The advent of integrated isolated dc-to-dc converters made many of these considerations more easily solvable by providing a compact, easy to use solution with documented safety certifications. Consider the scenario of a new project being approved, in which a previous design will be upgraded to higher performance metrics with additional features. The team members are immediately energized and are ready to dive into the work. But the technical project leader has to worry about all the different things that might go wrong and manage the increasing complexity under tighter budgetary and schedule constraints.
Among those project management challenges is meeting increasingly demanding electromagnetic-compatibility (EMC) requirements. More and more emerging applications and markets require compliance with numerous EMC specifications, and the bar continues to rise with more strenuous performance limits.
Existing discrete solutions, such as isolated flyback converters, have some benefits, including a low bill of material (BOM) cost, but there are drawbacks. A typical flyback design (Figure 1) contains a controller driving an isolation transformer, with rectification and filtering on the secondary, and an optically isolated feedback network. The error amplifier requires engineering effort to develop a compensation network to stabilize the voltage loop, and is dependent on optocoupler performance variability. Often seen as an inexpensive isolator for use in a power supply, an optocoupler has a variation in current transfer ratio (CTR) that will limit the voltage feedback performance and the effective operating temperature range. The CTR parameter is defined as the ratio of output transistor current to input LED current and is nonlinear, with substantial unit-to-unit variability. Optocouplers typically have a 2:1 uncertainty in initial CTR, which can degrade up to 50% after years of use in high temperature environments, such as those found in high power and high density power supplies. For a project manager, the discrete flyback approach seems better from a cost point of view, but there’s a trade-off in engineering effort and technical risk.