Call this the digital decade for power supplies. It starts with a creeping-digital (CD) phase in which more and more system control, monitoring, and reporting are being implemented digitally. The decade will end with another CD stage, completely digital, as the core power conversion becomes fully digital. Until then, system-level issues will continue to be more difficult to implement.
Semiconductor devices are becoming more sensitive to voltage changes, while imposing more demanding transient load conditions on the power source. Methods to sequence power sources are increasingly difficult, usually attempting to compensate for the inadequacies of power-management elements.
With the incessant implementation of DPA, there are many more power-management elements to sequence, monitor, and report on. Fault reporting is no longer sufficient. Power users want advanced pre-fault reporting so that faults are preempted. More discrete faults are being added to the report list, while some are morphing to operational condition values. These features are requested at minimum additional cost! In response, power-supply engineers are turning to higher-level microcontrollers, banks of EEPROMs, and I2C serial bus networks
As power-source designers adopt completely digital control, system-level features will be much easier to implement because most elements will be in the digital world already. This will enable enhanced features, such as wearout warning, pre-failure warning, satellite control with host system interface, smart transient recovery, and early transient warning.
Increased information reporting will encompass current, voltage, temperature operating values, and load balance data. Presently, complete digital control costs about three times its analog equivalent, so it's not yet affordable. However, that cost is dropping, and complete digital control should become mainstream in the next few years.
