
pretty helpful. im working on a renewable energy source pulse which i have to smooth and step down with minimal power loss from 100v to a smooth 24v dc line
Hi Dave, thanks for this amazing tutorial ! .... it is clear like water :)
Love your channel! Keep up the awesome work!
Very well explained Dave!
You can use 7806 linear regulator together with capacitors or just use LM317 adjustable voltage regulator.
Dave when you said "So what's all this bullshit I was talking about that SMPS aren't that much different from Linear Supplies?" I nearly burst out laughing.
Cheers man, always love watching your vids.
what a great teacher
Thanks Dave.
But in my opinion the reason new people are reluctant to using switching PSU's is not because they don't know how similar they are to linear ones, that doesn't matter that much.
They are reluctant to use them because of the inductor.
Resistors and capacities are easy to buy and use, but inductors aren't.
+Mac Cartier
Transistors are not binary in nature. In simple terms, when used digitally they are either cutoff or saturated. When used linearly they are biased at some point between cutoff and saturation so they can be used as amplifiers for example.
Useful and simple ! Thanks ! Please do upcoming videos like this !
@DagGirl Awesome!, that was my hope.
The first question that came to my mind after seeing Dave's linear regulator diagram is: how is a fixed reference voltage produced when the input to the regulator can be variable. e.g. how does a 7805 regulator produce a fixed 5 volt reference for the op-amp if the regulator input is anything between 7 and 20 volts?
In this video the SMPS is not much difference and still easy to understand because it is just Buck Converter, one of the easiest SMPS. What is hard and intimidating is the real power supply that supply high current from 220VAC, such as what you see in PC ATX Power Supply. The ATX Power Supply especially above 500W is intimidating and very complex because it is not just about Buck Converter, it is more than that: Power Factor Correction, Transient Damage protection, 3.3V, 5V, + and -12V regulated power supply each with overcurrent, overvoltage protection including step down transformer using Half Bridge topology. The circuit diagram is immensely complex. So, the real high power SMPS that we encounter in real life is not that simple.
hi sir , i want to thank you for this easy way to explain SMPS ..
Amazing and clear explanation. Would it be a decent idea to use a MOSFET instead of bipolar transistors? I'd love to know!
I noticed something strange: the basic schematic for an LDO linear regulator should have the + and - inputs switched, otherwise it creates a positive-feedback loop and is just a latch.
Thanks again Dave. Great stuff!
Inductor selection is a big thing also - as is the size of some of them.
Not that I'd want you to do a blog on inductors, just saying it's a 'Jedi' type skill
of analog designers maybe.
Linear regs burn a lot of energy as heat and are very inefficient (7805 is only about 60% efficient w/ 12vin, for example), while buck step down switchers can be up to 95% (and some even higher) efficient due to energy storage capabilities.
+EEVblog Where does the reference voltage come from in a linear regulator? Also, I thought MOSFETs were used in a SMPS.
they have an internal reference, mostly a zener diode oor so
It doesn't really matter what transistors they use, they still operate the same way. MOSFETs are typically used in SMPSes because they dissipate less power when fully on (milliohms D-S * I^2) than a bipolar does in saturation (about 0.3 volts C-E * I).
Well yes your right, but those are in the main part in the power supply.
Thank you so much! Man i really enjoyed you teaching this to me, please dont ever stop! :))))))
Thanks .EEVblog its beautiful and perfect explanation .Regards
please make video in detail on MOV ..how to select it based on package..
linear power supply often have a transformer in front,and they often are big and low
on efficiency,that eat a lot of power and work at fix frequencies like 50 or 60 Hz Depends on where you live.
I knew Dave would have a video on Switch Mode Power Supplies...
But Geez, I'm glad you got rid of that Leyland Brothers intro!!! ;)
I'm gonna go out on a limb and say that most SMPMs today don't use BJTs, but MOSFETs, due to the fact that junction behaves like a resistor and not a diode (Constant resistance instead of constant drop-out voltage) and has a really low resistance when saturated. Maybe not really relevant to mention in this video though. It's 17 minutes already.
That was a grate tutorial!
But I have a little question, that:
Capacitors are also the energy storing devices, so can we replace that series inductor in switching mode buck converter with a capacitor in shunt to the output (by the principle of duality) or similar in other topologies? Will it work? Thanks in advance!
It will theoretically work into a consumer which has a high-impedance power input. But you can only have high-impedance inputs if they are not current-carrying, because you'd be losing a lot of power there. It may work with compromises, if the device you're powering needs little power, the voltage drop you're trying to achieve is small, and the device doesn't actually care about sitting on a rather noisy power rail.
Consider, the output of your switching IC is basically a bunch of sharp current bursts. And you have designed a junction there where that current can go into your smoothing capacitor to ground, or it can go into your powered device. Naturally, you want your device to receive a smooth regulated voltage on the input, and you want the sharp edge of that burst to go into your smoothing capacitor. But the only way to accomplish that is if you can guarantee that the impedance of your capacitor to ground path is infinitely lower than the impedance of the power input of your powered device.
So a choke is the right choice here, it will limit the sharp current bursts regardless of what's behind it.
please do a 1hr blog on every type of switching regulator 😂😂
...continued...
Finally, keep in mind that a buck regulator's max. duty cycle is going to limit the maximum output voltage. If the regulator's max. duty cycle is 90% then the maximum output voltage is going to be (at best) 90% of the input voltage.
you explained this perfectly. although my only question is would the switch mode be more prone to wear and tear due to the constant switching on and off. both to the actual switch itself and the NPN?
Respected Sir, It was great experience when I watched the video. It definitely helped me a lot to understand the basics. I want to thank you for the same and want to wish you a very very long life so that you can give us your deep experience. But I hopped that you will show some oscilloscope live wave form of linear & switch mode which may be more exciting or not I think you may understand better. Best wishes and thanks.
I thought the hard part of SMPS was selection of external components; IE, the required inductor rating is dependent on both expected load and voltage (and must not be too big or too small), and that the capacitor must have a specific ESR and capacitance as well, and it must be able to tolerate current at the switching frequency, and then there's the parasitics from traces... and if anything is off even a slight amount, it will apparently cause a negative vacuum gradient that will destroy the entire universe. At least that's what I've been told.
Could you do a video on how to make a completele SMPS? Values on the output components and such? I've also read that SMPS might need some carefull board layout aswell
The switch mode regulators have 8 pins, but 2 are used for power and ground
Wow DAve.. You know your stuff! Talk about a Lightening Electronics talk!!)
Great work mate~!!
Daaaaamn, 9 years have passed.
Great information and learning from it..
You'd be a great mentor.
in 4:29, you explained that since Ref. Voltage was equal to Output Voltage, it will turn on NPN Switch to provide IN Voltage to OUT Voltage. But if we just apply input voltage ho to get FeedBack Voltage from OUT to open NPN?
Hi Dave.
The output voltage always has to be Vo(Op-Amp output voltage) - 0.7V(transistor vbe drop). This dictates the maximum Vout, since the Op-Amp's output voltage (Vo) swing can be = supply - a few volts at max. But Vout = Vin - vce(transistor). So Vout is always the min of these two constraints?
Do LDO regulators use a Szikali pair pass transistor stage? This will not need an inverse connection for the op-amp as the driver transistor is still a NPN transistor. I do get confused about the forward gain of the Szikali pair. I think it's mainly the driver transistor in this calculation.
4:35 thanks for this man 🙌🏻
You put an op amp there, but almost every representation of a linear regulator I've seen use a zener diode and there's no mention of an op amp. Also I understand the inductor part of the switching regulator, but I want to know more on the oscillator/switching. Everybody just put a magical switching box there and don't explain how it works!
The LDO can not be simple replace by a NPN to a PNP from normal regulator, the feed back path phase is reversed.
For a mosfet S.M.P.S., is there a way to modify the circuit by adding a "pot" to dim the output. The S.M.P.S. is being used to drive a c.o.b. led.
very informative I've learned...Can I ask a question??? I'm looking for a 1.5 dc voltage multiplier circuit..
Hey Dave, Good day.. if your switch mode power supply is trying to turn ON again after the input voltage have been removed? what might have been causing it?
on the other hand, on your LDO, if there's a fluctuation in your input voltage will it's output have same waveform as it's input?
thank you, very simple, easy for beginers
Hi Dave, On top of the question that Ricardo has asked below, I'd also like to ask why a PNP has a lower dropout compared to an NPN. Would it be due to the fact that in the case of the PNP, the voltage drop across the transistor Vbe junction is much higher ?
Thank you for being awesome
Fantastic work!
Me: "Switching mode power supply," enter
Me: _Scrolls_
Me: Yeah lets learn it from nerd Geddy Lee, click.
Love ya Dave.