Switch-Mode Power Supplies for Beginners: An Efficiency Primer Part 1

May 13, 2008 10:04 AM
By Daniel Wagner, Associate Member of the Technical Staff, Applications, and Roger M. Kenyon, Director, Customer Applications Engineering, Maxim Integrated Products, Chandler, Ariz.

The charge required by the Miller capacitance is denoted Q_GD, and as with the Miller capacitance, should be minimal for faster switching. Since MOSFET capacitance also varies with die size, a compromise between conduction and switching losses is usually considered, with careful attention given to the switching frequency of the circuit.

For the diode, forward voltage must be minimized, as losses due to it can be large. Forward voltage typically ranges between 0.7 V to 1.5 V for small, lower-rated diodes. Again, dimensions, process and voltage rating affect forward voltage and reverse-recovery time, with higher ratings and larger sizes exhibiting higher V_F and t_RR, resulting in greater loss.

Switching diodes geared toward high-speed applications are often categorized by speed, namely fast, super fast and ultrafast recovery diodes, with reverse-recovery time diminishing as speed increases. Fast diodes tend to have t_RR in the hundreds of nanoseconds, while ultrafast diodes tend to be in the few tens of nanoseconds.

While pn diodes tend to have large forward-voltage drops, they are also available with large voltage and current ratings, making them suitable for higher-power applications. But, even with optimized V_F and t_RR diodes, one will not typically see a high-speed recovery diode in a low-power or portable application since the energy penalties are just too great.

As a possible alternative to fast recovery diodes in low-power applications, Schottky diodes offer virtually non-existent recovery times and V_F that is nearly half that of fast recovery diodes (often from 0.4 V to 1 V), but are not available with voltage ratings as high as those of fast recovery diodes. Because of the benefits, Schottky diodes are widely used in low-power applications to greatly reduce the power loss associated with the switching diode, especially in low duty-cycle applications.

However, even with a low forward-voltage drop, the Schottky diode can present unacceptable conduction losses in low-voltage applications. Consider a stepdown output of 1.5 V, where a typical 0.5-V Schottky diode is used. This is 33% of the output voltage during the diode conduction time!

This high-loss situation can be improved by taking advantage of the low R_DSon of a MOSFET in a technique called synchronous rectification. Here, a MOSFET replaces the diode (compare Fig. 1 and Fig 2) and is synchronized with the other MOSFET so that both conduct alternately during their respective switching intervals. Now, the relatively high V_F of the diode is replaced by a much lower R_DSon voltage drop (depending on current) of the MOSFET, recouping efficiency lost to diode conduction.

However, synchronous rectification does have its tradeoffs, such as increased complexity and cost, and it may not prove considerable benefit for very high-current levels, since conduction loss of the MOSFET increases as the square of its current. Additionally, since power is expended while turning on the gate of the synchronous rectifier, the engineer must weigh the efficient penalty of the additional gate drive.

The Data Sheet

This far, power losses intrinsic to the two major components of the generic switch-mode power supply, the MOSFET and diode, have been discussed. Recalling the stepdown circuit in Fig. 1, a few important aspects of the controller IC that aid in its very efficient operation can be related by referring to its data sheet.

First, the switching components are integrated into the IC package, thus saving space and reducing parasitic losses. Second, low R_DSon MOSFETs are used. These are specified at 0.27 Ω (typical) and 0.19 Ω (typical) for the NMOS and PMOS, respectively. Third, synchronous rectification is employed. For a 50% duty cycle and 500-mA load, this reduces lower switch conduction loss from 250 mW if a 1-V diode were used, to approximately 34 mW due to the low R_DSon of the synchronous NMOS transistor.

While switching components greatly influence SMPS efficiency, there are more areas where the engineer can combat the invasive effects of power loss. In the second part of this article, passive component losses and important efficiency enhancing features of SMPS ICs will be reviewed.

References

Mohan, Ned; Undeland, Tore M.; and Robbins, William P. Power Electronics: Converters, Applications, and Design, chapters 2, 7, 20 and 22, John Wiley & Sons, third edition, 2003.

Want to use this article? Click here for options!
© 2007 Penton Media, Inc.