Power Module Designs Aim for Efficiency, Heat Management, and Flexibility

25 September 2009

Keith Nardone discusses Factorised Power Architecture, which breaks power conversion into flexible and scaleable power building blocks, and its advantages over other power system architectures.

As electronic systems continue the trend toward lower voltages and higher currents and as the speed of contemporary loads ― such as state-of-the-art processors and memory ― continues to increase, the power systems designer is challenged to provide small, cost effective, and efficient power system solutions that offer the requisite performance.

Even if they achieve all that, however, it may not be good enough. Suppliers can make an innovative power product, but, in the end, if it is not in a form factor or served up the way the OEM ― especially true for military or aerospace OEMs ― can easily bring it into their existing system and make use of it, then that product is not a solution for them.

Power module designs aim for efficiency, heat management, and flexibility, but as always, the application rules, and the full tool box of power design strategies is often likely to be needed. Such design strategies encompass not only those related to the product, but to the topologies and architectures as well.

One popular product strategy employed to improve efficiency is the use of synchronous rectification instead of diode rectification in a given DC-DC converter. In the case of diode rectification, the dissipated power is roughly proportional to the current through the diode. In the synchronous rectification case, which employs a MOSFET switch or switches at some additional cost and complexity, the power dissipated is roughly proportional to the square of the current. At lower currents, the MOSFET dissipates less heat than the diode. Another product strategy employs the planar face of a baseplate for removing or transferring heat and potting to provide an outstanding thermal interface around each component.

Historically, a variety of power system architectures have been and are still being used. The classic Centralised Power Architecture (CPA), which is simple and cost effective, continues to be applied wherever it is appropriate. CPA contains the entire power supply in one housing. It converts the line voltage to the number of DC voltages needed in the system and buses each to the appropriate load. Thermal management can be a special challenge with centralised architecture because the heat is all in one concentrated area. Large heat sinks and fans are often needed to prevent the power supply form overheating. System hot spots are a source of reduced reliability. CPA is inherently inflexible.

The high-density DC-DC converter was the enabling technology for distributed power architecture, which, in turn, enabled the busing of higher voltages and lower currents, to be converted at the load to a lower voltage at higher currents. This approach improved overall system efficiency by minimising I2R losses and overall system thermal management, because the power converters were spread throughout the system. Large conductors carrying lots of current back to the power source were eliminated, which eliminated noise and crosstalk potential. DC-DC converters, of course, provide isolation from the input to the output, transformation of the voltage, and regulation. As the power environment changed, with many systems requiring many different voltages, conventional DC-DC converters ― saddled with all the functions of isolation, transformation, and regulation ― developed a cost disadvantage.

Intermediate bus architecture, like the distributed bus, uses a front-end box that converts the incoming AC to a single bus voltage. However, instead of being fed directly to the Point-Of-Load (POL) converters, this bus voltage is converted to a lower, unregulated intermediate bus voltage, which then goes to non-isolated and relatively inexpensive POL converters. The intermediate bus architecture, however, involves another power processing stage, providing additional conversion efficiency losses.

Power module designs must be efficient to begin with and sufficiently flexible to manage the generated heat effectively. These dual requirements are met by Factorised Power Architecture (FPA), which breaks power conversion into flexible and scaleable power building blocks. One is a current multiplier chip that provides transformation and isolation. Another chip provides a regulated non-isolated output voltage – a ‘factorised bus’ – from an unregulated input source. These chips can typically exceed 96% efficiency depending upon input and output voltages.

An even more flexible way of using these chips is packaging them in small brick-like formats with flange-base plates, aluminium case, making the removal of heat even more simple. The chips themselves are surface-mountable, but the new packaging can be typical two-hole mount that’s compatible with lead-free, RoHS compliant soldering processes. It’s another tool in the designer’s kit to give them the answers they need for their own systems. These components can be mounted, can be attached to cold plates, very rugged type of mounting schemes with standoffs for withstanding severe environments, vibration.

Depending on the requirements for voltage regulation, load current, system cost, and other factors, the two brick modules facilitate a range of design configurations that include multiple outputs, high power arrays, high-current/low voltage, high voltage, and separation of regulation and transformation for optimal board space utilisation and thermal management. One brick, the Voltage Transformation Module (VTM) is a current multiplier and provides voltage transformation and isolation, and the other, the regulator module (PRM), provides regulation.

The PRM, which can be collocated with the VTM or be located a distance away from the VTM, provides regulation. In a variation of this simplest configuration, the output voltage of the VTM can be controlled with a choice of methods. The local-loop control method, connected to A, regulates the Factorised Bus voltage. The adaptive-loop control method, connected to B, improves regulation to within 1%. The remote-loop control method, connected to C, improves regulation to within 0.2%.

It’s also possible to operate one of the VTMs closed loop, with the PRM for the tightest voltage regulation at that load. The voltage regulation of the other VTMs will follow that of the VTM operating closed loop. PRMs can be paralleled to create high-power arrays, such as multi-kilowatt power systems.

Keith Nardone is the Director of Business Development, Aerospace & Defense at Vicor Corporation.

Contact Details and Archive...

Print this page | E-mail this page