The Forward and Flyback converters are two popular topologies widely used in isolated DC-DC power converters. These topologies are favored by designers for their simplicity, ability to handle multiple isolated outputs, and ease to optimize the duty cycle by selecting the transformer turns ratio.

Simplicity is partially based on the fact that conventional Forward and Flyback converters employ a single MOSFET switch, which is primary ground referenced for convenient gate drive implementation.

Editor's Note: The authors of this article will be presenting at the Embedded Power Conference to be held September 17-18 in San Jose, Ca. It will include courses by instructors from more than 20 companies on all aspects of embedded systems power management design as well as an experts panel. To attend, go to the Embedded Power Conference registration page.

However, the drawback to this single switch approach is that the voltage stress on the switch is the sum of the input voltage, the reflected transformer voltage and the turnoff voltage spike caused by leakage inductance.

Adding a second MOSFET switch on the high side results in the Two-Switch Forward or Flyback topology, of which the voltage stress on each MOSFET is clamped to the input voltage. The leakage inductance energy is also clamped and recycled back to the input to improve efficiency.

The dissipative snubber circuit that is often required in the single switch approach is no longer required. MOSFET switches with a rated voltage slightly higher than the input voltage can be employed in the two-switch topology, while a rating of greater than twice the input voltage is required for the single-switch topology.

For many applications the added complexity and part count of Two-Switch Forward and Flyback converters can be a small price to pay for the benefits received.

Figure 1. Two-Switch Forward Converter Topology

Two-Switch Forward Converter
Figure 1 above shows the Two-Switch Forward converter topology, which consists of the input capacitor CIN, two MOSFET switches Q1 and Q1, the power transformer T1, two clamp diodes D3 and D4, two rectifier diodes D1 and D2, and the output filter consisting of LO and Co.

Figures 2a and 2b below depict the operation of the Two-Switch Forward converter. Both Q1 and Q2 are turned on and off simultaneously. When they are on, as shown in Figure 2a, power is delivered to the load through the transformer and the output filter.

When the MOSFETs are turned off, as shown in Figure 2b, power flow in the primary circuit is cut off, and the voltage on the primary winding will reverse until the dot end is clamped to return by D3 and the non-dot end is clamped to VIN by D4. Therefore, each MOSFET will see a turnoff voltage stress magnitude of VIN.

Figure 2. Operating Modes of the Two-Switch Forward Converter

Not only is the energy from the transformer magnetizing inductance clamped but more importantly the leakage inductance energy is also clamped and returned to the input power bus through diodes D3 and D4. Energy stored in the leakage inductance during the on-time does not have to be dissipated in a resistive snubber or the MOSFETs themselves.

This advantage over a single switch approach reduces system power losses and reduces system noise, since the ringing normally associated with the release of the inductive energy is now clamped. Consequently, there is no need for snubber circuit and the EMI signature of the converter is greatly reduced.

Transformer core reset in a single switch Forward converter is normally accomplished with a tertiary reset winding. Generally the reset winding has the same number of turns as the primary winding. Thus, the core will always reset with a reset time equal to the on-time of 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 of the MOSFET switch to less than 50% the transformer core will always reset each cycle. The two-switch Forward converter resets the transformer in exactly the same way without the additional reset winding, because the conduction of D3 and D4 effectively applies the input voltage in reversed polarity to the power transformer 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 benefit can not be overstated. Peak voltage stress in a single switch approach is proportional to the value of leakage inductance, switching speed and circuit layout. Leakage inductance is difficult to control and can often vary even after the design goes into production.

At first glance, the series conduction loss of the high side MOSFET appears to be additional power dissipation. However, a study of MOSFET process characteristics reveals that the two-switch topology can actually results in a reduction of conduction losses. For a single-switch Forward converter with a 36V to 75V input application, a 200V MOSFET is often required provided the leakage inductance spike is controlled.

The die size, and hence the cost of 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 of the two MOSFETs usually can be smaller than a single switch with twice the voltage capability, for a given die size.

Gate drive losses are obviously higher with two switches, but with the lower Rds(ON) and the elimination of leakage inductance loss often results in a gain of conversion efficiency. The elimination of snubber components and control of the leakage inductance effects are big benefits of the two-switch topology especially at higher input voltages.

Higher input voltage applications often have more primary turns which tend to increase leakage inductance and loss. The benefits of the two-switch approach increase with increasing input voltage, but lower input voltage applications can often benefit as well.