Hi,

I understand how to calculate power dissipation when switched hard on using the info from the data sheet such as rsdon. but when it is switching there is a rise time due to the capacitance and this charging up which as it is turning on or off the mosfet is in it's non linear region. i think i am right in saying that this is due to a difference in resistance during this transition which uses more power. therefore to calculate the overall power you would need to add together the transition power plus the switched on power.

if the above is correct, how do you calculate the switching power dissipation?

Thanks

 

It's about doing an integral of the voltage across the mosfet times the current through the mosfet during the transition time. I haven't done it, but I believe you could calculate it from the datasheet information or watch the rise time on a scope. Perhaps somebody else could help with a better description of the method?

 

A simple approximation is to assume a linear rise-time and calculate the transition dissipation as being simply 1/2 V*I*\(\tau\)*f where V is the OFF source voltage, I is the ON current, \(\tau\) is the rise-time and f is the switching frequency. You do the same calculation for the fall-time and add the two together to get the total switching dissipation. That calculation will be in error by how much the rise and fall times are non-linear (e.g. exponential) but for most purposes it should be close-enough.

If that's not accurate enough then you need to do the numerical integration of the voltage times current over the rise/fall times from an oscilloscope measurement, as #12 noted. Some digital oscilloscopes may do this integration for you.

 

Spice will do the calculation for you if you have a trustworthy model of the FET.

 

Spice will do the calculation for you if you have a trustworthy model of the FET.

along with the various circuit parasitics.

 

Thanks for your help i will calulate it and use my scope. I think my scope has a intergral feature but i have never used. I also have some simulation software, but like it was said the modle for it wasnt any good

 

1/2VIFt (rise) + 1/2VIFt (fall) is the part I forgot. Thanks crutschow.

 

1/2VIFt (rise) + 1/2VIFt (fall)

one last quick question about this the above calculations (i understand are a rough estimate) what unit would the results be in? i am guessing as i was originally asking the question about power dissipation then the result would be in Watt's and also these calculations show V*I

sorry if this seems simple i just wanted to check.

 

Looks like watt seconds per second, and that boils down to watts.