• State
  • Locked Locked
  • Replies 3 replies
  • Subscribers 64 subscribers
  • Views 800 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Switching Mosfet Dissipation

A A2
Prodigy 60 points

Hi All,

              I'm trying to measure the thermal performance of an existing buck converter using hyperlynx thermal. To perform the simulation correctly I need to know the power dissipations of the critical components including the switching mosfets.

              First I tried using the power dissipation calculated in my spice software. This dissipation yielded a result that was too low, as confirmed by experiment. I then tried to measure the current through the mosfets by measuring the sense resistor voltage and the duty cycle. This still yielded temperatures that were too low. Then I simulated the circuit again and I multiplied the RMS current squared by the on resistance. This still yielded a low result.

              Next, I multiplied the RMS Ids and Vds from the simulation together, yielding a 5W dissipation. This 5W is way too high, because with the particular thetaJC a junction temperature of 1000C was found.

             Finally, I multiplied the average Vds and Ids together, which yielded the most realistic temps I have seen yet.

             Ultimately, my main question is can the power dissipation of a mosfet be calculated by multiplying the average VDS by the average IDS? Is this the most accurate power calculation for thermal purposes? From all of the different combos I've tried this is the method that has yielded the closest results.

  • 0
    Guru 22270 points

    Hello A,

    Are you performing the temperature measurements with or without heat-sinks? You will have to adjust your temperature reading to compensate for the thermal dissipation of the heat-sink plus the ambient temperature.

    Or you can remove the heat-sink and test, under moderate duty cycle off course.

    Thanks,

  • 0 in reply to Greenja
    Prodigy 60 points

    Hi greenja,

                   I'm performing the temperature measurements without any heatsinks. I'm using an IR gun with 2degC accuracy to measure the case temperatures of the SO8 mosfets. I'm trying to compare the results to what I get in hyperlynx thermal. Once I establish a good sim I can better choose a heatsink.

  • 0 in reply to A A2
    TI__Mastermind 18200 points

    A, 

    Multiplying average Vds by average Ids will not necessarily yield an accurate power dissipated value. Consider that when the FET is off, Vds is high and Ids is virtually zero. When the FET is on, Ids is high and Vds is virtually zero. 

    A more accurate method for measuring power dissipated is to use P=I^2*R. Thus you want to square you average Ids(on) through the FET and multiply this by the typical on resistance of the FET. If you want to be conservative, you can multiply by the max resistance of the FET, at your given Vgs, and then multiply this by the temperature coefficient for the temperature you are running at. Finally, multiply this value by your duty cycle to get the power dissipated by the FET. 

    Pdiss = (Duty Cycle) * (Ids(on) average)^2 * (Rds(on) @ Gate Voltage you are driving at) *(Normalized Temperature Coefficient @ the temp you are running at)

    All of these parameters should be available within the datasheet. Make sense?