POWER CONVERTERS
Understand two-switch orward/fyback con converters verters By Youhao Xi and Robert Bell National Semiconductor
The orward and yback converters are two popular topologies widely used in isolated DC-DC power converters. These topologies are avoured by designers or their simplicity, ability to handle multiple isolated outputs, and ease to optimise the duty cycle by selecting the transormer turns ratio. Simplicity is partially based on the act that conventional orward and yback converters employ a single MOSFET switch, which Figure 1: Two-switch orward converter topology is primary ground reerenced or convenient gate drive implemen- orward converter topology, tation. However, the drawback to which consists o the input cathis single switch approach is that pacitor CIN, two MOSFET switches the voltage stress on the switch is Q1 and Q1, the power transormer the sum o the input voltage, the T1, two clamp diodes D3 and D4, reected transormer voltage and two rectifer diodes D1 and D2, the turno voltage spike caused and the output flter consisting o by leakage inductance.’ LO and Co. Adding a second MOSFET Figures 2a and 2b depict the switch on the high side results in operation o the two-switch orthe two-switch orward or yback ward converter. Both Q1 and Q2 topology, o which the voltage are turned on and o simultanestress on each MOSFET is clamped ously. When they are on ( Figure to the input voltage. The leakage 2a), power is delivered to the load inductance energy is also clamped through the transormer and the and recycled back to the input to output flter. improve eciency. When the MOSFETs are turned The dissipative snubber circuit o (Figure 2b), power ow in the that is oten required in the single primary circuit is cut o, and the switch approach is no longer re- voltage on the primary winding quired. MOSFET switches with a will reverse until the dot end is rated voltage slightly higher than clamped to return by D3 and the the input voltage can be em- non-dot end is clamped to VIN by ployed in the two-switch topol- D4. Thereore, each MOSFET will ogy, while a rating o greater than see a turno voltage stress magtwice the input voltage is required nitude o VIN. or the single-switch topology. Not only is the energy rom the Figure 2: Operating Modes o the two-switch orward converter For many applications the transormer magnetizing inducadded complexity and part count tance clamped but more impor- dissipated in a resistive snubber or inductive energy is now clamped. o two-switch orward and yback tantly the leakage inductance en- the MOSFETs themselves. Consequently, there is no need or converters can be a small price to ergy is also clamped and returned This advantage over a single snubber circuit and the EMI sigpay or the benefts received. to the input power bus through switch approach reduces system nature o the converter is greatly diodes D3 and D4. Energy stored power losses and reduces system reduced. Two-switch orward converter in the leakage inductance during noise, since the ringing normally Transormer core reset in a Figure 1 shows the two-switch the on-time does not have to be associated with the release o the single switch orward converter
eetindia.com | EE Times-India
is normally accomplished with a tertiary reset winding. Generally the reset winding has the same number o turns as the primary winding. Thus, the core will always reset with a reset time equal to the on-time o the transistor. The voltage stress on the MOSFET switch will be twice the input voltage plus the spike caused by the leakage energy. By limiting the duty cycle o the MOSFET switch to less than 50 per cent the transormer core will always reset each cycle. The two-switch orward converter resets the transormer in exactly the same way without the additional reset winding, because the conduction o D3 and D4 eectively applies the input voltage in reversed polarity to the power transormer primary winding to reset the core. Since the maximum drain to source voltage across the MOSFETs is clamped to VIN, there is no uncertainty as to what the peak voltage stress will be. This beneft can not be overstated. Peak voltage stress in a single switch approach is proportional to the value o leakage inductance, switching speed and circuit layout. Leakage inductance is dicult to control and can oten vary even ater the design goes into production. At frst glance, the series conduction loss o the high side MOSFET appears to be additional power dissipation. However, a study o MOSFET process characteristics reveals that the twoswitch topology can actually results in a reduction o conduction losses. For a single-switch orward converter with a 36V to 75V input application, a 200V MOSFET is oten required provided the leakage inductance spike is controlled. The die size, and hence the cost o a MOSFET, are proportional to both the on-resistance (RdsON) and the voltage rating. While the two-switch approach requires two MOSFETs in series, the total resistance o the two MOSFETs usually can be smaller than a single switch with twice the voltage capability, or a given die size. Gate drive losses are obviously
eetindia.com | EE Times-India
Figure 3. Block diagram o high side gate drive circuit
Figure 4: Two-switch yback converter topology
higher with two switches, but with the lower Rds(ON) and the elimination o leakage inductance loss oten results in a gain o conversion eciency. The elimination o snubber components and control o the leakage inductance eects are big benefts o the two-switch topology especially at higher input voltages. Higher input voltage applications oten have more primary turns which tend to increase leakage inductance and loss. The benefts o the two-switch approach
increase with increasing input voltage, but lower input voltage applications can oten beneft as well. Historically, driving the high side MOSFET has been a challenge or the two-switch topology since the high side MOSFET requires a oating gate driver. New monolithic IC regulators eliminate the headache o the high side MOSFET gate drive through the use o a boot-strap capacitor technique controlled by a high speed level shit circuit. Figure 3 shows
the block diagram o the high side gate drive implementation. The advantages o two-switch orward become more remarkable in an integrated solution where the complete control circuit, gate drive or both high side and low side switches, and even the two high voltage MOSFETs, can all be integrated in the same IC. By clamping voltage stress on the MOSFETs, the maximum input voltage range o the power converter can approach the rated voltage o the MOSFETs, making
ull use o the MOSFET process capability. In contrast, the maximum input voltage range or a singleswitch orward converter is limited to less than hal the rated voltage o the MOSFET. A typical example o the ully integrated Two-Switch DC-DC regulator is National Semiconductor’s LM5015, which provides a high perormance low cost DC-DC regulator solution capable o a very wide input voltage range rom 4.25V to 75V. Two-switch fyback converter Figure 4 shows a Two-Switch
Flyback converter topology, which consists o two MOSFET switches Q1 and Q2, the power transormer T1, two clamp diodes D1 and D2, the secondary rectifer diode DO, the input flter capacitor CIN and the output flter capacitor CO.
eetindia.com | EE Times-India
Both MOSFET switches are All o the same benefts are turned on and o simultaneously, realised in the two-switch yback as in the two-switch orward as in the two-switch orward. The converter. The operation o the voltage stresses on the MOSFET Flyback transormer is best de- switches are clamped to VIN, and scribed as two-winding coupled the leakage inductance energy inductor. Energy is supplied to is returned to the input instead the inductor in the primary circuit o being dissipated in snubbers when the primary MOSFETs are ac- that are normally required in the tive, then the energy is released to single switch approach. the secondary when the primary The same techniques as shown MOSFETs turn o. in Fig. 3 can be used or the high The coupling between side MOSFET gate drive. The Twothe primary and secondary Switch Flyback can be operated in windings is never perect; this either discontinuous or continuleakage inductance can de- ous conduction mode just like the stroy the primary MOSFET in a Single-Switch Flyback converter. single switch approach i let unchecked. The clamp diodes Conclusion in the Two-Switch Flyback are By adding a high side MOSFET used to recover the leakage switch, the two-switch orward or energy back to the input, and to yback topology clamps the voltclamp the turn-o peak voltage age stress on each MOSFET switch across each MOSFET at VIN. to the input voltage. With the two-
switch approach, the leakage energy is recycled back to the input to improve eciency and there is no need or dissipative snubber circuit that is oten required in the single switch converters. The added complexity and part count o the two -switch orward and yback converters can be a small price to pay or the benefts received. The advantages o two-switch approach become more remarkable in an integrated solution in which the gate drive or both switches, and even the two MOSFET switches can be integrated in the same IC with the control circuit. The integrated solution allows the input voltage range to approach the rating o the MOSFET process capability providing a small orm actor, high perormance solution.