As the Internet of Things spreads its tentacles, the humble capacitor is set to play a starring role. The proliferation of low power, connected, microcontroller-based appliances – some of them always “on” — has spurred new interest in simple, small and economical power supplies.
Engineers are dusting off an old concept: The transformer-less capacitive power supply topology, now tweaked with innovations in circuits and components.
Conventional power supply designs aren’t well suited to such applications. Magnetics and MOSFETs are usually too large and too expensive compared to the components of the system they are powering; their conduction and switching losses are disproportionately high in relation to their outputs.
The solution? Use a “voltage dropping capacitor” in series with a resistor across the mains line; the capacitance functions as an impedance that causes a voltage drop from the main supply voltage to a voltage to be used by the application. Diodes for rectification, a smoothing capacitor to generate ripple-free DC and a Zener for voltage regulation are the other basic components. In theory, such capacitive voltage reduction makes for higher efficiencies and dramatically lower energy losses.
In practice it’s not that simple, of course. Some basic precautions have to be taken.
Since the capacitive power supply is directly connected to the mains, inrush current must be limited with a series resistance, which dissipates some power.
No transformer equals no galvanic isolation; care must therefore be taken to ensure that there’s never a direct path from input to output.
As the voltage dropping capacitor is directly connected to the power supply, very high demands are made on its reliability. It is therefore recommended that only X2 capacitors compliant with UL and ENEC be used. In fact, all passive components have to be carefully selected to ensure maximum circuit protection.