Multiple LED Wiring: Thought If You Lost One You Lost Both With This Method

[tehachapifan]

OK, guess I lost the LED on one side of the numberboard on one of my SP SW1500 bashes. The other side still works, however. I wired these with "lead A" of LED #1 going to "lead B" of LED #2, which I would call daisy-chained but I think that's not the correct term as I recall. Nonetheless, I thought if one burned out using this method, both LEDs would go out. Apparently, that's not the case? At any rate, this unit was boxed-up for several weeks and was just put back on the layout. Odd how these sort of things always seem to happen to units that have been stored for a while. Also odd is that these were also wired to a 68K resistor AND the output power was turned way down via CV settings, which seems like should really keep them from burning out. Obviously, something with my understanding with this is not correct.

[C855B]

The LED that is "burned out" is apparently shorted, so it's connecting through to the other side. When an incandescent bulb goes bad it breaks the connection, but LEDs can fail either "open" or "shorted", like any other semiconductor.

[tehachapifan]

Ok, thanks. Yeah, that makes sense. Odd that this suddenly happens after several months of trouble-free operation, where the 0402 sized SMD LED's have been secured in a mounting hole and held in place with either canopy glue or Krystal Klear away from any other components, etc. I'm guessing this failure could have been an internal component failure that is just random? Or, did the LED develop corrosion that lead to a short? Whatever the cause, this may never be repaired as I installed these early in the loco assembly phase, with the daisy-chain soldering done before they were installed (the leads are very short between the two LED's....the idea was to minimize the amount of wires needing to be tucked away when the shell was seated on the drive). I would have to pull both LED's to repair the one, as I won't attempt to do any soldering inside the shell (cab) at all. May just have to leave it alone.

...I believe this marks my first actual normal operation LED failure, not counting wiring errors, user-induced shorts, etc., which is pretty good in may respects but I guess also shows these things won't necessarily last forever.

[peteski]

The other possibility is that the wiring or the solder connections at that dark LED shorted together, thus bypassing the LED.  Like C855B mentioned, most common LED failure is an open circuit, but internal shots can also occur, just less frequently. That is why, whenever possible, I like to hook multiple LEDs in parallel.

Remember - daisy-chain in electrical terms is series-connection. Just liek the old-fashion Christmas tree lights. Since the lights/LEDs are in-series with each other. Or like a Television series where one episode follows the other in-series. 

[jagged ben]

I'm sure that with SP you can find a proto photo of an SW1500 with only one working numberboard light.

[tehachapifan]

I'm sure that with SP you can find a proto photo of an SW1500 with only one working numberboard light.

Yeah, trying to convince myself of that too. We'll see if it keeps bothering me. My guess is I will attempt an LED replacement at some point.

[tehachapifan]

Well, the LED decided to light up again tonight for whatever reason. Guess whatever is going on is an intermittent thing.