When choosing a MOS tube for a circuit, should I look at the power or Tj? I am confused.

小太阳yy

When choosing a MOS tube for a circuit, should I look at the power or Tj? I am confused. [Copy link]

 

When choosing a MOS tube for a circuit, should I look at the power or Tj? I am confused.

Let’s take the above circuit as an example. The circuit may not be correct. It is borrowed from a forum user. Let’s talk about how to choose this MOS?

Take the above MOS as an example. Its nominal power is 62W. Theoretically, there is a lot of margin. How come this MOS can also use 40W?

But if you use TJ to calculate, it is obvious that the highest TJ is 175, and the thermal resistance is 50K/W. If the ambient temperature is calculated at 25℃

(175-25)/50=3W. By this calculation, this MOS only uses 3W at most, which is a world of difference from the nominal 62W. What is going on? Can any expert tell me?

maychang

Note the test conditions.

The thermal resistance you marked is the thermal resistance from where to where, under what conditions?

maychang

It is best to post all the instructions. This thermal resistance value may be wrong. Usually the thermal resistance of power MOS tubes is not so large.

呜呼哀哉

The thermal resistance of the D2PAK package does not exceed 10℃/W under standard test conditions. Did you read this wrong?

小太阳yy

maychang posted on 2021-9-6 10:36 It is best to post all the instructions. This thermal resistance value may be wrong. Usually the thermal resistance of power MOS tubes is not so large.

I uploaded the instruction manual, please help me take a look

小太阳yy

Alas, published on 2021-9-6 10:43 The thermal resistance of the D2PAK package does not exceed 10℃/W under standard test conditions. Did you read this wrong?

I uploaded the instruction manual, please help me take a look

maychang

Xiao Taiyangyy posted on 2021-9-6 11:02 I uploaded the manual, please help me take a look

In the second post, I mentioned "pay attention to the test conditions". Can you see what the test conditions are for the value of 50K/W? This is copied from the datasheet you posted.

小太阳yy

maychang posted on 2021-9-6 12:03 In the second floor, I mentioned that we should pay attention to the test conditions. "Look at the test conditions for the value of 50K/W. This is from the datasheet you posted...

What does this test condition mean? Our company's design is calculated using this thermal resistance value.

maychang

Xiaoyangyy posted on 2021-9-6 12:24 What does this test condition mean? Our company's design is calculated using this thermal resistance value

The test conditions in the red box roughly mean: installed on the circuit board with the minimum contact area.

maychang

Xiaoyangyy posted on 2021-9-6 12:24 What does this test condition mean? Our company's design is calculated using this thermal resistance value

This installation method is almost the same as the tube being completely in the air, only slightly better than being completely in the air. The power calculated in this way can be said to be the dissipated power that the tube can withstand without any heat dissipation.

maychang

Xiaoyangyy posted on 2021-9-6 12:24 What does this test condition mean? Our company's design is calculated using this thermal resistance value

The 2.4K/W in the upper row is the thermal resistance from the die to the heat sink when the heat sink is infinite and the temperature is kept constant (for example, water cooling).

呜呼哀哉

That value is not quite appropriate for you.

The standard for defining θjA has been discussed in a post, which stipulates the PCB area and number of layers.

In fact, you can look at θjc and then estimate the thermal resistance from your PCB to other places, which is easy to estimate.

Accurate calculations using the finite element method are impossible without software.

小太阳yy

maychang posted on 2021-9-6 13:39 The 2.4K/W in the above line is the thermal resistance from the die to the heat sink when the heat sink is infinite and the temperature is kept constant (for example, water cooling).

I understand what you mean. The thermal resistance of this water cooling is not very realistic. This infinite heat dissipation situation cannot actually occur. So how do we consider it when selecting this period? When we use it, we stick it directly on the PCB board, and then punch some vias on the back. How do we consider the thermal resistance information? In our actual application, the MOS current does not exceed 250mA, so I don’t know how to calculate it?

小太阳yy

Alas, published on 2021-9-6 13:40 It is not appropriate for you to use that value. The standard for the definition of θjA has been discussed in a post, which stipulates the PCB area and number of layers. In fact, you can look at &t ...

There is no JC value or CA value in the manual.

呜呼哀哉

Xiao Taiyangyy posted on 2021-9-6 13:56 There is no JC value or CA value in the manual now - -

Aren't the parameters in the screenshot on the 7th floor the values from the junction to the bottom pad of the package? That is to say, you can control the temperature of the PCB where your MOS is soldered, and the junction temperature will not exceed 4 degrees Celsius per watt at that point.

maychang

小太阳yy posted on 2021-9-6 13:55 I understand what you mean. This kind of thermal resistance with water cooling doesn't seem very realistic. This kind of infinite heat dissipation situation can't actually happen. So we choose...

If your MOS tube is a constant current tube like the one in the first post, and the power supply voltage is 12V, then in extreme cases the power consumption of the tube is 12V*0.25A=3W. According to what you said, "When we use it, we stick it directly on the PCB board, and then punch some vias on the back", it is just enough. However, the margin is not large.

maychang

小太阳yy posted on 2021-9-6 13:55 I understand what you mean. This kind of thermal resistance with water cooling doesn't seem very realistic. This kind of infinite heat dissipation situation can't actually happen. So we choose...

"This kind of thermal resistance with water cooling doesn't seem very realistic. This kind of infinite heat dissipation situation cannot actually occur."

This situation is an ideal one, but of course it is impossible to happen in reality.

maychang

小太阳yy posted on 2021-9-6 13:55 I understand what you mean. This kind of thermal resistance with water cooling doesn't seem very realistic. This kind of infinite heat dissipation situation can't actually happen. So we choose...

In practical applications, a heat sink is often required. When calculating, you need to know the thermal resistance from the heat sink to the air, plus the thermal resistance of the tube itself. If an insulating pad is added between the tube and the heat sink, then the thermal resistance of the insulating pad must also be added.

Such calculation is approximate. As mentioned on the 12th floor, accurate calculation is very troublesome and unnecessary in practical applications.

annysky2012

In actual applications, after adding a heat sink, the power usage can be considered. However, the margin must be sufficient.

小太阳yy

maychang posted on 2021-9-6 16:03 In practical applications, a heat sink is often required. When calculating, you need to know the thermal resistance from the heat sink to the air, plus the thermal resistance of the tube itself. If the tube and the heat sink...

What kind of heat sink is actually added? What kind of heat sink is added to the MOS you use here? Can you make a picture for me to learn?