SDG #289 AC-DC Power Supply using Reactance of Capacitors

This episode tackles a practical design question: can a transformerless capacitive dropper supply run a pair of mains powered LoRa boards continuously and cheaply? Rather than guessing from datasheet figures, the boards are profiled with a bench power supply and current monitor (an EGteks mPower1203, which turned out to be a frustrating purchase because the software had to be tracked down from obscure sources and run in a virtual machine). That measurement step is the useful part for anyone designing a low power node: the receiver sits at roughly 30 mA with spikes to 36 mA, while the transmitter idles near 26 mA but jumps to around 100 mA for a full second during transmit.

The design constraint then becomes obvious. A capacitive dropper sets its current from the reactance of a film capacitor at 50 Hz, so peak demand has to come from stored energy rather than from the dropper itself. The LTspice work shows a 3.9 uF dropper feeding a bridge rectifier, bulk electrolytic and zener clamp, with a 0.47 F supercapacitor added on the transmitter to absorb the TX burst. The trade off is spelled out clearly: a bigger dropper capacitor charges the supercap faster (10 seconds instead of 80) but dumps far more current into the zener as continuous waste heat, which is exactly why zeners and series resistors are the parts that char on cheap PIR floodlight supplies.

Also covered is the reason a linear regulator cannot simply replace the zener at light load, plus the prototype PCB layout with dual film capacitor footprints and a fusible input resistor. Switch mode alternatives running directly off the mains are flagged for future episodes.