To increase efficiency in portable applications, many designers have turned to synchronous rectification for low-voltage switching power converters. This is usually implemented using a power MOSFET as the rectifier, either in parallel or in place of the Schottky diode.
To operate properly and attain maximum efficiency, the synchronous rectifier element needs to be turned on and off at the proper times during the switching cycle. If the rectifier is turned on at the wrong times, high cross-conduction currents can occur, creating noise spikes and possibly damaging components.
To complicate matters even further, synchronous converters operating in a discontinuous conduction mode need to turn off the rectifier when the current reaches zero. This point varies with load and input voltage, requiring additional circuitry to monitor the rectifier current.
In most cases, the drive signal for the rectifier is supplied by a PWM controller specifically designed for synchronous rectification. The controller provides the proper drive voltage and phasing, relative to the main power switch, to prevent cross-conduction. It also provides a means for detecting zero current. Some converters, however, are “self-driven.” Usually, these are transformer-coupled converters that use an additional winding on the secondary of the power transformer to provide the proper gate drive to the rectifying MOSFET.
Adding synchronous rectification after-the-fact, to improve the efficiency of a non-synchronous design, can be tricky. This often will require the use of charge pumps or highspeed level translation circuitry, along with circuits to provide the proper timing and prevent cross-conduction.
For simple, non-isolated boost, buck, or buck-boost topologies operating in continuous-conduction mode, the addition of synchronous rectification may be simplified by using a dual-winding inductor. This provides the proper gate-drive voltage to the synchronous rectifier, eliminating the need for charge pump and level translator circuitry. However, rectifier turn-off can still be an issue.
Consider the buck-boost circuit, for example, which generates 3.3 V from a lithium-ion battery (Fig. 1). One winding of a dual-winding inductor provides the gate drive to an n-channel MOSFET, which was added for purposes of synchronous rectification. When the main switch inside the UCC3954 controller turns off, the voltage across L1A reverses, and L1B provides a regulated 3.3 V to drive rectifier Q1 on. However, without a signal to turn the rectifier off at the beginning of the next cycle, a tug-of-war results between Q1 and the main switch when the main switch turns back on. A “Catch-22” condition exists: Q1 can’t turn off until the voltage on L1 reverses, and the voltage on L1 can’t reverse until Q1 turns off.
The circuit shown in Figure 2 offers a simple, low-cost solution to this problem. Adding a small ferrite bead in series with Q1 enables the voltage on L1 to reverse by momentarily blocking the current through the MOSFET when it tries to change direction, allowing the rectifier to be properly commutated. This results in a typical efficiency improvement of 7% at currents of a few hundred milliamps. The low-resistance beads are available in small surface-mount packages. Diode D1 provides a current path at the beginning of the rectifier conduction cycle, because the bead also delays the current through Q1 at turn-on.
Reprints
