Internet Explorer is not a supported browser for TI.com. For the best experience, please use a different browser.
Video Player is loading.
Current Time 0:00
Duration 6:36
Loaded: 0%
Stream Type LIVE
Remaining Time 6:36
 
1x
  • Chapters
  • descriptions off, selected
  • en (Main), selected

    Hi. I'm Robert Kollman. I'm a senior applications manager at Texas Instruments. Welcome to Power Tips.

    Hi. Welcome to Power Tip 28 and 29. Today we're going to talk about estimating the transient temperature rise in a MOSFET. The temperature rise of a semiconductor gets to be an issue in this typical kind of system.

    What we have here is a Hot Swap controller that controls the application of power to a downstream system and then also has to charge up a relatively large capacitor, Co. This capacitor gets charged through the M1 and a significant amount of power dissipation can be put into M1 for a short amount of time.

    One of the ways to understand the implications of this short, rapid, large amount of power being put into the transistor is to look at its SOA characteristics-- that's Safe Operating Area. What we have here are various regions of operation that are permitted by the datasheet.

    For instance, at 100 microsecond pulse width, we can put a significant amount of energy in to the transistor. We can put almost 100 amps, almost 10 volts into the transistor. We can put 1,000 watts into that transistor for 100 microseconds.

    And you see as the pulse width gets longer and longer, that limits the total power that we can put into the device. For instance, if we drop the pulse width down to 1 millisecond here on the curve, you'll see that we can put 10 volts and not even 10 amps into the device. So we've reduced the amount of power that we could dissipate into the device by a 10 to 1 variation in the period that we were doing it.

    This is usually all the information manufacturers provide to an engineer to make a decision. And there's significant things in the circuit that can impact how much power you can dissipate for these shorter periods of time. There are places in the system that you might be able to store some energy.

    And to address this, we're going to develop a thermal resistance/electrical analog here. And so on the left, we're showing a derived calculation based on the thermal resistance. And so it's very simple.

    The change in temperature is equal to the power times the thermal resistance. And the resistance is very analogous to the analog case. The resistance is just the resistivity of the material times the length of the path divided by the area.

    You can see that mirrors the electrical analog very well. If you replace temperature with voltage, power with current, and resistance with resistance, you'll have this very recognizable expression. And then the resistance calculation is very similar, also.

    The second analog that you can develop is one between the thermal capacitance of the device and the electrical capacitance within the circuit. So, if you start on the right side of the curve this time, you'll see that the change in voltage is simply 1 over the capacitance times the integral of the current over time. In the thermal system, the voltage is replaced by temperature.

    We develop a similar thermal capacitance, which is the product of the mass times the specific heat of the device, and then times the integral of the power over time. And so what you can do with those elements is you can look at the physical construction of your power semiconductor, and you can develop a thermal model. And that's what we've done here.

    We've replaced the power that will dissipate in the device with a current source. We've calculated what we believe to be the device's thermal capacitance. We've looked in the datasheet and gotten thermal resistance from the junction of the device to the case.

    And then we've also put in some thermal resistances and thermal capacitances for the leadframe itself, some potting compound, and then we've taken it, finally, to the ambient. One case, we included some in here for a heat sink. Another one we're showing actually a thermal resistance between the potting compound and the ambient temperature.

    And this is kind of an interesting analog here. You can look at some of the thermal time constants within the system. For instance, if you put one amp of current into the system, the first thing that you'll have to do is start to develop some voltage on the die, and that's really develop temperature rise in the die.

    And once you have some temperature rise, you can drive current on into the remainder of your system. And it will flow, first, into the leadframe, and then it kind of has a parallel path where it can flow through the heat sink to the ambient or through the potting compound to the ambient. And with this you can make a little better estimate of where you are on the safe operating curves.

    And that's what I've done in this chart. On the x-axis of this chart, I have time. And this is in milliseconds. And on the left axis, I have voltage. And this is the analog to temperature rise.

    And so you can see when we first put power into the device, the junction that defines raises its temperature quite rapidly. And you can see very little impact on the heat sink. For the first 10 milliseconds, all the temperature rise is within the device. And then as you continue out in time, you can see that the temperature rise on the heat sink starts to increase.

    And then also what you'll notice is that there's pretty much a fixed temperature rise between the junction of the device and the heat sink over time. So if you develop a pretty thorough thermal model of your semiconductor, you can develop a pretty good estimate of what the temperature rise is with any arbitrary power that you put into the device.

    So thank you for joining me with this Power Tip. There are many more Power Tips available on the double E times website. Just call up their web page and search on Power Tips. Or you can click on the link to all articles in the description section of this video. Thanks.

    This video is part of a series