WEBVTT 00:00:00.000 --> 00:00:01.988 align:middle line:90% [SQUEAKING] 00:00:01.988 --> 00:00:03.976 align:middle line:90% [RUSTLING] 00:00:03.976 --> 00:00:04.970 align:middle line:90% [CLICKING] 00:00:04.970 --> 00:00:12.193 align:middle line:90% 00:00:12.193 --> 00:00:13.610 align:middle line:84% DAVID PERREAULT: So I will just go 00:00:13.610 --> 00:00:19.340 align:middle line:84% through the old-fashioned paper syllabus. 00:00:19.340 --> 00:00:22.020 align:middle line:84% So the course is going to be taught on Canvas. 00:00:22.020 --> 00:00:24.410 align:middle line:84% So the Canvas site has been published. 00:00:24.410 --> 00:00:27.680 align:middle line:84% We will-- for homeworks, and things 00:00:27.680 --> 00:00:31.850 align:middle line:84% like that, they'll be handed out and turned in on paper, 00:00:31.850 --> 00:00:36.590 align:middle line:84% but you can download them from Canvas as well. 00:00:36.590 --> 00:00:40.670 align:middle line:84% The classes are going to be here Mondays, Wednesdays 00:00:40.670 --> 00:00:41.580 align:middle line:90% and Thursdays. 00:00:41.580 --> 00:00:45.920 align:middle line:84% So our usual class schedule is 1:00 PM. 00:00:45.920 --> 00:00:47.220 align:middle line:90% I've mentioned the textbook. 00:00:47.220 --> 00:00:50.630 align:middle line:84% There are some other textbooks that you might find handy, 00:00:50.630 --> 00:00:52.860 align:middle line:84% and I've listed those in the course handout. 00:00:52.860 --> 00:00:54.315 align:middle line:90% Those are completely optional. 00:00:54.315 --> 00:00:55.440 align:middle line:90% You don't have to buy them. 00:00:55.440 --> 00:00:59.060 align:middle line:90% The libraries also have copies. 00:00:59.060 --> 00:01:04.110 align:middle line:84% The way homeworks will work in this class is 00:01:04.110 --> 00:01:06.510 align:middle line:84% that will typically be issued on a Monday 00:01:06.510 --> 00:01:09.318 align:middle line:90% and due the following Monday. 00:01:09.318 --> 00:01:14.940 align:middle line:84% And, for the most part, you are allowed to collaborate and talk 00:01:14.940 --> 00:01:16.780 align:middle line:90% about the homework problems. 00:01:16.780 --> 00:01:20.310 align:middle line:84% So feel free to get together and discuss the problems. 00:01:20.310 --> 00:01:21.840 align:middle line:90% It's intended that you do so. 00:01:21.840 --> 00:01:24.060 align:middle line:84% The only constraint we have is that you 00:01:24.060 --> 00:01:25.600 align:middle line:90% must hand in your own solution. 00:01:25.600 --> 00:01:26.100 align:middle line:90% Right? 00:01:26.100 --> 00:01:27.373 align:middle line:90% So you can trade ideas. 00:01:27.373 --> 00:01:29.040 align:middle line:84% But, in the end, the thing you write up, 00:01:29.040 --> 00:01:31.120 align:middle line:84% it can't be a copy of your neighbor's. 00:01:31.120 --> 00:01:33.300 align:middle line:84% It has to be your solution, but you 00:01:33.300 --> 00:01:38.670 align:middle line:84% can base it on ideas you've exchanged with others. 00:01:38.670 --> 00:01:43.105 align:middle line:84% For those who don't know anybody, 00:01:43.105 --> 00:01:45.680 align:middle line:84% or who don't have access to study partners 00:01:45.680 --> 00:01:48.740 align:middle line:84% and may have questions, first of all, there will be office hours. 00:01:48.740 --> 00:01:52.850 align:middle line:84% [? Monse's ?] office hours will be Thursdays, 4:00 to 5:00 PM, 00:01:52.850 --> 00:01:57.800 align:middle line:84% and Fridays, 4:00 to 5:00 PM, in 10-178, in Building 10. 00:01:57.800 --> 00:02:02.180 align:middle line:84% But there are also a number of students in the Lees Laboratory, 00:02:02.180 --> 00:02:05.150 align:middle line:84% 10-050, who have had this class before and will be happy 00:02:05.150 --> 00:02:06.690 align:middle line:90% to answer any questions. 00:02:06.690 --> 00:02:08.570 align:middle line:84% And a few of those students are listed 00:02:08.570 --> 00:02:12.800 align:middle line:90% in the course description. 00:02:12.800 --> 00:02:16.880 align:middle line:84% We've also decided not to have exams this term, 00:02:16.880 --> 00:02:18.830 align:middle line:90% or in-class exams. 00:02:18.830 --> 00:02:22.310 align:middle line:84% Rather, we're going to have assessments, 00:02:22.310 --> 00:02:25.820 align:middle line:84% which are, essentially, take-home mini quizzes. 00:02:25.820 --> 00:02:27.550 align:middle line:90% They will generally be-- 00:02:27.550 --> 00:02:31.340 align:middle line:90% 00:02:31.340 --> 00:02:33.660 align:middle line:84% they'll be generally issued weekly. 00:02:33.660 --> 00:02:37.650 align:middle line:84% And this says they're going to be issued starting on March 1. 00:02:37.650 --> 00:02:39.080 align:middle line:90% I should check if that-- yeah. 00:02:39.080 --> 00:02:41.520 align:middle line:84% Starting on March 1, they'll be issued weekly. 00:02:41.520 --> 00:02:42.930 align:middle line:84% They'll be issued on a Wednesday. 00:02:42.930 --> 00:02:44.930 align:middle line:84% They will be due the next day, Thursday. 00:02:44.930 --> 00:02:48.290 align:middle line:84% And those will be submitted through Canvas, 00:02:48.290 --> 00:02:49.500 align:middle line:90% through Gradescope. 00:02:49.500 --> 00:02:53.050 align:middle line:90% 00:02:53.050 --> 00:02:55.840 align:middle line:84% As for homeworks, we can take late homeworks, 00:02:55.840 --> 00:02:57.670 align:middle line:84% if you arrange it ahead of time and there's 00:02:57.670 --> 00:02:59.810 align:middle line:84% some compelling reason why they're late. 00:02:59.810 --> 00:03:02.140 align:middle line:84% Assessments are a different story. 00:03:02.140 --> 00:03:05.840 align:middle line:84% Also, the collaboration policy for assessments is different. 00:03:05.840 --> 00:03:08.950 align:middle line:84% You cannot collaborate or discuss the problems at all 00:03:08.950 --> 00:03:09.890 align:middle line:90% for assessments. 00:03:09.890 --> 00:03:13.700 align:middle line:84% And we can't prove or not prove that you're doing so, 00:03:13.700 --> 00:03:16.613 align:middle line:84% but this is an exam, effectively. 00:03:16.613 --> 00:03:18.280 align:middle line:84% So don't discuss it with your neighbors, 00:03:18.280 --> 00:03:20.620 align:middle line:84% and don't discuss anything until the solutions are out 00:03:20.620 --> 00:03:22.870 align:middle line:84% because there may be people who have 00:03:22.870 --> 00:03:25.600 align:middle line:84% made some special arrangement because they're traveling, 00:03:25.600 --> 00:03:28.130 align:middle line:84% or for whatever reason, to hand it in late. 00:03:28.130 --> 00:03:31.340 align:middle line:84% So we're going to have homeworks and assessments. 00:03:31.340 --> 00:03:33.700 align:middle line:84% And the goal of the assessments, really, 00:03:33.700 --> 00:03:38.590 align:middle line:84% is, instead of a very few high-stakes opportunities 00:03:38.590 --> 00:03:40.690 align:middle line:84% to show your abilities, the assessments 00:03:40.690 --> 00:03:44.320 align:middle line:84% are distributed in low stakes and focused 00:03:44.320 --> 00:03:45.910 align:middle line:84% and gives you a better opportunity 00:03:45.910 --> 00:03:48.550 align:middle line:90% to show what you really know. 00:03:48.550 --> 00:03:51.670 align:middle line:84% So please complete them entirely on your own. 00:03:51.670 --> 00:03:52.922 align:middle line:90% No consultation. 00:03:52.922 --> 00:03:54.380 align:middle line:84% The only thing you're allowed to do 00:03:54.380 --> 00:03:56.870 align:middle line:84% is ask the course staff clarifying questions, 00:03:56.870 --> 00:03:59.527 align:middle line:84% just the way you might in an exam. 00:03:59.527 --> 00:04:04.070 align:middle line:84% And, as for the assessments, you can use 00:04:04.070 --> 00:04:05.960 align:middle line:90% any of the course materials-- 00:04:05.960 --> 00:04:07.405 align:middle line:84% read the book, whatever you want. 00:04:07.405 --> 00:04:09.980 align:middle line:90% 00:04:09.980 --> 00:04:15.140 align:middle line:84% Do not go outside and try to use the world wide web. 00:04:15.140 --> 00:04:18.680 align:middle line:84% And for that matter, the use of bibles is also prohibited. 00:04:18.680 --> 00:04:23.260 align:middle line:84% I know that there's collections of old 6.334 materials floating 00:04:23.260 --> 00:04:24.050 align:middle line:90% around. 00:04:24.050 --> 00:04:27.310 align:middle line:84% You're not supposed to be consulting those. 00:04:27.310 --> 00:04:30.700 align:middle line:84% This is just supposed to be a measure of what you've learned. 00:04:30.700 --> 00:04:33.560 align:middle line:84% So the grading will be based on three components. 00:04:33.560 --> 00:04:35.890 align:middle line:90% Homeworks are going to be 40%. 00:04:35.890 --> 00:04:39.010 align:middle line:90% These assessments will be 50%. 00:04:39.010 --> 00:04:43.460 align:middle line:84% And there's also a final project, which is 10%. 00:04:43.460 --> 00:04:45.790 align:middle line:84% And the final project sounds like it's only 10%, 00:04:45.790 --> 00:04:48.340 align:middle line:84% but it's the last thing we look at when we're going 00:04:48.340 --> 00:04:50.030 align:middle line:90% to assign a grade to everybody. 00:04:50.030 --> 00:04:51.820 align:middle line:84% And it really is your opportunity 00:04:51.820 --> 00:04:53.770 align:middle line:84% to put together knowledge that you've 00:04:53.770 --> 00:04:56.470 align:middle line:84% learned throughout the class into a real-- 00:04:56.470 --> 00:04:57.710 align:middle line:90% it's on paper. 00:04:57.710 --> 00:05:01.103 align:middle line:84% It's not a physical converter you will construct, 00:05:01.103 --> 00:05:03.520 align:middle line:84% but it's a paper design-- but it's really your opportunity 00:05:03.520 --> 00:05:05.350 align:middle line:84% to show us how you've synthesized 00:05:05.350 --> 00:05:07.420 align:middle line:84% all this knowledge, to be able to really design 00:05:07.420 --> 00:05:08.900 align:middle line:90% power electronics. 00:05:08.900 --> 00:05:12.200 align:middle line:84% And I should say, just as an aside, in this class, 00:05:12.200 --> 00:05:17.480 align:middle line:84% we only have a paper design, but this complements nicely 00:05:17.480 --> 00:05:20.570 align:middle line:84% the undergraduate power electronics class, which 00:05:20.570 --> 00:05:23.300 align:middle line:84% has a lot of really nice lab activities 00:05:23.300 --> 00:05:24.630 align:middle line:90% and design activities there. 00:05:24.630 --> 00:05:26.180 align:middle line:84% So even if you're a graduate student, 00:05:26.180 --> 00:05:27.860 align:middle line:84% it can be a pretty good thing to take, 00:05:27.860 --> 00:05:32.725 align:middle line:84% in terms of rounding out your lab skill set in this area. 00:05:32.725 --> 00:05:36.430 align:middle line:90% 00:05:36.430 --> 00:05:40.820 align:middle line:84% If you have any necessary technical accommodations-- 00:05:40.820 --> 00:05:44.500 align:middle line:84% don't have access to iPads, or whatever else 00:05:44.500 --> 00:05:46.530 align:middle line:84% you need for Gradescope-- please let us know. 00:05:46.530 --> 00:05:48.780 align:middle line:90% We'll try to assist with that. 00:05:48.780 --> 00:05:54.600 align:middle line:90% 00:05:54.600 --> 00:05:57.550 align:middle line:84% So, with that, are there any questions 00:05:57.550 --> 00:06:00.860 align:middle line:84% about anything associated with the course mechanics? 00:06:00.860 --> 00:06:04.690 align:middle line:90% 00:06:04.690 --> 00:06:06.085 align:middle line:90% OK. 00:06:06.085 --> 00:06:12.890 align:middle line:84% So let me give you a sense of what this course is 00:06:12.890 --> 00:06:13.920 align:middle line:90% going to be about. 00:06:13.920 --> 00:06:17.775 align:middle line:84% And this is one of my favorite photos. 00:06:17.775 --> 00:06:21.262 align:middle line:84% It is actually one that Nikola Tesla mocked up. 00:06:21.262 --> 00:06:23.970 align:middle line:84% He wasn't really sitting next to his Tesla coil when he did this, 00:06:23.970 --> 00:06:26.070 align:middle line:90% or he might have gotten killed. 00:06:26.070 --> 00:06:27.690 align:middle line:90% He kind of double exposed this. 00:06:27.690 --> 00:06:31.940 align:middle line:84% But, more to the point, the quote from him is, 00:06:31.940 --> 00:06:34.665 align:middle line:84% "If we could produce electrical effects of the required quality, 00:06:34.665 --> 00:06:37.040 align:middle line:84% this whole planet, and the conditions of existence on it, 00:06:37.040 --> 00:06:38.430 align:middle line:90% could be transformed." 00:06:38.430 --> 00:06:41.960 align:middle line:84% And I think the more than 100 years 00:06:41.960 --> 00:06:45.830 align:middle line:84% since he said that, or well more than 100 years since he 00:06:45.830 --> 00:06:49.430 align:middle line:90% said that, have borne that out. 00:06:49.430 --> 00:06:53.000 align:middle line:84% But it's also true that, even today, 00:06:53.000 --> 00:06:55.430 align:middle line:84% there's really revolutions happening 00:06:55.430 --> 00:06:57.430 align:middle line:90% in the way we use energy. 00:06:57.430 --> 00:07:00.470 align:middle line:84% Everything's being electrified-- from vehicles, 00:07:00.470 --> 00:07:02.720 align:middle line:84% to transportation, to power generation-- 00:07:02.720 --> 00:07:04.940 align:middle line:90% from renewable resources. 00:07:04.940 --> 00:07:09.440 align:middle line:84% And handling all that requires some means 00:07:09.440 --> 00:07:11.880 align:middle line:84% of processing, controlling, and converting energy. 00:07:11.880 --> 00:07:13.820 align:middle line:84% And that's really what we're about, 00:07:13.820 --> 00:07:17.620 align:middle line:84% processing, controlling, and converting electrical energy. 00:07:17.620 --> 00:07:23.170 align:middle line:84% If you look at what the IEEE, which is the governing body 00:07:23.170 --> 00:07:26.440 align:middle line:84% of electrical engineering, says about power electronics, 00:07:26.440 --> 00:07:28.450 align:middle line:84% it says, "This technology encompasses the use 00:07:28.450 --> 00:07:31.540 align:middle line:84% of electronic components, the application of circuit theory 00:07:31.540 --> 00:07:34.780 align:middle line:84% and design techniques, and the development of analytical tools 00:07:34.780 --> 00:07:37.210 align:middle line:84% towards efficient electronic conversion control 00:07:37.210 --> 00:07:39.050 align:middle line:84% and conditioning of electric power." 00:07:39.050 --> 00:07:40.960 align:middle line:84% And that's what we're really about here. 00:07:40.960 --> 00:07:43.010 align:middle line:84% So we're going to do circuit theory. 00:07:43.010 --> 00:07:44.690 align:middle line:84% We're going to learn design techniques. 00:07:44.690 --> 00:07:46.960 align:middle line:84% We're going to learn about all the components you 00:07:46.960 --> 00:07:47.630 align:middle line:90% need to do this. 00:07:47.630 --> 00:07:49.130 align:middle line:84% We're going to learn about controls. 00:07:49.130 --> 00:07:52.360 align:middle line:84% How do you put it all together, to make energy conversion 00:07:52.360 --> 00:07:54.736 align:middle line:90% systems? 00:07:54.736 --> 00:07:58.120 align:middle line:84% So, as I mentioned, the primary function of power electronics 00:07:58.120 --> 00:08:01.180 align:middle line:84% is to take electrical energy in one form 00:08:01.180 --> 00:08:04.720 align:middle line:84% and convert it into some other form you need. 00:08:04.720 --> 00:08:07.880 align:middle line:84% It's really a core technology in the electrical infrastructure. 00:08:07.880 --> 00:08:11.085 align:middle line:84% It used to be that the AC grid was generators, 00:08:11.085 --> 00:08:12.460 align:middle line:84% and you'd connect it up to things 00:08:12.460 --> 00:08:15.160 align:middle line:84% like-- directly to things like motors, or lighting, 00:08:15.160 --> 00:08:15.890 align:middle line:90% or whatever. 00:08:15.890 --> 00:08:18.310 align:middle line:84% But that's pretty much changed, at this point. 00:08:18.310 --> 00:08:19.850 align:middle line:90% Lighting is LED lighting. 00:08:19.850 --> 00:08:22.750 align:middle line:84% You need power supplies to go between the grid 00:08:22.750 --> 00:08:23.750 align:middle line:90% and the lighting. 00:08:23.750 --> 00:08:24.530 align:middle line:90% Same thing. 00:08:24.530 --> 00:08:28.550 align:middle line:84% Heavily loads computers, motors, everything else you need. 00:08:28.550 --> 00:08:31.930 align:middle line:84% Energy tends to flow through one, or even several, layers 00:08:31.930 --> 00:08:36.280 align:middle line:84% of power conversion circuitry, from the principal source 00:08:36.280 --> 00:08:38.050 align:middle line:90% to the final usage. 00:08:38.050 --> 00:08:42.190 align:middle line:84% And so the power electronics-- first of all, the efficiency 00:08:42.190 --> 00:08:43.419 align:middle line:90% of that is very important. 00:08:43.419 --> 00:08:47.570 align:middle line:84% But also, how you do it impacts the quality of the final system. 00:08:47.570 --> 00:08:49.330 align:middle line:84% So the power electronics can really 00:08:49.330 --> 00:08:53.020 align:middle line:84% be a major factor impacting what you can and can't do 00:08:53.020 --> 00:08:54.105 align:middle line:90% and how well it works. 00:08:54.105 --> 00:08:57.435 align:middle line:90% 00:08:57.435 --> 00:09:01.910 align:middle line:84% So if you showed up 100 years ago, 00:09:01.910 --> 00:09:04.760 align:middle line:84% this is what power electronics would have looked like-- 00:09:04.760 --> 00:09:07.040 align:middle line:84% some vacuum tubes, and some transformers, 00:09:07.040 --> 00:09:08.510 align:middle line:90% and that kind of thing. 00:09:08.510 --> 00:09:12.870 align:middle line:84% Interestingly-- of course, it's nothing like that today. 00:09:12.870 --> 00:09:15.768 align:middle line:84% But, with the techniques you're going to learn in this course, 00:09:15.768 --> 00:09:17.810 align:middle line:84% you could actually go back and analyze this thing 00:09:17.810 --> 00:09:19.120 align:middle line:90% and figure out what it did. 00:09:19.120 --> 00:09:19.620 align:middle line:90% Right? 00:09:19.620 --> 00:09:21.080 align:middle line:84% So this is some foundational ideas 00:09:21.080 --> 00:09:24.990 align:middle line:84% that we're going to come back to, which can be 100 years old. 00:09:24.990 --> 00:09:28.615 align:middle line:84% But there's also elements that are extremely new. 00:09:28.615 --> 00:09:32.790 align:middle line:84% And today-- this was fancy 100 years ago. 00:09:32.790 --> 00:09:35.510 align:middle line:84% Today, power electronics is everywhere, from-- 00:09:35.510 --> 00:09:37.950 align:middle line:84% I say, from milliwatts to gigawatts. 00:09:37.950 --> 00:09:40.425 align:middle line:84% And it does, actually, use switched-mode power conversion 00:09:40.425 --> 00:09:41.550 align:middle line:90% down to those power levels. 00:09:41.550 --> 00:09:44.180 align:middle line:84% This is actually a multi-watt power supply. 00:09:44.180 --> 00:09:46.260 align:middle line:84% And this is, literally, at the gigawatt scale. 00:09:46.260 --> 00:09:50.720 align:middle line:84% So if you go out to the Sandy Pond terminal, 00:09:50.720 --> 00:09:53.540 align:middle line:84% there is a power converter that takes 2 gigawatts, coming down 00:09:53.540 --> 00:09:55.700 align:middle line:84% from Canada Hydro, and converts it 00:09:55.700 --> 00:09:59.180 align:middle line:84% to AC, to power homes and everything else around here. 00:09:59.180 --> 00:09:59.900 align:middle line:90% Right? 00:09:59.900 --> 00:10:03.030 align:middle line:84% And the techniques that we're going to learn in this class 00:10:03.030 --> 00:10:06.150 align:middle line:90% really span the entire range. 00:10:06.150 --> 00:10:08.680 align:middle line:84% So some of the details change, and we'll learn about that. 00:10:08.680 --> 00:10:10.410 align:middle line:84% But there's underlying principles 00:10:10.410 --> 00:10:13.680 align:middle line:84% that cut across all kinds of electrical energy conversion 00:10:13.680 --> 00:10:14.180 align:middle line:90% systems. 00:10:14.180 --> 00:10:16.887 align:middle line:90% 00:10:16.887 --> 00:10:17.970 align:middle line:90% What kind of applications? 00:10:17.970 --> 00:10:19.095 align:middle line:90% Well, portable electronics. 00:10:19.095 --> 00:10:21.790 align:middle line:84% This is slightly older, an iPhone 5. 00:10:21.790 --> 00:10:24.510 align:middle line:84% And you think, OK. iPhone's got radio transmitters, 00:10:24.510 --> 00:10:26.010 align:middle line:84% and displays, and other stuff in it. 00:10:26.010 --> 00:10:30.150 align:middle line:84% But it turns out that a large fraction of the volume in board 00:10:30.150 --> 00:10:33.540 align:middle line:84% area is actually associated with energy conversion in the thing. 00:10:33.540 --> 00:10:35.040 align:middle line:84% Because no matter what you're doing, 00:10:35.040 --> 00:10:37.320 align:middle line:84% you're processing energy to process information. 00:10:37.320 --> 00:10:41.220 align:middle line:84% So something like 40% of the motherboard, in this example, 00:10:41.220 --> 00:10:43.830 align:middle line:84% is associated with power conversion. 00:10:43.830 --> 00:10:46.080 align:middle line:84% Likewise, at some point, you're going 00:10:46.080 --> 00:10:49.890 align:middle line:84% to charge your phone, or your iPad, or your computer, 00:10:49.890 --> 00:10:51.330 align:middle line:90% into the wall. 00:10:51.330 --> 00:10:53.560 align:middle line:84% That's mostly power electronics, too. 00:10:53.560 --> 00:10:54.670 align:middle line:90% All kinds of computers. 00:10:54.670 --> 00:10:56.520 align:middle line:84% If you're in a data center, there's 00:10:56.520 --> 00:10:59.910 align:middle line:84% several layers of power conversion between the AC grid 00:10:59.910 --> 00:11:03.782 align:middle line:90% and the final set of processors. 00:11:03.782 --> 00:11:07.590 align:middle line:84% This is more of what it looks like inside your home computer. 00:11:07.590 --> 00:11:11.220 align:middle line:84% We will actually learn exactly about those kind of converters 00:11:11.220 --> 00:11:14.350 align:middle line:84% and about all the components that are in them. 00:11:14.350 --> 00:11:17.650 align:middle line:84% If you're going to communicate, the transmitters-- 00:11:17.650 --> 00:11:20.380 align:middle line:84% we tend to think of this as analog circuits 00:11:20.380 --> 00:11:21.860 align:middle line:90% to make RF transmitters. 00:11:21.860 --> 00:11:24.550 align:middle line:84% But, in fact, power electronics are heavily 00:11:24.550 --> 00:11:27.280 align:middle line:84% embedded in any real communication systems 00:11:27.280 --> 00:11:31.110 align:middle line:84% to increase the efficiency of transmission. 00:11:31.110 --> 00:11:33.990 align:middle line:84% All kinds of commercial applications, whether you're 00:11:33.990 --> 00:11:36.750 align:middle line:90% doing LED lighting or-- 00:11:36.750 --> 00:11:41.670 align:middle line:84% this is actually from some water purity device, 00:11:41.670 --> 00:11:44.240 align:middle line:84% but, of course, it requires a power supply. 00:11:44.240 --> 00:11:48.790 align:middle line:84% So almost any use of energy these days 00:11:48.790 --> 00:11:51.150 align:middle line:90% requires a power supply. 00:11:51.150 --> 00:11:52.390 align:middle line:90% Even in your home. 00:11:52.390 --> 00:11:55.407 align:middle line:84% I mentioned that it used to be that you'd connect motors up 00:11:55.407 --> 00:11:57.240 align:middle line:84% to the grid, and they'd run, and maybe you'd 00:11:57.240 --> 00:11:58.990 align:middle line:84% turn them on and off, something like that. 00:11:58.990 --> 00:11:59.580 align:middle line:90% But no longer. 00:11:59.580 --> 00:12:00.080 align:middle line:90% Right? 00:12:00.080 --> 00:12:01.890 align:middle line:84% If you want high performance, you 00:12:01.890 --> 00:12:04.150 align:middle line:84% need to be able to modulate that energy. 00:12:04.150 --> 00:12:05.920 align:middle line:90% So two examples here. 00:12:05.920 --> 00:12:08.820 align:middle line:84% This is for an air conditioning unit, 00:12:08.820 --> 00:12:12.270 align:middle line:84% and it uses an inverter, a DC to AC converter, 00:12:12.270 --> 00:12:15.690 align:middle line:84% inside it to drive the motor much more quietly and much more 00:12:15.690 --> 00:12:18.610 align:middle line:84% modulated, for higher overall system efficiency. 00:12:18.610 --> 00:12:21.090 align:middle line:84% Even your dishwasher, these days, 00:12:21.090 --> 00:12:25.020 align:middle line:84% has power converters in it because it's more efficient-- 00:12:25.020 --> 00:12:26.130 align:middle line:90% in this case, quieter-- 00:12:26.130 --> 00:12:28.770 align:middle line:90% to do it that way. 00:12:28.770 --> 00:12:32.700 align:middle line:84% Ideas that you might not think of as 00:12:32.700 --> 00:12:35.380 align:middle line:84% power converters-- medical applications. 00:12:35.380 --> 00:12:39.495 align:middle line:84% This is, actually, a magnetic stimulator. 00:12:39.495 --> 00:12:43.410 align:middle line:84% It generates 5,000 amps pulse trains in a transducer coil 00:12:43.410 --> 00:12:46.260 align:middle line:84% to throw out magnetic fields that can trigger nerves. 00:12:46.260 --> 00:12:47.530 align:middle line:90% This is an interesting one. 00:12:47.530 --> 00:12:48.730 align:middle line:90% It's actually homework 0. 00:12:48.730 --> 00:12:50.897 align:middle line:84% So the thing you're analyzing in the first homework, 00:12:50.897 --> 00:12:55.700 align:middle line:84% just to break the rust off, is actually this box, right here. 00:12:55.700 --> 00:12:57.320 align:middle line:90% Scientific applications. 00:12:57.320 --> 00:12:59.510 align:middle line:84% And you may not be processing energy, 00:12:59.510 --> 00:13:01.910 align:middle line:84% but even if you just need to generate electric-- 00:13:01.910 --> 00:13:03.920 align:middle line:84% high electrical fields, for whatever reason, 00:13:03.920 --> 00:13:05.420 align:middle line:90% or magnetic fields-- 00:13:05.420 --> 00:13:07.010 align:middle line:84% as in the magnetic stimulator-- you 00:13:07.010 --> 00:13:09.680 align:middle line:84% need energy conversion circuits to do it. 00:13:09.680 --> 00:13:11.540 align:middle line:90% Then there's the more-- 00:13:11.540 --> 00:13:15.800 align:middle line:84% maybe, the applications you might think of, transportation. 00:13:15.800 --> 00:13:19.945 align:middle line:84% Let's say, the electric vehicle, or a hybrid vehicle. 00:13:19.945 --> 00:13:23.860 align:middle line:84% You need power electronics to drive the energy conversion. 00:13:23.860 --> 00:13:25.110 align:middle line:90% And this is not a small thing. 00:13:25.110 --> 00:13:26.960 align:middle line:84% First of all, you need the power electronics 00:13:26.960 --> 00:13:29.146 align:middle line:90% to drive these things. 00:13:29.146 --> 00:13:33.770 align:middle line:84% Secondly-- in fact, in one example, 00:13:33.770 --> 00:13:36.620 align:middle line:84% they redesigned the power converter 00:13:36.620 --> 00:13:40.700 align:middle line:84% for a Prius, the powertrain for the Prius. 00:13:40.700 --> 00:13:44.450 align:middle line:84% The fuel economy went up by 5%, just by redesigning the power 00:13:44.450 --> 00:13:45.860 align:middle line:90% electronics to be better. 00:13:45.860 --> 00:13:50.430 align:middle line:84% So it has a huge impact on the overall application. 00:13:50.430 --> 00:13:54.030 align:middle line:84% And, of course, that's electric vehicles. 00:13:54.030 --> 00:13:58.440 align:middle line:84% But traction, trains, that's higher power, 00:13:58.440 --> 00:13:59.480 align:middle line:90% but the same issue. 00:13:59.480 --> 00:14:01.100 align:middle line:90% Actually, even, future trains. 00:14:01.100 --> 00:14:03.090 align:middle line:84% It's a little hard to see behind this railing. 00:14:03.090 --> 00:14:06.620 align:middle line:84% That's a maglev-- magnetically levitated-- train. 00:14:06.620 --> 00:14:08.990 align:middle line:84% Along the wayside, over in the back corner, 00:14:08.990 --> 00:14:10.310 align:middle line:90% you see this big building here. 00:14:10.310 --> 00:14:12.560 align:middle line:84% Inside that big building is these racks 00:14:12.560 --> 00:14:14.080 align:middle line:90% of power electronics. 00:14:14.080 --> 00:14:16.580 align:middle line:84% Now, you better have pretty reliable power electronics, 00:14:16.580 --> 00:14:19.640 align:middle line:84% if your vehicle is flying along at 400 kilometers an hour, 00:14:19.640 --> 00:14:21.500 align:middle line:84% floating on that far off the ground. 00:14:21.500 --> 00:14:22.000 align:middle line:90% Right? 00:14:22.000 --> 00:14:24.620 align:middle line:84% So not only do you need efficiency, 00:14:24.620 --> 00:14:28.160 align:middle line:84% but you need reliability and precision. 00:14:28.160 --> 00:14:33.410 align:middle line:84% Even stranger things-- this is an example of a drone, 00:14:33.410 --> 00:14:37.020 align:middle line:90% just powered by high voltage. 00:14:37.020 --> 00:14:38.640 align:middle line:90% You just apply high voltage. 00:14:38.640 --> 00:14:40.910 align:middle line:84% It breaks down the air, accelerates ions, 00:14:40.910 --> 00:14:43.128 align:middle line:84% and you can use that for propulsion. 00:14:43.128 --> 00:14:44.670 align:middle line:84% That's the first demonstration of it. 00:14:44.670 --> 00:14:47.000 align:middle line:84% But it actually-- it's very similar, 00:14:47.000 --> 00:14:49.800 align:middle line:84% in some regards, to what people use for space propulsion. 00:14:49.800 --> 00:14:53.360 align:middle line:84% You've heard of ion engines, the twin ion engine, TIE, fighter? 00:14:53.360 --> 00:14:58.620 align:middle line:84% Or what, more practically, they use to reposition satellites. 00:14:58.620 --> 00:15:01.340 align:middle line:84% Those require power electronics to generate high voltages 00:15:01.340 --> 00:15:04.660 align:middle line:90% and accelerate ions. 00:15:04.660 --> 00:15:08.485 align:middle line:84% Power transmission and generation. 00:15:08.485 --> 00:15:10.360 align:middle line:84% So you're getting your energy from somewhere, 00:15:10.360 --> 00:15:12.950 align:middle line:84% and, increasingly, we're getting it from renewable resources. 00:15:12.950 --> 00:15:15.940 align:middle line:84% Well, generally, the way things are trending, 00:15:15.940 --> 00:15:20.260 align:middle line:84% you take some mechanical, or solar, or other source 00:15:20.260 --> 00:15:22.750 align:middle line:84% of energy, and you transition it, 00:15:22.750 --> 00:15:25.870 align:middle line:84% not only through a generator, but through power electronics, 00:15:25.870 --> 00:15:26.420 align:middle line:90% to get there. 00:15:26.420 --> 00:15:29.320 align:middle line:84% And that's true for terrestrial things, things like-- 00:15:29.320 --> 00:15:32.090 align:middle line:84% this is a house rooftop PV system. 00:15:32.090 --> 00:15:33.345 align:middle line:90% This is a microinverter. 00:15:33.345 --> 00:15:36.320 align:middle line:90% 00:15:36.320 --> 00:15:38.270 align:middle line:90% Automotive systems. 00:15:38.270 --> 00:15:44.540 align:middle line:84% Or even much smaller things, like power harvesting, 00:15:44.540 --> 00:15:45.990 align:middle line:90% energy harvesting, techniques. 00:15:45.990 --> 00:15:49.160 align:middle line:84% You need power electronics in it. 00:15:49.160 --> 00:15:51.440 align:middle line:84% Also, all kinds of industrial applications, 00:15:51.440 --> 00:15:55.580 align:middle line:84% whether you're doing plasma processing, for semiconductor 00:15:55.580 --> 00:15:59.990 align:middle line:84% processing, or you want to refine metals, 00:15:59.990 --> 00:16:04.400 align:middle line:84% like a DC arc furnace, you need power electronics there too. 00:16:04.400 --> 00:16:07.100 align:middle line:84% So that's just a long way of saying, 00:16:07.100 --> 00:16:09.830 align:middle line:84% power electronics are in almost everything 00:16:09.830 --> 00:16:12.440 align:middle line:84% you care about these days, and it's only 00:16:12.440 --> 00:16:14.960 align:middle line:84% getting more so because we have to be better about how 00:16:14.960 --> 00:16:15.625 align:middle line:90% we use energy. 00:16:15.625 --> 00:16:19.732 align:middle line:90% 00:16:19.732 --> 00:16:22.130 align:middle line:84% So what's inside a power converter? 00:16:22.130 --> 00:16:25.370 align:middle line:84% And we're going to talk about this in much greater detail. 00:16:25.370 --> 00:16:28.900 align:middle line:84% But if you want to think about it, what we have 00:16:28.900 --> 00:16:34.150 align:middle line:84% is, typically, some kind of energy storage elements. 00:16:34.150 --> 00:16:36.870 align:middle line:84% These could be inductors, or capacitors, or sometimes 00:16:36.870 --> 00:16:38.172 align:middle line:90% other things. 00:16:38.172 --> 00:16:39.630 align:middle line:84% And what we're going to do is we're 00:16:39.630 --> 00:16:42.030 align:middle line:84% going to use semiconductor switches, 00:16:42.030 --> 00:16:44.860 align:middle line:84% and we're going to draw energy from some source. 00:16:44.860 --> 00:16:46.480 align:middle line:84% We're going to manipulate it somehow. 00:16:46.480 --> 00:16:48.480 align:middle line:84% So we store it in these energy storage elements, 00:16:48.480 --> 00:16:50.355 align:middle line:84% and then we're going to put it to the output. 00:16:50.355 --> 00:16:52.680 align:middle line:90% And we're going to repeat that. 00:16:52.680 --> 00:16:57.060 align:middle line:84% So draw, transform, put it to the output. 00:16:57.060 --> 00:17:00.120 align:middle line:84% Because we do this on a cyclic basis, 00:17:00.120 --> 00:17:03.330 align:middle line:84% we generate sort of ripple and, potentially, noise, 00:17:03.330 --> 00:17:06.369 align:middle line:84% that could interfere with things like television or electronics. 00:17:06.369 --> 00:17:10.260 align:middle line:84% So you generally need filters to do that, too. 00:17:10.260 --> 00:17:13.390 align:middle line:84% We also, of course, need to control that flow of energy. 00:17:13.390 --> 00:17:14.928 align:middle line:90% We can't be stupid about it. 00:17:14.928 --> 00:17:17.220 align:middle line:84% We have to-- especially for something high performance, 00:17:17.220 --> 00:17:19.137 align:middle line:84% like a microprocessor, which might be changing 00:17:19.137 --> 00:17:24.672 align:middle line:84% its operation very, very fast, we need to control that flow. 00:17:24.672 --> 00:17:25.880 align:middle line:90% So there's control circuitry. 00:17:25.880 --> 00:17:27.547 align:middle line:84% And we're going to be learning about all 00:17:27.547 --> 00:17:31.605 align:middle line:84% of the aspects of designing all of these kinds of elements. 00:17:31.605 --> 00:17:35.058 align:middle line:90% 00:17:35.058 --> 00:17:40.280 align:middle line:84% So we're leveraging everything from circuit design, 00:17:40.280 --> 00:17:43.590 align:middle line:84% semiconductor devices, passive components and materials. 00:17:43.590 --> 00:17:47.570 align:middle line:84% Increasingly, packaging, and cooling, and controls 00:17:47.570 --> 00:17:48.790 align:middle line:90% also matter. 00:17:48.790 --> 00:17:51.290 align:middle line:84% And things have been getting better, and better, and better, 00:17:51.290 --> 00:17:56.870 align:middle line:84% both because the ways we can manufacture things get better 00:17:56.870 --> 00:18:01.200 align:middle line:84% and because the devices which we have 00:18:01.200 --> 00:18:04.690 align:middle line:84% access to, in order to implement these things, get better. 00:18:04.690 --> 00:18:07.410 align:middle line:84% And so, we're going to look at some of the different ways 00:18:07.410 --> 00:18:08.970 align:middle line:90% you can do that. 00:18:08.970 --> 00:18:13.780 align:middle line:90% I'll just give you one example. 00:18:13.780 --> 00:18:16.560 align:middle line:84% This is what you would find something like in a data center. 00:18:16.560 --> 00:18:17.700 align:middle line:90% So they come in. 00:18:17.700 --> 00:18:18.880 align:middle line:90% They rectify the voltage. 00:18:18.880 --> 00:18:23.260 align:middle line:84% They get it 400 volts, for reasons we'll go into. 00:18:23.260 --> 00:18:26.200 align:middle line:84% This converter, here, was designed to take that 400 volt 00:18:26.200 --> 00:18:28.990 align:middle line:90% and bring it down to 12 volts. 00:18:28.990 --> 00:18:31.840 align:middle line:84% That tiny little thing, a little bit bigger than a penny, 00:18:31.840 --> 00:18:34.240 align:middle line:84% there, actually, can handle a kilowatt. 00:18:34.240 --> 00:18:36.730 align:middle line:84% Then you need more converters, to take it down from there 00:18:36.730 --> 00:18:40.660 align:middle line:84% to the 1 volt that the processor uses. 00:18:40.660 --> 00:18:46.890 align:middle line:84% So a long way of saying all kinds of things 00:18:46.890 --> 00:18:49.620 align:middle line:84% are limited by energy and how we can control it, 00:18:49.620 --> 00:18:51.610 align:middle line:90% across all these applications. 00:18:51.610 --> 00:18:53.070 align:middle line:84% And what we're usually trying to do 00:18:53.070 --> 00:18:55.980 align:middle line:84% is figure out how to make them smaller and lighter. 00:18:55.980 --> 00:18:58.290 align:middle line:84% If it's taking up 40% of your iPhone, 00:18:58.290 --> 00:19:00.285 align:middle line:84% you want to make that thing smaller, 00:19:00.285 --> 00:19:01.410 align:middle line:90% so it doesn't take that up. 00:19:01.410 --> 00:19:04.080 align:middle line:84% And you can replace that volume either with nothing 00:19:04.080 --> 00:19:06.600 align:middle line:84% or with things that are more interesting to you. 00:19:06.600 --> 00:19:08.970 align:middle line:84% We need higher efficiency, both for the converters 00:19:08.970 --> 00:19:09.670 align:middle line:90% and the systems. 00:19:09.670 --> 00:19:12.722 align:middle line:84% How we process energy can matter to the-- not only 00:19:12.722 --> 00:19:14.430 align:middle line:84% the ultimate efficiency of the converter, 00:19:14.430 --> 00:19:15.940 align:middle line:84% but the efficiency of the system. 00:19:15.940 --> 00:19:17.910 align:middle line:84% How it uses energy can be determined 00:19:17.910 --> 00:19:19.350 align:middle line:90% by the power electronics. 00:19:19.350 --> 00:19:21.390 align:middle line:90% We want higher performance. 00:19:21.390 --> 00:19:25.180 align:middle line:84% That could mean higher bandwidth or other aspects. 00:19:25.180 --> 00:19:29.670 align:middle line:84% And then, there's all kinds of means that we can use-- 00:19:29.670 --> 00:19:32.830 align:middle line:84% better electrical processing-- to enable new applications. 00:19:32.830 --> 00:19:35.250 align:middle line:84% So these are the kind of things that we're 00:19:35.250 --> 00:19:37.650 align:middle line:90% going to learn about this term. 00:19:37.650 --> 00:19:38.605 align:middle line:90% I'll pause there. 00:19:38.605 --> 00:19:39.730 align:middle line:90% I've been going on a while. 00:19:39.730 --> 00:19:41.280 align:middle line:84% Are there any questions about any of this, 00:19:41.280 --> 00:19:42.130 align:middle line:90% before we get going? 00:19:42.130 --> 00:19:44.140 align:middle line:84% And this is just to orient you as to what we're 00:19:44.140 --> 00:19:45.620 align:middle line:90% going to focus the term on. 00:19:45.620 --> 00:19:50.370 align:middle line:90% 00:19:50.370 --> 00:19:51.900 align:middle line:90% I can put the slides on Canvas. 00:19:51.900 --> 00:19:54.000 align:middle line:90% Sure. 00:19:54.000 --> 00:19:55.240 align:middle line:90% Any other questions? 00:19:55.240 --> 00:19:58.340 align:middle line:90% 00:19:58.340 --> 00:19:58.840 align:middle line:90% OK. 00:19:58.840 --> 00:20:02.120 align:middle line:90% 00:20:02.120 --> 00:20:08.330 align:middle line:84% I will just give you a slight notion. 00:20:08.330 --> 00:20:09.230 align:middle line:90% OK? 00:20:09.230 --> 00:20:13.400 align:middle line:84% Just-- this is a favorite thing I like to bring in, 00:20:13.400 --> 00:20:16.130 align:middle line:84% just to show you the kind of things 00:20:16.130 --> 00:20:18.380 align:middle line:84% we're going to look at this term. 00:20:18.380 --> 00:20:20.977 align:middle line:84% This is a piece of commercial hardware, 00:20:20.977 --> 00:20:23.060 align:middle line:84% from a company called MKS Instruments, that I just 00:20:23.060 --> 00:20:25.310 align:middle line:90% happen to have in my office. 00:20:25.310 --> 00:20:27.740 align:middle line:84% This input piece is a filter, so you 00:20:27.740 --> 00:20:29.510 align:middle line:84% don't create electromagnetic interference 00:20:29.510 --> 00:20:32.800 align:middle line:84% and you don't generate noise you don't want. 00:20:32.800 --> 00:20:36.910 align:middle line:84% Then this piece, here, takes energy from the 60 hertz grid 00:20:36.910 --> 00:20:38.230 align:middle line:90% and generates DC from it. 00:20:38.230 --> 00:20:41.355 align:middle line:84% And it has to do it while making the whole thing 00:20:41.355 --> 00:20:42.730 align:middle line:84% look like a resistor to the grid, 00:20:42.730 --> 00:20:44.540 align:middle line:84% so that you don't mess up the grid. 00:20:44.540 --> 00:20:45.040 align:middle line:90% OK? 00:20:45.040 --> 00:20:46.400 align:middle line:90% That's what this part does. 00:20:46.400 --> 00:20:49.090 align:middle line:84% Then it has these isolated DC to DC converters, that 00:20:49.090 --> 00:20:52.360 align:middle line:84% can take that DC voltage and generate other DC 00:20:52.360 --> 00:20:55.040 align:middle line:84% voltages, referenced in other ways, 00:20:55.040 --> 00:21:00.230 align:middle line:84% and without worrying about currents going back to ground. 00:21:00.230 --> 00:21:00.730 align:middle line:90% OK? 00:21:00.730 --> 00:21:02.690 align:middle line:90% It's galvanically isolated. 00:21:02.690 --> 00:21:07.010 align:middle line:84% And then, it takes that energy, and it goes back DC to AC. 00:21:07.010 --> 00:21:07.510 align:middle line:90% OK? 00:21:07.510 --> 00:21:10.930 align:middle line:84% Now, in some systems, you might go DC to AC for 60 hertz 00:21:10.930 --> 00:21:13.580 align:middle line:84% to generate into the grid, or for a motor, or something else. 00:21:13.580 --> 00:21:15.850 align:middle line:84% This particular one has these outputs 00:21:15.850 --> 00:21:18.910 align:middle line:84% that are at 13 megahertz, for driving plasmas 00:21:18.910 --> 00:21:21.915 align:middle line:90% to process semiconductors. 00:21:21.915 --> 00:21:24.280 align:middle line:84% But you get all these kinds of functions-- 00:21:24.280 --> 00:21:26.920 align:middle line:90% AC to DC, DC to DC, DC to AC. 00:21:26.920 --> 00:21:30.700 align:middle line:84% And we're going to learn the-- first of all, the underlying 00:21:30.700 --> 00:21:34.520 align:middle line:84% principles of doing all these things and how to design them. 00:21:34.520 --> 00:21:41.090 align:middle line:84% So the real goal is, by the time you walk out of here, 00:21:41.090 --> 00:21:44.240 align:middle line:84% you should have all the tools to go off and design 00:21:44.240 --> 00:21:48.570 align:middle line:84% power electronics for all kinds of applications. 00:21:48.570 --> 00:21:51.590 align:middle line:84% And, in fact, graduates of this class 00:21:51.590 --> 00:21:54.980 align:middle line:84% have worked on electric vehicles, battery chargers, 00:21:54.980 --> 00:22:01.970 align:middle line:84% solar, communication systems, data centers. 00:22:01.970 --> 00:22:05.780 align:middle line:84% The-- in fact, the power supply I usually use for my laptop 00:22:05.780 --> 00:22:09.110 align:middle line:84% and, in fact, the wireless charger for my iWatch 00:22:09.110 --> 00:22:11.480 align:middle line:84% were all designed, or the teams were led, 00:22:11.480 --> 00:22:13.260 align:middle line:90% by graduates of this class. 00:22:13.260 --> 00:22:16.460 align:middle line:84% So this-- the goal is to give you the tools you need to go 00:22:16.460 --> 00:22:19.420 align:middle line:90% do these things. 00:22:19.420 --> 00:22:27.440 align:middle line:84% So, with that, let me start by just giving you 00:22:27.440 --> 00:22:35.300 align:middle line:84% a sense of what goes on inside power electronics. 00:22:35.300 --> 00:22:41.400 align:middle line:90% 00:22:41.400 --> 00:22:45.080 align:middle line:84% So let's think of the simplest case. 00:22:45.080 --> 00:22:48.150 align:middle line:84% I have some DC input voltage, and I'd 00:22:48.150 --> 00:22:50.260 align:middle line:90% like some other DC voltage. 00:22:50.260 --> 00:22:56.250 align:middle line:84% So suppose I have some input that's-- 00:22:56.250 --> 00:23:00.540 align:middle line:90% maybe it's 9 volts to 16 volts. 00:23:00.540 --> 00:23:03.210 align:middle line:84% I'm making that up arbitrarily, but that's, roughly, 00:23:03.210 --> 00:23:07.320 align:middle line:84% what you might get in a typical vehicle, out of the cigarette 00:23:07.320 --> 00:23:09.885 align:middle line:90% lighter. 00:23:09.885 --> 00:23:14.490 align:middle line:84% And suppose I want-- and I'll call this V in-- 00:23:14.490 --> 00:23:16.800 align:middle line:84% and suppose I want-- here's some load, 00:23:16.800 --> 00:23:19.160 align:middle line:84% that I'm going to represent with a resistor. 00:23:19.160 --> 00:23:21.240 align:middle line:90% I'm going to call that V out. 00:23:21.240 --> 00:23:25.900 align:middle line:90% And suppose I want Vg 5 volts. 00:23:25.900 --> 00:23:26.400 align:middle line:90% Right? 00:23:26.400 --> 00:23:30.440 align:middle line:84% To power something that takes 5 volts. 00:23:30.440 --> 00:23:34.180 align:middle line:84% What's the most obvious and simple way 00:23:34.180 --> 00:23:37.605 align:middle line:84% to get 5 volts at my output from some higher voltage? 00:23:37.605 --> 00:23:42.660 align:middle line:90% 00:23:42.660 --> 00:23:43.570 align:middle line:90% Voltage divider. 00:23:43.570 --> 00:23:45.000 align:middle line:90% Precisely. 00:23:45.000 --> 00:23:47.790 align:middle line:84% I could come up here, and I could say, OK. 00:23:47.790 --> 00:23:54.030 align:middle line:84% Let me put in some variable resistor, 00:23:54.030 --> 00:23:57.960 align:middle line:84% and I'll drop down this voltage, to give me the voltage I want. 00:23:57.960 --> 00:23:59.230 align:middle line:90% And I'm done. 00:23:59.230 --> 00:24:03.300 align:middle line:84% And, in fact, a lot of systems, that's exactly, more or less, 00:24:03.300 --> 00:24:04.450 align:middle line:90% what they have inside them. 00:24:04.450 --> 00:24:08.730 align:middle line:84% In fact, inside most integrated circuits, 00:24:08.730 --> 00:24:10.780 align:middle line:84% there's lots of these things going on. 00:24:10.780 --> 00:24:11.280 align:middle line:90% OK? 00:24:11.280 --> 00:24:12.700 align:middle line:90% Well, OK. 00:24:12.700 --> 00:24:14.170 align:middle line:90% They don't do it quite this way. 00:24:14.170 --> 00:24:15.670 align:middle line:90% What do they really do? 00:24:15.670 --> 00:24:20.990 align:middle line:90% They-- we'll take V in. 00:24:20.990 --> 00:24:24.050 align:middle line:84% And what they will do is they'll use some kind of transistor, 00:24:24.050 --> 00:24:27.260 align:middle line:84% a MOSFET or a bipolar transistor, 00:24:27.260 --> 00:24:32.210 align:middle line:84% and they will treat that, essentially, 00:24:32.210 --> 00:24:34.010 align:middle line:90% as a variable resistor. 00:24:34.010 --> 00:24:35.720 align:middle line:84% So I implement the variable resistor 00:24:35.720 --> 00:24:38.840 align:middle line:84% with a controlled transistor, and I will come 00:24:38.840 --> 00:24:40.800 align:middle line:84% and I'll have some feedback loop. 00:24:40.800 --> 00:24:44.060 align:middle line:84% I'll feed in a reference voltage. 00:24:44.060 --> 00:24:46.430 align:middle line:84% And I'll measure the output voltage. 00:24:46.430 --> 00:24:48.920 align:middle line:84% And I'll control the gate of this transistor 00:24:48.920 --> 00:24:50.780 align:middle line:84% and, essentially, make that transistor 00:24:50.780 --> 00:24:53.150 align:middle line:84% look like a variable resistor and control 00:24:53.150 --> 00:24:55.940 align:middle line:84% the voltage division, so that, even if this voltage 00:24:55.940 --> 00:24:59.275 align:middle line:84% varies between 9 and 16 volts, I always get my 5 volts output. 00:24:59.275 --> 00:25:01.814 align:middle line:90% 00:25:01.814 --> 00:25:06.135 align:middle line:84% And, like I said, that's very common. 00:25:06.135 --> 00:25:07.510 align:middle line:84% But what's the problem with this? 00:25:07.510 --> 00:25:09.640 align:middle line:84% Well, there's a whole lot of problems with this. 00:25:09.640 --> 00:25:16.720 align:middle line:84% Why wouldn't you want to do this in typical operation? 00:25:16.720 --> 00:25:17.800 align:middle line:90% AUDIENCE: Low efficiency. 00:25:17.800 --> 00:25:19.342 align:middle line:84% DAVID PERREAULT: Terrible efficiency. 00:25:19.342 --> 00:25:21.510 align:middle line:90% Yes, exactly. 00:25:21.510 --> 00:25:23.760 align:middle line:90% So let's think about this. 00:25:23.760 --> 00:25:26.700 align:middle line:90% If I accept the fact-- 00:25:26.700 --> 00:25:35.705 align:middle line:84% let me call this current I in and this current I out. 00:25:35.705 --> 00:25:39.510 align:middle line:90% 00:25:39.510 --> 00:25:43.950 align:middle line:84% If this thing does act exactly like a variable resistor, 00:25:43.950 --> 00:25:46.330 align:middle line:84% the input currents equal the output current. 00:25:46.330 --> 00:25:48.322 align:middle line:84% Now, this-- actually, a real system 00:25:48.322 --> 00:25:50.280 align:middle line:84% might have some current that actually also goes 00:25:50.280 --> 00:25:52.560 align:middle line:84% to the ground, but let's take the best case, 00:25:52.560 --> 00:25:55.710 align:middle line:84% where the output current is equal to the input current. 00:25:55.710 --> 00:25:58.300 align:middle line:84% What would be the efficiency of this beast? 00:25:58.300 --> 00:26:07.340 align:middle line:84% Well, the efficiency-- the efficiency 00:26:07.340 --> 00:26:11.660 align:middle line:84% is equal to the output power over the input power. 00:26:11.660 --> 00:26:14.630 align:middle line:84% I'm drawing energy, maybe, from my car battery, 00:26:14.630 --> 00:26:16.460 align:middle line:84% or whatever my battery source is, 00:26:16.460 --> 00:26:18.190 align:middle line:84% and I'm delivering it to the output. 00:26:18.190 --> 00:26:20.190 align:middle line:84% But not all of it's getting to the output. 00:26:20.190 --> 00:26:25.700 align:middle line:84% So the output power is V out times I out, 00:26:25.700 --> 00:26:30.020 align:middle line:84% and my input power is V in times I in. 00:26:30.020 --> 00:26:33.410 align:middle line:84% And I just told you that I out was equal to I 00:26:33.410 --> 00:26:35.130 align:middle line:90% in, in the best case. 00:26:35.130 --> 00:26:38.270 align:middle line:90% So that's V out over V in. 00:26:38.270 --> 00:26:39.060 align:middle line:90% Right? 00:26:39.060 --> 00:26:45.170 align:middle line:84% So if I'm coming in from 15 volts, and I'm getting 5 volts, 00:26:45.170 --> 00:26:47.240 align:middle line:90% that's 33% efficiency. 00:26:47.240 --> 00:26:51.420 align:middle line:84% I've taken 2/3 of my energy and thrown it away. 00:26:51.420 --> 00:26:55.600 align:middle line:84% Now, if I've got lots of energy floating around, 00:26:55.600 --> 00:26:58.890 align:middle line:84% and I'm processing a microwatt, maybe I don't care. 00:26:58.890 --> 00:27:00.850 align:middle line:84% Maybe I'll live with that because this thing's 00:27:00.850 --> 00:27:01.390 align:middle line:90% pretty simple. 00:27:01.390 --> 00:27:01.890 align:middle line:90% Right? 00:27:01.890 --> 00:27:04.970 align:middle line:84% It's a transistor-- it's a power transistor and some controls. 00:27:04.970 --> 00:27:07.690 align:middle line:84% And they can, often, put all that on one integrated circuit, 00:27:07.690 --> 00:27:10.660 align:middle line:84% or as even as a sub-block on an integrated circuit. 00:27:10.660 --> 00:27:12.410 align:middle line:84% I might still need some filtering. 00:27:12.410 --> 00:27:18.000 align:middle line:84% It's not quite as easy as it sounds, but pretty much. 00:27:18.000 --> 00:27:20.550 align:middle line:90% It's relatively simple. 00:27:20.550 --> 00:27:22.120 align:middle line:90% But my efficiency is miserable. 00:27:22.120 --> 00:27:27.360 align:middle line:84% Now, if I thought about in your computer, your desktop computer, 00:27:27.360 --> 00:27:29.910 align:middle line:84% typically, the intermediate supply that you're 00:27:29.910 --> 00:27:32.477 align:middle line:84% getting after, it's sort of come in from the grid 00:27:32.477 --> 00:27:33.560 align:middle line:90% and been transformed down. 00:27:33.560 --> 00:27:35.370 align:middle line:90% You get 12 volts. 00:27:35.370 --> 00:27:37.770 align:middle line:84% And let's say your computer is at sort of-- it 00:27:37.770 --> 00:27:41.890 align:middle line:84% depends on what it's doing, but say it's operating at a volt. 00:27:41.890 --> 00:27:45.022 align:middle line:84% So then you're less than 10% efficient. 00:27:45.022 --> 00:27:48.400 align:middle line:84% And if you think that you can have a microprocessor that's 00:27:48.400 --> 00:27:51.910 align:middle line:84% taking 200 amps at a volt, so a few hundred watts, 00:27:51.910 --> 00:27:56.020 align:middle line:84% suddenly, if I've got less than 10% efficiency on 200 watts, 00:27:56.020 --> 00:27:58.515 align:middle line:84% that means I need to put in a few kilowatts. 00:27:58.515 --> 00:28:00.390 align:middle line:84% Well, you can't even plug that into the wall, 00:28:00.390 --> 00:28:03.057 align:middle line:84% never mind the fact that you've just made a massive room heater. 00:28:03.057 --> 00:28:08.380 align:middle line:84% That's, like-- that'd be great for heating your dorm room, 00:28:08.380 --> 00:28:13.295 align:middle line:84% but pretty terrible for the overall system. 00:28:13.295 --> 00:28:18.180 align:middle line:84% So the number one reason we don't like this solution 00:28:18.180 --> 00:28:20.200 align:middle line:90% is efficiency. 00:28:20.200 --> 00:28:21.310 align:middle line:90% Good thing, it's simple. 00:28:21.310 --> 00:28:23.860 align:middle line:84% Bad thing, it's terrible efficiency. 00:28:23.860 --> 00:28:28.140 align:middle line:84% This technique, I should have noted, 00:28:28.140 --> 00:28:30.660 align:middle line:84% is what's known, for historical reasons, 00:28:30.660 --> 00:28:36.460 align:middle line:84% as a, quote unquote, linear power supply. 00:28:36.460 --> 00:28:37.990 align:middle line:90% Why linear? 00:28:37.990 --> 00:28:42.010 align:middle line:84% I think only because analog circuits, kind 00:28:42.010 --> 00:28:45.130 align:middle line:84% of a category of them, became known as, quote unquote, 00:28:45.130 --> 00:28:46.370 align:middle line:90% linear circuits. 00:28:46.370 --> 00:28:48.370 align:middle line:84% There's nothing really that linear about it, 00:28:48.370 --> 00:28:51.040 align:middle line:84% but this would be called a linear power converter, 00:28:51.040 --> 00:28:57.670 align:middle line:84% or sometimes a linear regulator is what it would be called. 00:28:57.670 --> 00:29:00.452 align:middle line:84% For obvious reasons, I hate throwing away energy. 00:29:00.452 --> 00:29:02.410 align:middle line:84% We're not going to talk about linear regulators 00:29:02.410 --> 00:29:03.970 align:middle line:84% at all in this class, and there's 00:29:03.970 --> 00:29:06.020 align:middle line:84% plenty of classes where you can learn about that. 00:29:06.020 --> 00:29:08.800 align:middle line:84% We want to do things some better way, 00:29:08.800 --> 00:29:13.090 align:middle line:84% that's not going to burn lots of energy. 00:29:13.090 --> 00:29:19.960 align:middle line:84% So, ideally, in my world, the goal is to take as much 00:29:19.960 --> 00:29:24.000 align:middle line:84% of the input energy in and put it to the output. 00:29:24.000 --> 00:29:25.690 align:middle line:90% And that's important because-- 00:29:25.690 --> 00:29:26.690 align:middle line:90% think about it this way. 00:29:26.690 --> 00:29:29.350 align:middle line:84% I showed you a photograph of something 00:29:29.350 --> 00:29:34.930 align:middle line:84% that was really tiny, maybe that big and that thick, 00:29:34.930 --> 00:29:36.970 align:middle line:84% and it was processing a kilowatt. 00:29:36.970 --> 00:29:37.820 align:middle line:90% Right? 00:29:37.820 --> 00:29:40.000 align:middle line:84% If I don't do that at really high efficiency, 00:29:40.000 --> 00:29:41.666 align:middle line:90% that thing will burn up. 00:29:41.666 --> 00:29:44.650 align:middle line:84% So, in order to make something small, 00:29:44.650 --> 00:29:47.540 align:middle line:90% I need to make it efficient. 00:29:47.540 --> 00:29:49.690 align:middle line:84% So we want almost all of the energy 00:29:49.690 --> 00:29:51.120 align:middle line:84% at the input to get to the output. 00:29:51.120 --> 00:29:53.600 align:middle line:90% Well, how can we do that? 00:29:53.600 --> 00:29:56.180 align:middle line:84% Let's think of a completely different way 00:29:56.180 --> 00:30:00.410 align:middle line:84% we might achieve this same function. 00:30:00.410 --> 00:30:01.430 align:middle line:90% And here's the idea. 00:30:01.430 --> 00:30:02.125 align:middle line:90% Here's my input. 00:30:02.125 --> 00:30:09.910 align:middle line:90% 00:30:09.910 --> 00:30:12.660 align:middle line:84% I'm going to create a switch here, 00:30:12.660 --> 00:30:14.840 align:middle line:84% a single pole double throw switch. 00:30:14.840 --> 00:30:16.990 align:middle line:84% Now, we're not going to go get physical switches. 00:30:16.990 --> 00:30:19.032 align:middle line:84% We're probably going to get semiconductor devices 00:30:19.032 --> 00:30:20.620 align:middle line:84% and make it do something like this. 00:30:20.620 --> 00:30:24.030 align:middle line:84% But we could use anything-- any technology that 00:30:24.030 --> 00:30:26.940 align:middle line:90% was practical-- as a switch. 00:30:26.940 --> 00:30:28.903 align:middle line:90% And let me just define. 00:30:28.903 --> 00:30:30.570 align:middle line:84% I'm going to define a switching function 00:30:30.570 --> 00:30:32.250 align:middle line:90% that I'm going to call q of t. 00:30:32.250 --> 00:30:37.680 align:middle line:84% If q of t is 1, I'll connect the switch to the input. 00:30:37.680 --> 00:30:44.670 align:middle line:84% If q of t is 0, I'll connect the switch to the ground. 00:30:44.670 --> 00:30:48.150 align:middle line:90% So this notion, this voltage-- 00:30:48.150 --> 00:30:50.340 align:middle line:90% maybe I'll call this voltage Vx. 00:30:50.340 --> 00:30:56.040 align:middle line:90% 00:30:56.040 --> 00:30:58.770 align:middle line:84% So one thing I could do is I could say, OK. 00:30:58.770 --> 00:31:01.260 align:middle line:84% Let me just hook this up to my load. 00:31:01.260 --> 00:31:04.800 align:middle line:84% Here's my resistive load, whatever it is, 00:31:04.800 --> 00:31:06.125 align:middle line:90% and I'll call this V out. 00:31:06.125 --> 00:31:08.670 align:middle line:90% 00:31:08.670 --> 00:31:09.847 align:middle line:90% Well, all right. 00:31:09.847 --> 00:31:10.930 align:middle line:90% What would that look like? 00:31:10.930 --> 00:31:13.640 align:middle line:84% Well, let me define my switching function. 00:31:13.640 --> 00:31:23.170 align:middle line:90% 00:31:23.170 --> 00:31:25.490 align:middle line:84% I'm going to put the switch in the up 00:31:25.490 --> 00:31:28.280 align:middle line:84% position for a little while, and then I'll 00:31:28.280 --> 00:31:31.790 align:middle line:84% put it in the down position by setting the switching 00:31:31.790 --> 00:31:33.000 align:middle line:90% function to 0. 00:31:33.000 --> 00:31:34.500 align:middle line:90% And then I'll just repeat that. 00:31:34.500 --> 00:31:36.375 align:middle line:84% So I said we're going to operate in some kind 00:31:36.375 --> 00:31:37.880 align:middle line:90% of repetitive fashion here. 00:31:37.880 --> 00:31:40.190 align:middle line:90% And so forth. 00:31:40.190 --> 00:31:44.515 align:middle line:84% Let me operate with some period T. 00:31:44.515 --> 00:31:46.140 align:middle line:84% That's going to be my switching period. 00:31:46.140 --> 00:31:49.820 align:middle line:84% In this example, I'm showing it as a fixed switching period. 00:31:49.820 --> 00:31:53.570 align:middle line:84% And I will keep the switch in the up position some fraction 00:31:53.570 --> 00:31:56.030 align:middle line:90% of the time that I'll call DT. 00:31:56.030 --> 00:31:57.980 align:middle line:90% So D is a fraction. 00:31:57.980 --> 00:31:59.830 align:middle line:90% 0 is less than D is less than 1. 00:31:59.830 --> 00:32:04.410 align:middle line:90% 00:32:04.410 --> 00:32:07.700 align:middle line:84% So if I do that, then what do I get for Vx? 00:32:07.700 --> 00:32:10.000 align:middle line:84% Vx is going to look something like this. 00:32:10.000 --> 00:32:14.520 align:middle line:90% 00:32:14.520 --> 00:32:17.870 align:middle line:84% When the switch is in the up position, Vx equals V in. 00:32:17.870 --> 00:32:22.300 align:middle line:90% 00:32:22.300 --> 00:32:23.860 align:middle line:90% So this is Vx. 00:32:23.860 --> 00:32:26.925 align:middle line:84% When the switch is in the down position, Vx is equal to 0. 00:32:26.925 --> 00:32:32.582 align:middle line:90% 00:32:32.582 --> 00:32:35.350 align:middle line:84% And I rinse and repeat, and I get this. 00:32:35.350 --> 00:32:39.310 align:middle line:84% Now, I have this pulsating voltage Vx. 00:32:39.310 --> 00:32:43.410 align:middle line:84% What's the average value of that voltage Vx? 00:32:43.410 --> 00:32:49.010 align:middle line:84% I can take simple integration, 1 over t, the integral of Vx of t, 00:32:49.010 --> 00:32:54.620 align:middle line:84% over a period T. And what I would find 00:32:54.620 --> 00:33:02.820 align:middle line:84% is the average voltage of Vx is equal to D times V in. 00:33:02.820 --> 00:33:08.020 align:middle line:84% So I can create a waveform here whose average value 00:33:08.020 --> 00:33:10.540 align:middle line:84% is something different than V in, just 00:33:10.540 --> 00:33:15.460 align:middle line:90% by controlling this timing, D. 00:33:15.460 --> 00:33:22.560 align:middle line:84% So now, if my load resistor here was a space heater-- you know, 00:33:22.560 --> 00:33:25.410 align:middle line:84% maybe this is some load resistance RL-- 00:33:25.410 --> 00:33:28.620 align:middle line:84% and I wanted to modulate the power to that load resistance 00:33:28.620 --> 00:33:31.840 align:middle line:84% by controlling the average voltage on the load resistance, 00:33:31.840 --> 00:33:36.050 align:middle line:84% then this technique would work great. 00:33:36.050 --> 00:33:39.320 align:middle line:84% If, on the other hand, my load was a microprocessor, 00:33:39.320 --> 00:33:43.285 align:middle line:84% and I start pulsing 12 volts-- between 12 volts and 0 on it, 00:33:43.285 --> 00:33:44.660 align:middle line:84% I'm probably going to blow it up. 00:33:44.660 --> 00:33:46.220 align:middle line:90% So that's no good. 00:33:46.220 --> 00:33:50.930 align:middle line:84% But this notion is at least that I can control an average voltage 00:33:50.930 --> 00:33:54.250 align:middle line:90% by pulsing a switch. 00:33:54.250 --> 00:33:58.310 align:middle line:84% And this would be known as PWM, or pulse width modulation, 00:33:58.310 --> 00:34:00.860 align:middle line:84% because I control the average voltage 00:34:00.860 --> 00:34:03.290 align:middle line:84% by the fraction of the time the switch is in one 00:34:03.290 --> 00:34:06.350 align:middle line:90% position versus the other. 00:34:06.350 --> 00:34:10.489 align:middle line:84% So that's the basic concept we're going to be using. 00:34:10.489 --> 00:34:16.070 align:middle line:84% How do I fix this little problem of-- in practice? 00:34:16.070 --> 00:34:19.310 align:middle line:84% What I wanted was a DC voltage, and what I got 00:34:19.310 --> 00:34:21.050 align:middle line:84% was a pulsating voltage, that just 00:34:21.050 --> 00:34:23.370 align:middle line:84% happened to have the right average value. 00:34:23.370 --> 00:34:25.560 align:middle line:84% Well, I could go do something like this. 00:34:25.560 --> 00:34:27.710 align:middle line:90% Maybe I'll go back and say, OK. 00:34:27.710 --> 00:34:32.880 align:middle line:84% Let me throw in a filter and extract out 00:34:32.880 --> 00:34:34.239 align:middle line:90% the component I want. 00:34:34.239 --> 00:34:37.050 align:middle line:90% I want the average value of Vx. 00:34:37.050 --> 00:34:39.960 align:middle line:84% So maybe I'll come back here and say, OK. 00:34:39.960 --> 00:34:42.659 align:middle line:90% Let me throw in a filter. 00:34:42.659 --> 00:34:50.179 align:middle line:84% And I'll use an inductor here, an L. And if I want, 00:34:50.179 --> 00:34:55.300 align:middle line:84% optionally, I can put a capacitor here, C. OK? 00:34:55.300 --> 00:35:01.340 align:middle line:84% And I think people can look at this filter block 00:35:01.340 --> 00:35:04.490 align:middle line:84% and recognize that as a low-pass filter. 00:35:04.490 --> 00:35:08.060 align:middle line:84% The DC component of Vx passes through the filter 00:35:08.060 --> 00:35:13.190 align:middle line:84% to the output, and the AC component of Vx 00:35:13.190 --> 00:35:15.990 align:middle line:84% gets rejected by the filter and doesn't get to the output. 00:35:15.990 --> 00:35:19.258 align:middle line:84% So, in this case, I might get an output voltage V out 00:35:19.258 --> 00:35:20.550 align:middle line:90% that looks something like this. 00:35:20.550 --> 00:35:23.960 align:middle line:84% I'm going to sort of make this up, but-- 00:35:23.960 --> 00:35:25.640 align:middle line:90% I'm amplifying the ripple. 00:35:25.640 --> 00:35:28.550 align:middle line:84% But, eventually, it's going to filter 00:35:28.550 --> 00:35:30.210 align:middle line:90% the energy content of that. 00:35:30.210 --> 00:35:33.538 align:middle line:84% And the fundamental and higher harmonic terms of Vx 00:35:33.538 --> 00:35:36.080 align:middle line:84% are going to go away, and the DC term is going to go through. 00:35:36.080 --> 00:35:39.320 align:middle line:84% And I get an output voltage V out 00:35:39.320 --> 00:35:42.870 align:middle line:84% that's very close to whatever value I want. 00:35:42.870 --> 00:35:46.730 align:middle line:84% And if I, basically, make the filter cutoff hard enough, 00:35:46.730 --> 00:35:50.780 align:middle line:84% I can't distinguish between V out and the average value of Vx, 00:35:50.780 --> 00:35:54.240 align:middle line:90% and I get exactly what I wanted. 00:35:54.240 --> 00:36:06.000 align:middle line:84% So we've, essentially, now created a voltage converter that 00:36:06.000 --> 00:36:09.270 align:middle line:84% lets me use this pulse width modulation, by controlling 00:36:09.270 --> 00:36:13.470 align:middle line:84% this duty cycle D, to regulate the output just the way 00:36:13.470 --> 00:36:14.580 align:middle line:90% I wanted. 00:36:14.580 --> 00:36:18.360 align:middle line:84% So instead of-- here, I'm just changing the gate voltage 00:36:18.360 --> 00:36:20.350 align:middle line:84% on my transistor to control the output. 00:36:20.350 --> 00:36:21.880 align:middle line:90% Here I'm changing timing. 00:36:21.880 --> 00:36:23.960 align:middle line:90% I'm going to control timing. 00:36:23.960 --> 00:36:27.270 align:middle line:84% And by controlling timing, I control average value. 00:36:27.270 --> 00:36:29.210 align:middle line:90% And then I get what I want. 00:36:29.210 --> 00:36:29.710 align:middle line:90% All right. 00:36:29.710 --> 00:36:32.280 align:middle line:90% Any questions about that? 00:36:32.280 --> 00:36:34.750 align:middle line:84% AUDIENCE: What's the efficiency of the supply? 00:36:34.750 --> 00:36:37.260 align:middle line:84% DAVID PERREAULT: That is an excellent question. 00:36:37.260 --> 00:36:45.370 align:middle line:84% The answer is, ideally, theoretically, 00:36:45.370 --> 00:36:48.020 align:middle line:90% the efficiency can be 100%. 00:36:48.020 --> 00:36:50.130 align:middle line:90% In reality, it can't be. 00:36:50.130 --> 00:36:55.430 align:middle line:84% Why do I say the efficiency can ideally be 100%? 00:36:55.430 --> 00:37:01.290 align:middle line:84% Well, how would I implement this box in the real world? 00:37:01.290 --> 00:37:04.190 align:middle line:84% Usually, I don't get semiconductor single pole 00:37:04.190 --> 00:37:05.550 align:middle line:90% double throw switches. 00:37:05.550 --> 00:37:09.410 align:middle line:84% The way I would usually build this beast is like this. 00:37:09.410 --> 00:37:10.160 align:middle line:90% OK? 00:37:10.160 --> 00:37:13.730 align:middle line:84% I would usually have a first switch and a second switch 00:37:13.730 --> 00:37:15.500 align:middle line:90% implemented like this. 00:37:15.500 --> 00:37:19.310 align:middle line:84% So I close this when q of t is equal to 1, 00:37:19.310 --> 00:37:24.410 align:middle line:84% and I close this one when q of t is equal to 0. 00:37:24.410 --> 00:37:26.785 align:middle line:90% And then I build my filter. 00:37:26.785 --> 00:37:29.970 align:middle line:90% 00:37:29.970 --> 00:37:33.960 align:middle line:90% And then I put my load on here. 00:37:33.960 --> 00:37:38.020 align:middle line:84% So what would be the efficiency of this thing? 00:37:38.020 --> 00:37:39.230 align:middle line:90% Well, let's think about this. 00:37:39.230 --> 00:37:42.150 align:middle line:90% 00:37:42.150 --> 00:37:47.310 align:middle line:84% This is voltage V sub x and this is voltage V sub out. 00:37:47.310 --> 00:37:47.810 align:middle line:90% All right. 00:37:47.810 --> 00:37:51.530 align:middle line:90% 00:37:51.530 --> 00:37:56.030 align:middle line:84% What power is theoretically dissipated in my switch? 00:37:56.030 --> 00:38:00.170 align:middle line:84% If I have an ideal switch, a 0 resistance when it's on 00:38:00.170 --> 00:38:03.280 align:middle line:84% and it has infinite resistance when it's off. 00:38:03.280 --> 00:38:04.110 align:middle line:90% What's that? 00:38:04.110 --> 00:38:05.020 align:middle line:90% 0 power. 00:38:05.020 --> 00:38:05.820 align:middle line:90% Why? 00:38:05.820 --> 00:38:07.900 align:middle line:90% Because the power dissipated. 00:38:07.900 --> 00:38:10.680 align:middle line:84% The power that goes into this box-- 00:38:10.680 --> 00:38:16.080 align:middle line:84% let me call this V switch, and let me call this I switch. 00:38:16.080 --> 00:38:17.010 align:middle line:90% OK? 00:38:17.010 --> 00:38:21.030 align:middle line:84% Well, P switch, the power going into the switch, the power being 00:38:21.030 --> 00:38:23.400 align:middle line:84% dissipated in the switch, is going 00:38:23.400 --> 00:38:27.660 align:middle line:90% to be V switch times I switch. 00:38:27.660 --> 00:38:28.800 align:middle line:90% OK? 00:38:28.800 --> 00:38:34.020 align:middle line:84% Well, if it's an ideal switch, then, If the switch is on, 00:38:34.020 --> 00:38:35.995 align:middle line:90% V switch is 0. 00:38:35.995 --> 00:38:36.495 align:middle line:90% Right? 00:38:36.495 --> 00:38:39.340 align:middle line:84% It's ideal, so it has no voltage drop when it's on. 00:38:39.340 --> 00:38:43.260 align:middle line:90% So the power when it's on is 0. 00:38:43.260 --> 00:38:45.580 align:middle line:84% When the switch is off, it has infinite resistance. 00:38:45.580 --> 00:38:47.830 align:middle line:90% So the switch is current 0. 00:38:47.830 --> 00:38:51.390 align:middle line:84% So, basically, the power going into the switch, 00:38:51.390 --> 00:38:56.790 align:middle line:84% if it's an ideal switch, is, ideally, 0. 00:38:56.790 --> 00:39:02.910 align:middle line:84% So these elements, this ideal switch, is a lossless element. 00:39:02.910 --> 00:39:06.780 align:middle line:90% Likewise, I wasn't-- 00:39:06.780 --> 00:39:10.250 align:middle line:84% I didn't just randomly choose any filter here. 00:39:10.250 --> 00:39:12.440 align:middle line:90% I chose an LC filter. 00:39:12.440 --> 00:39:14.460 align:middle line:90% Why did I choose an LC filter? 00:39:14.460 --> 00:39:16.670 align:middle line:84% I chose an LC filter because inductors and capacitors 00:39:16.670 --> 00:39:18.420 align:middle line:90% are energy storage elements. 00:39:18.420 --> 00:39:20.180 align:middle line:84% If they're ideal, then they store energy, 00:39:20.180 --> 00:39:22.790 align:middle line:90% but they don't dissipate energy. 00:39:22.790 --> 00:39:28.160 align:middle line:84% So, basically, everything in the white box here-- 00:39:28.160 --> 00:39:30.470 align:middle line:84% everything between the input and the output-- 00:39:30.470 --> 00:39:32.360 align:middle line:90% is a lossless element. 00:39:32.360 --> 00:39:35.960 align:middle line:84% So then, one would assume that, if every element is lossless, 00:39:35.960 --> 00:39:41.100 align:middle line:84% any energy walking in to the left comes out to the right. 00:39:41.100 --> 00:39:45.350 align:middle line:84% And that's how that kilowatt-- little 400 volt to 12 volt 00:39:45.350 --> 00:39:51.000 align:middle line:84% converter I showed you in the photo is about 97% efficient. 00:39:51.000 --> 00:39:55.022 align:middle line:84% It's not 100% because the wires have some resistance, 00:39:55.022 --> 00:39:56.980 align:middle line:84% and the switches have some on-state resistance. 00:39:56.980 --> 00:39:59.313 align:middle line:84% There's a whole bunch of things that contribute to loss, 00:39:59.313 --> 00:40:01.250 align:middle line:90% and we'll talk about that. 00:40:01.250 --> 00:40:04.520 align:middle line:84% But I can make it really close to 100%, 00:40:04.520 --> 00:40:07.710 align:middle line:84% even though I might be doing a huge step down. 00:40:07.710 --> 00:40:09.750 align:middle line:84% So if I tried to build a linear regulator that 00:40:09.750 --> 00:40:13.350 align:middle line:84% was going from 400 to 12 volts, that would be about 2.5% 00:40:13.350 --> 00:40:14.040 align:middle line:90% efficient. 00:40:14.040 --> 00:40:16.800 align:middle line:84% So if I want to generate a kilowatt at 2.5% efficiency, 00:40:16.800 --> 00:40:17.610 align:middle line:90% what is that? 00:40:17.610 --> 00:40:19.420 align:middle line:90% 40 kilowatt input? 00:40:19.420 --> 00:40:25.620 align:middle line:84% And instead, what I get is, sort of, like, 1.03 kilowatts input 00:40:25.620 --> 00:40:27.720 align:middle line:90% to generate a kilowatt output. 00:40:27.720 --> 00:40:30.420 align:middle line:84% So the whole magic that we're going 00:40:30.420 --> 00:40:33.570 align:middle line:84% to talk about this term is, how can I 00:40:33.570 --> 00:40:38.160 align:middle line:84% use perfectly lossless elements to draw energy in, process it, 00:40:38.160 --> 00:40:39.990 align:middle line:90% and put it out the other side? 00:40:39.990 --> 00:40:42.150 align:middle line:84% And, by the way, I should say, not 00:40:42.150 --> 00:40:44.580 align:middle line:84% only did I say inductors and capacitors are 00:40:44.580 --> 00:40:48.420 align:middle line:84% lossless in principle, lossless elements, 00:40:48.420 --> 00:40:53.660 align:middle line:84% it wasn't an accident that we put an inductor here. 00:40:53.660 --> 00:40:56.161 align:middle line:90% 00:40:56.161 --> 00:40:57.340 align:middle line:90% Because think about it. 00:40:57.340 --> 00:41:01.920 align:middle line:84% When this switch is closed, I have 00:41:01.920 --> 00:41:03.750 align:middle line:84% V in on this side of the inductor 00:41:03.750 --> 00:41:05.890 align:middle line:84% and V out on this side of the inductor. 00:41:05.890 --> 00:41:09.075 align:middle line:84% And I have current flowing this way. 00:41:09.075 --> 00:41:12.520 align:middle line:84% Well, what's happening, when this switch is closed? 00:41:12.520 --> 00:41:15.952 align:middle line:84% Basically, there's voltage across this inductor and current 00:41:15.952 --> 00:41:16.660 align:middle line:90% going through it. 00:41:16.660 --> 00:41:19.110 align:middle line:84% We are storing energy in that inductor. 00:41:19.110 --> 00:41:22.050 align:middle line:84% So the difference in voltage between the input and the output 00:41:22.050 --> 00:41:25.020 align:middle line:84% is, basically, putting energy in the inductor. 00:41:25.020 --> 00:41:28.290 align:middle line:84% In the other part of the cycle, when 00:41:28.290 --> 00:41:32.140 align:middle line:84% I turn this switch off and this switch on, 00:41:32.140 --> 00:41:33.640 align:middle line:90% basically, I'm taking-- 00:41:33.640 --> 00:41:36.448 align:middle line:84% now I have a negative voltage across the inductor. 00:41:36.448 --> 00:41:37.990 align:middle line:84% I'm taking energy out of the inductor 00:41:37.990 --> 00:41:39.198 align:middle line:90% and putting it in the output. 00:41:39.198 --> 00:41:42.550 align:middle line:84% So, essentially, I'm using this inductor as a filter, 00:41:42.550 --> 00:41:44.350 align:middle line:84% but it's also an intermediate store 00:41:44.350 --> 00:41:48.760 align:middle line:84% of energy, that lets me take energy from the input 00:41:48.760 --> 00:41:52.000 align:middle line:84% and transfer it to the output, with a voltage conversion, 00:41:52.000 --> 00:41:55.840 align:middle line:84% without losing any of the energy. 00:41:55.840 --> 00:41:57.050 align:middle line:90% Excellent question. 00:41:57.050 --> 00:41:58.890 align:middle line:90% Long answer to a short question. 00:41:58.890 --> 00:42:00.260 align:middle line:90% Any other questions? 00:42:00.260 --> 00:42:05.000 align:middle line:90% 00:42:05.000 --> 00:42:05.500 align:middle line:90% OK. 00:42:05.500 --> 00:42:11.260 align:middle line:84% So, as I said, my goal is to, first 00:42:11.260 --> 00:42:13.910 align:middle line:84% of all, teach you a lot of the underlying principles. 00:42:13.910 --> 00:42:16.930 align:middle line:84% This is the world's, if you will, simplest 00:42:16.930 --> 00:42:18.200 align:middle line:90% switching power converters. 00:42:18.200 --> 00:42:19.610 align:middle line:90% So we use this technique. 00:42:19.610 --> 00:42:21.600 align:middle line:84% We also often say we're switched-mode. 00:42:21.600 --> 00:42:25.060 align:middle line:90% 00:42:25.060 --> 00:42:25.630 align:middle line:90% All right? 00:42:25.630 --> 00:42:28.720 align:middle line:84% With the notion that we're going to use switches and energy 00:42:28.720 --> 00:42:30.200 align:middle line:84% storage elements to process energy. 00:42:30.200 --> 00:42:33.920 align:middle line:84% And that's sort of what sits at the core of power electronics. 00:42:33.920 --> 00:42:34.420 align:middle line:90% OK? 00:42:34.420 --> 00:42:38.030 align:middle line:84% And, as I said, we're going to look at all of the aspects. 00:42:38.030 --> 00:42:39.610 align:middle line:90% How do you design these things? 00:42:39.610 --> 00:42:40.610 align:middle line:90% How do you control them? 00:42:40.610 --> 00:42:42.320 align:middle line:84% How do you design the components? 00:42:42.320 --> 00:42:43.945 align:middle line:84% And, by the time we're done, you should 00:42:43.945 --> 00:42:46.750 align:middle line:84% be able to put it all together and start designing 00:42:46.750 --> 00:42:48.370 align:middle line:90% power electronics of your own. 00:42:48.370 --> 00:42:52.770 align:middle line:84% And you will, for the final project. 00:42:52.770 --> 00:42:55.800 align:middle line:90% So, any final questions? 00:42:55.800 --> 00:42:58.940 align:middle line:90% 00:42:58.940 --> 00:42:59.440 align:middle line:90% OK. 00:42:59.440 --> 00:43:03.600 align:middle line:84% We'll wrap up today, and I will see everybody on Wednesday. 00:43:03.600 --> 00:43:22.000 align:middle line:90%