[Design Application]
Modular Converters Speed Power Designs
Though Easy To Use, Modular DC-DC Converters Must Still Be Chosen Carefully To Get A Jump On Noise, Heat, And Safety Issues.
Modular dc-dc converters, in numerous standard designs, provide reliable, field-proven, power solutions. While users can easily specify inputs, outputs, and power levels, the main advantage of these devices is that a user need not be an expert in power conversion to design them into a pc-board application. Their successful use, however, does require the designer to carefully consider noise, heat, and safety issues. In addition, the design choices made by dc-dc converter manufacturers can have a profound impact on the modules' design-in process. This makes it imperative that the designer select carefully among these relatively standard products.
These issues have always been at the forefront of power-system design. However, user demands for higher power densities, higher efficiencies, and smaller packages continue to raise the ante.
The main and most obvious reason behind opting for a modular solution is design simplicity. A complete power system using modular components can be implemented with an ac-dc front end, dc-dc converters for each of the outputs needed, and a few discrete components. More output voltages can be obtained with additional dc-dc converters and filters. A simplified set of inputs and outputs is all a designer needs to consider.
With thousands of dc-dc converters and multiple manufacturers to choose from, the process of selecting the right module can be a major challenge. What's more, each supplier is, of course, seeking a competitive edge. As a result, one supplier provides the highest efficiency, another the smallest footprint, and yet another a new high for power density, and so on. Unfortunately, each achievement likely comes with trade-offs elsewhere in the specification.
Noise And Topology Noise can vary widely among converters from supplier to supplier and model to model. The reasons range from the fundamental technology employed, to simple differences in design choices, to variations in intended applications.
Many converter topologies are used to produce the output voltage, power, and regulation needed by electronic equipment. These topologies reduce to essentially two classespulse-width modulation (PWM) and quasi-resonant designs, such as zero-current-switching (ZCS).
In switch-mode converters, common-mode conducted noise is a function of the dv/dt across the main switch in the converter, and the effective input-to-output capacitance of the converter. It's not always easy to identify the specific noise generator, so here are some typical sources, derived from real-life modules:
Topology: Noise is highly dependent on the topology. A PWM topology, for example, often produces noise at high frequencies (above 5 MHz). The most likely source of that noise is construction-method parasitics generated by the high dv/dt and di/dt associated with this topology. Components such as diodes and MOSFETs generate heat, so they are mounted on an insulating ceramic substrate, which is mounted to the aluminum baseplate of the module. Because ceramic is a dielectric, there is capacitance from the diode and the FET to the baseplate. So, while this construction facilitates heat removal, it also produces parasitic capacitance, which generates noise.
Switching harmonics: Multiples of the 300-kHz switching frequency up to 9 MHzor 30 times the switching frequencyhave been found in some modules. Such converters, used with a typical EMI filter, have been known to fail VDE0871 B requirements for conducted noise.
Packaging and circuit-design: High capacitive coupling, common in metal pc boards and planar-transformer designs, can produce noise as much as 25 times higher than a typical module. Although isolation is normally thought of as a safety issue, the high capacitive coupling associated with nonisolated converters, or those with low isolation voltage, also contributes to higher system noise.
A comparison of the noise produced by converter modules of different design is shown (Fig. 1a and 1b). The fundamental comparison in this case is technological: one module employs pulse-width modulation (where the frequency is fixed and the duty cycle is variable) (1a), while the other uses a quasi-resonant topology (where the pulse-repetition rate is variable) (1b).
A partial explanation for the difference in noise is the relative ease or difficulty with which each topology filters harmonics of its pulse-repetition rate or operating frequency. In PWM converters, most of the energy is found at the fixed frequency or at an odd harmonic of it. A 100-kHz PWM converter will have most of its conducted noise at 100 kHz, and some at 300 and 500 kHz. They also have significant harmonics at or above 1 to 2 MHz due to non-zero-current-switching (high di/dt). The input conducted filter has to be sized to handle maximum power at 100 kHz.
Quasi-resonant converters simplify the design of the conducted line filter because the energy that needs to be filtered is spread between 1/T2 (where T2 is the pulse repetition rate) and approximately 2 MHz. For example, if the converter is operating at its maximum frequency of 1 MHz, all of the energy is contained in a narrow band. This band is easily filtered due to its high frequency. If the converter is operating at a relatively low 100 kHz, the energy is spread between 100 kHz and 2 MHz. In the case of energy spread, for example, by a factor of 10, the peak amplitudes of the harmonics are reduced by a factor of 10.
Please refresh the page if you have trouble reading this text.
Search Electronic Design
Email Newsletter
Sponsored By:
The Find Power Products monthly newsletter brings you the most important new developments within the world of power design. The newsletter includes exerpts from industry leader Sam Davis's exclusive blog, as well as overviews of the latest new products.
Enter Email to Subscribe
Web Seminar
Sponsored By:
Title: Exploring How Good GUIs Drive Adoption in the Digital Power Management Space