Question

How much switching energy does each additional ohm of gate resistance cost on a common power MOSFET at a fixed load, and is the relationship linear?

Hypothesis

Switching loss should grow roughly linearly with gate resistance, since the gate charge is fixed and the charging current scales inversely with R.

Setup

An IRFZ44N switching a 12 V, 2 A resistive load at 50 kHz, driven by a gate driver through swappable gate resistors: 10, 47, 100, 150, and 220 Ω.

SUPPLY12.04 V
LOAD2.01 A
F(SW)50.0 kHz
T(AMB)23.1 °C

Method

For each gate resistor, capture gate and drain waveforms, integrate V·I across the switching transitions, and average over 64 cycles.

FIG. 01Gate waveform at 10 Ω vs. 100 Ω gate resistance. The Miller plateau stretches as R grows.
Oscilloscope capture comparing gate voltage rise at 10 ohm and 100 ohm gate resistance; the 100 ohm trace rises visibly slower with a pronounced Miller plateau.

Results

R(gate) t(on) t(off) E(sw) per cycle
10 Ω 38 ns 52 ns 8.2 µJ
47 Ω 96 ns 118 ns 19.5 µJ
100 Ω 180 ns 215 ns 38.9 µJ
150 Ω 261 ns 300 ns 57.4 µJ
220 Ω 370 ns 428 ns 83.1 µJ
FIG. 02Switching energy vs. gate resistance. Near-linear, as hypothesized.
Plot of switching energy versus gate resistance showing five measured points climbing from about 8 to about 83 microjoules, close to a straight line.

Interpretation

The relationship is close to linear across this range - the fixed gate charge model holds. The interesting cost is thermal: at 50 kHz the 220 Ω resistor turns 4.2 W of switching loss into heat that the 10 Ω case simply does not.

Limitations

Resistive load only - an inductive load would shift losses toward turn-off. Probe compensation was checked but gate-loop inductance was not controlled.

Follow-Up

Repeat with an inductive load and a proper Kelvin gate connection.