A Study of ZVS, ZCS and ZVZCS techniques in Reducing the Energy Loss and Improving the Soft Switching Range

This paper presents a comparative analysis of the ZVZCS soft switching technique with the ZVS and the ZCS counterpart. The generalization of the voltage-current crossover or the energy loss factor obtained from simulation of the prototype converter shows that the ZVZCS significantly reduces the loss and helps to improve the efficiency of the converter as compared to the ZVS or the ZCS. On the other hand, it is also found that the soft switching range of operation of the ZVS and the ZCS are largely affected by the maximum switch voltage and switch current respectively. In contrary, these factors have a negligible effect on the ZVZCS operation which results in an extended range of soft switching operation. Additionally, a detailed LTPICE simulation is performed for selected ZVS, ZCS and ZVSCS topologies from the recent literature and the switching losses in the main switches of the converters are measured. It is observed that the energy losses in the ZVZCS mode are reduced on average by approximately 26 % at turn on and 20 % at the turn off as compared to the ZVS and the ZCS. Besides, the low standard deviation in this mode confirms a stable low loss profile which renders extended soft switching range. An experimental test is also conducted by building the prototype converter to verify the simulation results. It is found that the switching losses are minimum while the converter is operated in the ZVZCS mode. Besides, the efficiency drop remains consistently low as compared to the ZVS and the ZCS in the whole operating range. Resultantly, the simulation and the experimental results are both found to be consistent.


Introduction
The switching losses are one of the primary contributors to the efficiency reduction of the dc-dc converters. This is particularly critical when the converter is operated at low power level and high switching frequency. To alleviate this problem, various soft switching techniques are being proposed [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15], which includes the zero voltage switching (ZVS), zero current switching (ZCS) and zero voltage zero current switching (ZVZCS). When operating in these modes, the voltage and current transients are manipulated to reduce the voltage-current crossover, otherwise defined as the switching loss. In the ZVS case, the voltage transients at the turn on and turn off instants are restricted to reduce large crossover with the current. On the other hand, for the ZCS, the current transientsrather than the voltage are controlled in order to achieve the same purpose. In the ZVZCS [16][17][18][19][20][21][22], both the voltage and current transients are simultaneously controlled to reduce the crossover at both the turn on and turn off instants. This results in significant improvement in the switching losses, hence higher efficiency. Particularly, this is a direct consequence of simultaneous manipulation of both the voltage and the current transients to reduce the crossover losses. However, the existing papers [16][17][18][19][20][21][22] do not explicitly cover this issue and analyze the ZVZCS from this point of view. Hence, considering these factors, it would be interesting to investigate the relative standing of the ZVZCS to the other soft switching techniques (i.e. the ZVS and ZCS) in terms of loss reduction capability and soft switching range of operation. Eventually, a generic conclusion regarding the superiority or inferiority of the ZVZCS technique over the ZVS and the ZCS can be reached. Additionally, this would further enhance the understanding of the soft switching techniques and their implementation into the dc-dc converters.
Based on this literary gap, this paper presents a generalized analysis to investigate the relative status of the ZVZCS over the ZVS and the ZCS techniques. For this purpose, the switching transients of the ZVS, ZCS and the ZVZCS are linearized from the simulation results of the prototype converter. A geometrical analysis to determine the crossover energy and the soft switching range of operation is then performed as demonstrated in Section II and III respectively. It is found that the crossover energy is reduced in the ZVZCS as compared to the ZVS and the ZCS. In addition, the dependency of the soft switching performance on the input voltage and the switch current is lowered. Next, the simulation for comparative analysis is performed. A large number of converters are selected (to retain a large sample size) for simulation and the results are presented in Section IV. The experimental analysis is then performed by building a prototype converter which is operable in different soft switching modes (i.e. the ZVS, ZCS and ZVZCS) independently. The experimental results are demonstrated in Section V. It is found that the ZVZCS reduce the loss margin further to improve the efficiency and extends the soft switching range by lowering the dependency on the switch voltage and current. Consequently, the superiority of the ZVZCS over the other soft switching techniques in terms of energy loss reduction capability and extending the soft switching range of operation is proved. Finally, Section VI draws the conclusion.

Analysis of the soft switching modes for energy loss measurement
The soft switching modes are analyzed to find the gross voltage-current crossover area induced at the turn on instant for each mode and the respective improvement from the hard switching counterpart by simulating the prototype ZVZCS converter [22]. Only the turn on instant is selected for the analysis as the turn off situation can be replicated by replacing simple parameters. To perform the theoretical analysis, the following assumptions are made: a) all the voltage and current transients are considered linearized b) the converter operates in the CCM mode, c) the standard ZVS and ZCS conditions are assumed where the switch voltage and the switch current remains always positive. The specifications are given in Table 1. The PWM switching is described in Section V. It is to be mentioned that the switching transients are typical for a MOSFET. An identical analysis can be performed for an IGBT. In this case, the energy loss would be higher at the turn off because of the tail current. Aside from this, the comparative results will be identical as long as the comparison is made based on energy loss (or the crossover area).

The Hard Switching Mode
The crossover area that is formed due to the voltage and current rise and fall times (at the switch turn on instant) is shown in Fig. 1. In Fig. 1 (a), the switching waveform pattern for PSPICE simulation is shown to justify the approximation in Fig. 1 (b) for theoretical analysis to be followed. Hence, the crossover area can be approximated by the area of the triangle  ABC Where, (t3-t1) is the base and h is the height of the triangle, respectively. Consequently, for switching frequency fs, the total energy loss at turn on instant is It can be assumed that the fall time of the switch voltage (tf) and the rise time of the switch current (tr) are equal. Hence, (2) can be rewritten as Now the switch voltage v can be described as where, t1<t<t3. The voltage and the current transients intersect at A located at t2 on the x-axis. Hence, the voltage v at t2 can be obtained by replacing t=t2 in (4), 21 max max where, vmax is the maximum voltage applied across the switch. Hence, by replicating the concept in (2), the total energy loss at turn on becomes,   In this mode, the sharp voltage fall at turn on is restricted by emulating zero voltage across the switch. This is done by turning on the body diode of the switch and discharging the passive snubber element. For computation purpose, it is assumed that the voltage transient is linear. The original hard switching curve for v is shown by the broken line. Hence, as obvious from Fig. 2 (b) and justifiable through Fig. 2 (a), the fall time for v is reduced and the voltage-current intersection is shifted left to a new position D from the hard switching point at A. The new intersection point D is located in between t2 and t1 and let it is at t′2. Consequently, the switch voltage v can be expressed as max 11 where, t1<t<t′3 . If voltage is v′p at t′2, then from (9) it can be obtained as Consequently, the energy loss in the ZVS turn on can be derived from (1) where, tf, ZVS is the revised switch voltage fall time for the ZVS mode and t′3-t1=tf, ZVS. If t′2-t1=t′, then (9) becomes Now, geometrically to maintain the ZVS mode switching, the following inequality must be true.
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 5 August 2020 doi:10.20944/preprints202008.0123.v1 The smaller the value of (tf, ZVS-t′), the better the ZVS execution. Hence, from (7) , (12) and (13), it can be deduced that the energy loss in the ZVS mode is reduced as compared to that in the hard switching counterpart. That is, ,, ZVS t on loss HS t on loss For turn off instant, (11) can be rewritten as

The Zero Current Switching Mode
In contrary to voltage restriction, the ZCS applies current restriction techniques to reduce the crossover area. For this purpose, the current slope is minimized while the voltage transient remains similar as that in hard switching mode. This is shown in Fig. 3. As evident from Fig. 3 (a), the current rise time (tr) and the voltage fall time (tf) are different for the ZCS operation. Accordingly, the switch voltage v in Fig. 3 (b) can be derived as where, t1<t<t3 . The voltage and current intersection now shift to the right from A to E located at t′′2 on the x-axis. Hence, v′′p at t′′2 can be defined as Hence, the energy loss becomes Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 5 August 2020 doi:10.20944/preprints202008.0123.v1 Here, t3-t1=tf=tf,ZCS. Now, as E is located far right to the point A in the ZCS operation as obvious from Fig. 3, the following inequality must be true. Consequently, Hence, from (7) and (18), it can be deduced that, That is, the energy loss is reduced in the ZCS mode as compared to the hard switching counterpart. At turn off (18) becomes,

The True Zero Voltage Zero Current Switching
In the ZVZCS mode, both the voltage and current transients are altered to minimize the crossover as obvious from Fig. 4 (a). For this case, as obvious from Fig. 4 (b), the voltage and current intersection where, t1<t<t′3 . Now, v′′′p at t′′′2 can be defined as Consequently, the energy loss can be defined as Here, t′3-t1=tf,ZVZCS=tf,ZVS and t′2-t1=t′. Moreover, x is defined as As F is located at the right of D, x must be positive. Hence, it is obvious from (22) and (25) that the energy loss in the ZVZCS mode is further truncated as compared to the ZVS or ZCS. For the turn off instant, (26) can be rewritten as Hence, from (7), (11), (18) and (25), it becomes obvious that the ZVZCS mode switching is more efficient in reducing the energy loss at the turn on point than the ZVS or the ZCS. More specifically, as compared to the ZVS, the ZVZCS reduces the loss in percentage by ,, Roughly the net loss reduction in percentage at the turn on and turn off should be, As compared to the ZCS, the loss reduction can be defined as ,,

ZCS t on loss ZVZCS t on loss t on loss reduction
ZCS t on loss Again, the net loss reduction as compared to the ZCS in percentage should be, Hence, it can be concluded that the ZVZCS is capable of reducing the energy loss at the switch transition points as compared to that induced in the HS, ZVS or the ZCS mode.

The Zero Voltage Switching (ZVS)
In the ZVS, the turn on loss is defined in terms of input voltage as Or in terms of switch current, it can be rewritten as It can be observed that the ZVS turn on loss largely depends on the maximum applied switch voltage vmax which is the input voltage for a forward mode converter. This dependency can be reduced if switching frequency (fs) is decreased. However, it is not an effective solution to reducing the switching frequency would increase the size of the passive components, while, the term (tf,ZVS-t′) is constant. On the other hand, the dependency on the switch current is much weaker as obvious from (32). Hence, this large dependency on the switch voltage makes the ZVS operation largely vulnerable to the input voltage and duty cycle ratio. Subsequently, the ZVS operation is difficult to maintain for the wide operating window.

The Zero Current Switching (ZCS)
In the ZCS, the energy loss in terms of maximum switch voltage (vmax) and current (imax) are described as In (34), as obvious from Fig. 3, Hence, the maximum switch voltage (vmax) has minimal effect on the ZCS operation. On the contrary, the ZCS operation is largely affected by imax as can be seen in (35). To reduce this effect, fs has to be minimized as the term (t2′′′/t4) is constant. However, as mentioned earlier, it is not an effective solution as the size of the passive components can increase substantially. The dependency on the switch current makes the ZCS operation vulnerable to the converter loading conditions. Consequently, for wide load variation, the ZCS is difficult to achieve. In other words, the soft switching range of operation is affected.

The True Zero Voltage Zero Current Switching
In this mode, the energy loss terms are Subsequently, as obvious from (38), the dependency of the ZVZCS operation on the switch current is negligible. Hence, the ZVZCS is not largely affected by the input voltage and loading conditions. In return, it becomes capable of providing wider soft switching range by remaining operational irrespective of the input voltage and load current conditions.

Simulation Results for comparative analysis
To prove the theoretical statement, a PSPICE simulation is performed for selected topologies from the recent literature. The converters are simulated in the boost mode and the following specifications are maintained: Furthermore, the converters are intentionally operated at a low power level (i.e. 200 W) for better visualization of the switching losses. The main switches of the converters are taken into consideration to measure the incurred turn on and turn off losses while the converters are operated either in a) the ZVS mode or b) the ZCS mode or c) the ZVZCS mode. The measurement is performed in different states (State A to State F) of the circuit. Each state denotes a unique combination of switching frequency (fs), the input voltage (VL) and the load current. The variation of the operational states makes the results unbiased to the operating conditions of the converter. Furthermore, the measured turn on and turn off losses for different soft switching conditions are averaged out to achieve more accurate results. The simulation results are demonstrated in Table 3 below. To provide a better visual understanding, the average turn on and turn off losses induced in the ZVS, the ZCS or in the ZVZCS mode are plotted in Fig. 5 Table 4 below. As can be observed, the ZVZCS incurs the lowest turn on and turn off losses on average. Besides, the standard deviation (SD) shows that the deviation from the average is fairly lower for the ZVZCS mode as compared to others. This signifies stable low loss profile throughout all six states and proves the superiority of the ZVZCS soft switching performance in the whole operating range.

Experimental verification
To verify the theoretical claim, an experimental test is performed. A prototype ZVZCS converter [22] is built to evaluate the switching losses under various soft switching conditions. The simulation waveforms for the prototype converter are demonstrated in Figs. (1) to (4) for different switching states. In this section, the experimental test procedure is discussed. The description of the converter operational modes, intervals and results are to be followed in the rest of this section.

Converter Topology and PWM Switching
The converter (as shown in Fig. 7) consists of the main switches S1 and S2, the auxiliary switches Sr1 and Sr2, the main inductor LM, the auxiliary inductors Lr1 and Lr2 and the auxiliary capacitor Cr1. It is bidirectional and operable in the ZVS, the ZCS and the ZVZCS mode separately. For the experimental verification, the converter is tested in the boost mode only and the turn on and turns off losses are calculated for the main switch S1. For this purpose, the circuit is operated in four different stages, namely Stage 1: the ZVZCS mode Stage 2: the ZVS mode Stage 3: the ZCS mode and Stage 4: the hard switching (HS) mode Fig. 7. The schematic of the ZVZCS converter under test [22] In Stage 1, the turn on and turn off of S1 is conducted in the ZVZCS mode. To achieve this, both the auxiliary switches (Sr1 and Sr2) are turned on as shown in Fig. 8 (a). This leads to the near zero switching loss at the transitions. In Stage 2, Sr1 is turned on and Sr2 is turned off as shown in Fig. 8 (b). Consequently, S1 is operated in the ZVS mode. The sharp voltage rise across the switch S1 is controlled by the ZVS mode switching by turning on the respective body diode. In Stage 3, to achieve the ZCS at S1, the Sr1 is turned off and Sr2 is turned on as in Fig. 8 (c). This forces the body diode of S1 to remain turned off. Simultaneously, the Lr2 and the capacitor Cr1 realize the ZCS operation in S1. For this purpose, Lr1 and Lr2 control the sharp rise of current through S1. Finally in Stage 4, for the hard switching stage, both the auxiliary switches (i.e. Sr1 and Sr2) are turned off as shown in Fig. 8 (d). To give a general idea on the operation of the circuit, the operational intervals are shortly described next.

Converter Operational Intervals
The operational intervals are important to understanding the basic operation of the circuit. Hence, the operational intervals are described in short which should be sufficient to understand the soft switching operation. At t1, the body diode of S1 is turned on and Lr3 starts to discharge through it. This, in return, enables S1 to be turned on in the ZVS mode at t2. At t2, the current through the switch Sr1 reaches zero and Sr1 is turned off in ZCS mode. To allow Lr3 to discharge completely, t21 has to be selected as Combining (1) and (2), the following relationship is obtained.  (1) and (2) are satisfied, the ZVS condition is achievable. However, to avoid unnecessary losses in the auxiliary components and the body diode of the main switch, the delay should be kept equal to or lower than twice the minimum limit. Hence, the functional range of t20 should be At t2, the switch S1 is turned on in the ZVZCS mode. To achieve this purpose, the inductor Lr2 in series with S1, restricts the sharp rise of current through the switch. In general, as the value of the inductor Lr2 increases, the slope di/dt decreases as obvious from (5). However, excessively large Lr3 would affect the ZVS transition of S1. At t3, Sr2 is turned on for a very short period. Consequently, Cr1 is charged by the reverse recovery current irr. To this purpose, Cr1 is designed to be able to accommodate the reverse recovery charge Qrr. Throughout this interval, the inductor current iLM continues to increase. At t4, S1 is turned off in the ZVZCS mode. This interval is kept short enough to avoid any unnecessary current stress on the power switches and simultaneously to avoid the short circuit condition while both S1 and S2 are switched on.

Results and Discussion
For all cases, the switching losses are measured in switch S1. In Fig. 10, the voltage and current waveforms of S1 are shown for different switching conditions. The measured losses are given in Table  6. It can be observed from Table 4 that the ZVS and ZCS provide improve the switching states and reduce the losses as compared to the hard switching condition by more than 30 %. However, the improvement induced by the ZVZCS operation from the HS counterpart is more than 80 % which surpasses other soft switching conditions. Thus, the efficiency is largely improved.  The efficiency improvement in the ZVZCS mode can be easily identified which is stated in Table 7.
As can be seen, the efficiency drop is reduced from 2.30 % in the hard switching mode to 0.23 % in the ZVZCS mode. This signifies an average of 2 % improvement in the efficiency of the converter  11 (a) and (b) respectively. As obvious, the efficiency drop in the ZVZCS mode remains consistently lower and is more stable throughout the operating conditions (i.e. variation in fs and VL). For example, as can be seen in Fig. 11 (a), Δ Ed= 0.2 % and 0.1 % for the ZVS, ZCS and the ZVZCS respectively. Similarly, it is 0.15 % and 0.05 % for the ZVS, ZCS and the ZVZCS respectively as shown in Fig. 11 (b). Consequently, wider soft switching operating range in the ZVZCS mode is ensured.

Discussion
In this paper, the ZVZCS soft switching technique is analyzed. A theoretical study is provided to demonstrate the superiority of the ZVZCS over the ZVS and the ZCS in reducing the switching losses and improving the soft switching range of operation. It is found that the improvement is primarily because of the ZVZCS not being largely affected by the factors that affect the ZVS turn on (i.e. the input voltage) and the ZCS (i.e. the load current) turn off operations. To support the theoretical claim, a simulation test is performed on several soft switching converters from recent literature. The converters are tested in different operational states to ensure unbiased results. The turn on and turn off losses are then measured in the main switches. It is observed that the induced losses are reduced for the ZVZCS operation which amounts to 26 % and 20 % reduction from the ZVS and the ZCS counterpart respectively. Besides, the low standard deviation (SD) for the ZVZCS operation denotes minimal deviation from the average loss profile throughout the operational states. Thus, a stable low loss profile rendering wider soft switching range is ensured. On top of that, a prototype soft switching converter is built to perform the experimental test. The converter is operable in different soft switching conditions (i.e. the HS, the ZVS, the ZCS, and the ZVZCS) independently. Subsequently, Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 5 August 2020 doi:10.20944/preprints202008.0123.v1 the incurred turn on and turns off losses in the main switch are measured under these conditions. Overall, it is observed that the ZVZCS mode reduces the switching losses as compared to the HS, the ZVS or the ZCS and help to improve the converter efficiency further. Moreover, it is found that the efficiency drop remains consistently low and stable while measured against the variation of switching frequency and input voltage. This proves the wider operability of the ZVZCS over the ZVS and the ZCS modes.