Transistors｜Evaluation

The Importance of the Reverse Recovery Characteristics of Switching Elements in Inverter Circuits Basic Operation of 3-Phase Modulation Inverter Circuits

2023.12.14

Points of this article

・During operation of an inverter circuit, reverse recovery currents occur in body diodes.

・If reverse recovery times are long and reverse recovery currents are large, increased losses result; this is one disadvantage of inverter circuits.

・By using MOSFETs with short reverse recovery times and small reverse recovery current peaks, losses in inverter circuits can be reduced, and the risk of MOSFET destruction can be alleviated.

This article explains the second topic, “Basic operation of 3-phase modulation inverter circuits”. As mentioned in the previous article, from this point explanations will use as an example the sinusoidal driving (3-phase modulation) method, which is widely employed in motor driving.

■Types of inverter circuits and energization methods
■Basic operation of 3-phase modulation inverter circuits
■Comparison of losses in a PrestoMOS™ MOSFET and a standard SJ MOSFET using double-pulse tests (actual measurement results)
■Comparison of efficiency of a PrestoMOS™ MOSFET and a standard SJ MOSFET in a 3-phase modulation inverter circuit (simulations)

Basic Operation of 3-Phase Modulation Inverter Circuits

Fig. 6 is a timing chart for the U phase of a 3-phase modulation inverter circuit. During U phase positive polarity, the high side switch (Q1) performs energizing, and therefore as the U phase current peak is approached the gate driving signal duty increases, and the closer the approach to negative polarity, the more the duty decreases; during negative polarity, freewheeling operation occurs. The opposite is the case during U phase negative polarity; the low side (Q2) performs energizing, and freewheeling operation occurs with positive polarity.

In this driving pattern, PWM operation and freewheeling operation are similarly occurring in the V and W phases as well, and so a feature of this circuit is the fact that switching is occurring in all three phases, regardless of the AC output timing; for this reason, it is called 3-phase modulation operation.

The duty D(t) as a function of time for switching in each phase is represented by the following equation using the AC frequency f of the inverter output and the phase difference θ.

Here D_max is the duty of the AC output peak, and is called the modulation factor.

Fig. 7 shows the phase current waveform for the U phase near the U phase current peak (positive polarity), and the gate driving waveforms for the transistors of each phase (Q1/Q2, Q3/Q4, Q5/Q6).

In the vicinity of the U phase current peak, the section in which the U phase high side switch (Q1), which is an energizing switch to accumulate energy in the inductor, is switched from on to off and again to on, can be explained by dividing the section into operation modes labeled (1) to (13). The diagrams shown below indicate the changes in current paths for the U phase.

	Mode（1）・Q1, Q4 and Q6 are on, Q2, Q3 and Q5 are off. ・The drain potential of Q4 and Q6 becomes 0 V. ・The Q2 drain potential becomes V_in, and V_in is applied to the U phase inductor L_U. ・・By passing an energizing current through L_U, energy is accumulated in L_U.
	Mode（2）・Q1 and Q6 are on, Q2, Q3, Q4 and Q5 are off. ・The Q6 drain potential remains in the 0 V state. ・L_V, which was energized by the L_U energizing current, has a freewheeling current flowing through the body diode of Q3 due to Q4 being turned off. ・Due to this freewheel current and the energizing current flowing to L_W, an energizing current and freewheel current flow in L_U.
	Mode（3）・Q1, Q3 and Q6 are on, Q2, Q4 and Q5 are off. ・The Q6 drain potential remains in the 0 V state. ・Q3 turns on, and the freewheel current which had been flowing in the Q3 body diode flows in the Q3 channel, so that synchronous rectification operation occurs. ・An energizing current and freewheeling current continue to flow in L_U.
	Mode（4）・Q1 and Q3 are on, Q2, Q4, Q5 and Q6 are off. ・L_V, which was energized by the L_U energizing current, has a freewheeling current flowing through the body diode of Q3 due to Q4 being turned off. ・As a result, L_V and L_W enter freewheeling states, and the sum total of the freewheeling currents cause the current flowing in L_U to be maintained.
	Mode（5）・Q1, Q3 and Q5 are on, Q2, Q4 and Q6 are off. ・Q5 turns on, and the freewheeling current which had been flowing in the Q5 body diode flows in the Q5 channel, so that synchronous rectification operation occurs. ・At this time, L_V and L_W remain in freewheeling states, and the current flowing in L_U is maintained.
	Mode（6）・Q1 and Q3 are on, Q2, Q4, Q5 and Q6 are off. ・First Q5 is turned off, so that a freewheeling current again flows in the Q5 body diode, and freewheeling currents flow in current paths similar to Mode (4).
	Mode（7）・Q1, Q3 and Q6 are on, Q2, Q4 and Q5 are off. ・Upon again turning on Q6, the Q6 drain potential is lowered to 0 V. ・By lowering the Q6 drain potential, the V_in voltage is again applied at the L_U pin. ・Current paths are similar to Mode（3）, and a freewheeling current and energizing current flow in L_U.
	Mode（8）・Q1 and Q6 are on, Q2, Q3, Q4 and Q5 are off. ・Upon again turning off Q3, a freewheeling current flows in the Q3 body diode. ・Current paths are similar to Mode（2）, and a freewheeling current and energizing current continue to flow in L_U.
	Mode（9）・Q1, Q4 and Q6 are on, Q2, Q3 and Q5 are off. ・Upon again turning off Q4, the Q4 drain potential is lowered to 0 V. ・Current paths are similar to those in Mode (1), and freewheeling currents no longer flow. ・L_U, L_V, and L_W are in energized states, a large energizing current again flows in L_U, and energy is accumulated in L_U.
	Mode（10）・Q4 and Q6 are on, Q1, Q2, Q3 and Q5 are off. ・Upon turning off Q1, the energizing current that had been flowing in L_U ceases to flow. ・At this time, L_U is retaining energy, and so a freewheeling current flows in the Q2 body diode.
	Mode（11）・Q2, Q4 and Q6 are on, Q1, Q3 and Q5 are off. ・Upon turning on Q2, the freewheeling current that had been flowing in the Q2 body diode flows in the Q2 channel, so that synchronous rectification operation occurs. ・The freewheeling current continues to flow due to the energy accumulated in L_U.
	Mode（12-1）・Q4 and Q6 are on, Q1, Q2, Q3 and Q5 are off. ・Upon turning off Q2, a freewheeling current again flows in the Q2 body diode. ・The freewheeling current continues to flow due to the energy accumulated in L_U.
	Mode（12-2）・Q4 and Q6 are on, Q2, Q3 and Q5 are off. ・In this mode Q1 is turned on from the off state. ・Q1 is turned on while a freewheeling current is flowing in the Q2 body diode, so a reverse recovery current occurs in the Q1 channel and the Q2 body diode. ・Due to this reverse recovery current, a turn-on switching loss occurs. ・When the reverse recovery current stops flowing, there is a transition to Mode (13).
	Mode（13）・Q1, Q4 and Q6 are on, Q2, Q3 and Q5 are off. ・When the reverse recovery current stops flowing, current flows in paths similar to those of Mode (1). ・Through the energizing current flowing in L_U, energy is again accumulated in L_U.

From this operation, a reverse recovery current occurs in body diodes, for example in Mode (12-2). Such reverse recovery currents occur in the body diodes of all the switches Q1 to Q6, and so the relative merits of the reverse recovery characteristics in such an inverter circuit are extremely important. Adverse effects of these reverse recovery currents include the following.

●Cases in which the reverse recovery current (peak current) is large
For example, as in Mode (12-2), when Q1 turns on a reverse recovery current of Q2 flows. If the reverse recovery current peak I¬rr¬ is large, an excessive current flows in Q1. At this time, if the MOSFET rating is exceeded (the current density is high), a drain-source short-circuit destruction may occur to result in an arm short-circuit state, with the possible destruction of both the Q1 and Q2 MOSFETs.

●Cases in which the reverse recovery time is long
When a reverse recovery current of a body diode flows, if in Mode (12-2) the body diode of Q2 is conducting, a voltage due to V_in is applied across the Q1 drain and source. At this time, the turn-on switching waveform is as shown in Fig. 11. The longer the reverse recovery time t_rr, the longer are the time during which the drain current I¬D(t) of the drain of the turn-on Q1 is flowing and the time during which the drain-source voltage V_DS(t) is applied. The switching loss P_SW at this time is expressed by the following equation, where TS is one switching period.

From equation (2), the area of the product of I_D(t) and V_DS(t) multiplied by time is the loss energy E_ON. Therefore, we see that the slower the reverse recovery, the greater is the increase in the switching loss. In the case of an inverter circuit, the current flowing in inductors changes sinusoidally, so that depending on the switching timing, the turn-on reverse recovery current also changes. In other words, the closer to the sine wave peak, the greater is the reverse recovery current. Hence in switching near the sine wave peak, losses due to the reverse recovery current are larger, so that proper caution is particularly necessary.

In this way, a long reverse recovery time and a large reverse recovery current act as disadvantages in an inverter circuit. By using MOSFETs with short reverse recovery times and small reverse recovery current peaks, losses in an inverter circuit can be reduced, and the risk of switching device destruction can be alleviated.

In general, double pulse tests are used in single-arm evaluations of inverter circuits. In the next article, double pulse tests are used to compare the losses for MOSFETs with superior reverse recovery characteristics and standard SJ MOSFETs.

Transistors

Basic

Evaluation