EXCELLENT NON POP 3PDT.

[Ben Lyman]

Cap needs to be "upstream" from the LED anode and always connected to supply - that's all that matters. The Cap acts as an extra power reserve for when the LED is turned on so that the sudden draw of current doesn't have to come from the main supply which affects the circuit. So the cap needs to be able to stay charged up when the LED is off. If it's disconnected from the supply with the path to ground through the LED intact, it will discharge through the LED (the LED would have a fade out effect).

Whether the whole shebang is between supply and switch, or the cap is between supply and switch with the LED between switch and ground - doesn't matter IMO.

Excellent description, thanks Gord! I guess now I know why mine wasn't working optimally. I will have to take it apart again and move that whole doo-dad to the other side of the LED so it's "upstream" of the anode.

[GGBB]

Now my LED surges very bright for a fraction of a second when I switch it on.

I was a bit puzzled by that when I first started using the AMZ anti-pop too. Here's what I think is happening. Current flowing out of the cap to the LED is limited only by the smaller resistor, so for as long as the cap can supply current (a fraction of a second depending on cap size) the LED will draw whatever it can through that resistor alone. When the the cap is depleted, current from the main supply through the larger resistor kicks in and is far lower because of the greater resistance. At some point the system stabilizes until the LED is turned off, whence the cap becomes unloaded and charges up again. I think I'm probably missing some technical details but that is basically what's going on IMO.

So the second resistor's value is key - too low and you could pop the LED, but making it higher (with the other resistor going lower to keep the same total) moves it closer to the main supply load-wise and therefore less effective at isolating the main supply during the surge. I suppose that's the point of the variation that uses a trimmer - find the exact point where the pop is no longer audible in order to minimize the brightness surge.

Another tip I read somewhere is to connect the LED supply directly to the power source and not after any current limiting resistor in the main supply. This provides just a bit more isolation.

While the AMZ anti-pop is simple and useful, it isn't perfect. I suspect a better solution could be had by taking advantage of the slow turn-on of the Millennium 1 bypass but used "off circuit." Slightly more complex but possibly totally silent. However, different switch wiring would be required since turning the LED on requires breaking a connection to ground rather than making a connection, and I can't see a way to have both PCB input and output grounded when bypassed: