The field of power conversion is going through a transformation akin to that produced by the first bipolar transistors; only today, it’s a combination of semiconductor and packaging technology that’s driving the field to high-density integration. With the consumer market demanding ever-smaller power designs, semiconductor companies are delivering integrated modules. These modules can provide an entire power subsystem within a single high-power, multichip package.
There are several reasons why this type of integration has not been easily achieved in the past. First, the power losses in switching devices were high, so integrating several devices in one package posed a difficult thermal challenge. But over time, power semiconductors have improved significantly. Their advanced power efficiency allows packaging of several power devices in a single module without violating the maximum die temperature limits, which are usually around 150°C to 175°C. The new semiconductor devices also make possible the price versus the performance needed for the market to seriously consider the modular approach.
The second barrier to modular integration has always been the conceived risk taken by the design engineer when using a new technology. Designers always ask, “Is this new technology mature enough to be as reliable as the discrete solution?” Recent advances in high-power packaging have paved the way toward this — up until recently — elusive goal of integrating several power devices, drivers and control ICs with no sacrifice in reliability or performance.
This approach offers a wide range of advantages to the power designer to be able to produce designs not achievable using discrete components alone. First and foremost is that the semiconductor companies making these modules now have the chance — maybe for the first time — to design a set of power devices, drivers and control ICs as a group such that they offer the best performance possible for a given application.
Designing power devices to give the best performance possible for the application allows for the optimization of die size and, hence, price. Matching these power devices with the most suitable drivers then allows for extracting the last drop of performance from the driver IC-power device combination. That optimization leads to further reduction of the subsystem price without any sacrifice to performance, but rather, in many cases, with improved performance. Lastly, the control ICs are designed to take advantage of this optimized power train.
The modular approach means that the system designer does not have to spend precious time looking for the best MOSFETs from various manufacturers, the best drivers for those specific MOSFETs — a daunting task in and of itself — and, ultimately, the control IC to tie all these components together in a functioning subsystem. The module manufacturer has done all the selection and optimization. The designer now has to select the module, add a few inductors and/or transformers and capacitors to produce a functioning power system that, when done correctly, cannot be replicated in discrete solutions. Other advantages are smaller pc-board footprint, fewer components in inventory and a much faster time to market.
The complexity and sophistication of some of the integrated power modules available today from several companies is a very good indication of how impressive this approach is. Several large consumer-equipment manufacturers all over the world are adopting this approach, making this field a very exciting one for both the semiconductor subsystem designer and the end system designers.