I never sleep.
Ghurd,
Would you explain your formula a little bit more for me.
Sure.
I do not do it very conventionally. Someone will do it fancier than I do it. (Ross)
This is simplified.
The fet data sheet will have
Rds listed
Resistance from
Drain to
Source in Ohms while fully saturated.
Fully saturated, to me, is either 6V for logic level gates, or 12V for standard fets. Usually.
Depending on how old you are, or how old your books are, 'I' is Amps.
Ohm's Law say Amps multiplied by Amps multiplied by Ohm's = Watts.
For this discussion, Watts is heat in the fet, and I like it below 1W in a TO-220 package (physical size of the fet).
I^2 = Amps Squared = Amps times Amps = A x A
Watts of heat in the fet are I^2 x Ohms. We know the amps. The data sheet shows the Ohms while fully saturated.
The 1W is what I use for a T0-220 package. A bit more for a fet with a small heat sink, a bit less without.
Example A.
IRFZ44N mosfet.
The data sheet, right there at the top, show the resistance as 17.5milliohms. 0.0175 Ohms.
The desired load is 4A.
Watts (heat) in the fet are I^2*R. Amps x Amps x Ohms.
4A x 4A x 0.0175 Ohms = 0.28 Watts
Example B.
Same fet. Load is 8A.
8A x 8A x 0.0175 Ohms = 1.12 Watts
The Point, Part #1:
A small increase in amps increases the watts in the fet exponentially.
Doubling the amps increases the watts in the fet exponentially.
The Point, Part #2:
My fancy $2 fet, FDP7045L, has a resistance of 0.0035 Ohms.
17A x 17A x 0.0035 Ohms = 1 Watt
1 Watt is OK in my rule of thumb math, but 17A in those little tiny fet legs is an awful lot of amps.
Another issue regarding heat in the fet is frequency.
Fets have higher resistance while in 'between' fully On and fully Off. Fast switching fets switch more often, so they spend more time in a higher resistance state.
The time a fet spends in the "between" state is almost a fixed value.
A fet switching a 6A load at 500,000 cycles per second at a 50% duty cycle will run hotter than a fet switching 6A at 1 cycle per second at a 50% duty cycle, because it is spending 500,000 times more time in a high resistance state.
I slowed the ghurd controller down to a lower frequency so it is not an issue.
That's it in a nut shell.
G-