Transistor max power?

Hi,

I’ll use the irfz44n as an example to explain what i would like to know.
On the datasheet it sais max current is 49a, max voltage is 55v, resistance is 17.5mOhms and max power dissipation is 94w. When you calculate heat being produced, you get about 42w. I’m pretty sure that’s more than enough to make that transistor release its magic smoke so i’m wondering what the power dissipation means (it’s obviously not the amount of heat the thing can dissipate because that seems way too much) and how i can calculate max power.

Cecemel

Hi,

Yes, 49A @ 17.5mOhms gives 42W.
Datasheet says 94W.

Both are specified @ package (metal plate) temperature of 25°C.
For sure you need a powerful heatsink (in real world about impossible) to ensure this 25°C at a power dissipation of 94W.

Let's say datasheet says the r_th_jc = 1.5K/W.
This means at 94W there is a temperature difference of 1.5 x 94 = 141K.
With a case temp of 25°C ... this gives a junction temperature of 25 + 141 = 166 °C.
This almost meets the max junction temperature of 175°C.

Why there is a deviation of 94W an 42W... I can only guess.
My idea is, when the Mosfet is very low ohmic then the current distribution of the chip is not equal, I assume close to the connectons there is increased power dissipation...this creates some hot spot on the chip. Thus the limit of 42W.

But in case of non saturated operation... let's say with V_DS of several volts the heat distribution over the whole chip is more equal.
No extreme hot spots anymore. --> Higher overall power dissipation

Klaus

94W is 94W, the rating has nothing to do with operating in the saturated region. In the linear region it could dissipate a lot more than 94W without exceeding the voltage or current ratings.

Hi,

See Note (1)

It will be for a nano- or microseconds, I suppose.





Note (1) Test signal, 300nS and you probably should read the small print. Look for those "See note 1" in the datasheet electrical characteristics and it will say how long the mosfet was tested at 94W. Another way of figuring out how far you can actually go and for how long is perusing the SOA graph - a quick calculation of the axes junctions on the graph will/should show under what conditions 94W are possible.

Right the wattage number is typically just an indication of the Tjc parameter assuming an ideal heatsink. As such it’s not a very useful number.

Hi,

94 W are continous power...as long as the case temperature is kept to 25°C max.
Calculations give a max junction temperature of about 170°C in thus case.

Klaus

Thanks for the replies, cleared up a lot. So basically the 94 watts is the max output of R*I(squared) assuming you have an ideal heatsink.
I’ve heard that the switching of the transistor generates heat as well, is there an easy way to calculate this?

Hi,

nothing special
Heat is always dissipated power = V x I.
During switching V is rising(falling) and I is rising (falling)
The integral of V x I (dt) during the switching is the switching loss. (Here the swiching energy. If you multiply the energy with the switching frequency you get the switching power loss)

In general: The longer the switching time the higher the switching loss. The higher the switching frequency the higher the switching loss.

Klaus

KlausST said:

Hi,

nothing special
Heat is always dissipated power = V x I.
During switching V is rising(falling) and I is rising (falling)
The integral of V x I (dt) during the switching is the switching loss. (Here the swiching energy. If you multiply the energy with the switching frequency you get the switching power loss)

In general: The longer the switching time the higher the switching loss. The higher the switching frequency the higher the switching loss.

Klaus

Click to expand...

Thanks but i don’t fully understand, could you give an example with a circuit like this?

Hi,

No, it doesn´t work this way.
You need the switching waveform of voltage (across CE) and waveform of current and timing.

Klaus

Look at it this way: when the MOSFET is off, there's maximum voltage across the device but zero current==> power equals zero. When the MOSFET is fully on, there's minimal voltage (I*R) and maximum current==> Power =I^2*R; if R is small, power is small. In between those two limits the voltage across the device goes from maximum to minimum so power dissipation changes. If you were to plot the power vs gate voltage during the switching transition you would see that the power goes from zero to a maximum and then down to a minimum when it's fully on.

Cecemel said:

Thanks but i don’t fully understand, could you give an example with a circuit like this?

Click to expand...

It is well known that for resistive switching, the Esw loss is estimated=1/6 *I*V*tsw. However, in order to compute the "tsw" time, one needs more work.
Computing switching time is another story.
The easiest way is to get them via simulation.