Renesas
PWM Duty Cycle, Dead Time, and Switching Loss
-
Question 1 of 11
What does duty cycle mean in a motor drive?
Reveal answerEach motor wire connects to a pair of switches called a half bridge. The top switch connects the wire to the supply. The bottom switch connects it to ground. Duty cycle is the share of each cycle that the top switch stays on. The motor winding smooths out the fast pulses, so the motor sees the average voltage.
Average voltage = duty cycle × supply voltage
At 40% duty on a 48 V supply: 0.4 × 48 V = 19.2 VRenesas motor firmware works with the same idea. The 1-shunt sensorless vector control application note describes a modulation rate based on the command voltage and the supply voltage (R01AN6840EJ0121, Section 7.10, p. 48).
-
Question 2 of 11
Why do motor drives use center-aligned PWM?
Reveal answerIn edge-aligned PWM, every pulse starts at the same instant. In center-aligned PWM, the timer counts up and then back down, so each pulse is centered in the cycle. This spreads the switching edges apart. It also creates a quiet window at the bottom of the count when all the bottom switches are on. That window is the best time to measure motor current.
The 3-shunt sensorless vector control application note takes its current readings at that point. It syncs the analog-to-digital converter with the bottom, or trough, of the PWM carrier (R01AN6839EJ0121, Section 8.2, p. 51).
-
Question 3 of 11
What is complementary PWM?
Reveal answerA three-phase motor drive has six switches in three pairs. In complementary PWM, the two switches in each pair always do the opposite of each other. When the top switch is on, the bottom is off, and the reverse is also true. The timer creates both signals from one setting, so they stay in sync.
The Renesas Flexible Software Package, or FSP, controls all three pairs with one driver. The FSP three-phase driver for the general PWM timer, or GPT, sets up complementary PWM and adds dead time to each pair (FSP v6.6.0 User’s Manual, r_gpt_three_phase, Overview section).
-
Question 4 of 11
What is shoot-through, and how does dead time prevent it?
Reveal answerShoot-through happens when both switches in a pair are on at the same time. That creates a direct short across the power supply. Current spikes, and the switches can fail in an instant. The risk comes from timing. Switches take longer to turn off than to turn on. Dead time is a short pause where both switches are held off, so one fully stops before the other starts.
The 1-shunt sensorless vector control application note explains the purpose directly. The inverter sets a dead time with both switches off to keep the top and bottom from creating a short circuit (R01AN6840EJ0121, Section 7.9, p. 47).
-
Question 5 of 11
How do you choose a dead time value?
Reveal answerDead time must be long enough for the outgoing switch to fully turn off before the incoming one turns on. Pull the timing numbers from the transistor and gate driver datasheets. Then add a safety margin for temperature and part-to-part variation.
Dead time ≥ turn-off delay + fall time − turn-on delay + driver mismatch + margin
Example: 60 ns + 30 ns − 15 ns + 5 ns = 80 ns. Add 50% margin and you get about 120 ns.
Gate driver timing comes from the driver datasheet. The RRP68150 and RRP68151 half-bridge driver datasheet lists typical delays of 19 ns and 17 ns. It also notes the driver has no built-in dead time, so the controller must provide it (R16DS0489EU0102, Sections 3.5 and 5.6, pp. 8 and 14).
-
Question 6 of 11
How does dead time affect motor performance?
Reveal answerDuring dead time, both switches are off, but the motor current keeps flowing through a diode. The voltage on the motor wire then depends on which way the current is flowing, not on the PWM command. This creates a small voltage error every cycle.
Voltage error ≈ supply voltage × dead time × switching frequency
24 V × 2 µs × 10 kHz = 0.48 VAt low speed, the commanded voltage is small, so this error matters more. It can cause rough running and distorted current.
The same 1-shunt application note explains that the error depends on current direction, dead time, and the power switches. Its solution adds an opposite correction to the voltage command (R01AN6840EJ0121, Section 7.9, p. 47).
-
Question 7 of 11
What are conduction loss and switching loss?
Reveal answerPower switches waste energy in two ways. Conduction loss happens while a switch is fully on, because it has a small resistance. Switching loss happens during each on or off transition, when voltage and current overlap for a moment.
Conduction loss = current² × on-resistance × duty cycle
(20 A)² × 5 mΩ × 0.5 = 1 WSwitching loss = ½ × voltage × current × transition time × switching frequency
½ × 48 V × 20 A × 40 ns × 20 kHz = about 0.38 WSwitching loss grows with frequency. Doubling the frequency doubles it. Higher frequencies make the motor quieter and smoother, so designers balance noise against heat.
The IGBT loss calculation application note splits losses the same way. It defines switching loss as switching frequency times the energy lost each time the switch turns on and off (R07AN0046EJ0100, Section 2.2, p. 3).
-
Question 8 of 11
Does dead time add loss too?
Reveal answerYes. During dead time, current flows through a diode instead of a switch. A silicon diode drops roughly 0.7 to 1 V, which wastes more power than a switch that is fully on. This happens twice per cycle.
Dead-time loss = 2 × diode voltage × current × dead time × switching frequency
2 × 0.8 V × 20 A × 500 ns × 20 kHz = 0.32 W per phaseThe same IGBT loss calculation application note treats the diode as its own source of loss. It calculates diode conduction loss from forward voltage and current (R07AN0046EJ0100, Sections 3.1 and 3.2, p. 4).
-
Question 9 of 11
Should dead time be created in hardware or software?
Reveal answerSoftware can create dead time by setting two slightly different timing values for each switch pair. That works, but one bug or late update can remove the gap and short the supply. Hardware dead time is built into the timer. You set one duty value, and the timer adds the gap automatically every time. That is why most motor control MCUs include it.
An RH850/U2x application note compares the two approaches. One example creates three-phase PWM without the dead time module. The other uses the module to add 2.5 µs at an 8 kHz carrier (R01AN6371EJ0110, Sections 2.1 and 2.2.1, pp. 4 and 10).
-
Question 10 of 11
How does an MCU shut off the motor quickly during a fault?
Reveal answerSoftware checks current once per control cycle, which can be too slow for a short circuit. Hardware protection connects a fault signal directly to the PWM outputs. When a fault occurs, the outputs shut off right away without waiting for the CPU.
On Renesas RA MCUs, the port output enable for GPT, or POEG, handles this. The FSP POEG driver can trigger from an input pin, a comparator, or matching output levels (FSP v6.6.0 User’s Manual, r_poeg, Features section). The 3-shunt application note uses POEG to put the PWM outputs into a high-impedance state when overcurrent is detected (R01AN6839EJ0121, Section 8.2, p. 51).
-
Question 11 of 11
How do GaN transistors change dead time?
Reveal answerGallium nitride, or GaN, transistors switch far faster than silicon. That means dead time can shrink from microseconds to tens of nanoseconds. Shorter dead time means less wasted energy and less voltage error. It also means the timing must be set precisely, because GaN loses more energy for every extra nanosecond of dead time.
The RRP68150 and RRP68151 datasheet covers 100 V half-bridge drivers designed for GaN. Its design example uses just 20 ns of dead time (R16DS0489EU0102, Table 3, p. 16).