We've long assumed our second floor hall bathroom was haunted, specifically by a Poltergeist. How else to explain the shower mysteriously turning on in the middle of the night? Or the faucets that were carefully turned off, only to start dripping once your back
was turned?
OK, maybe it's the "we don't see many of those anymore" combination of steel valve housings and brass valve stems that were used, combined with rubber stop washers that have hardened over time to turn into perfect thermal-time-delay systems. Turn the valve off when hot
water's been running through, and when they cool off they open again. But, even after this advanced education in the perversity of mechanical objects, I was not prepared for the demonstration yesterday.
My daughter came out of the bathroom after taking her morning shower, and as she turned off the ceiling light.....it started flashing. Switch on, steady light; switch off, one blink every few
seconds. Huh?
Ok... diagnostics time: broken switch, or defective bulb? Switched in an incandescent bulb, all is well. Replace with another CFL, same results. I then took apart the fixture, but found nothing obviously wrong; no shorted wires, no damaged insulation. But, as I put it back together, I noticed that the ceiling was wet, the result of the muggy weather and steam from the shower. Hmmmm.......
A 100 Watt incandescent bulb draws a little under 1 amp at 120VAC. Thus, a 12 watt CFL draws something like 100ma. Still in a low-impedance happy place, as far as leakage currents and home wiring goes.
But, what's inside that bulb? I hypothesize there's a diode bridge going to a small capacitor to smooth out the current, followed by an offline switch-mode power converter, that chops the sorta-DC into 10-20 kHz, the better to allow a small inductor to serve as a current limiter and startup voltage booster. Turn on the AC, the DC voltage rises to the point that the switcher starts, which fires the lamp, and bada-bing, you're good to go. Turn off the AC, the DC voltage drops across the cap, the switcher shuts down, the light goes out.
But, what if there's some leakage current on the input, say from a ground-side light switch paralleled by the enclosure touching a damp surface? (Remember: back at the fuse box the neutral wire coming from that wall switch goes to the same ground bus that the grounding wire without a switch goes to. Input current via the leakage path is tiny, but on each cycle of the AC input some charge still transfers to the cap, the voltage of which rises until the switcher starts up, only to drain out all the juice and shut off again. Figure a duty cycle of maybe 1 msec on and seconds off, and you can imagine a few 10s of microAmps allowing the light to blink at partial brightness.
So, a warning to us all. If the next generation of ultra-efficient light sources remain incandescent compatible and are 10x as efficient as CFL, they may never turn off! After all, my home wiring is
insulated with COTTON DIPPED IN INDIA RUBBER, with wall switches that have 70 years' of flash-deposited metal spattered around the switch contacts. Once you start considering microAmp leakage currents, you'll see them everywhere. So, what to do? Refit all the wiring with teflon insulation, ceramic High-Z insulators, and leakage-current guard rings around all of the neutral line connections? The expense! But, otherwise we might die from lack of sleep, as our lights conspire against us to shine into our eyes all night long!
(There's a wee bit of exaggeration here, in case anyone is taking my latter ranting seriously. :-)