Someone please explain this to me

[funkylab@mastodon.social]

@azonenberg does it do the same when you remove the three other bulbs?

[azonenberg@ioc.exchange]

@funkylab haven't tried, already replaced this one (there was another 100% dead bulb in the fixture I replaced before filming this)

[lasse@social.tchncs.de]

@azonenberg bulb is weeping because the white balance is off

[cliffle@hachyderm.io]

@azonenberg I have a couple of dimmer switches that appear to not turn all the way off, presumably because they don't have neutral and must continue passing some current to satisfy their own needs. This usually manifests as one LED bulb in a fixture flickering and staying alive long past the others, though the flicker pattern I get is different.

Is the switch a dimmer?

[funkylab@mastodon.social]

@azonenberg ah cool, yeah clearly the other three bleed their offline caps nice and smoothly through the LEDs after turning off probably when detecting an undervoltage, whereas the haunted one seems to have a bad day when it comes to that… hence me wondering whether the input filtering on the other three might be causing spurs in the spookybulb's input detection.

[azonenberg@ioc.exchange]

@cliffle nope no dimmers anywhere in the house

[tezoatlipoca@mas.to]

@azonenberg incandescent bulbs are, electrically simple. E flows from the knob to the can of the bulb stem thru the filament. Switch go off, no current, no light.

Led bulbs have little driver boards in them that amongst other things converts 60 or 50 Hz ac into a low voltage dc. So there are capacitors in there. Some capacitors - in conjunction with a bad board design (leakage current etc) - don't discharge properly and still hold excess power for a few seconds.

[azonenberg@ioc.exchange]

@lasse lol this is the only fixture in the house that takes screw base bulbs anymore so I'm just using whatever I have lying around. It's an original 1970s bathroom with a leaky shower we need to rip our and replace anyway so aesthetics of lighting are not high on my radar

[cliffle@hachyderm.io]

@azonenberg Neat. Kinda wanna put a scope on that circuit then.

[azonenberg@ioc.exchange]

@tezoatlipoca yeah but usually I see a nice smooth ramp down. Also the bulb is running when power is on... Usually.

This one actually was dead, i only looked at it because i saw a dark spot on the fixture with no light coming from this position.

Somehow shutting it off creates an inductive spike or something in the SMPS that makes it start running until it's drained the capacitors.

[marshray@infosec.exchange]

@azonenberg Looks like you got a diesel bulb mixed in with the rest.

[johnjburnsiii@kzoo.to]

@azonenberg

All those electrons running around trying to stay alive... collecting together in their last refuge until they slowly & silently illum their last...

[petrillic@hachyderm.io]

@cliffle @azonenberg now imagining a van full of hardware nerds showing up and unpacking 6 crates full of test gear to identify the problem.

[ldcd@social.treehouse.systems]

@azonenberg if they're designed to be dimmer switch compatible maybe a cap popped off or otherwise failed in the duty cycle detection circuit and it went unstable?

[ldcd@social.treehouse.systems]

@azonenberg I assume the way those work is by integrating the rectified voltage or something similar

[azonenberg@ioc.exchange]

@gme all four bulbs are in parallel and the other 3 turn off fine. It's definitely discharging stored energy in a capacitor

[mcr314@todon.nl]

I guess that you are noticing the one bulb that stays on for awhile: @azonenberg Might be like: https://www.youtube.com/watch?v=1uEmX5XClPY the coil of wire formed a capacitor.

[azonenberg@ioc.exchange]

@gme the whole circuit is switched by a single 120V SPST switch controlling all four bulbs wired in parallel. There is to my knowledge no other switch.

And the capacitors are inside the LED bulbs as part of the SMPS circuit.