NerdKits - electronics education for a digital generation

You are not logged in. [log in]

NEW: Learning electronics? Ask your questions on the new Electronics Questions & Answers site hosted by CircuitLab.

Basic Electronics » Power MOSFET - Understanding and headaches

July 01, 2009
by sector0
sector0's Avatar

So, I basically understand the operation of the MOSFET (i.e. 2n7000) for switching as explained in the servo example and nerd kit guide. However, what I don't get is (and maybe the question itself is wrong) how much current is required to operate the gate. I kind of understand that voltage is what controls the gate, but what about current? When I push that gate to +5v I can get current to flow from drain to source but what kind of current is getting used at the gate side (or required)? I have tried LTSpice to simulate and got really confused... It pertains to a specific circuit I am making, that works in real life, but I don't get why or how much I should be worried ;)

July 02, 2009
by lockwor
lockwor's Avatar

Hey there! In a 'real life' setting the gate will pass current, but in an ideal sense the gate current should be close to zero! A MOSFET is a metal on Semi-Conductor Field Effect Transistor. The 'metal' refers to the gate and between it and the semi-conductor is an insulator, usually glass since it's just oxidized Si. Anyways, the resistance from gate to drain should be on the order of a mega ohm. Very small currents should be observed. This is quite different from a bi-polar junction transistor where control current is more significant. Check out the MOSFET Wiki on Wikipedia for lots more info and images.

July 02, 2009
by mrobbins
(NerdKits Staff)

mrobbins's Avatar

Lockwor is right about the static gate current being close to zero. But notice that I say static -- meaning when the voltage V_GS is not changing. There's another current that may or may not be important, which is the dynamic gate current. That is, as you are raising the voltage V_GS, you will have to put charge into the gate terminal to charge up the gate capacitance and raise the voltage, which means that there is a non-zero current flowing into the gate when the MOSFET is being switched on. Similarly, when you are transitioning the gate voltage from high to low, you have to remove that charge from the gate, which leads to a non-zero current flowing out of the gate terminal.

For example, for the 2N7000, the gate charge at V_GS = 5 volts is about 1.0 nC (see 2N7000 datasheet Figure 10). Let's do some back of the envelope math: if you very roughly assume that the microcontroller output pin can source around 50mA, then it will take (1.0e-9)/(50e-3) = 20ns for the gate voltage to make its rise. Bigger power MOSFETs will have a bigger gate capacitance, and in general the problems that arise here are because for that 20ns or whatever the transition time is, the MOSFET is neither fully on nor fully off. It is passing some current, but still has substantial non-zero V_DS, which means that it's dissipating a lot of power as heat. This leads to the MOSFET getting hot, and possibly failing.

So in summary, it's not the gate current itself that can lead to trouble, but the fact that the current (as limited by the microcontroller or whatever is driving the gate) leads to the MOSFET taking a "long time" to switch, and this leads to excess heating and possibly failure. Since this heating happens every time the MOSFET gate is charged and discharged, the power dissipation to heat is proportional to the switching frequency.

Actually, this dynamic current is exactly why your microcontroller requires any current at all! Since the logic is essentially made up entirely of MOSFETs, there is zero current flowing when everything is at steady state. But the fact that the MOSFETs from one stage have to charge and discharge the capacitance of the next stage is exactly why your microcontroller needs current to run, or why your computer CPU gets hot. If you're curious, look at Figures 29-1 and 29-2 on page 316 and 317 of the ATmega168 datasheet. They show that the current consumed by the microcontroller is basically proportional to the operating clock frequency.

If you're just switching low power stuff, this dynamic current is almost certainly not an issue, but for big motor controllers, it often is.

Hope that helps!

Mike

July 07, 2009
by sector0
sector0's Avatar

That is very useful indeed :) I kept thinking of the MOSFET like a relay (mechanically) so was assuming a large amount of current like a coil. After setting a meter in there I could not register any current since it was not sensitive enough to pick it out. Then, after simulating it in LTSpice I got confused by seeing how low the current was when switching states.

The cool thing is, using the MOSFET and two resistors I was able to create an always on MOSFET that could be triggered off like a reset switch (like the pins on a NC relay) to do a reset on my circuit I have posted in a different thread. Basically, I have the resistors set as a voltage divider with the second resistor hooked to the output of the 555 timer. When the timer goes high or is not connected it acts normally with a bit of resistance and 5 volts. When the timer goes low it causes the second resistor to connect to ground and voltage is split between the two resistors. This causes the MOSFET (hooked in the middle) to get a tiny amount of voltage and turn off. Works much nicer than the relay.

Post a Reply

Please log in to post a reply.

Did you know that a microcontroller can measure an RC time constant? Learn more...