So much important as to turn on the Power MosFET (charge the gate) with 10 to 12V and fast current (think about 100mA or more), is also the fast discharge when you want to cut off the MosFET. Any slow charge or discharge will make the transistor go into the linear region of the condution, and it will heat.
This is why people develop several solutions with one or two transistors to charge/discharge the gate, or even using a Power MosFET driver.
No kidding, you can destroy a Power MosFET easily if going constantly into the linear region with a high current load, due high temperature in the junction.
We also use a zener diode between gate and source to limit charging voltage on the gate. It is easy to destroy the gate with excessive voltage. Sometimes 18V between Gate and Source can destroy a transistor designed to work with a maximum of 12VGS.
The following is a simple N Channel driver, but it lacks the good discharge transistor, it uses a simple resistor. To speed up the discharge, R4 must be a low value, no higher than 150 or 180Ω, by itself it will constantly drain current from R3 and Q2 when it is feeding the gate. Also, for Q2 to be able to supply almost all +VBat to the gate, R3 must be much lower than R4, so, lots of current is wasted in this driver when feeding the MosFET's gate. As you can see, a zener diode is used to protect the gate against high VGS.
View attachment 64037
The following circuit uses a medium power MosFET (IRF510), but look at the R1 resistor, 10kΩ, when Q2 is not conducting, the maximum peak gate charging current is 1.2mA, this is ridiculous, the poor IRF510 will work in the linear region all the time, and will require a large heatsink. By the other side, Q2 does a pretty good job to discharge the gate, since the 2N3903 can sink hundreds of mA at once from the gate. The IRF510 has a medium "virtual capacitor" at the gate, it can store quite good charge. For this circuito to work nicely, R1 should allow 80 to 100mA flowing from +12V to the gate. This would only be possible, if R1 would be from 50 to 100Ω, but no more than that. Of course, when Q2 conducts, it will not only discharge the IRF510 gate capacitor, but will also sucks current from +12V via R1, so make sure to not exagerate. If R1 = 100Ω, Q2 will sink 120mA, with a possible peak of 400mA or more, based on the IRF510 gate capacitor charged. During the sinking, Q2 may dissipate from 100mW to 250mW, be careful to not use lower values at R1.
View attachment 64038
Below a terrible configuration for a Power MosFET driver, in totem-pole configuration.
Due Q1 being a NPN, the MosFET gate will be loaded with the input pulse voltage, less 0.6V, that's nasty.
If the MosFET requires 10 to 12VGS, the input pulse at the bases of the bipolar transistores must be at least that voltage plus 0.6V.
It doesn't make any sense use such configuration when driving the transistors with a microcontroller, powered by 5V, and the Power MosFET requiring more than 10V at the gate.
View attachment 64039
The following shows a better way to do it, when input is zero, Q1 is not conducting, R1 + R2 feeds current to the Q3's base, it conducts and feeds almost +VCC to the MosFET's gate. When input is high or floating, Q1 conducts, connecting R2 to ground, cutting off Q3 and turning on Q2, that will discharge MosFET's gate fast. This works nicely. R3 in the circuit is to avoid "ringing waves" between the transistors and the MosFET's gate, it must be a value 10Ω or lower.
View attachment 64040
Below another driver using NPN + PNP to drive the Power MosFET's gate. Note the use of reversed NPN (bottom) and PNP (top). The left transistor, a NPN, works as a level translator. Its base is connected to the microcontroller's VCC, and the control is done at the NPN emitter. This works very, very well when voltages feeding the NPN+PNP pair and the base resistor is way above the microcontroller logic VCC. When the input is zero, the NPN conducts, since VBE is practically VCC, then PNP conducts and feeds the maximum possible current and voltage to the MosFET's gate. When the input is high, close to the microcontroller VCC, the NPN cuts off, now is the NPN of the pair that conducts and sinks all the MosFET gate's charged capacitor. This is one of the best and fast drive, but, with a price, it doesn't work very well for high frequencies, when the square wave starts to looks like round in the corners, and the transistors pair get lazy and do not switch fast enough, feeding linear voltage to the gate, MosFET will heat.
View attachment 64041
The following is one of the best drivers you can built with discrete components.
When the input is high (above 3V), the 3904 conducts, also the 3906 conducts. This will load the MosFET gate's capacitor very fast, via the 1N4148 diode. As the VBE of the bottom 3906 is reversed (more positive at its base when compared to its emitter, caused by the voltage drop over the 1N4148 diode), the transistor does not conduct. The MosFET's VGS is limited by the 12V zener. You could include a 10Ω resistor between the 1N4148 and the zener to ensure the zener will not blow up if +V is way above it. When the input is low, the 3904 will cut off, the 3906 will cut off as well, now there is no more current to charge the MosFET's gate, but the MosFET gate's capacitor still charged with +12V. Now this gate's charge can't not go back via the 1N4148 diode, so the base of the bottom 3906 is at the ground level via the 1k5 resistor, while its emitter is at +12V from the MosFET gate's charge voltage. It makes the bottom 3906 conducts and discharge the gate in high speed and current. This is a fast self discharge system and a very effective way to drive the Power MosFET, at the cost of several components, but believe me, it worths build it.
View attachment 64042