Random LED Failures

Ask and answer questions, share tips and resources for installing lighting and other electronics in your models.

Moderators: Sparky, Moderators

Post Reply
User avatar
Disillusionist
Posts: 1134
Joined: Wed Apr 09, 2003 10:11 pm
Location: Closer than you think

Random LED Failures

Post by Disillusionist »

Just wondering if anyone else has experienced this. I've had several random LED failures during my current project (PL refit). The first couple of times it was easy enough to repair. This last time is going to require the building of a new nacelle.

I know my resistor values are correct. I've even measured my currents with an ammeter, and everything checks out good. I've given all of the lighting an overnight burn-in before sealing up the sub-assemblies. However, every so often when I power things up for a test, one random LED doesn't come on. When I disassemble to check things out, the wiring always looks good. No bad solder connections or anything of that nature. Just a defective LED. I'm using a high quality bench-top power supply, so I don't think that's the problem either. I'm really at a loss...

Usually, when this sort of thing happens, I tend to think that it's something I'm doing. But I'm not so sure this time. I'm starting to wonder now if I've gotten hold of some sub-standard parts from an online vendor. But how does one know? How do you know if you're receiving high quality parts, or the bottom of the barrel cheapies? If I'm going to be sealing these up inside of a model, I'd like some reassurance that I'm getting the "top-binned" parts. Every vendor claims to sell the best, but how do you really know?

Any suggestions?
User avatar
Kylwell
Moderator
Posts: 29650
Joined: Sun Nov 02, 2003 9:25 pm
Location: Lakewood, CO
Contact:

Post by Kylwell »

Do you check the LEDs separately before assembly?
Abolish Alliteration
User avatar
Disillusionist
Posts: 1134
Joined: Wed Apr 09, 2003 10:11 pm
Location: Closer than you think

Post by Disillusionist »

Yeah, everything is checked individually.
User avatar
Kylwell
Moderator
Posts: 29650
Joined: Sun Nov 02, 2003 9:25 pm
Location: Lakewood, CO
Contact:

Post by Kylwell »

er...power supply producing enough amperage?
Abolish Alliteration
User avatar
Disillusionist
Posts: 1134
Joined: Wed Apr 09, 2003 10:11 pm
Location: Closer than you think

Post by Disillusionist »

:D It's a 30 amp power supply. It should be able to supply all the current these little circuits can suck. I did pull out the o-scope, just in case, and checked it for noise or excessive ripple.

I think I can safely rule it out.
jwrjr

Post by jwrjr »

A. some led suppliers are more reliable than others. B. check the specs of the power supply. Switching power supplies frequently do not have good regulation if the current being drawn from them is less than some minimum value - typically 10% of max.
User avatar
Disillusionist
Posts: 1134
Joined: Wed Apr 09, 2003 10:11 pm
Location: Closer than you think

Post by Disillusionist »

jwrjr wrote:A. some led suppliers are more reliable than others. B. check the specs of the power supply. Switching power supplies frequently do not have good regulation if the current being drawn from them is less than some minimum value - typically 10% of max.
Can you name a good supplier? I believe the LEDs that are failing were ordered from SuperbrightLEDs a couple of years ago. I've seen both good and bad comments about them, so I'm not sure what to think.

The power supply is linear, not switch-mode.
jwrjr

Post by jwrjr »

I have heard that the chinese led suppliers that are found on Ebay sometimes have higher than average failure rates. So the obvious answer would be to run all led+resistors a while before closing up a model. (My refit has over 200 leds, and I had to replace a few.)
As for the power supply, I did jump to a conclusion, there. But on the other hand one doesn't find many linear 30 amp power supplies. They would be rather heavy. Most are switching supplies tp reduce size and heat generation.
User avatar
Disillusionist
Posts: 1134
Joined: Wed Apr 09, 2003 10:11 pm
Location: Closer than you think

Post by Disillusionist »

Yeah, it's a heavy brute. It won't power the completed model, but I use it on the bench for all of my testing.

I've noticed that most of the online vendors don't seem to disclose the manufacturer of the LEDs they're selling. So, who knows where they're actually coming from. I would have thought that burning everything in overnight would have weeded out the bad ones, but apparently I was wrong. Oh well, live and learn.
User avatar
Pat Amaral
Posts: 3730
Joined: Fri Jul 12, 2002 2:44 pm
Location: Ok, I'm here. You can start now

Post by Pat Amaral »

The problem could actually be much simpler. If you spend too much time with the soldering iron actually in contact with the LEDs' leads, you run the risk of causing heat damage which can shorten the little guys' lives. An LED is actually a tiny junction between two chemicals which are usually very sensitive to heat. It doesn't take much. That's one of the reasons I use wire wrapping.

If you have to solder, you shouldn't need a soldering iron that's any hotter than say, 30 amps. A clean, sharp, well tinned tip will conduct heat best. You should always use a heat sink to wick the heat AWAY from the device. Using flux will also make things go a little quicker. The iron should only be in contact with the lead for a second.
Pat A.
=============================
50% Nerd, 50% Geek, 100% Cool
macfrank
Posts: 8726
Joined: Fri Oct 11, 2002 6:55 pm
Location: Los Angeles, CA
Contact:

Post by macfrank »

Besides what Pat said, which was my first guess, are you using white LEDs? Some white LEDs (and blue as well) are static sensitive. If it's very dry, or your soldering iron isn't grounded it may be damaging some of the LEDs. Enough so that after just a little use, the junction fails.
User avatar
Disillusionist
Posts: 1134
Joined: Wed Apr 09, 2003 10:11 pm
Location: Closer than you think

Post by Disillusionist »

Ok, I had an interesting development as I rebuilt my port nacelle. So here goes.

There are two purple LEDs wired in series for the warp drive, with 17 mA flowing through the circuit. I let it run overnight for the past two nights with no problems. After completing all the wiring in the new nacelle, I switched on to make sure all was well before sealing it up. Lo and Behold, one of the new LED's was dark. Then as I sat there contemplating what to do next (cry scream, stomp my feet), it came on! I know this can't be a bad connection because if that were the case, both LEDs would be out since they're in series. I'm pretty confident I didn't ESD, or overheat the parts. I'm using a well grounded MetCal soldering station with a low temp tip, so I know it isn't the cause of the problem. Plus, I used alligator clips for heat sinks while I soldered.

So now I'm pretty sure I've got some bad LEDs. I'm not sure why it would fail intermittent, but that appears to be the case. What irks me the most is I'm not sure I can trust the rest of the LEDs in the model to function long term. Lesson learned. For now on I'll be buying from a different supplier with known brand name parts. It does seem all LEDs are not created equal.
Post Reply