I’m ‘prototyping’ a mains-powered bug zapper, based on a Cockcroft-Walton multiplier, and would be interested to hear opinions on how to make it safe and effective. From the start, I think the earth/ground reference should be removed with a small isolating transformer. Do commercial units have these? I can’t find a schematic. Surely a safety-rated device can’t rely totally on physically preventing the user from touching live parts. The multiplier is set up as below, with the output being taken between the left-hand node (ground symbol) and Vo. I’ve ensured that the neutral is connected at the bottom, but an isolating transformer would make that moot. Vi is 240V ac; Cs are 1000V, 150 nF; Ds are 1N4007. I can’t measure the output voltage directly, but adding up that across each cap suggests 1440V. This chimes with the theoretical value of 2N(Vp), N being the number of stages (2 here) and Vp the peak input voltage of 330V or so. Short circuit current is around 10 mA. For the grid, I cut an oven shelf in half across the bars, then mounted and interleaved the two halves in wood blocks drilled for the purpose. A small, convenient fluorescent lamp (not UV) is lit behind the grid. My high voltage tester shows flashover with this set-up occurs at 3500V, with almost no leakage beforehand, so we’re well inside that. Questions: 1. Voltage suitable? 2. Current suitable? 3. Safety recommendations? I will pot the multiplier and live connections, and enclose the grid, of course. A very small fuse will protect the device. My main worry is a failed-short component, especially a diode. Both caps and diodes are, I think, well rated. (Each cap, except C1, will be charged to 2Vp, or about 660V). 4. What discharge resistor would be best to ensure the unit drains properly when unplugged? The total capacitance (1/C=1/C1+1/C2...) is 37.5 nF. With a 1.5M resistor, the RC constant will be 0.06 seconds. But at 1440 V, the leakage will be about 1 mA and the power dissipation 1.5 W. I think a higher value would be better.