Talk:Capacitive power supply

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia


This article omits the ac mains source voltage and frequency. This is a VERY UNSAFE OMISSION. Much of the world has 220V RMS or higher mains voltage. Peak-to-peak this is 345V, which will cause a 250V rated capacitor C1 as shown to fail, perhaps to explode.

There seems to be a diagram (with a linear regulator) omitted.

Capacitive dropper is used in applications where minimizing the cost is the highest priority. Linear regulators, not so much. I would especially not expect to see a regulator in the power supply for an LED "light bulb." LEDs don't really care about regulated voltage: What an LED needs is limited (not necessarily regulated) current. 74.111.96.172 (talk) 16:12, 15 April 2022 (UTC)[reply]

It could be helpful to include a formula for calculating C1.

Given the risk to life, perhaps the opening comments about safety should be formatted in bold and red? The article implies that someone reading it might build a power supply of this type (use of "you can"). Perhaps add a note to use extreme caution including an RCD. Tonyjmcgregor (talk) 22:24, 16 February 2015 (UTC)[reply]

GFCI / RCD[edit]

Its a great safety idea if one is fooling around with such a circuit to use a GFCI protected outlet. But as far as an LED lightbulb using this circuit, assuming it is properly insulated, there is no reason it needs to be on an GFCI/RCD protected circuit. — Preceding unsigned comment added by 185.23.113.221 (talk) 16:14, 1 January 2017 (UTC)[reply]

One of the External Links contains "fundamental errors"[edit]

There are currently two links to Microchip "design notes". The second of these documents contains the following text warning of "fundamental errors" in the first (AN954):

"The equations in the Microchip Application Note AN954, Transformerless Power Supplies: Resistive and Capacitive, have fundamental errors and the reader is encouraged to use the corrected equations provided here" etc.

The errors are in fact pretty obvious. I feel strongly that the link to AN954 should be removed, and preferably replaced with a more accurate reference. I will look for one. I also recommend not adding new links to Microchip documents, since this source is apparently not reliable. LyleHoward (talk) 14:32, 28 March 2024 (UTC)[reply]