Best Guess for Cause of flickering LED?

[tehachapifan]

I have an SMD LED permanently installed in the shell of a loco that begins to flicker terribly after a short period of illumination. If the light has not been turned on in quite some time, I might be able to get a couple minutes of steady operation out of it but, once things are I guess warmed up, it will only be a few seconds before it starts. What's you best guess as to the culprit? Does it sound like a problem the LED itself, an in-line resistor failure or perhaps the decoder itself? Again, heat seems to be a factor. This is a standard SMD LED I use in all my lighting installs and usually have excellent results. I can't find any bad wiring connections or potential shorting points. I'm actually hoping it's a decoder failure as the LED, which I had to permanently install in the shell, cannot be replaced. I should mention that I did accidentally rake the end of a tiny screwdriver accross the decoder at one point during installation , after which I could not spot any obvious damage. It seems like I asked a similar question quite some time ago but can't spot anything doing a quick search.

[peteski]

I agree that it is most likely the LED.  While those "solid-state" LEDs are supposed to  last forever (ora  very long time), sometimes they do fail, even under normal operation (not overloaded).  I have observed flickering LEDs in full size vehicle CHMSL units (the 3rd center-mounted stop lights). Those are often made up from strings of LEDs, and I observed that one or more LEDs flicker.

As I see it, it is very unlikely that it is the decoder and even less likely, the resistor. Easy way to test this would be to wire a test LED and a resistor to the output of the decoder (in parallel with the original LED/resistor).

Such problems are exactly why I always try to make my installs easy to service. It makes replacing failed components easy.  But I realize that in some instances this just is not possible.  That is why I'm suggesting a parallel-installed test LED hoked up to the decoder.  That way you will know if it is the decoder or the LED/resistor (again, unlikely it is the resistor, so LED is likely the problem).

[tehachapifan]

Being the LED itself is what I'm afraid of, but you're probably right. Of course it would be the permantly-installed LED. I typically install LED's in no more than a semi-permanent fashion but this install required a more permanent one. It's the LED for the gyralight atop the light package sandwich board in my SP SW1200 seen here...



To replace the LED, I would have to try to cut a hole in the rear of the board which I'm not sure I would want to attempt.

Holding on hope that there's another cause, I will try the test LEDs as you suggest.

Thanks!

[GM50 4164]

I hope you can get this figured out Russ. I would like to see it in action!

[tehachapifan]

I hope you can get this figured out Russ. I would like to see it in action!

Me too! I did get it t work OK for a while by playing with the lighting effect CV's, including changing it from a gyralight effect to Mars as the Mars effect spends less time at max brightness (the theory being a heat related issue with the LED itself or some other component where max brightness seems to be a contributing factor). Doing so, I was able to get it to work for a few minutes before the flickering would start but now its degraded back to flickering after only a couple seconds of operation. I have another decoder on order that I will try out in hopes that I did, indeed, damage a component with the accidental screwdriver slip/rake I mentioned above. I'm sort of hoping that I can will it to be something other than the LED causing the problem, but the reality is it probably is the LED....which would be extremely difficult to switch out at this point. I would have to try to carefully cut out the back of the light cluster housing pictured above to do it.

[Steveruger45]

I really really like the SP number boards and light clusters, Russ.  I too think the issue is the led itself or one of the wires soldered to it inside the cluster.  Hope you can get it fixed without too much pain, it’s too nice an install not to.

[tehachapifan]

Thanks, Steve! Yeah, I'm sure you and Peteski are right on the cause. Of course the back of the cluster is currently perfectly smooth and blemish free and was also adequately prepped so that there is no light bleed-thru at all back there. Going to really be hard to tear into it, also worrying about breaking loose the whole light sandwich board in the process (imagine how much of a pain it was to get that thing mounted straight and centered in the first place! ). For these reasons, I may actually elect to just leave it be. The gyralight works fine in my other SW1200 and maybe I'll just leave them lashed up with this unit doing most of the trailing. We'll see I guess.

[Steveruger45]

Yeah. I see your point and operating in consist solution.
 I’m thinking, if it was me (this glitch would bother me no end) I would consider making a complete replacement light cluster first, fully test it afterward and if all ok then swap the doggy unit out complete for the replacement. 
You can still operate it in the meantime so no loco in the shop for any long length of time.
Just a thought.

[tehachapifan]

Good thought indeed, Steve. I do think this will be something that will drive me nuts.

Tonight I went ahead and really tried to figure out the cause of the problem and it is, indeed and unfortunately, the LED itself. As a last ditch effort to find a way to get it working again, I changed the resistor from a 1K to a 2.2K in hopes that it would keep the LED cool enough. This worked like a charm....for a quite a while actually and then the flickering began to creep back. It's not (currently) occurring as quickly as before but I fear that the problem with this LED is a progressive and degenerative one, rather than something that I can try to keep below its fail threshold. Tearing apart the light housing may be my only remaining option.

[Steveruger45]

Glad you found the cause and I like your solution orientation technique vis-a-vis putting a larger resistor in. Brilliant idea. Shame it didn’t work though in this case.
Just keep saying “model railroading is fun”, over and over while you fix it.
I find it helps. 

[jagged ben]

An SP crud with a malfunctioning gyralight wouldn't be too unprototypical, would it?

[tehachapifan]

Glad you found the cause and I like your solution orientation technique vis-a-vis putting a larger resistor in. Brilliant idea. Shame it didn’t work though in this case.
Just keep saying “model railroading is fun”, over and over while you fix it.
I find it helps. 

Lol! Yeah, sometimes I gotta remind myself this is fun (same with golf!) At any rate, the larger resistor seems to be doing the trick, at the moment anyway, although I'm only activating the gyralight periodically as to not push my luck.

An SP crud with a malfunctioning gyralight wouldn't be too unprototypical, would it?

Ha! I actually thought about that too!