WEBVTT 00:00:00.000 --> 00:00:01.940 align:middle line:90% [SQUEAKING] 00:00:01.940 --> 00:00:04.365 align:middle line:90% [RUSTLING] 00:00:04.365 --> 00:00:07.760 align:middle line:90% [CLICKING] 00:00:07.760 --> 00:00:12.580 align:middle line:90% 00:00:12.580 --> 00:00:14.170 align:middle line:84% PROFESSOR: Why don't we get started. 00:00:14.170 --> 00:00:19.870 align:middle line:84% So before going back to talking about more kinds of power 00:00:19.870 --> 00:00:21.790 align:middle line:84% electronic circuits, I wanted to spend 00:00:21.790 --> 00:00:26.740 align:middle line:84% one lecture talking about heat transfer 00:00:26.740 --> 00:00:30.410 align:middle line:84% and how you model and design thermal systems. 00:00:30.410 --> 00:00:36.220 align:middle line:84% And this is principles of power electronics, chapter 25. 00:00:36.220 --> 00:00:38.980 align:middle line:84% Now, to be clear, you can spend an entire course-- 00:00:38.980 --> 00:00:42.610 align:middle line:84% or for that matter, multiple entire courses studying 00:00:42.610 --> 00:00:44.780 align:middle line:84% heat transfer and the ways to do that. 00:00:44.780 --> 00:00:47.540 align:middle line:84% And clearly, we don't have time to do that here, 00:00:47.540 --> 00:00:50.920 align:middle line:84% so what I'm going to attempt to do in this lecture 00:00:50.920 --> 00:00:55.040 align:middle line:84% is give you just enough information to be dangerous. 00:00:55.040 --> 00:00:57.400 align:middle line:84% In that a lot of the things we need 00:00:57.400 --> 00:00:59.500 align:middle line:84% to do as a power electronics designer, 00:00:59.500 --> 00:01:04.629 align:middle line:84% let us offset some of the detailed computational 00:01:04.629 --> 00:01:09.550 align:middle line:84% or deep analytical modeling onto heat 00:01:09.550 --> 00:01:11.350 align:middle line:90% sink designers and other people. 00:01:11.350 --> 00:01:14.405 align:middle line:84% And what we need to know is enough information about how 00:01:14.405 --> 00:01:16.280 align:middle line:84% to model these things so that we can predict, 00:01:16.280 --> 00:01:18.110 align:middle line:84% things like temperature rises and select 00:01:18.110 --> 00:01:19.710 align:middle line:90% the kind of correct components. 00:01:19.710 --> 00:01:23.000 align:middle line:84% And that's what we're going to try to get to today. 00:01:23.000 --> 00:01:29.790 align:middle line:84% Now, if we think about heat transfer, there's sort of-- 00:01:29.790 --> 00:01:32.110 align:middle line:84% and first of all, why am I obsessed about this? 00:01:32.110 --> 00:01:35.640 align:middle line:84% It's because despite our best attempts, 00:01:35.640 --> 00:01:37.860 align:middle line:84% we're going to dissipate some power in our devices 00:01:37.860 --> 00:01:39.190 align:middle line:90% and our components. 00:01:39.190 --> 00:01:41.730 align:middle line:84% And if I want to make something really high power density, 00:01:41.730 --> 00:01:43.590 align:middle line:84% I've really, first of all, got to minimize 00:01:43.590 --> 00:01:45.360 align:middle line:90% how much heat I dissipate. 00:01:45.360 --> 00:01:49.370 align:middle line:84% But in the end, I've also got to take whatever heat I dissipate 00:01:49.370 --> 00:01:50.820 align:middle line:90% and get it out of the system. 00:01:50.820 --> 00:01:53.810 align:middle line:84% And how good a job I do of getting it away 00:01:53.810 --> 00:01:57.080 align:middle line:84% from the components and keeping their temperature rise low 00:01:57.080 --> 00:02:00.410 align:middle line:84% says a lot about how high performance a power converter 00:02:00.410 --> 00:02:02.060 align:middle line:90% I can build. 00:02:02.060 --> 00:02:04.070 align:middle line:84% Now, let's assume we're going to dissipate 00:02:04.070 --> 00:02:07.520 align:middle line:84% a certain amount of power and ask, how do I get the heat out 00:02:07.520 --> 00:02:08.630 align:middle line:90% of my system? 00:02:08.630 --> 00:02:11.720 align:middle line:84% And there's really three fundamental ways 00:02:11.720 --> 00:02:13.520 align:middle line:90% you can do that. 00:02:13.520 --> 00:02:15.166 align:middle line:84% You can do it through conduction. 00:02:15.166 --> 00:02:18.550 align:middle line:90% 00:02:18.550 --> 00:02:22.930 align:middle line:84% Conduction is just simply, if I think of heat 00:02:22.930 --> 00:02:25.090 align:middle line:84% as being vibrations of particles, 00:02:25.090 --> 00:02:27.340 align:middle line:84% I can let the heat conduct through some heat conductor 00:02:27.340 --> 00:02:28.280 align:middle line:90% and be carried away. 00:02:28.280 --> 00:02:30.280 align:middle line:84% So some physical element, like a piece of metal, 00:02:30.280 --> 00:02:33.150 align:middle line:90% will carry heat along it. 00:02:33.150 --> 00:02:35.546 align:middle line:90% I can do it by convection. 00:02:35.546 --> 00:02:38.780 align:middle line:90% 00:02:38.780 --> 00:02:41.660 align:middle line:84% I transfer heat to some moving fluid-- that 00:02:41.660 --> 00:02:44.780 align:middle line:84% could be air or water-- and I let 00:02:44.780 --> 00:02:47.730 align:middle line:90% the fluid carry the heat away. 00:02:47.730 --> 00:02:49.950 align:middle line:90% Or I can do it by radiation. 00:02:49.950 --> 00:02:52.560 align:middle line:90% 00:02:52.560 --> 00:02:56.250 align:middle line:84% If I make something hot, it emits photons 00:02:56.250 --> 00:02:57.510 align:middle line:90% at whatever wavelength. 00:02:57.510 --> 00:03:00.300 align:middle line:84% It emits electromagnetic radiation 00:03:00.300 --> 00:03:03.840 align:middle line:84% that carries the heat away from my device. 00:03:03.840 --> 00:03:07.830 align:middle line:84% And depending upon the application, all three of these 00:03:07.830 --> 00:03:09.180 align:middle line:90% can be important. 00:03:09.180 --> 00:03:16.380 align:middle line:84% However, radiation, if I looked at the rate of heat transfer-- 00:03:16.380 --> 00:03:20.010 align:middle line:84% so this is heat power being transferred-- 00:03:20.010 --> 00:03:24.600 align:middle line:84% tends to be proportional to the source thing that's 00:03:24.600 --> 00:03:27.540 align:middle line:84% delivering the heat away to the fourth power 00:03:27.540 --> 00:03:31.920 align:middle line:84% minus the environment around it to the fourth power. 00:03:31.920 --> 00:03:35.910 align:middle line:84% So this high to the fourth power means 00:03:35.910 --> 00:03:40.170 align:middle line:84% that when things get really hot, radiation really 00:03:40.170 --> 00:03:41.790 align:middle line:90% dominates heat transfer. 00:03:41.790 --> 00:03:45.590 align:middle line:84% But for most terrestrial applications, 00:03:45.590 --> 00:03:48.170 align:middle line:90% it's not the highest-- 00:03:48.170 --> 00:03:50.420 align:middle line:84% it's not the most important thing. 00:03:50.420 --> 00:03:52.610 align:middle line:84% So if I have an incandescent light bulb 00:03:52.610 --> 00:03:56.150 align:middle line:84% and it's really, really hot, then yeah, 00:03:56.150 --> 00:03:58.100 align:middle line:84% a large part of the energy may get out of it 00:03:58.100 --> 00:04:02.960 align:middle line:84% by just radiating away from it as blackbody radiation. 00:04:02.960 --> 00:04:07.910 align:middle line:84% If I'm in a satellite, and there's 00:04:07.910 --> 00:04:10.430 align:middle line:84% no way for heat to conduct out of the thing, 00:04:10.430 --> 00:04:12.680 align:middle line:84% and there's no air or anything for the heat 00:04:12.680 --> 00:04:14.360 align:middle line:84% to go into, well, then eventually, 00:04:14.360 --> 00:04:16.980 align:middle line:84% all the heat is going to go away from that thing by radiation. 00:04:16.980 --> 00:04:18.920 align:middle line:84% So in those kind of applications, yeah, 00:04:18.920 --> 00:04:22.160 align:middle line:84% I really may care very heavily about radiation. 00:04:22.160 --> 00:04:26.450 align:middle line:84% But in most cases, it's only a small-ish fraction 00:04:26.450 --> 00:04:30.380 align:middle line:84% of the heat I'm getting rid of, because we don't 00:04:30.380 --> 00:04:31.860 align:middle line:90% get crazy high temperatures. 00:04:31.860 --> 00:04:33.770 align:middle line:84% We don't have thousands of degrees Kelvin 00:04:33.770 --> 00:04:34.975 align:middle line:90% to remove the heat. 00:04:34.975 --> 00:04:38.140 align:middle line:90% 00:04:38.140 --> 00:04:43.840 align:middle line:84% So let's go back and look at convection and conduction, 00:04:43.840 --> 00:04:46.580 align:middle line:90% and I'll start with conduction. 00:04:46.580 --> 00:04:53.530 align:middle line:84% Suppose I took-- actually, I'll just go over here. 00:04:53.530 --> 00:04:56.900 align:middle line:90% Suppose I took some-- 00:04:56.900 --> 00:05:02.440 align:middle line:84% I'll keep everything 1D, and I'll get some thing that 00:05:02.440 --> 00:05:04.990 align:middle line:90% I'll put at temperature T1. 00:05:04.990 --> 00:05:16.580 align:middle line:84% And then I'll have some block of material, 00:05:16.580 --> 00:05:21.460 align:middle line:84% and I will have a second surface that I 00:05:21.460 --> 00:05:24.920 align:middle line:90% will keep a temperature T2. 00:05:24.920 --> 00:05:27.953 align:middle line:84% And in between them, I have some material through which 00:05:27.953 --> 00:05:28.870 align:middle line:90% heat is going to flow. 00:05:28.870 --> 00:05:30.580 align:middle line:90% It's going to conduct. 00:05:30.580 --> 00:05:36.610 align:middle line:84% And this material, whatever it is, 00:05:36.610 --> 00:05:41.890 align:middle line:84% might have some cross-sectional area A, 00:05:41.890 --> 00:05:47.290 align:middle line:84% and it might be transferring heat through some length l. 00:05:47.290 --> 00:05:51.610 align:middle line:84% And in this kind of situation, with 1D heat transfer, 00:05:51.610 --> 00:05:55.330 align:middle line:84% for many materials, what you get is the following relationship. 00:05:55.330 --> 00:05:58.102 align:middle line:90% 00:05:58.102 --> 00:06:02.080 align:middle line:84% Q-- that's the amount of heat transferred-- 00:06:02.080 --> 00:06:07.750 align:middle line:84% is proportional to the temperature difference T1 00:06:07.750 --> 00:06:18.300 align:middle line:84% minus T2 times the cross-sectional area divided 00:06:18.300 --> 00:06:22.920 align:middle line:84% by the thermal resistance, RTh and divided by the length. 00:06:22.920 --> 00:06:24.660 align:middle line:84% So if I have more cross-sectional area, 00:06:24.660 --> 00:06:26.050 align:middle line:90% more heat flows. 00:06:26.050 --> 00:06:29.490 align:middle line:84% If I have to transfer it through a thicker piece of material, 00:06:29.490 --> 00:06:32.140 align:middle line:90% the heat flow goes down. 00:06:32.140 --> 00:06:34.050 align:middle line:84% And this proportionality constant 00:06:34.050 --> 00:06:38.100 align:middle line:84% is called the thermal resistivity of the material. 00:06:38.100 --> 00:06:45.150 align:middle line:84% I could rewrite this as T1 minus T2 divided 00:06:45.150 --> 00:06:49.410 align:middle line:84% by RTh times l divided by A. That 00:06:49.410 --> 00:06:52.590 align:middle line:84% would be another way to write this. 00:06:52.590 --> 00:07:02.860 align:middle line:84% So this would reflect how heat-- this would be reflect conduction 00:07:02.860 --> 00:07:05.810 align:middle line:90% of heat in steady state. 00:07:05.810 --> 00:07:13.010 align:middle line:90% 00:07:13.010 --> 00:07:17.010 align:middle line:84% This relationship-- and this is sort of a bulk relationship-- 00:07:17.010 --> 00:07:22.410 align:middle line:84% I could create an incremental relationship, 00:07:22.410 --> 00:07:30.250 align:middle line:84% which would be Q is equal to minus A over rho Th dT dx. 00:07:30.250 --> 00:07:33.250 align:middle line:90% 00:07:33.250 --> 00:07:35.940 align:middle line:84% That would be if I treated everything as a little increment 00:07:35.940 --> 00:07:36.850 align:middle line:90% in space. 00:07:36.850 --> 00:07:39.480 align:middle line:84% And if I want to start doing Fourier heat transfer, 00:07:39.480 --> 00:07:46.750 align:middle line:84% I get into these kind of differential equations 00:07:46.750 --> 00:07:47.350 align:middle line:90% descriptions. 00:07:47.350 --> 00:07:51.884 align:middle line:84% But for a bulk piece of element, we can write it like this. 00:07:51.884 --> 00:07:58.060 align:middle line:84% This relationship, where Q is my heat flow-- 00:07:58.060 --> 00:08:02.020 align:middle line:84% so Q is whatever heat is flowing this way-- 00:08:02.020 --> 00:08:07.710 align:middle line:84% suggests a circuit analog where-- here's my circuit model. 00:08:07.710 --> 00:08:14.000 align:middle line:84% I have some temperature source T1 and another temperature 00:08:14.000 --> 00:08:17.515 align:middle line:90% source or temperature sink T2. 00:08:17.515 --> 00:08:19.730 align:middle line:90% So temperature now means-- 00:08:19.730 --> 00:08:23.765 align:middle line:84% a voltage source here means constant temperature. 00:08:23.765 --> 00:08:27.770 align:middle line:84% And then there's between them some thermal resistance 00:08:27.770 --> 00:08:35.159 align:middle line:84% RTh which is equal to Rho Th times l over A. 00:08:35.159 --> 00:08:39.469 align:middle line:84% And that has the same form as an electrical resistor. 00:08:39.469 --> 00:08:45.020 align:middle line:84% And then my heat flow Q is the current flowing 00:08:45.020 --> 00:08:45.920 align:middle line:90% around this loop. 00:08:45.920 --> 00:08:53.230 align:middle line:84% And here, I have some point to which these temperatures 00:08:53.230 --> 00:08:54.460 align:middle line:90% are referenced. 00:08:54.460 --> 00:08:57.550 align:middle line:84% Ideally speaking, this should be absolute 0. 00:08:57.550 --> 00:09:01.300 align:middle line:84% In practice, people often just reference it to ambient 00:09:01.300 --> 00:09:03.800 align:middle line:84% and pretend ambient is the equivalent of absolute 0. 00:09:03.800 --> 00:09:08.590 align:middle line:84% It'd be kind of like having an offset from absolute 0. 00:09:08.590 --> 00:09:11.270 align:middle line:84% Any questions about that structure? 00:09:11.270 --> 00:09:13.750 align:middle line:84% You get-- if you analyze this circuit, 00:09:13.750 --> 00:09:17.780 align:middle line:84% you get exactly this relationship. 00:09:17.780 --> 00:09:20.780 align:middle line:84% The reason this works is temperature 00:09:20.780 --> 00:09:22.345 align:middle line:84% is conserved going around any loop. 00:09:22.345 --> 00:09:24.470 align:middle line:84% If I measure temperature differences around a loop, 00:09:24.470 --> 00:09:25.730 align:middle line:90% I ought to get 0. 00:09:25.730 --> 00:09:30.560 align:middle line:84% And if I consider heat inputs and heat transfers, 00:09:30.560 --> 00:09:34.850 align:middle line:84% conservation of power is the equivalent 00:09:34.850 --> 00:09:36.810 align:middle line:90% of KCL in this circuit model. 00:09:36.810 --> 00:09:38.627 align:middle line:84% So I'm showing you a simplified circuit 00:09:38.627 --> 00:09:40.710 align:middle line:84% model for a very particular kind of heat transfer, 00:09:40.710 --> 00:09:44.480 align:middle line:84% but you can use this model to represent heat transfers 00:09:44.480 --> 00:09:46.810 align:middle line:90% in a variety of conditions. 00:09:46.810 --> 00:09:47.940 align:middle line:90% Any questions about that? 00:09:47.940 --> 00:09:53.220 align:middle line:90% 00:09:53.220 --> 00:09:59.600 align:middle line:84% So I should say, by the way, Q is in watts. 00:09:59.600 --> 00:10:02.060 align:middle line:90% That's the heat flow in watts. 00:10:02.060 --> 00:10:08.870 align:middle line:84% T is in Kelvin, or often, people use it in degrees Celsius. 00:10:08.870 --> 00:10:15.320 align:middle line:84% And RTh ends up having units of degree C per watt 00:10:15.320 --> 00:10:16.600 align:middle line:90% or Kelvin per watt. 00:10:16.600 --> 00:10:23.820 align:middle line:90% 00:10:23.820 --> 00:10:26.310 align:middle line:84% In this regard, thermal resistance 00:10:26.310 --> 00:10:29.040 align:middle line:84% is very like regular resistance, because if I 00:10:29.040 --> 00:10:31.530 align:middle line:84% think about a regular resistor, if I 00:10:31.530 --> 00:10:33.960 align:middle line:84% make the resistor proportionally longer, 00:10:33.960 --> 00:10:36.100 align:middle line:90% the thermal resistance goes up. 00:10:36.100 --> 00:10:37.920 align:middle line:84% If I increase the cross-sectional area, 00:10:37.920 --> 00:10:39.430 align:middle line:84% the thermal resistance goes down. 00:10:39.430 --> 00:10:42.060 align:middle line:84% And here, all I have is thermal resistivity instead 00:10:42.060 --> 00:10:43.510 align:middle line:90% of electrical resistivity. 00:10:43.510 --> 00:10:49.180 align:middle line:84% So the form of the relationships are exactly the same. 00:10:49.180 --> 00:10:55.300 align:middle line:84% So basically, we can assume all of our devices in steady state 00:10:55.300 --> 00:10:57.320 align:middle line:90% have some thermal resistances. 00:10:57.320 --> 00:10:59.090 align:middle line:84% We can put power into the system. 00:10:59.090 --> 00:11:01.000 align:middle line:84% If we know some fixed temperatures, 00:11:01.000 --> 00:11:03.430 align:middle line:84% like some ambient temperatures, we can create a circuit 00:11:03.430 --> 00:11:05.775 align:middle line:84% and then analyze what our heat flows will look like. 00:11:05.775 --> 00:11:09.490 align:middle line:90% 00:11:09.490 --> 00:11:11.280 align:middle line:90% This is conduction. 00:11:11.280 --> 00:11:16.740 align:middle line:84% What might my other means of transferring heat energy 00:11:16.740 --> 00:11:17.790 align:middle line:90% look like? 00:11:17.790 --> 00:11:21.090 align:middle line:84% Well, my other means of heat transferring energy 00:11:21.090 --> 00:11:22.830 align:middle line:90% might be something like this. 00:11:22.830 --> 00:11:29.760 align:middle line:84% I've got some structure which has heat going into it 00:11:29.760 --> 00:11:35.250 align:middle line:84% and then some fluid going over it. 00:11:35.250 --> 00:11:38.010 align:middle line:84% This could be a flow of air, or a flow of water, 00:11:38.010 --> 00:11:39.370 align:middle line:90% or something like that. 00:11:39.370 --> 00:11:42.390 align:middle line:84% And what we expect to happen is the heat 00:11:42.390 --> 00:11:47.190 align:middle line:84% is going to come out of whatever the item is up into the air, 00:11:47.190 --> 00:11:49.553 align:middle line:84% and then the air is going to carry it away. 00:11:49.553 --> 00:11:51.970 align:middle line:84% So we're going to transfer the heat into the moving fluid, 00:11:51.970 --> 00:11:54.070 align:middle line:84% and the moving fluid kind of walks away with it. 00:11:54.070 --> 00:11:57.465 align:middle line:90% 00:11:57.465 --> 00:11:58.880 align:middle line:90% What does this look like? 00:11:58.880 --> 00:12:02.850 align:middle line:90% 00:12:02.850 --> 00:12:05.370 align:middle line:84% Heat transfer for this case looks 00:12:05.370 --> 00:12:10.050 align:middle line:84% like Q is the heat transfer is equal to hA 00:12:10.050 --> 00:12:14.286 align:middle line:84% times the temperature of the surface 00:12:14.286 --> 00:12:17.610 align:middle line:84% minus the temperature of the fluid. 00:12:17.610 --> 00:12:22.730 align:middle line:84% So if I have some fluid at T fluid, 00:12:22.730 --> 00:12:28.535 align:middle line:84% and I have some temperature at the surface T surface, 00:12:28.535 --> 00:12:31.630 align:middle line:84% the heat transfer again in some range 00:12:31.630 --> 00:12:33.580 align:middle line:84% is going to be proportional to the temperature 00:12:33.580 --> 00:12:34.930 align:middle line:84% difference between the surface of this thing 00:12:34.930 --> 00:12:36.850 align:middle line:84% and the fluid that's carrying the heat away. 00:12:36.850 --> 00:12:41.415 align:middle line:84% And this is the so-called wetted area. 00:12:41.415 --> 00:12:44.110 align:middle line:90% 00:12:44.110 --> 00:12:46.660 align:middle line:84% This is the area at the interface 00:12:46.660 --> 00:12:48.910 align:middle line:84% that the fluid is flowing over to grab the heat 00:12:48.910 --> 00:12:50.270 align:middle line:90% and carry it away. 00:12:50.270 --> 00:12:53.660 align:middle line:84% And then this is the heat-transfer coefficient. 00:12:53.660 --> 00:12:58.700 align:middle line:90% 00:12:58.700 --> 00:13:04.690 align:middle line:84% Now, here's where it's-- right here is essentially where 00:13:04.690 --> 00:13:07.180 align:middle line:90% I'm burying all the ugliness. 00:13:07.180 --> 00:13:11.360 align:middle line:84% This is where I'm massively oversimplifying it, 00:13:11.360 --> 00:13:14.080 align:middle line:84% and the reason you can take whole courses in this-- 00:13:14.080 --> 00:13:17.470 align:middle line:84% cool courses in this topic is because what is this heat 00:13:17.470 --> 00:13:18.560 align:middle line:90% transfer coefficient? 00:13:18.560 --> 00:13:24.530 align:middle line:84% Well, it depends upon is the flow laminar or turbulent? 00:13:24.530 --> 00:13:27.630 align:middle line:84% It depends upon all kinds of stuff. 00:13:27.630 --> 00:13:30.250 align:middle line:84% And it can even be a function of temperatures 00:13:30.250 --> 00:13:32.404 align:middle line:90% in the general case. 00:13:32.404 --> 00:13:38.020 align:middle line:90% So I have to tell you-- 00:13:38.020 --> 00:13:39.610 align:middle line:84% I'm telling you this magic number. 00:13:39.610 --> 00:13:42.530 align:middle line:84% How you find that magic number is tricky in a lot of cases. 00:13:42.530 --> 00:13:44.780 align:middle line:84% But what we're going to see is, as designers, we often 00:13:44.780 --> 00:13:46.520 align:middle line:90% don't have to worry about that. 00:13:46.520 --> 00:13:51.230 align:middle line:84% So in this case, I could have the same model where 00:13:51.230 --> 00:13:58.820 align:middle line:84% I would just say, RTh is simply equal to hA inverse. 00:13:58.820 --> 00:14:00.440 align:middle line:84% If I told you the area through which I 00:14:00.440 --> 00:14:02.565 align:middle line:84% was transferring heat and I knew that heat transfer 00:14:02.565 --> 00:14:04.810 align:middle line:90% coefficient, great. 00:14:04.810 --> 00:14:09.500 align:middle line:84% I've got my answer, and I go on my merry way. 00:14:09.500 --> 00:14:12.320 align:middle line:84% So putting aside that it's difficult to find this number, 00:14:12.320 --> 00:14:16.220 align:middle line:84% or in some cases, this equivalent number, 00:14:16.220 --> 00:14:17.990 align:middle line:84% we have a means of predicting what it is, 00:14:17.990 --> 00:14:20.630 align:middle line:84% and I can just model that heat transfer as a thermal resistor 00:14:20.630 --> 00:14:21.620 align:middle line:90% again. 00:14:21.620 --> 00:14:23.500 align:middle line:90% Any questions about that? 00:14:23.500 --> 00:14:26.620 align:middle line:90% 00:14:26.620 --> 00:14:30.310 align:middle line:90% So how might we use this? 00:14:30.310 --> 00:14:36.070 align:middle line:84% Well, in the typical case, I'm going 00:14:36.070 --> 00:14:38.990 align:middle line:84% to dissipate some power in my device. 00:14:38.990 --> 00:14:41.600 align:middle line:84% I've got to connect that device through some heat transfer 00:14:41.600 --> 00:14:45.080 align:middle line:84% means, and the heat transfer is going to carry it away. 00:14:45.080 --> 00:14:48.740 align:middle line:90% So here's a typical structure. 00:14:48.740 --> 00:14:50.700 align:middle line:90% Maybe I have some power device. 00:14:50.700 --> 00:14:52.380 align:middle line:90% This could be a MOSFET. 00:14:52.380 --> 00:14:56.930 align:middle line:84% And basically, somewhere inside this device, 00:14:56.930 --> 00:14:58.740 align:middle line:84% I'm dissipating a bunch of power. 00:14:58.740 --> 00:15:02.090 align:middle line:90% So that's a source of heat. 00:15:02.090 --> 00:15:05.030 align:middle line:84% And then what I'm going to do is I'm going to say, 00:15:05.030 --> 00:15:12.020 align:middle line:84% I need to connect that down onto a heat sink. 00:15:12.020 --> 00:15:14.840 align:middle line:84% And here, when I connect this device, 00:15:14.840 --> 00:15:16.190 align:middle line:90% I'm going to have some thermal-- 00:15:16.190 --> 00:15:19.580 align:middle line:84% I'm going to have some interface here. 00:15:19.580 --> 00:15:22.220 align:middle line:90% What is that interface? 00:15:22.220 --> 00:15:24.050 align:middle line:84% I'll talk about that in a second, 00:15:24.050 --> 00:15:25.880 align:middle line:84% but that's going to be between the device, 00:15:25.880 --> 00:15:28.300 align:middle line:90% and then maybe a heat sink. 00:15:28.300 --> 00:15:30.290 align:middle line:84% And maybe the heat sink has a bunch 00:15:30.290 --> 00:15:34.170 align:middle line:84% of fins that's designed to help me cool this thing. 00:15:34.170 --> 00:15:36.410 align:middle line:84% And so the heat is going to come from the device 00:15:36.410 --> 00:15:39.320 align:middle line:84% through this interface into the heat sink, and then maybe 00:15:39.320 --> 00:15:42.920 align:middle line:84% the heat sinks, transferring it off to-- 00:15:42.920 --> 00:15:48.420 align:middle line:84% by convection past the heat sink out into space around it, 00:15:48.420 --> 00:15:50.120 align:middle line:84% and then I forget about the heat. 00:15:50.120 --> 00:15:52.490 align:middle line:90% It's just going out somewhere. 00:15:52.490 --> 00:15:55.790 align:middle line:90% Well, what is this interface? 00:15:55.790 --> 00:15:59.300 align:middle line:84% This device here-- and let me just 00:15:59.300 --> 00:16:02.410 align:middle line:90% illustrate just as an example. 00:16:02.410 --> 00:16:06.310 align:middle line:90% 00:16:06.310 --> 00:16:09.235 align:middle line:90% Here is a typical MOSFET. 00:16:09.235 --> 00:16:14.050 align:middle line:84% This is in a TO-220 FET package, which is maybe not the most 00:16:14.050 --> 00:16:17.480 align:middle line:84% modern of packages, but they're still used in a lot of cases. 00:16:17.480 --> 00:16:20.510 align:middle line:84% And so the device is inside this plastic package. 00:16:20.510 --> 00:16:22.480 align:middle line:84% These leads, gate drain and source 00:16:22.480 --> 00:16:25.893 align:middle line:84% are connected to the device, and then one side of the device 00:16:25.893 --> 00:16:28.060 align:middle line:84% is connected down onto this pad through which you're 00:16:28.060 --> 00:16:29.600 align:middle line:90% going to transfer heat. 00:16:29.600 --> 00:16:31.370 align:middle line:84% So that's the case of the device. 00:16:31.370 --> 00:16:33.970 align:middle line:84% So inside this package somewhere, I'm dissipating heat, 00:16:33.970 --> 00:16:36.580 align:middle line:84% and it goes to the case of the device. 00:16:36.580 --> 00:16:40.330 align:middle line:84% Then I'm going to mount the case of the device down to some heat 00:16:40.330 --> 00:16:42.475 align:middle line:90% sink. 00:16:42.475 --> 00:16:46.960 align:middle line:84% But there is some interface of the heat sink. 00:16:46.960 --> 00:16:51.670 align:middle line:84% Now, in a practical device, it depends on the device type. 00:16:51.670 --> 00:16:54.420 align:middle line:84% In this particular device-- this is a power MOSFET-- 00:16:54.420 --> 00:16:58.338 align:middle line:84% that heat sink-- or that terminal of the device through 00:16:58.338 --> 00:17:00.630 align:middle line:84% which all the heat is going to go is actually connected 00:17:00.630 --> 00:17:02.040 align:middle line:90% to the drain of the device-- 00:17:02.040 --> 00:17:05.140 align:middle line:84% to the high side, the top side of the MOSFET. 00:17:05.140 --> 00:17:05.986 align:middle line:90% Why? 00:17:05.986 --> 00:17:08.069 align:middle line:84% Because it's a vertical MOSFET, and that's the way 00:17:08.069 --> 00:17:10.119 align:middle line:90% they could connect it. 00:17:10.119 --> 00:17:14.040 align:middle line:84% And so if that drain is, say, for a boost converter, 00:17:14.040 --> 00:17:16.885 align:middle line:84% that may be flapping up and down in voltage. 00:17:16.885 --> 00:17:18.510 align:middle line:84% So it's kind of unfortunate that you're 00:17:18.510 --> 00:17:20.468 align:middle line:84% trying to transfer heat out of something that's 00:17:20.468 --> 00:17:22.079 align:middle line:90% flying around in voltage. 00:17:22.079 --> 00:17:30.300 align:middle line:84% So in that case, this thing might be a silicone insulation 00:17:30.300 --> 00:17:30.800 align:middle line:90% pad. 00:17:30.800 --> 00:17:35.800 align:middle line:90% 00:17:35.800 --> 00:17:37.180 align:middle line:90% What does that thing do? 00:17:37.180 --> 00:17:40.510 align:middle line:84% Essentially, it's electrically insulating, but reasonably 00:17:40.510 --> 00:17:43.630 align:middle line:84% thermally conductive, so it lets heat transfer through, but not 00:17:43.630 --> 00:17:44.360 align:middle line:90% electrons. 00:17:44.360 --> 00:17:46.810 align:middle line:84% And by the way, that's not a very common thing. 00:17:46.810 --> 00:17:49.670 align:middle line:84% Most things that are good thermal conductors 00:17:49.670 --> 00:17:51.170 align:middle line:84% are also good electrical conductors, 00:17:51.170 --> 00:17:53.712 align:middle line:84% but you can find some materials that will help you out there. 00:17:53.712 --> 00:17:58.660 align:middle line:84% So this thing might be some electrical insulation pad 00:17:58.660 --> 00:18:01.570 align:middle line:84% to let you connect this thing to a heat sink thermally 00:18:01.570 --> 00:18:04.040 align:middle line:90% but not electrically. 00:18:04.040 --> 00:18:09.730 align:middle line:84% You can find-- a typical brand name for this is like a Sil-Pad. 00:18:09.730 --> 00:18:12.503 align:middle line:84% That's Bergquist's trade name for it. 00:18:12.503 --> 00:18:14.170 align:middle line:84% There's a lot of companies that make it, 00:18:14.170 --> 00:18:16.210 align:middle line:84% and they come in varying degrees of quality 00:18:16.210 --> 00:18:19.870 align:middle line:84% and electrical insulation capability and so forth. 00:18:19.870 --> 00:18:22.300 align:middle line:84% In some applications, maybe you don't 00:18:22.300 --> 00:18:26.130 align:middle line:90% need electrical insulation. 00:18:26.130 --> 00:18:31.200 align:middle line:84% If you happen to have a package that was already electrically 00:18:31.200 --> 00:18:32.910 align:middle line:84% insulated-- and they do make some FETs 00:18:32.910 --> 00:18:35.010 align:middle line:84% that way-- or the package happened 00:18:35.010 --> 00:18:36.510 align:middle line:84% to be connected to ground, maybe you 00:18:36.510 --> 00:18:40.380 align:middle line:84% could connect that directly down to the heatsink. 00:18:40.380 --> 00:18:43.740 align:middle line:84% But often, you will have some other interface 00:18:43.740 --> 00:18:45.330 align:middle line:84% between your device and your heatsink. 00:18:45.330 --> 00:18:46.120 align:middle line:90% Why? 00:18:46.120 --> 00:18:49.830 align:middle line:84% Because if you think about surface roughness, 00:18:49.830 --> 00:18:52.410 align:middle line:84% maybe here's the back package of the device 00:18:52.410 --> 00:18:55.020 align:middle line:84% and here's the surface of the heatsink. 00:18:55.020 --> 00:18:58.510 align:middle line:84% You don't want any air in between it. 00:18:58.510 --> 00:19:01.660 align:middle line:84% So in that case, what you might do is use thermal grease. 00:19:01.660 --> 00:19:06.035 align:middle line:84% So this could also be instead thermal grease. 00:19:06.035 --> 00:19:10.770 align:middle line:90% 00:19:10.770 --> 00:19:12.190 align:middle line:90% Why would I use thermal grease? 00:19:12.190 --> 00:19:14.580 align:middle line:84% Because thermal grease has a lot better heat 00:19:14.580 --> 00:19:17.280 align:middle line:84% transfer characteristics than air does. 00:19:17.280 --> 00:19:20.340 align:middle line:84% So you wouldn't want any scratches or open areas 00:19:20.340 --> 00:19:23.670 align:middle line:84% in the back of your device limiting heat flow, 00:19:23.670 --> 00:19:25.920 align:middle line:84% so you fill that with thermal grease. 00:19:25.920 --> 00:19:29.190 align:middle line:84% Of course, thermal grease has way worse thermal conductivity 00:19:29.190 --> 00:19:30.070 align:middle line:90% than metal. 00:19:30.070 --> 00:19:32.920 align:middle line:84% So what you would do is you just put enough thermal grease on it. 00:19:32.920 --> 00:19:36.240 align:middle line:84% And sometimes, if you've ever mounted, say, a CPU down 00:19:36.240 --> 00:19:38.340 align:middle line:84% to the CPU cooler, they'll give you 00:19:38.340 --> 00:19:40.435 align:middle line:90% this material to put in there. 00:19:40.435 --> 00:19:42.060 align:middle line:84% And the idea is you put it on, and then 00:19:42.060 --> 00:19:43.960 align:middle line:84% you basically try to squeegee it all off. 00:19:43.960 --> 00:19:45.340 align:middle line:90% It's just to fill in cracks. 00:19:45.340 --> 00:19:48.120 align:middle line:84% You don't want like a thick layer of this stuff, 00:19:48.120 --> 00:19:52.650 align:middle line:84% because it will add thermal resistance to this path. 00:19:52.650 --> 00:19:54.330 align:middle line:84% The Sil-Pad adds thermal resistance. 00:19:54.330 --> 00:19:55.292 align:middle line:90% So does thermal grease. 00:19:55.292 --> 00:19:57.750 align:middle line:84% So you put as little as on as you can get away with it just 00:19:57.750 --> 00:20:01.650 align:middle line:84% to make it essentially flat, and then mount that device down 00:20:01.650 --> 00:20:03.840 align:middle line:90% to the heatsink. 00:20:03.840 --> 00:20:05.302 align:middle line:90% Any questions about that? 00:20:05.302 --> 00:20:07.510 align:middle line:84% So it depends whether you need electrical insulation, 00:20:07.510 --> 00:20:09.135 align:middle line:84% or whether you just use thermal grease, 00:20:09.135 --> 00:20:10.740 align:middle line:84% or you use an insulating pad, or you 00:20:10.740 --> 00:20:13.470 align:middle line:84% use some kind of non-insulating thermal pad 00:20:13.470 --> 00:20:15.910 align:middle line:90% that'll be compliant. 00:20:15.910 --> 00:20:18.464 align:middle line:90% 00:20:18.464 --> 00:20:21.410 align:middle line:84% And that's the physical structure. 00:20:21.410 --> 00:20:25.970 align:middle line:84% What does this thing look like as a circuit model? 00:20:25.970 --> 00:20:33.950 align:middle line:84% Well, here, I will have some Pdiss. 00:20:33.950 --> 00:20:37.100 align:middle line:84% That's the power I'm dissipating in the device. 00:20:37.100 --> 00:20:39.620 align:middle line:84% And ideally, that power dissipation 00:20:39.620 --> 00:20:42.360 align:middle line:90% is connected to absolute 0. 00:20:42.360 --> 00:20:45.200 align:middle line:84% But maybe I'll just call this T ambient 00:20:45.200 --> 00:20:50.560 align:middle line:84% and recognize that, oh, maybe this 00:20:50.560 --> 00:20:55.130 align:middle line:90% is connected to really 0 Kelvin. 00:20:55.130 --> 00:20:57.310 align:middle line:84% So I can either treat this as 0 Kelvin 00:20:57.310 --> 00:21:00.490 align:middle line:84% or call it whatever my ambient temperature is 00:21:00.490 --> 00:21:01.870 align:middle line:90% and work from there. 00:21:01.870 --> 00:21:04.450 align:middle line:90% 00:21:04.450 --> 00:21:06.290 align:middle line:90% That's the power dissipation. 00:21:06.290 --> 00:21:12.670 align:middle line:84% And we often say this temperature is at T junction 00:21:12.670 --> 00:21:13.510 align:middle line:90% from-- 00:21:13.510 --> 00:21:16.120 align:middle line:84% so it's either-- this is T junction 00:21:16.120 --> 00:21:19.930 align:middle line:84% to ambient or T junction absolute depending upon how 00:21:19.930 --> 00:21:21.230 align:middle line:90% I decide to write this. 00:21:21.230 --> 00:21:23.828 align:middle line:84% The junction is just my notion of wherever the power is 00:21:23.828 --> 00:21:25.120 align:middle line:90% being dissipated in the device. 00:21:25.120 --> 00:21:27.560 align:middle line:84% It may be an actual physical semiconductor junction. 00:21:27.560 --> 00:21:30.520 align:middle line:84% It may just be wherever in this device is where 00:21:30.520 --> 00:21:34.360 align:middle line:90% the heat is being dissipated. 00:21:34.360 --> 00:21:37.780 align:middle line:84% Then I've got to transfer the heat 00:21:37.780 --> 00:21:40.510 align:middle line:84% to the surface of the device, because I 00:21:40.510 --> 00:21:46.720 align:middle line:84% said the heat's actually being dissipated inside, 00:21:46.720 --> 00:21:48.760 align:middle line:84% and there's some thermal resistance junction 00:21:48.760 --> 00:21:50.620 align:middle line:84% between where the heat is being dissipated 00:21:50.620 --> 00:21:53.660 align:middle line:84% inside the semiconductor to the actual case. 00:21:53.660 --> 00:21:59.790 align:middle line:84% So there's going to be some RTh junction to case. 00:21:59.790 --> 00:22:02.546 align:middle line:90% 00:22:02.546 --> 00:22:06.700 align:middle line:84% And then there's going to be some other thermal resistance, 00:22:06.700 --> 00:22:13.870 align:middle line:84% and so this temperature here would be the T case. 00:22:13.870 --> 00:22:18.110 align:middle line:84% Then there's going to be some other thermal resistance, 00:22:18.110 --> 00:22:23.040 align:middle line:90% RTh case to sink. 00:22:23.040 --> 00:22:25.480 align:middle line:84% That depends upon is this grease, 00:22:25.480 --> 00:22:27.550 align:middle line:90% is it a thermal pad, what is it? 00:22:27.550 --> 00:22:29.880 align:middle line:84% And I care because that's going to contribute 00:22:29.880 --> 00:22:31.300 align:middle line:90% to temperature rise. 00:22:31.300 --> 00:22:35.972 align:middle line:84% And then over here, I have my heat sink. 00:22:35.972 --> 00:22:39.030 align:middle line:84% And that heat sink, if I include what's 00:22:39.030 --> 00:22:42.430 align:middle line:84% going to transfer off it by conduction and convection, 00:22:42.430 --> 00:22:52.330 align:middle line:84% maybe I say there is some sink to ambient. 00:22:52.330 --> 00:22:54.880 align:middle line:84% And I connect this back to ambient, 00:22:54.880 --> 00:23:00.130 align:middle line:84% and that's my thermal model, and this is Tsink. 00:23:00.130 --> 00:23:03.210 align:middle line:84% So this is my model for the whole thing. 00:23:03.210 --> 00:23:05.480 align:middle line:90% Any questions about that? 00:23:05.480 --> 00:23:09.360 align:middle line:84% And I should say this is in thermal steady state. 00:23:09.360 --> 00:23:14.220 align:middle line:84% Now, fortunately, if you get a lot of typical devices, 00:23:14.220 --> 00:23:17.680 align:middle line:84% and you look carefully in the device datasheet, here we go. 00:23:17.680 --> 00:23:21.010 align:middle line:84% What they tell me is they've done some measurements. 00:23:21.010 --> 00:23:24.150 align:middle line:84% Somebody with a lot of thermal modeling experience has said-- 00:23:24.150 --> 00:23:25.650 align:middle line:84% done some measurements to figure out 00:23:25.650 --> 00:23:27.390 align:middle line:84% how hot the junctions are going to get 00:23:27.390 --> 00:23:30.310 align:middle line:84% from above the case for a given amount of dissipation. 00:23:30.310 --> 00:23:32.760 align:middle line:84% And they give me a thermal resistance junction to case. 00:23:32.760 --> 00:23:35.760 align:middle line:84% And they say here, it's 2.5 degrees C per watt. 00:23:35.760 --> 00:23:37.530 align:middle line:84% And then they tell me, oh, if you 00:23:37.530 --> 00:23:40.430 align:middle line:84% go get just the right amount of thermal grease 00:23:40.430 --> 00:23:42.930 align:middle line:84% and you do a really good job of mounting this down to a heat 00:23:42.930 --> 00:23:46.080 align:middle line:84% sink so we're not worrying about any additional insulation pad 00:23:46.080 --> 00:23:48.930 align:middle line:84% here, it's another half a degree C per watt 00:23:48.930 --> 00:23:51.220 align:middle line:90% between the case and the sink. 00:23:51.220 --> 00:23:53.230 align:middle line:84% That's if I don't need electrical isolation. 00:23:53.230 --> 00:23:57.060 align:middle line:84% If I do, I have to look up the thermal resistance of my Sil-Pad 00:23:57.060 --> 00:23:59.270 align:middle line:84% or whatever else I'm putting in there. 00:23:59.270 --> 00:24:03.390 align:middle line:84% And then I'm-- usually, as the designer, 00:24:03.390 --> 00:24:05.610 align:middle line:90% I'm trying to pick this. 00:24:05.610 --> 00:24:09.920 align:middle line:84% But if I went out and picked a particular heat sink, 00:24:09.920 --> 00:24:14.570 align:middle line:84% I can look up its value of RTh sink to ambient. 00:24:14.570 --> 00:24:19.010 align:middle line:84% So then I can find delta T junction 00:24:19.010 --> 00:24:25.460 align:middle line:84% to ambient is simply going to be Pdiss, which is this power 00:24:25.460 --> 00:24:35.060 align:middle line:84% dissipation, times RTh junction to case plus RTh case 00:24:35.060 --> 00:24:42.530 align:middle line:84% to sink plus RTh sink to ambient, and then I'm done. 00:24:42.530 --> 00:24:45.950 align:middle line:90% If, in my example-- 00:24:45.950 --> 00:24:50.960 align:middle line:84% typical example numbers that I might use-- 00:24:50.960 --> 00:25:00.465 align:middle line:84% I said T ambient might be 40 degrees C. I don't know. 00:25:00.465 --> 00:25:02.590 align:middle line:84% That's not-- that's a very that's actually a pretty 00:25:02.590 --> 00:25:04.155 align:middle line:90% cool ambient temperature. 00:25:04.155 --> 00:25:06.530 align:middle line:84% Some industrial stuff you design, they assume the ambient 00:25:06.530 --> 00:25:08.110 align:middle line:90% is going to be 100 degrees C. 00:25:08.110 --> 00:25:13.540 align:middle line:84% But 40 degrees C is hot but not crazy hot if you're in a room. 00:25:13.540 --> 00:25:16.355 align:middle line:90% 00:25:16.355 --> 00:25:20.710 align:middle line:84% And maybe I would say this kind of device, 00:25:20.710 --> 00:25:24.790 align:middle line:84% Pdiss might be equal to 10 watts, 00:25:24.790 --> 00:25:28.000 align:middle line:84% This is something I got from my circuit design. 00:25:28.000 --> 00:25:31.540 align:middle line:84% So here, I could say, that's 10 watts times-- 00:25:31.540 --> 00:25:36.790 align:middle line:84% junction to case I said was 2.5 degrees C per watt, 00:25:36.790 --> 00:25:41.950 align:middle line:84% plus 0.5 degrees C per watt was the-- 00:25:41.950 --> 00:25:44.090 align:middle line:84% if I just had thermal degrees in there. 00:25:44.090 --> 00:25:48.400 align:middle line:84% And if I went out and I bought a Redpoint thermal km 51 00:25:48.400 --> 00:25:51.010 align:middle line:84% heatsink-- that's just one of these convection-cooled heat 00:25:51.010 --> 00:25:54.140 align:middle line:84% sinks to which I might mount this device-- 00:25:54.140 --> 00:25:57.070 align:middle line:90% that is 4.8 degrees C per watt. 00:25:57.070 --> 00:26:00.830 align:middle line:90% 00:26:00.830 --> 00:26:05.390 align:middle line:84% Which if I say this is plus T ambient-- 00:26:05.390 --> 00:26:07.430 align:middle line:84% so this would be just T junction-- would 00:26:07.430 --> 00:26:09.150 align:middle line:90% be T ambient plus this stuff. 00:26:09.150 --> 00:26:11.742 align:middle line:90% 00:26:11.742 --> 00:26:17.730 align:middle line:84% This would give me 118 degrees C, and maybe I'd be a happy guy. 00:26:17.730 --> 00:26:24.210 align:middle line:84% Typical devices that you can buy in plastic packages, 150 to 175 00:26:24.210 --> 00:26:26.220 align:middle line:84% is the maximum temperature you're 00:26:26.220 --> 00:26:28.530 align:middle line:90% allowed to get the junction. 00:26:28.530 --> 00:26:30.600 align:middle line:84% You can get devices that are specially packaged 00:26:30.600 --> 00:26:33.660 align:middle line:84% that will go higher, or if they're not going to-- 00:26:33.660 --> 00:26:35.380 align:middle line:84% bad stuff won't happen to the device. 00:26:35.380 --> 00:26:37.080 align:middle line:84% It's often limited by the packaging. 00:26:37.080 --> 00:26:40.300 align:middle line:84% But typically, 150 or 175 is what you can get, 00:26:40.300 --> 00:26:42.670 align:middle line:84% and this just gives me some margin. 00:26:42.670 --> 00:26:43.570 align:middle line:90% Any-- 00:26:43.570 --> 00:26:45.700 align:middle line:84% AUDIENCE: Is that on the data sheet? 00:26:45.700 --> 00:26:47.440 align:middle line:84% PROFESSOR: Yes, usually, a data sheet 00:26:47.440 --> 00:26:50.920 align:middle line:84% will tell you what the maximum allowable temperature-- here we 00:26:50.920 --> 00:26:52.000 align:middle line:90% go. 00:26:52.000 --> 00:26:55.210 align:middle line:84% Operating and storage temperature range-- 00:26:55.210 --> 00:26:57.490 align:middle line:84% operating junction and storage temperature range. 00:26:57.490 --> 00:27:00.310 align:middle line:90% 150 degrees C for this part. 00:27:00.310 --> 00:27:03.292 align:middle line:84% A lot of more modern devices might be 175. 00:27:03.292 --> 00:27:04.750 align:middle line:84% It just depends what you're buying. 00:27:04.750 --> 00:27:07.240 align:middle line:84% So what would you be-- really job? 00:27:07.240 --> 00:27:09.970 align:middle line:84% Usually, what you do is-- you know all this stuff. 00:27:09.970 --> 00:27:11.890 align:middle line:84% You know your ambience of specification, 00:27:11.890 --> 00:27:14.110 align:middle line:84% you know your dissipation because you calculated it, 00:27:14.110 --> 00:27:15.950 align:middle line:90% you know these things. 00:27:15.950 --> 00:27:19.540 align:middle line:84% And your job is to go pick the heat sink or maybe 00:27:19.540 --> 00:27:21.760 align:middle line:90% the thermal interface material. 00:27:21.760 --> 00:27:24.370 align:middle line:84% And you have to pick this so that this number doesn't 00:27:24.370 --> 00:27:25.510 align:middle line:90% get too high. 00:27:25.510 --> 00:27:27.260 align:middle line:84% So I don't know this stuff in advance. 00:27:27.260 --> 00:27:31.870 align:middle line:84% This is the thing I'm choosing as the heat sink designer. 00:27:31.870 --> 00:27:34.850 align:middle line:90% What happens if you can't do it? 00:27:34.850 --> 00:27:37.720 align:middle line:84% Well, you either got to go reduce your dissipation, 00:27:37.720 --> 00:27:40.600 align:middle line:84% or find a better package with a lower junction to case, 00:27:40.600 --> 00:27:43.098 align:middle line:84% or some other combination of things. 00:27:43.098 --> 00:27:44.890 align:middle line:84% And you're often pushed up against the wall 00:27:44.890 --> 00:27:47.390 align:middle line:84% if you're-- especially if you're trying to go for high power 00:27:47.390 --> 00:27:49.070 align:middle line:90% density. 00:27:49.070 --> 00:27:50.770 align:middle line:90% Questions? 00:27:50.770 --> 00:27:53.440 align:middle line:84% AUDIENCE: So it seems like a lot to do for one 00:27:53.440 --> 00:27:55.833 align:middle line:90% component in your system. 00:27:55.833 --> 00:27:57.250 align:middle line:84% What do you do for a workflow when 00:27:57.250 --> 00:28:00.610 align:middle line:84% you have a lot of components in the circuit [INAUDIBLE]? 00:28:00.610 --> 00:28:03.310 align:middle line:84% PROFESSOR: So if you're a power converter designer, usually, 00:28:03.310 --> 00:28:05.830 align:middle line:84% you only have a few sources of thermal dissipation 00:28:05.830 --> 00:28:07.840 align:middle line:84% that are the most important, and they're often 00:28:07.840 --> 00:28:09.640 align:middle line:90% your power devices. 00:28:09.640 --> 00:28:13.810 align:middle line:84% At a broader scale, if you have to consider 00:28:13.810 --> 00:28:17.800 align:middle line:84% the whole thermal design, well, either you get a bigger network, 00:28:17.800 --> 00:28:22.330 align:middle line:84% or you start to pull out computational finite element 00:28:22.330 --> 00:28:23.960 align:middle line:90% tools or something like that. 00:28:23.960 --> 00:28:26.577 align:middle line:84% But very often, for a power converter design, 00:28:26.577 --> 00:28:28.660 align:middle line:84% you think-- you've got your-- in a buck converter, 00:28:28.660 --> 00:28:30.880 align:middle line:84% you might have your MOSFET and your diode 00:28:30.880 --> 00:28:33.100 align:middle line:84% as being big sources of loss, and then 00:28:33.100 --> 00:28:35.290 align:middle line:90% maybe your inductor is next. 00:28:35.290 --> 00:28:37.670 align:middle line:84% And you just model those into the system 00:28:37.670 --> 00:28:42.000 align:middle line:84% and figure out how you're going to heatsink it. 00:28:42.000 --> 00:28:44.430 align:middle line:84% I should say, by the way, that a very typical thing 00:28:44.430 --> 00:28:46.990 align:middle line:90% that you would do is-- 00:28:46.990 --> 00:28:49.718 align:middle line:90% 00:28:49.718 --> 00:28:51.010 align:middle line:90% a very typical thing to do is-- 00:28:51.010 --> 00:28:52.670 align:middle line:84% I'm showing you one picture here, 00:28:52.670 --> 00:28:56.080 align:middle line:84% but it may be true that you come up 00:28:56.080 --> 00:29:03.300 align:middle line:90% and you might do the following. 00:29:03.300 --> 00:29:09.920 align:middle line:84% You might have a heat sink, and then your diode and your MOSFET 00:29:09.920 --> 00:29:12.775 align:middle line:84% are right next to each other and mounted to the same heatsink. 00:29:12.775 --> 00:29:13.650 align:middle line:90% So what would you do? 00:29:13.650 --> 00:29:20.510 align:middle line:84% You'd say, here's PFET and here's TJ-FET. 00:29:20.510 --> 00:29:23.720 align:middle line:90% 00:29:23.720 --> 00:29:26.180 align:middle line:90% And that would be RTh. 00:29:26.180 --> 00:29:28.200 align:middle line:84% You'd have junction to case, case to sink. 00:29:28.200 --> 00:29:30.050 align:middle line:90% So maybe that would RTh-- 00:29:30.050 --> 00:29:32.240 align:middle line:84% junction to sink would be the total. 00:29:32.240 --> 00:29:36.020 align:middle line:84% And you'd do the same thing for RTh junction 00:29:36.020 --> 00:29:38.603 align:middle line:90% to sink for your diode. 00:29:38.603 --> 00:29:43.860 align:middle line:90% 00:29:43.860 --> 00:29:48.480 align:middle line:90% And here's Tj diode. 00:29:48.480 --> 00:29:54.420 align:middle line:84% And then you'd go out and get one RTh sink to ambient, 00:29:54.420 --> 00:29:56.520 align:middle line:84% and you'd go calculate what the junction 00:29:56.520 --> 00:29:58.020 align:middle line:84% temperature of the diode is and what the junction 00:29:58.020 --> 00:29:59.103 align:middle line:90% temperature of the FET is. 00:29:59.103 --> 00:30:01.800 align:middle line:84% And you'd have to pick these things low enough, 00:30:01.800 --> 00:30:05.490 align:middle line:84% and you can expand out from there. 00:30:05.490 --> 00:30:06.490 align:middle line:90% Questions? 00:30:06.490 --> 00:30:11.120 align:middle line:90% 00:30:11.120 --> 00:30:15.650 align:middle line:84% So that's one way to think about it. 00:30:15.650 --> 00:30:19.280 align:middle line:90% 00:30:19.280 --> 00:30:23.660 align:middle line:84% This is the steady state thermal model. 00:30:23.660 --> 00:30:27.432 align:middle line:84% This is, I plug my converter in, I let it run for a long time. 00:30:27.432 --> 00:30:29.390 align:middle line:84% Everything heats up to its maximum temperature, 00:30:29.390 --> 00:30:30.990 align:middle line:84% and I find out if it blows up or not. 00:30:30.990 --> 00:30:34.726 align:middle line:84% Did the thing get too hot when everything settled out? 00:30:34.726 --> 00:30:38.420 align:middle line:84% But this isn't really a transient thermal model. 00:30:38.420 --> 00:30:42.740 align:middle line:84% I mean, if I put my tea kettle on, 00:30:42.740 --> 00:30:45.380 align:middle line:84% and I turn on the power, which is basically dumping heat 00:30:45.380 --> 00:30:46.820 align:middle line:84% into the water from my tea kettle, 00:30:46.820 --> 00:30:49.730 align:middle line:84% it doesn't instantly get up to maximum temperature. 00:30:49.730 --> 00:30:52.010 align:middle line:84% Eventually, it'll get up to-- hopefully, 00:30:52.010 --> 00:30:55.502 align:middle line:84% if I have-- pouring enough heat in to get up to boiling, 00:30:55.502 --> 00:30:57.810 align:middle line:90% but it doesn't do it instantly. 00:30:57.810 --> 00:31:05.090 align:middle line:84% This model does not tell you at all about instantaneous behavior 00:31:05.090 --> 00:31:06.810 align:middle line:90% or transient behavior. 00:31:06.810 --> 00:31:11.570 align:middle line:84% And sometimes that's quite important, because we sometimes 00:31:11.570 --> 00:31:13.760 align:middle line:84% don't even wait for things to get to steady state, 00:31:13.760 --> 00:31:17.040 align:middle line:84% or we want to know, how quickly is my BIEST going to heat up? 00:31:17.040 --> 00:31:21.920 align:middle line:84% Well, it turns out you can measure that as well, 00:31:21.920 --> 00:31:23.700 align:middle line:90% or you can model that as well. 00:31:23.700 --> 00:31:26.690 align:middle line:84% And what we do with it is to model things 00:31:26.690 --> 00:31:37.430 align:middle line:84% with a thermal capacitance which has units of joules per degree C 00:31:37.430 --> 00:31:40.280 align:middle line:90% or joules per Kelvin. 00:31:40.280 --> 00:31:43.430 align:middle line:84% How many joules do I have to put into this structure 00:31:43.430 --> 00:31:46.100 align:middle line:84% to raise it one degree Kelvin just by heating up 00:31:46.100 --> 00:31:48.236 align:middle line:90% the mass of this thing? 00:31:48.236 --> 00:31:50.240 align:middle line:84% And so if I know the heat capacitance 00:31:50.240 --> 00:31:55.160 align:middle line:84% for some material per volume or per mass, 00:31:55.160 --> 00:31:58.010 align:middle line:84% I can multiply that and get the heat capacitance 00:31:58.010 --> 00:31:59.842 align:middle line:90% of the whole thing. 00:31:59.842 --> 00:32:02.210 align:middle line:84% So maybe what I would do is I would say, 00:32:02.210 --> 00:32:05.550 align:middle line:84% suppose I have some surface here, 00:32:05.550 --> 00:32:10.370 align:middle line:84% which I know is going to be at some T ambient-- 00:32:10.370 --> 00:32:12.148 align:middle line:90% or let me call this T2. 00:32:12.148 --> 00:32:16.910 align:middle line:90% 00:32:16.910 --> 00:32:18.950 align:middle line:90% I guess I did call it T ambient. 00:32:18.950 --> 00:32:21.200 align:middle line:90% So suppose it's T ambient. 00:32:21.200 --> 00:32:27.200 align:middle line:84% And here, I have some block of material-- 00:32:27.200 --> 00:32:30.220 align:middle line:90% piece of element of my system. 00:32:30.220 --> 00:32:34.890 align:middle line:84% And let me characterize this by its average temperature T1. 00:32:34.890 --> 00:32:37.840 align:middle line:84% And now, I'm going to go pour heat into this thing, 00:32:37.840 --> 00:32:41.280 align:middle line:84% Q. This is the power dissipation. 00:32:41.280 --> 00:32:42.510 align:middle line:90% Maybe it's at this surface. 00:32:42.510 --> 00:32:44.302 align:middle line:84% This is going in and heating this thing up, 00:32:44.302 --> 00:32:46.420 align:middle line:84% and this thing is transferring energy to ground. 00:32:46.420 --> 00:32:47.620 align:middle line:90% How would I model that? 00:32:47.620 --> 00:32:53.721 align:middle line:84% Well, the steady state model would look like this. 00:32:53.721 --> 00:32:55.800 align:middle line:90% So this is Pdiss-- 00:32:55.800 --> 00:32:58.020 align:middle line:90% or this is Q, I guess. 00:32:58.020 --> 00:32:59.430 align:middle line:90% What did I call it in my note? 00:32:59.430 --> 00:33:08.470 align:middle line:84% I guess I called it Q. Then I have some-- 00:33:08.470 --> 00:33:17.590 align:middle line:84% if this is T ambient, this is RTh and this is T1. 00:33:17.590 --> 00:33:22.440 align:middle line:84% So this is the thermal model in steady state for this thing. 00:33:22.440 --> 00:33:24.630 align:middle line:84% How do I account for the fact that this thing 00:33:24.630 --> 00:33:27.350 align:middle line:84% has some thermal mass that takes time to heat up? 00:33:27.350 --> 00:33:30.180 align:middle line:84% Well, I find out what is thermal capacitance is 00:33:30.180 --> 00:33:33.590 align:middle line:90% and I put it right here. 00:33:33.590 --> 00:33:34.090 align:middle line:90% CTh. 00:33:34.090 --> 00:33:38.080 align:middle line:90% 00:33:38.080 --> 00:33:42.790 align:middle line:84% And now, this gives me a model so that if at some time-- 00:33:42.790 --> 00:33:47.590 align:middle line:84% suppose here's T. If I come in here and I say, 00:33:47.590 --> 00:33:53.230 align:middle line:90% here is Q of T. It steps. 00:33:53.230 --> 00:33:57.710 align:middle line:84% This is Q of T, and at T equals 0, it turns on. 00:33:57.710 --> 00:34:02.950 align:middle line:84% What I'm going to get is a T1 with respect 00:34:02.950 --> 00:34:06.460 align:middle line:84% to ambient that's going to do this. 00:34:06.460 --> 00:34:11.208 align:middle line:90% 00:34:11.208 --> 00:34:17.159 align:middle line:84% Eventually, I'm going to get the final value is going to be-- 00:34:17.159 --> 00:34:24.000 align:middle line:84% T1 minus t ambient is going to be equal to Q times RTh. 00:34:24.000 --> 00:34:25.739 align:middle line:84% That's this final value, and it's 00:34:25.739 --> 00:34:29.340 align:middle line:84% going to rise with a time constant tau that's 00:34:29.340 --> 00:34:32.719 align:middle line:90% equal to Rh Th-- 00:34:32.719 --> 00:34:37.310 align:middle line:90% Rh-- RTh CTh. 00:34:37.310 --> 00:34:40.040 align:middle line:90% So it'll have a time constant. 00:34:40.040 --> 00:34:42.050 align:middle line:84% It'll rise up, so it won't immediately get hot. 00:34:42.050 --> 00:34:44.739 align:middle line:90% It'll just rise over time. 00:34:44.739 --> 00:34:51.179 align:middle line:84% Now, I should say, fundamentally, 00:34:51.179 --> 00:34:55.199 align:middle line:84% a thermal capacitance should technically always 00:34:55.199 --> 00:34:58.910 align:middle line:90% be referenced to absolute 0. 00:34:58.910 --> 00:35:01.153 align:middle line:84% If I'm just dealing with some ambient, 00:35:01.153 --> 00:35:02.570 align:middle line:84% and I want to call it an offset, I 00:35:02.570 --> 00:35:04.740 align:middle line:84% can reference it to a fixed temperature. 00:35:04.740 --> 00:35:06.740 align:middle line:84% But thermal capacitance is in terms 00:35:06.740 --> 00:35:08.840 align:middle line:90% of modeling them heating up. 00:35:08.840 --> 00:35:11.450 align:middle line:84% I really have to-- in terms of modeling 00:35:11.450 --> 00:35:14.180 align:middle line:84% the physics, I always have to have that thermal capacitance 00:35:14.180 --> 00:35:17.720 align:middle line:84% to my reference point, which would ideally 00:35:17.720 --> 00:35:21.270 align:middle line:90% be 0 degrees Kelvin. 00:35:21.270 --> 00:35:23.020 align:middle line:84% And I'll come back to that in a moment. 00:35:23.020 --> 00:35:24.395 align:middle line:84% Any questions about that, though? 00:35:24.395 --> 00:35:27.225 align:middle line:90% 00:35:27.225 --> 00:35:29.600 align:middle line:84% AUDIENCE: So in this case where we have multiple devices, 00:35:29.600 --> 00:35:33.690 align:middle line:84% how would we insert the thermal capacitances? 00:35:33.690 --> 00:35:34.760 align:middle line:90% PROFESSOR: That's right. 00:35:34.760 --> 00:35:39.012 align:middle line:84% So in this case, if I was talking-- 00:35:39.012 --> 00:35:41.220 align:middle line:84% and you've got to say the thermal capacitance of what 00:35:41.220 --> 00:35:43.190 align:middle line:90% now-- the heat sink, the device? 00:35:43.190 --> 00:35:45.170 align:middle line:84% If I was talking about the thermal capacitance 00:35:45.170 --> 00:35:48.850 align:middle line:84% of the device, I would put that right here. 00:35:48.850 --> 00:35:52.670 align:middle line:84% So this device package might have a thermal capacitance, 00:35:52.670 --> 00:35:54.890 align:middle line:84% and this device might have a thermal capacitance, 00:35:54.890 --> 00:35:59.040 align:middle line:84% and I ought to place it to the reference potential. 00:35:59.040 --> 00:36:01.040 align:middle line:84% The heat sink might have a thermal capacitance, 00:36:01.040 --> 00:36:03.560 align:middle line:84% and then I get a higher order thing. 00:36:03.560 --> 00:36:06.650 align:middle line:84% Now, when I get to T equals infinity 00:36:06.650 --> 00:36:08.262 align:middle line:84% and I have some step response here, 00:36:08.262 --> 00:36:10.345 align:middle line:84% it all settles out to the same steady state model. 00:36:10.345 --> 00:36:14.070 align:middle line:90% 00:36:14.070 --> 00:36:17.800 align:middle line:84% Along those lines, what happens if I'm not 00:36:17.800 --> 00:36:20.620 align:middle line:84% too happy with the resolution I have here? 00:36:20.620 --> 00:36:24.370 align:middle line:84% I'm basically treating T1 of this whole block 00:36:24.370 --> 00:36:27.825 align:middle line:90% as being the same thing. 00:36:27.825 --> 00:36:30.480 align:middle line:84% But in reality, I recognize if the heat's really 00:36:30.480 --> 00:36:35.030 align:middle line:84% coming in over here-- this ends hotter, this ends cooler. 00:36:35.030 --> 00:36:37.440 align:middle line:84% It's not got-- spatially, it doesn't 00:36:37.440 --> 00:36:39.690 align:middle line:84% have the same temperature everywhere, 00:36:39.690 --> 00:36:41.518 align:middle line:84% and maybe even temporally, it doesn't 00:36:41.518 --> 00:36:43.060 align:middle line:84% have the same temperature everywhere. 00:36:43.060 --> 00:36:44.460 align:middle line:84% If I'm not happy with this model, 00:36:44.460 --> 00:36:46.290 align:middle line:84% I can break it up into more lumps. 00:36:46.290 --> 00:36:52.290 align:middle line:84% So I can come over and say here's my BIEST. 00:36:52.290 --> 00:36:58.530 align:middle line:84% Let me break this up into T1, T2. 00:36:58.530 --> 00:37:04.768 align:middle line:84% Here's T ambient, here's Q. And now, I 00:37:04.768 --> 00:37:06.060 align:middle line:90% have a different thermal model. 00:37:06.060 --> 00:37:08.820 align:middle line:90% I have Q coming in. 00:37:08.820 --> 00:37:18.000 align:middle line:84% I could say, here's CTh1 for just this region. 00:37:18.000 --> 00:37:24.660 align:middle line:84% Here's RTh 1 to 2, which is the thermal resistance here. 00:37:24.660 --> 00:37:33.960 align:middle line:84% Here's CTh for this block, and here's the thermal resistance 00:37:33.960 --> 00:37:35.830 align:middle line:90% coming back to ambient here. 00:37:35.830 --> 00:37:40.480 align:middle line:84% And this is going to be T2, this is going to be T1. 00:37:40.480 --> 00:37:45.450 align:middle line:84% And now I have something that's more spatially accurate. 00:37:45.450 --> 00:37:49.700 align:middle line:84% If all things were equal, RTh, CTh would be a factor of 4 00:37:49.700 --> 00:37:50.200 align:middle line:90% smaller. 00:37:50.200 --> 00:37:53.580 align:middle line:84% I'd get better time resolution and better spatial resolution 00:37:53.580 --> 00:37:56.180 align:middle line:84% as to how I'm modeling my system. 00:37:56.180 --> 00:37:59.150 align:middle line:84% If I carry this to an extreme, I get-- basically, 00:37:59.150 --> 00:38:00.710 align:middle line:84% I'm back to Fourier heat transfer 00:38:00.710 --> 00:38:02.460 align:middle line:84% and partial differential equations. 00:38:02.460 --> 00:38:06.890 align:middle line:84% So this is your best lumped model approximation 00:38:06.890 --> 00:38:08.035 align:middle line:90% to Fourier heat transfer. 00:38:08.035 --> 00:38:10.770 align:middle line:90% 00:38:10.770 --> 00:38:11.920 align:middle line:90% Any other questions? 00:38:11.920 --> 00:38:17.650 align:middle line:90% 00:38:17.650 --> 00:38:27.720 align:middle line:84% How do people model this in, say, the thermal capacitance 00:38:27.720 --> 00:38:28.740 align:middle line:90% aspect? 00:38:28.740 --> 00:38:30.930 align:middle line:84% Sometimes you can look up the thermal-- 00:38:30.930 --> 00:38:33.517 align:middle line:84% if you have a material, you can go look up its heat capacity, 00:38:33.517 --> 00:38:35.100 align:middle line:84% and then multiply by the mass, and you 00:38:35.100 --> 00:38:37.710 align:middle line:84% get a thermal capacitance, and you can do it 00:38:37.710 --> 00:38:39.400 align:middle line:90% from fundamentals this way. 00:38:39.400 --> 00:38:43.230 align:middle line:84% That's not always the way they give it to you for a device, 00:38:43.230 --> 00:38:45.830 align:middle line:84% because if I think about it, a device is kind 00:38:45.830 --> 00:38:47.080 align:middle line:90% of a complicated thing, right. 00:38:47.080 --> 00:38:49.860 align:middle line:84% There's the silicon, and then maybe inside there's 00:38:49.860 --> 00:38:52.980 align:middle line:84% some header, and there's different stuff going on. 00:38:52.980 --> 00:38:55.530 align:middle line:84% Sometimes, what they will do is they 00:38:55.530 --> 00:39:01.460 align:middle line:90% will give you a lumped model. 00:39:01.460 --> 00:39:04.700 align:middle line:84% And unfortunately, this is often called 00:39:04.700 --> 00:39:19.730 align:middle line:84% a transient thermal impedance, or ZTh. 00:39:19.730 --> 00:39:22.400 align:middle line:84% And it's a-- I say unfortunately because it's not 00:39:22.400 --> 00:39:24.170 align:middle line:84% an impedance in the way you think of. 00:39:24.170 --> 00:39:28.488 align:middle line:84% An impedance is a frequency domain thing, 00:39:28.488 --> 00:39:31.630 align:middle line:84% but this is just the language that has stuck. 00:39:31.630 --> 00:39:33.680 align:middle line:84% It's unfortunate, but that's what they call it. 00:39:33.680 --> 00:39:35.430 align:middle line:84% But it's not an impedance at all. 00:39:35.430 --> 00:39:36.710 align:middle line:90% Here is an example. 00:39:36.710 --> 00:39:38.680 align:middle line:90% This is the transient-- 00:39:38.680 --> 00:39:40.630 align:middle line:84% or here, they actually they call it 00:39:40.630 --> 00:39:44.260 align:middle line:84% the maximum effective transient thermal impedance 00:39:44.260 --> 00:39:46.290 align:middle line:90% junction to case. 00:39:46.290 --> 00:39:48.070 align:middle line:90% So what does this thing mean? 00:39:48.070 --> 00:39:51.490 align:middle line:84% This thing means if I took the device case, 00:39:51.490 --> 00:39:54.580 align:middle line:84% and I mounted it against a really mega good heat 00:39:54.580 --> 00:39:56.900 align:middle line:84% sink that's just going to hold the temperature exactly 00:39:56.900 --> 00:39:58.600 align:middle line:90% at ambient-- 00:39:58.600 --> 00:40:00.710 align:middle line:90% a perfect heat sink-- 00:40:00.710 --> 00:40:04.150 align:middle line:84% then I stepped the power into it. 00:40:04.150 --> 00:40:07.590 align:middle line:84% And when they say a single pulse, what this really is, 00:40:07.590 --> 00:40:09.820 align:middle line:90% is a plot of the step response. 00:40:09.820 --> 00:40:13.140 align:middle line:84% It's basically this plot for the junction temperature 00:40:13.140 --> 00:40:17.230 align:middle line:84% above the case, where the case is held exactly at constant. 00:40:17.230 --> 00:40:19.470 align:middle line:84% So what this is saying is if I put 00:40:19.470 --> 00:40:22.950 align:middle line:84% if I put one pulse into this device, 00:40:22.950 --> 00:40:26.340 align:middle line:84% and I waited 10 to the minus 4 seconds, 00:40:26.340 --> 00:40:32.680 align:middle line:84% I would have a 0.1 degree C per watt of that step rise. 00:40:32.680 --> 00:40:36.490 align:middle line:84% So if I had a 10 watt step, I'd have a 1 degree C rise 00:40:36.490 --> 00:40:40.040 align:middle line:90% after 100 microseconds. 00:40:40.040 --> 00:40:45.410 align:middle line:84% After a millisecond, I'd have a higher value, et cetera. 00:40:45.410 --> 00:40:49.340 align:middle line:84% So this curve-- oh, you-- yeah, if you can see my cursor, 00:40:49.340 --> 00:40:55.760 align:middle line:84% this curve is really a log log plot of the step response. 00:40:55.760 --> 00:40:58.430 align:middle line:84% If it has one time constant, it would look like one thing. 00:40:58.430 --> 00:41:00.200 align:middle line:84% If it has multiple time constants, 00:41:00.200 --> 00:41:02.420 align:middle line:90% the shape of this curve changes. 00:41:02.420 --> 00:41:04.650 align:middle line:84% So a transient thermal impedance, 00:41:04.650 --> 00:41:07.940 align:middle line:84% what they're really telling you is essentially, 00:41:07.940 --> 00:41:11.120 align:middle line:84% this step response for the device assuming it's 00:41:11.120 --> 00:41:14.300 align:middle line:90% against the perfect heatsink. 00:41:14.300 --> 00:41:16.070 align:middle line:84% What they also sometimes tell you 00:41:16.070 --> 00:41:18.092 align:middle line:90% is if I now have little pulses. 00:41:18.092 --> 00:41:20.300 align:middle line:84% I step it up, and I step it down, and I wait a while. 00:41:20.300 --> 00:41:22.260 align:middle line:84% I step it up, and I step it down, and I wait a while. 00:41:22.260 --> 00:41:24.170 align:middle line:84% They will tell you the maximum temperature 00:41:24.170 --> 00:41:28.500 align:middle line:84% it will get during the step for different duty ratios of pulses. 00:41:28.500 --> 00:41:29.870 align:middle line:90% Why do they tell you this? 00:41:29.870 --> 00:41:35.330 align:middle line:84% Because look, if I put 100 watts into this device continuously, 00:41:35.330 --> 00:41:38.210 align:middle line:84% and I put the case at 0 degrees C 00:41:38.210 --> 00:41:40.060 align:middle line:84% and I held it there continuously, 00:41:40.060 --> 00:41:42.740 align:middle line:84% what this would say is 100 times 2.5 00:41:42.740 --> 00:41:45.260 align:middle line:84% would give me 250 degrees C rise. 00:41:45.260 --> 00:41:48.490 align:middle line:84% Well, that would blow the thing up, 00:41:48.490 --> 00:41:50.680 align:middle line:84% because this said 150-degree C device. 00:41:50.680 --> 00:41:53.170 align:middle line:84% But if I only did it for a millisecond, 00:41:53.170 --> 00:41:59.110 align:middle line:84% this thing would tell me that I'd have only 0.1, 0.2-- 00:41:59.110 --> 00:42:00.700 align:middle line:90% I'd be right here on the curve. 00:42:00.700 --> 00:42:03.760 align:middle line:84% Like 0.3 degrees C per watt, and maybe it 00:42:03.760 --> 00:42:06.310 align:middle line:84% would only go up 30 degrees C. So I 00:42:06.310 --> 00:42:08.800 align:middle line:84% can have a 1 millisecond pulse that's really big, 00:42:08.800 --> 00:42:10.780 align:middle line:84% and it won't kill the device because 00:42:10.780 --> 00:42:16.120 align:middle line:84% of the thermal capacitance of the device in its package. 00:42:16.120 --> 00:42:18.790 align:middle line:84% So thermal-- transient thermal impedance 00:42:18.790 --> 00:42:23.310 align:middle line:84% is a step response plotted on a log log plot. 00:42:23.310 --> 00:42:26.230 align:middle line:90% 00:42:26.230 --> 00:42:28.240 align:middle line:84% The other thing I should tell you 00:42:28.240 --> 00:42:32.470 align:middle line:84% is that you could somehow extract that model 00:42:32.470 --> 00:42:35.000 align:middle line:84% and come up with a thermal model for the device. 00:42:35.000 --> 00:42:39.360 align:middle line:84% Maybe it's only one RC that you approximate it with. 00:42:39.360 --> 00:42:42.070 align:middle line:84% But in a complicated big module, it 00:42:42.070 --> 00:42:44.080 align:middle line:84% might be a more complicated model, 00:42:44.080 --> 00:42:46.570 align:middle line:84% because there might be the thermal capacitance 00:42:46.570 --> 00:42:48.940 align:middle line:84% of the silicon, the thermal capacitance of the header 00:42:48.940 --> 00:42:49.640 align:middle line:90% and so forth. 00:42:49.640 --> 00:42:52.762 align:middle line:90% 00:42:52.762 --> 00:42:56.370 align:middle line:84% You will often see thermal models-- 00:42:56.370 --> 00:42:58.620 align:middle line:84% this is called a Cauer form network. 00:42:58.620 --> 00:43:00.560 align:middle line:84% You will often see thermal models 00:43:00.560 --> 00:43:02.720 align:middle line:84% in datasheets, or app notes, or something 00:43:02.720 --> 00:43:10.380 align:middle line:84% that are something like this and that are done like this. 00:43:10.380 --> 00:43:13.730 align:middle line:84% The thermal resistances and thermal capacitances, 00:43:13.730 --> 00:43:18.000 align:middle line:84% that's sort of a foster form of network. 00:43:18.000 --> 00:43:20.430 align:middle line:84% That's just wrong, unfortunately. 00:43:20.430 --> 00:43:23.220 align:middle line:84% You can kind of try to take this kind of model 00:43:23.220 --> 00:43:25.240 align:middle line:84% and map it over to that kind of model, 00:43:25.240 --> 00:43:27.240 align:middle line:84% but any real thermal capacitance, if you're 00:43:27.240 --> 00:43:29.790 align:middle line:84% going to represent Fourier heat transfer, 00:43:29.790 --> 00:43:34.373 align:middle line:84% thermal capacitances ought to always be down to-- 00:43:34.373 --> 00:43:36.540 align:middle line:84% we can use ambient, but it should always technically 00:43:36.540 --> 00:43:37.720 align:middle line:90% be absolute 0. 00:43:37.720 --> 00:43:40.590 align:middle line:84% These are referencing the thermal capacitance 00:43:40.590 --> 00:43:43.430 align:middle line:84% to other places in the thermal path 00:43:43.430 --> 00:43:45.820 align:middle line:84% where the temperature is changing. 00:43:45.820 --> 00:43:47.900 align:middle line:84% So it's basically just not right, 00:43:47.900 --> 00:43:51.160 align:middle line:84% but people do it because it's convenient. 00:43:51.160 --> 00:43:55.155 align:middle line:84% And I suppose if the thermal capacitance 00:43:55.155 --> 00:43:57.280 align:middle line:84% here or the thermal resistance here were super low, 00:43:57.280 --> 00:44:00.530 align:middle line:84% maybe it's almost like this was at Earth, 00:44:00.530 --> 00:44:02.600 align:middle line:84% so they can get away with it sometimes. 00:44:02.600 --> 00:44:03.650 align:middle line:90% But it's just a warning. 00:44:03.650 --> 00:44:05.560 align:middle line:84% If you want to stick with the physics, 00:44:05.560 --> 00:44:07.680 align:middle line:84% this is the correct kind of model to use. 00:44:07.680 --> 00:44:11.710 align:middle line:90% 00:44:11.710 --> 00:44:14.150 align:middle line:84% So that's just giving you a heads up. 00:44:14.150 --> 00:44:17.110 align:middle line:84% I want to just tell you about two more things. 00:44:17.110 --> 00:44:19.788 align:middle line:84% What I've told you so far is if I have a big package, 00:44:19.788 --> 00:44:21.830 align:middle line:84% then I'm going to mount it against the heat sink, 00:44:21.830 --> 00:44:24.190 align:middle line:84% and that's a very typical thing you have to do. 00:44:24.190 --> 00:44:27.460 align:middle line:84% In some cases-- and in a lot of designs, 00:44:27.460 --> 00:44:28.580 align:middle line:90% it doesn't work that way. 00:44:28.580 --> 00:44:30.038 align:middle line:84% You might have something like this. 00:44:30.038 --> 00:44:33.550 align:middle line:84% Here's my device that I might mount-- instead of mounting 00:44:33.550 --> 00:44:37.670 align:middle line:84% my device directly to a heat sink, 00:44:37.670 --> 00:44:40.550 align:middle line:84% you will often have a circuit board. 00:44:40.550 --> 00:44:43.015 align:middle line:90% Here's my circuit board. 00:44:43.015 --> 00:44:45.670 align:middle line:84% And I'll have some device, and maybe this device 00:44:45.670 --> 00:44:48.550 align:middle line:90% will have a heat pad. 00:44:48.550 --> 00:44:49.717 align:middle line:90% Here's my package. 00:44:49.717 --> 00:44:51.550 align:middle line:84% It'll have a bunch of electrical connections 00:44:51.550 --> 00:44:54.200 align:middle line:84% off the device onto the circuit board. 00:44:54.200 --> 00:44:56.710 align:middle line:84% So the first thing I do is I mount my device 00:44:56.710 --> 00:44:59.640 align:middle line:90% to the circuit board. 00:44:59.640 --> 00:45:04.540 align:middle line:84% And my electrical connections would come off of that. 00:45:04.540 --> 00:45:05.963 align:middle line:90% I'd run my gate and my source. 00:45:05.963 --> 00:45:07.630 align:middle line:84% I'd have a bunch of wires on the circuit 00:45:07.630 --> 00:45:08.870 align:middle line:90% board going different places. 00:45:08.870 --> 00:45:10.990 align:middle line:84% So I've made my electrical connection to device, 00:45:10.990 --> 00:45:13.650 align:middle line:84% but now I also need to get the heat out of it. 00:45:13.650 --> 00:45:16.210 align:middle line:90% So what do I do? 00:45:16.210 --> 00:45:18.857 align:middle line:84% Often, what you do is, instead of trying 00:45:18.857 --> 00:45:20.940 align:middle line:84% to mount the device directly against the heatsink, 00:45:20.940 --> 00:45:22.740 align:middle line:84% you mount the device to the board, 00:45:22.740 --> 00:45:24.300 align:middle line:84% carry the heat through the board, 00:45:24.300 --> 00:45:26.020 align:middle line:84% and then to a heatsink on the other side. 00:45:26.020 --> 00:45:29.610 align:middle line:84% So maybe what I would do is, on the other side of the board, 00:45:29.610 --> 00:45:34.150 align:middle line:84% maybe I would put a thermal interface pad and then 00:45:34.150 --> 00:45:36.422 align:middle line:90% a heat sink. 00:45:36.422 --> 00:45:38.630 align:middle line:84% The heatsinks mounted to the other side of the board. 00:45:38.630 --> 00:45:42.380 align:middle line:90% 00:45:42.380 --> 00:45:45.980 align:middle line:84% And I'm going to try to get heat out of the device 00:45:45.980 --> 00:45:48.830 align:middle line:84% through the package, through the board, 00:45:48.830 --> 00:45:53.110 align:middle line:84% through the interface material into the heat sink. 00:45:53.110 --> 00:45:56.320 align:middle line:84% And this is necessary because you 00:45:56.320 --> 00:45:58.420 align:middle line:84% want to connect the device to the board 00:45:58.420 --> 00:46:02.455 align:middle line:84% for the electrical connection-- is the most important thing. 00:46:02.455 --> 00:46:04.690 align:middle line:84% So one of the things you often do 00:46:04.690 --> 00:46:07.163 align:middle line:84% is you will often put thermal vias in the board. 00:46:07.163 --> 00:46:09.080 align:middle line:84% So what you'll do is, you'll go to the board-- 00:46:09.080 --> 00:46:11.180 align:middle line:84% and if I looked at this from the side, 00:46:11.180 --> 00:46:13.180 align:middle line:84% you'll drill a bunch of holes through the board. 00:46:13.180 --> 00:46:15.430 align:middle line:84% You'll put holes in the board like this, 00:46:15.430 --> 00:46:18.730 align:middle line:84% and you'll put what's known as a via farm. 00:46:18.730 --> 00:46:20.740 align:middle line:84% If I look straight down on top of it, 00:46:20.740 --> 00:46:24.130 align:middle line:84% you'd see a bunch of holes drilled in the board like this. 00:46:24.130 --> 00:46:25.100 align:middle line:90% Why? 00:46:25.100 --> 00:46:27.700 align:middle line:84% Because those holes will then be plated with copper, 00:46:27.700 --> 00:46:29.360 align:middle line:84% which is a good thermal conductor. 00:46:29.360 --> 00:46:32.470 align:middle line:84% And I-- it'll help me transfer heat through down 00:46:32.470 --> 00:46:35.200 align:middle line:84% through the board to the interface material 00:46:35.200 --> 00:46:36.250 align:middle line:90% to the heatsink. 00:46:36.250 --> 00:46:40.360 align:middle line:84% Sometimes, you fill those holes with copper in a fancy board, 00:46:40.360 --> 00:46:42.310 align:middle line:84% or you fill it with thermally conductive epoxy 00:46:42.310 --> 00:46:44.560 align:middle line:84% to help the heat transfer through the board. 00:46:44.560 --> 00:46:46.870 align:middle line:84% Sometimes, you'll try to use a thin board so that you 00:46:46.870 --> 00:46:49.730 align:middle line:90% have lower thermal resistance. 00:46:49.730 --> 00:46:55.070 align:middle line:84% This app note that you'll be provided 00:46:55.070 --> 00:46:57.080 align:middle line:84% with in the handouts talks about, 00:46:57.080 --> 00:47:00.050 align:middle line:84% if I have thermal vias of a certain length, 00:47:00.050 --> 00:47:02.210 align:middle line:84% how many degrees C per watt for a 12 00:47:02.210 --> 00:47:05.940 align:middle line:84% mil via going through a 63 mil board and so forth? 00:47:05.940 --> 00:47:07.640 align:middle line:84% You can figure out your thermal model 00:47:07.640 --> 00:47:11.850 align:middle line:84% transferring your heat through the circuit board. 00:47:11.850 --> 00:47:13.560 align:middle line:84% So that's just a heads up about how 00:47:13.560 --> 00:47:16.980 align:middle line:84% you might do things for PC board-mounted components. 00:47:16.980 --> 00:47:18.570 align:middle line:84% The last thing I'd like to do for you 00:47:18.570 --> 00:47:21.445 align:middle line:84% today-- any questions about that? 00:47:21.445 --> 00:47:23.570 align:middle line:84% The last thing I wanted to do for you today is just 00:47:23.570 --> 00:47:26.450 align:middle line:84% show you a thermal transfer example. 00:47:26.450 --> 00:47:27.840 align:middle line:90% So here we go. 00:47:27.840 --> 00:47:29.880 align:middle line:84% Here's the thing I'm going to show you. 00:47:29.880 --> 00:47:33.300 align:middle line:84% This is for a particular matching network design 00:47:33.300 --> 00:47:36.600 align:middle line:84% that we did, and this thing has these four transistors 00:47:36.600 --> 00:47:42.035 align:middle line:84% that are mounted between the circuit and ground. 00:47:42.035 --> 00:47:44.900 align:middle line:84% Each of these transistors can dissipate something 00:47:44.900 --> 00:47:46.640 align:middle line:90% like 30 watts in the worst case. 00:47:46.640 --> 00:47:48.620 align:middle line:90% It's a high power design. 00:47:48.620 --> 00:47:51.953 align:middle line:84% So four devices times 30 watts is like 100 and-- well, 00:47:51.953 --> 00:47:55.627 align:middle line:84% it's 33 watts, so roughly speaking, it's like 140 watts-- 00:47:55.627 --> 00:47:56.460 align:middle line:90% something like that. 00:47:56.460 --> 00:47:58.055 align:middle line:90% 130 watts. 00:47:58.055 --> 00:48:00.430 align:middle line:84% That's a lot of heat to pull out of these little devices. 00:48:00.430 --> 00:48:03.940 align:middle line:90% 00:48:03.940 --> 00:48:05.770 align:middle line:90% Here's the devices. 00:48:05.770 --> 00:48:07.150 align:middle line:90% These are the package. 00:48:07.150 --> 00:48:10.060 align:middle line:84% One side of this has these electrical connections 00:48:10.060 --> 00:48:12.070 align:middle line:84% to the devices, and the other side of this 00:48:12.070 --> 00:48:13.572 align:middle line:84% has a big thermal pad out of which 00:48:13.572 --> 00:48:14.780 align:middle line:90% the heat is supposed to come. 00:48:14.780 --> 00:48:17.500 align:middle line:84% So this is a two-sided device, where 00:48:17.500 --> 00:48:19.450 align:middle line:84% you have electrical connections on one side 00:48:19.450 --> 00:48:21.820 align:middle line:84% and a thermal pad for heat transfer on the other side 00:48:21.820 --> 00:48:23.930 align:middle line:90% of the device. 00:48:23.930 --> 00:48:28.030 align:middle line:84% So one side of this device, the side that's showing flat here, 00:48:28.030 --> 00:48:30.520 align:middle line:84% is going to be connected to the circuit board. 00:48:30.520 --> 00:48:32.170 align:middle line:84% On the other side of the device, we're 00:48:32.170 --> 00:48:35.020 align:middle line:84% going to want to connect to the heatsink. 00:48:35.020 --> 00:48:36.610 align:middle line:90% So what do we do? 00:48:36.610 --> 00:48:37.870 align:middle line:90% Here you can see one side. 00:48:37.870 --> 00:48:39.110 align:middle line:84% This is the top side of the circuit board. 00:48:39.110 --> 00:48:40.450 align:middle line:84% This is not the side of the circuit board 00:48:40.450 --> 00:48:41.630 align:middle line:90% where the devices are on. 00:48:41.630 --> 00:48:44.020 align:middle line:84% The devices are mounted to the bottom side of the circuit 00:48:44.020 --> 00:48:46.312 align:middle line:84% board, and part of the bottom side of the circuit board 00:48:46.312 --> 00:48:48.200 align:middle line:84% is this over ground plane you see here. 00:48:48.200 --> 00:48:49.700 align:middle line:90% These are the devices. 00:48:49.700 --> 00:48:52.102 align:middle line:84% The side of the device you cannot see is what's got 00:48:52.102 --> 00:48:54.310 align:middle line:84% the electrical connections for the drain and the gate 00:48:54.310 --> 00:48:55.900 align:middle line:90% and the source and the drivers. 00:48:55.900 --> 00:48:59.240 align:middle line:84% And the drivers are all on this side of the board. 00:48:59.240 --> 00:49:02.350 align:middle line:84% So what's exposed here on the bottom side of the board 00:49:02.350 --> 00:49:05.380 align:middle line:84% is the heat pad through which I'm 00:49:05.380 --> 00:49:07.940 align:middle line:90% going to try to transfer heat. 00:49:07.940 --> 00:49:09.270 align:middle line:90% So what do we do? 00:49:09.270 --> 00:49:13.460 align:middle line:84% We created a heat spreader, because we're 00:49:13.460 --> 00:49:16.610 align:middle line:84% going to flip this board over and [SMACKS FIST] bang, down 00:49:16.610 --> 00:49:17.760 align:middle line:90% on the heat spreader. 00:49:17.760 --> 00:49:19.220 align:middle line:84% You notice that the heat spreader 00:49:19.220 --> 00:49:22.980 align:middle line:90% has some area milled out. 00:49:22.980 --> 00:49:28.360 align:middle line:84% So if you look at this, there's a space for the devices 00:49:28.360 --> 00:49:33.115 align:middle line:84% to sit in this space in these four spaces for the four devices 00:49:33.115 --> 00:49:35.740 align:middle line:84% so that the circuit board will sit flat with the heat spreader, 00:49:35.740 --> 00:49:37.157 align:middle line:84% and then the devices will sit down 00:49:37.157 --> 00:49:41.400 align:middle line:84% and sit flat against the little milled out area. 00:49:41.400 --> 00:49:45.750 align:middle line:84% These little circles here are just areas for extra solder 00:49:45.750 --> 00:49:48.030 align:middle line:90% to flow out of the device. 00:49:48.030 --> 00:49:50.370 align:middle line:84% So what we're going to do is we have the device soldered 00:49:50.370 --> 00:49:52.328 align:middle line:84% to the board, and then we're going to solder it 00:49:52.328 --> 00:49:54.540 align:middle line:90% to this heat spreader. 00:49:54.540 --> 00:49:56.320 align:middle line:90% How do we make that work? 00:49:56.320 --> 00:49:59.010 align:middle line:84% Well, the first thing we do is we solder the circuit-- 00:49:59.010 --> 00:50:02.340 align:middle line:84% all the chips to the board with high temperature solder. 00:50:02.340 --> 00:50:05.310 align:middle line:84% So then I've mounted my device to the board. 00:50:05.310 --> 00:50:06.720 align:middle line:90% Now, I put thermal-- 00:50:06.720 --> 00:50:08.220 align:middle line:90% I put solder ball-- 00:50:08.220 --> 00:50:10.015 align:middle line:90% I put solder paste on it. 00:50:10.015 --> 00:50:11.890 align:middle line:84% Here, you can see the dots, the solder paste. 00:50:11.890 --> 00:50:14.010 align:middle line:84% I flip this thing over under this heat sink, 00:50:14.010 --> 00:50:16.030 align:middle line:90% and then I reflow it again. 00:50:16.030 --> 00:50:18.220 align:middle line:84% And this solder is low temperature solder. 00:50:18.220 --> 00:50:21.330 align:middle line:84% So I don't unsolder the device to the board 00:50:21.330 --> 00:50:24.580 align:middle line:84% while I do solder it to the heat spreader. 00:50:24.580 --> 00:50:27.010 align:middle line:90% Make sense to everybody? 00:50:27.010 --> 00:50:31.240 align:middle line:84% So we got this hierarchy of temperatures of my solder. 00:50:31.240 --> 00:50:32.510 align:middle line:90% How do I cool this device? 00:50:32.510 --> 00:50:34.720 align:middle line:84% Now I got this heat spreader, which has got, 00:50:34.720 --> 00:50:37.600 align:middle line:84% really low thermal resistance now. 00:50:37.600 --> 00:50:42.470 align:middle line:84% So I've gone from my chip directly to the heat spreader. 00:50:42.470 --> 00:50:44.870 align:middle line:90% And by the way, I said these-- 00:50:44.870 --> 00:50:47.170 align:middle line:84% this thermal pad on the device is actually 00:50:47.170 --> 00:50:49.700 align:middle line:84% connected to source in this particular device, 00:50:49.700 --> 00:50:52.100 align:middle line:84% and source is ground in the particular circuit. 00:50:52.100 --> 00:50:53.860 align:middle line:84% So I can basically make this whole heat 00:50:53.860 --> 00:50:56.380 align:middle line:84% spreader and the whole thing electrical ground. 00:50:56.380 --> 00:50:58.320 align:middle line:90% That's kind of nice. 00:50:58.320 --> 00:51:01.960 align:middle line:84% Here's my-- so I'm going to connect the heat 00:51:01.960 --> 00:51:06.740 align:middle line:84% spreader onto this Dynatron copper heatsink. 00:51:06.740 --> 00:51:08.830 align:middle line:84% And here's what you can see about this heat sink. 00:51:08.830 --> 00:51:11.280 align:middle line:84% Here's the thermal resistance to ambient of the heat 00:51:11.280 --> 00:51:14.100 align:middle line:90% sink versus air flow. 00:51:14.100 --> 00:51:18.090 align:middle line:84% So this is where I said, yeah, to really know-- 00:51:18.090 --> 00:51:20.430 align:middle line:84% this heat-transfer coefficient's a lot of work, 00:51:20.430 --> 00:51:23.580 align:middle line:84% and to design the heat sink to get a low heat transfer 00:51:23.580 --> 00:51:25.920 align:middle line:84% coefficient or a high heat transfer coefficient is hard. 00:51:25.920 --> 00:51:27.730 align:middle line:84% But somebody else did that for me. 00:51:27.730 --> 00:51:29.160 align:middle line:84% All I got to know is how much air 00:51:29.160 --> 00:51:30.660 align:middle line:84% I'm going to blow across this, and I 00:51:30.660 --> 00:51:33.190 align:middle line:84% can get down to crazy low thermal resistances. 00:51:33.190 --> 00:51:36.300 align:middle line:84% If this wasn't low enough, then I'd have to go to liquid cooling 00:51:36.300 --> 00:51:38.145 align:middle line:90% and get even lower ones. 00:51:38.145 --> 00:51:39.980 align:middle line:84% But we were able to do it with air. 00:51:39.980 --> 00:51:42.810 align:middle line:84% So all we did was then we basically mounted 00:51:42.810 --> 00:51:44.295 align:middle line:84% the device to the heat spreader-- 00:51:44.295 --> 00:51:46.290 align:middle line:84% the heat spreader of the heat sink 00:51:46.290 --> 00:51:47.980 align:middle line:90% and then blew air across it. 00:51:47.980 --> 00:51:49.925 align:middle line:84% Here, you can see another view of it. 00:51:49.925 --> 00:51:53.200 align:middle line:84% Here's the heat sink, the device mounted to it, 00:51:53.200 --> 00:51:54.880 align:middle line:84% and we're to blow a lot of air to it. 00:51:54.880 --> 00:51:57.660 align:middle line:84% Here's the actual BIEST So you can 00:51:57.660 --> 00:52:03.750 align:middle line:84% see here's the circuit mounted to the heat spreader. 00:52:03.750 --> 00:52:06.440 align:middle line:84% The heat sink is now all silver because we-- 00:52:06.440 --> 00:52:09.485 align:middle line:84% it's silver-colored because we silver-plated it. 00:52:09.485 --> 00:52:14.340 align:middle line:84% And here, you can see the fan blowing air through the fins. 00:52:14.340 --> 00:52:19.730 align:middle line:84% The result of that-- as I said, we had like 100-and-something 00:52:19.730 --> 00:52:22.350 align:middle line:84% watts of heat dissipation in the worst case. 00:52:22.350 --> 00:52:25.670 align:middle line:84% But if you look at the temperature rises of the thing, 00:52:25.670 --> 00:52:28.170 align:middle line:84% the heat sink surface only goes up 20 C. Why? 00:52:28.170 --> 00:52:30.900 align:middle line:90% Because we got crazy heat flow. 00:52:30.900 --> 00:52:33.250 align:middle line:84% So this is the kind of thing you do. 00:52:33.250 --> 00:52:37.950 align:middle line:84% And we use exactly these models to essentially figure out 00:52:37.950 --> 00:52:41.200 align:middle line:84% what's the thermal resistance from the junction to the case. 00:52:41.200 --> 00:52:44.520 align:middle line:84% Then we have the case going down to the heat spreader. 00:52:44.520 --> 00:52:46.270 align:middle line:84% And it was soldered, so that's pretty low. 00:52:46.270 --> 00:52:47.730 align:middle line:84% And then the heat spreader to the heat sink, 00:52:47.730 --> 00:52:50.580 align:middle line:84% and the heat sink to ambient we looked up from the data sheet. 00:52:50.580 --> 00:52:52.140 align:middle line:84% And boom, we have our model, and we 00:52:52.140 --> 00:52:55.620 align:middle line:84% have our nice thing where we can pour lots of dissipated power 00:52:55.620 --> 00:52:59.390 align:middle line:90% into it and still keep it cool. 00:52:59.390 --> 00:53:00.490 align:middle line:90% So I'm out of time. 00:53:00.490 --> 00:53:02.480 align:middle line:90% Are there any final questions. 00:53:02.480 --> 00:53:04.063 align:middle line:84% AUDIENCE: Will these slides be posted? 00:53:04.063 --> 00:53:05.897 align:middle line:84% PROFESSOR: Yeah, we can put up these slides. 00:53:05.897 --> 00:53:06.410 align:middle line:90% Yep. 00:53:06.410 --> 00:53:08.988 align:middle line:90% So this-- yeah? 00:53:08.988 --> 00:53:11.030 align:middle line:84% AUDIENCE: I'm just wondering what the device was. 00:53:11.030 --> 00:53:13.890 align:middle line:84% PROFESSOR: What the dev-- it was a GaN system. 00:53:13.890 --> 00:53:16.220 align:middle line:90% There were four GaN systems. 00:53:16.220 --> 00:53:19.770 align:middle line:84% GaN on silicon 650-volt transistors. 00:53:19.770 --> 00:53:22.290 align:middle line:90% 00:53:22.290 --> 00:53:25.950 align:middle line:84% So yeah, that's a lightning introduction to quick and dirty 00:53:25.950 --> 00:53:27.240 align:middle line:90% heat sinking. 00:53:27.240 --> 00:53:29.020 align:middle line:84% I still encourage you to actually take 00:53:29.020 --> 00:53:31.493 align:middle line:84% much more sophisticated classes on heat transfer. 00:53:31.493 --> 00:53:32.910 align:middle line:84% I'm not saying this replaces that, 00:53:32.910 --> 00:53:36.540 align:middle line:84% but it's enough to get you to build some pretty cool stuff. 00:53:36.540 --> 00:53:39.460 align:middle line:84% So next time, we'll go back to talking more about circuits. 00:53:39.460 --> 00:53:41.330 align:middle line:90% Have a great day. 00:53:41.330 --> 00:53:50.000 align:middle line:90%