I just finished modification of my second MTH light tower to 8mm 100 ma wide angle LEDs. The top pair is wired in series with a 56 ohm resistor, as is the bottom pair. They are connected in parallel to 12 vdc. I momentarily touched them to power before shrinkwrapping all the solder joints, and miraculously all lit up. Then I did the shrinkwrap , tucked away wires, and powered them again. The top two would only stay on for about 20 seconds. In another 20 seconds they would come back on, and would continue this cycle. Changed the resistor; no joy. Changed one LED; no change. The second LED solved the problem and the lights now all stay on. I hooked the problem LED to 3vdc and it would go on and off. I thought LEDs either work or not work; this one acts like it has a resettable fuse. Is there a good explanation for this?
Replies sorted oldest to newest
Maybe not a GOOD explanation, but perhaps the culprit is a blinking LED. I've used red ones as tail marker lights on passenger cars. I could be wrong, and usually am. Ask my wife...
John, not to hijack your thread since I am not helping you on your problem, but I have been thinking about doing this too. I assume this is the four light tower with bulbs. How did you secure the leds and replace the bulbs? Pictures welcome and thanks.
It does sound like a blinking led got into the mix.
hokie71 posted:John, not to hijack your thread since I am not helping you on your problem, but I have been thinking about doing this too. I assume this is the four light tower with bulbs. How did you secure the leds and replace the bulbs? Pictures welcome and thanks.
John H has some build photos here. Laidoffsick has some here.
As for the blinking, I think it's a defective LED with some kind of thermally induced bonding fault. It warms up and something breaks an internal connection. It cools down and the connection is re-established. I suppose it would be amusing to see if you can change the blink rate by changing the voltage from 2.9 to 3.0 to 3.1 or whatever. Or if you can slow or stop the blink rate by dipping the lens in a glass of ice-cold beer (or whatever)...or heat it up with a heat-gun. Or just toss it and be done with it...
Thanks Stan, sorry I missed those. Perfect!
Thanks Stan. In my (not so) vast experience I hadn't seen it before. It's way too long an interval to be a blinking LED, so your explanation makes the most sense. I did have the ice cold beer there, but it was gone by the time you posted. Good luck, Hokie71. These are the lights I used: 151187143491 from ebay.
I always have a can of this around to trouble-shoot stuff like this.
Attachments
Is that what you use to get the beer ice-cold to dunk it in?
Well, to cool the beer, I typically use the CO2 Fire Extinguisher, it works quicker.
I finally took the floodlight tower to the bench today. I changed the two leds that go out with new ones and swapped the 56 ohm resistor with a new identical one. I put 12 vdc to it and the top two go out after a couple minutes, then come back on. The bottom two stay on, wired the same. I might try a 68 ohm resistor I have on hand, but the bottom pair work with a 56 ohm like they should.
John H posted:.... Changed one LED; no change. The second LED solved the problem and the lights now all stay on. I hooked the problem LED to 3vdc and it would go on and off.
John H posted:I finally took the floodlight tower to the bench today. I changed the two leds that go out with new ones and swapped the 56 ohm resistor with a new identical one. I put 12 vdc to it and the top two go out after a couple minutes, then come back on. The bottom two stay on, wired the same. I might try a 68 ohm resistor I have on hand, but the bottom pair work with a 56 ohm like they should.
?? I thought 9 months ago you swapped out the blinking LED (the 2nd LED of the top pair) and all was well.
So are you saying in the last 9 months, the top pair have started acting up again? The first bad pair blinked after 20 seconds of power. Are you saying the new bad pair blinks after a couple of minutes?
100 mA x ~3V = 300 mW of power which one would think is OK for an 8mm dome LED package. I still think it's a thermal issue with one "bad" LED in the pair failing, and breaking the series connection so both LEDs in a pair go out and come back together. I figure it's impractical to heatsink the LED package to keep it cool, so perhaps the 68 ohms resistor will lower the power enough to keep the temperature in check.
Stan, I'm sorry. I forgot I did that. Anyway, it started going out again. I should check the rest of the package with 3 volts. I tried another resistor, then a 68 ohm, and then two 56 ohm in series. It has stayed on for over an hour with the series combo. They actually don't look any less bright than the other two that have had no problem.
I rarely try to run LED's at or close to their rated current, as for most model train applications, it really isn't needed. I'm perfectly happy with standard 20ma LED's running them at 10-15 ma in my designs.
Presumably you have a meter? Why not simply confirm the DC current flow into each pair. If too much hassle cutting wires or what-not, then with circuit powered measure DC voltage across the resistor and "do the math". Current = Voltage / Resistance.
While at it, if convenient I'd measure the DC forward-voltage (presumably just over 3V) across each LED in the pair.
I need to find some wire penetrating test leads to do that, as I have everything sealed after it worked this last time. I'll see if I can duplicate it on a breadboard and try to measure there. That is if I can figure out how to use a breadboard. I measured ~8 v after the first led, ~4 v after the second, and 0 after the resistor before I sealed the wires. My Fluke doesn't do current. I didn't actually know what it meant; I was just trying to measure something. When the lights went out, there was no voltage after the first led. I had 11.7 v going in.
If it's all bundled up I wouldn't bother. I'm convinced it was a thermal issue solved by increasing the resistor. I figure all this info might help other guys contemplating a tower bulb-to-LED conversion...and the pitfalls of dutifully plugging numbers into "LED calculators".
So to be clear, you ended up using a 12V DC source, and two 56 Ohm resistors (effectively 112 Ohms) for each pair of 8mm LEDs? As we all know, even eBay listing numbers go stale after a while. So I pasted together this from your number above - should provide enough info a year from now; it appears the seller is even in the US.
Attachments
BTW, I think you can change the title of a thread - perhaps update the title to something like:
Odd LED problem - MTH tower light conversion...or whatever. All this for the greater good that is.
I'm in on the all for the greater good part; I like good. Just for the record, the bottom pair has worked from the getgo with only one 56 ohm resistor, as has the first tower I converted the same way.