Why the smaller the series resistance of the MOS transistor gate, the better

“We know that MOSFET devices are voltage-controlled devices. Different from bipolar triodes, the conduction of MOSFETs only needs to control the voltage of the gate to exceed its turn-on threshold voltage, and does not require gate current. So essentially, there is no need to connect any resistor in series with the gate of the MOS tube studio.

“

Just ask any seasoned electrical engineer — like Professor Gureux in our story today — what to put in front of the MOSFET gate, and you’ll likely hear “a resistor of about 100 Ω.”

Although our answer to this question is very positive, you may continue to ask – “Why? What is his specific function? Why is the resistance value 100 Ω?”

With curiosity, the young professor Neubean conducted an experimental study on the usual practice of connecting a 100Ω resistor in series with the gate of the MOSFET, and came to a conclusion that could appease his doubtful mind, but he felt that he still did not really grasp the problem.

We know that MOSFET devices are voltage-controlled devices. Different from bipolar triodes, the conduction of MOSFETs only needs to control the voltage of the gate to exceed its turn-on threshold voltage, and does not require gate current. So essentially, there is no need to connect any resistor in series with the gate of the MOS tube studio.

For ordinary bipolar triodes, it is a current controlled device. Its base series resistance R1 is to limit the size of the base current. Otherwise, for the driving signal source, the base of the triode is equivalent to a diode to the ground, which will affect the front driving circuit.

For a MOS tube, since its gate is insulated from the drain and source, there is no need for a series resistor on the gate to limit current.

On the contrary, considering the parasitic capacitance of the gate of the MOS tube, in order to speed up the turn-on and turn-off of the MOS tube and reduce the loss of the MOS tube during the turn-on and turn-off process, the equivalent resistance on its gate should be as small as possible. OK, 0.

However, in many practical MOSFET circuits, the resistance connected in series on the gate of the MOS tube is almost everywhere. It seems that everyone has forgotten that the existence of this resistance will prolong the turn-on and turn-off time of the MOS and increase unnecessary losses.

The following is a few MOS tube circuits randomly found on the network. You can see that the red circle indicates that their gates have series resistance.

In the above circuit examples, two of them put the MOSFET in the feedback circuit, and the MOS tube does not work in the switching state, so the resistors connected in series in these two circuits should be in essence matched with the gate capacitance of the MOS tube to reduce the circuit. frequency response, increasing the phase stability margin of the circuit.

This is similar to the situation discussed by Professor Neubean’s experiment in the tweet “Why 100Ω? The real professor found the simple problem behind the simple conclusion”, but Professor Neubean was more ruthless. In the discussion, he increased the MOS gate series resistance to 1MΩ, And through simulation, it is found that increasing the gate resistance will cause output instability in his circuit instead. So the lower the gate series resistance, the better. Below is the circuit of Neubean.

It’s a pity that Professor Neubean just got the gate resistance Rgate as small as possible, but why not reduce it to 0 ohms? Why do you need to reserve 100 ohms? For this issue, his mentor Gureux just said confidently, just choose 100 ohms.

Regarding the gate series resistance of the MOS tube, especially the effect of the gate series resistance of the MOSFET working in the switching state, the usual explanation is to prevent the oscillating waveform during the switching process of the MOSFET, which will not only increase the switching loss of the MOSFET, if the oscillation is too large , it will also cause the MOS tube to be broken down.

The following is a simulation experiment of this phenomenon.In the following circuit, the gate of the MOS transistor has a resistor R3 in series, its drain load is an inductive load, and also includes a 10nH line distributed inductance

In the experiment, the R3 is set to 1 ohm, 10 ohm, and 50 ohm respectively for simulation experiments. It can be seen that when R3 is 1 ohm, there is a high-frequency oscillation signal on the output voltage Vds.

When R3 is increased to 10 ohms, the high frequency oscillation signal of the output Vds is obviously attenuated.

When R3 increases to 50 ohms, the rising edge of Vds becomes slower. On its gate voltage, a step due to the Miller capacitance effect between drain-gate is also evident. At this time, the power consumption of the corresponding MOS tube is greatly increased.

From the above simulation experiment results, the resistance connected in series on the gate of the MOS tube needs to be determined according to the specific MOS tube and the stray inductance of the circuit distribution. If its value is small, it will cause output ringing. If it is too large, it will increase the switching transition time of the MOS tube, thereby increasing the power consumption.

However, the simulation circuit above does not reveal the real role played by the gate resistor. The figure below shows the real effect of the gate series resistance of the power MOS tube given by Infineon semiconductor in eliminating switching oscillations.

In the driving circuit loop of the power MOS tube, there will be various distributed inductances Lp, and they will form a resonance circuit with the Cgd and Cge of the MOSFET. They will resonate the high-frequency harmonic components in the switch drive signal, which will cause the output voltage of the power tube to fluctuate.

The series resistance Rg at the gate of the MOS tube can increase the loss in the driving loop of the MOS tube, reduce the Q value of the resonant loop, and make the inductance and capacitance resonance phenomenon attenuate as soon as possible.

There is a series resistance Rg among them, so that the driving loop is just in a critical damping state. At this time, the driving signal of the MOS transistor is in a compromise between a small oscillation and a fast switching state. Therefore, Rg should be related to the distributed inductance of the MOSFET drive loop and its stray capacitance. Therefore, it is not a fixed value and needs to be determined experimentally.

Speaking of this, let’s re-read the discussion in the tweet “Why 100Ω? The more serious professor found the simple problem behind the simple conclusion”, although Neubean proved that the smaller the series resistance of the MOS transistor gate in his circuit, the better, but Neubean’s mentor, Professor Gureux, should not be so confident when it comes to 100 ohms.