Ha, very good point made, I honestly had breezed right past that as I am still learning to read datasheets.
Going back and looking at the graph you posted of the drain current vs Vgs curve we can see that the curve is exponential, we also see that at 25*C @ 5v it is eclipsing 10 to the '2nd' power, thus 100Amps.
But your correct that at 3.3v gate voltage you would be able to drive ~25-30Amps.
Very good stuff, you really made me go back and study that datasheet graph a little closer.
Thanks, you helped me learn a little more then I knew this morning!
But as Tom pointed out, I think we are looking at 2 different Mosfets
IRF vs IRL
Now in this case, 3.3v gate drive would not even be enough to drive the Mosfet and it would also be correct that it would not be suitable for a logic level Mosfet.
One more point: the gate is a capacitor (IRLZ44 = 3300 pF typically) and that acts as a dead short to ground during turn on.
So the port output voltage drops significantly due to the output impedance. An engineer might assume 2.5V output, until the MOSFET gate is fully charged (0.5 us). At 2.5V drive the IRLZ44 can typically support only 7 A drain current.
The IRLZ44 is indeed a "logic-level" MOSFET. What you need to remember is that logic level here (at least when the parts was introduced) typically meant CMOS logic. A CMOS logic high is 5V which if you look at the Id vs Vgs for the IRL, you'd see hitting max at 5V. Given that most Arduinos are still 5V devices, the IRLZ44 is a good choice for Arduino based designs where you're switching large loads with just an output pin.
So, the issue of concern with the IRLZ44 is whether the total turn on/off time will work for you. You can spend time agonizing over it from gate charge, but the datasheet provides the other parameters in ns so why bother. It's not symmetric with turn on/off (as we should expect) so for a quick swag, we'll use the higher turn on time and round to 250ns. This means your driving waveform should have a frequency of 2Mhz or less.
There isn't an Rdson vs Vgs plot so the safe thing to do is further de-rate your max switching speed, say to 200kHz or 20kHz to reduce the amount of time you're spending in the "high" power dissipation (area between Vgsth and full on) region (which is still pretty small). Given that in Arduino, PWM tops out at 1kHz, you'd be completely fine using the IRLZ44 without any worries about power dissipation in the motor control application presented here.
