Help with 36-LED 12V light

[Josey]

I recently purchased a very expensive 36-LED light for my 12V solar-powered cabin. ($165) The light has 36 LEDs in 12 3-LED circuits. Each 3-LED circuit has its own 82-ohm resistor.

Now two of the rows (6 lights) don't work. All the resistors check out. Continuity checks out. I didn't know how to test the LEDs, but I noticed that when I measured resistance with the positive lead on the positive terminal of the LED, the LED gets a very faint light.

On the two bad ciruits, the inside LED on each doesn't light up with the resistance test done this way.

Can two LEDs go bad so fast?

Do I just have to replace the two LEDs?

On such a small circuit board, how do I keep the solder from running together and shorting out the LED?

The customer service was so bad from the company that makes these lights that it took several months to get my order because they typically sent the wrong item. So I don't want to go through that again.

Thanks --- Josey

 

[FNinjaP90]

They are white LED's right? If they are, normal white LED's look yellow when you look directly at it. When the die burns out, it's magnified by the shape of the LED and the LED will look yellow with a brown speck in it.

Also, just wondering, how bright is the light? Can you give me a comparison with any regular Surefire?

 

[Josey]

Thanks FNinjaP90:

Yes, they are white LEDs. The two I suspect are bad look bright yellow from above just like the others. No brown speck.

This light is 31 Lux compared to 44 lux from my SF L4 (both are flood), as measured from 39 inches.

Josey

 

[MrAl]

Hi there Josey,

How long did you have this light?

In the last EE Course part i've outlined a procedure for
understanding completely what can happen with parallel
LED's. This doesnt mean it always happens, but just that
it can happen.

If your circuit has 82 ohms per 3 LED's (in parallel)
then it means they intend to divide about 100ma up
three ways to the three LED's. A better way would have
been to supply one resistor per LED.

You're on the right track to test them, and you can
get maybe a little better test with a battery and
resistor.
To replace one, you have to use a fine tipped iron.

Perhaps you can post a pic too and that would help
understand how to solder your board a little better.


Take care,
Al

 

[Josey]

Thanks MrAl:

The whole circuit board (the head of the light) is just 2.125 inches in diameter; so 36 LEDs and 12 resistors don't leave a lot of room for soldering. I don't have the ability to post a photo, but if you saw the soldering you would be amazed at how bad it is.

It seems odd to me that the two bad LEDs, if I have it right, are both next to each other and both at the head of their respective series. (Am I explaining this right: 12 3-LED series circuits in parallel with each other.

I'll get back to your course to see if I can pick up some more clues. Thanks.

Josey

Edit: The light is just a couple of months old. Shouldn't these LEDs last for 100,000 hours?

 

[evan9162]

Mr Al,

Assuming a Vf of 3.6V, the 82 ohms for 3 in series would yield less than 20mA through each series string (14 mA for 12V in)

Josey,

Does your DMM have a diode test function? If so, try using it on each of the LEDs in the dead string. You should find the dead one since it won't light up.

The reason you saw dim light from them with the resistance test is that a DMM puts a small voltage across what you're testing for resistance, then measures the current through it - so the small voltage would cause the LED to light up a bit.

The diode test function works better, since it typically pushes a small current through the device (1mA?) so it should light up the LEDs better.

LEDs can go bad - what can cause them to go bad would be voltage spikes and heat.

 

[MrAl]

Hello there,

Oh ok i see now. There are multiple sets of 3 LED's in series, all connected in parallel.
That sounds better, but makes it harder to understand how
one could have blown.
Perhaps as you noticed, if the solder connections are bad
maybe one worked loose. In that case you'd only have to
maybe melt the solder again and let it flow a little
and then remove the iron.

It does sound rather hard to solder, but i guess you'll
either have to try or else send the board (alone) to me
and ill try myself if you like...you'll have to cover
postage and insurance im afraid...but no other charges :-)

If one actually blew out i'd then have to ask what
kind of power source is driving this? Is it similar
to an automobile electrical system with spikes and
surges and stuff like that? If it's solar alone,
what is the max output voltage of the panel and/or
regulator?

I guess you used the meter across each individual LED, which
was a good idea. If you use a battery and resistor you'll
see a little brighter light. It makes sense that if one
LED blew out because of current then all three could have
blown, but it's not that likely i dont think. Perhaps there
is only one LED per series string that went out. Check
again with battery and resistor, across each individual
LED. A resistor of 100 ohms and battery of 6 volts
should work well. Each LED should light up bright.


Take care,
Al

 

[Josey]

Thanks evan9162: I did use the diode tester and the LEDs did light very faintly, except for the two that I am now sure are bad.

Thanks MrAl: I'm pretty sure now that just one LED in each of the two bad circuits are bad. I'm believe that the soldered connections are not the problem because I am able to bend the bad lights to test their legs on the top of the board and they don't light up like all the good ones. And they are not shorted at the bottom of the board.

The voltage these lights see is between 12.2 and 14.5 VDC. My battery bank (8 Trojam L16s at a total of 1600 amp hours at 12V) is going bad, so the swings of late have been larger than normal.

I'm pretty dubious that I can do this fine of soldering. If you still are willing, MrAl, I'll pay all costs and make a $20 donation to CPF in your name. Just pm me with your address. I asked an electronics shop in the nearest town, but it was "not something he did."

Thanks, everyone.

Oh, one interesting thing. The second string to go bad didn't just die. It strobed (all three lights in the circuit), for a day or two.

Josey

 

[NewBie]

Happen to know the brand of this, that way we can avoid their "designs" like the plague?

 

[Doug Owen]

I'm not sure what the problem with the design might be, seems solid and conservative enough for most needs. The quality of the parts they use might be an issue, but bad parts happen. FWIW, I bet if there's two already, there's more on the way......

Another troubleshooting tip is to use your volt meter to read the voltage across the LEDs. Bad ones will have 12 Volts, more or less. No LED with a Vf so high will be of much use to you. You don't have to pull it, you can solder a good one across it and all's Jake.

Doug Owen

 

[MrAl]

Hi Josey,

Your generous offer to donate to CPF on my behalf
means i'll have to insist on paying for the return
shipping should you decide to send it to me for
repair :-) but please read on before you decide
one way or the other...

As Newbie said, we'd all like to know the company so
we can avoid them too :-)

I agree with Doug Owen 100% too, in that it seems like
a good *electrical* design. Unfortunately, i also fear
the same as he: if there are two gone, there's more on the
way.

With your new info, im afraid the *mechanical* design aspect
of the light comes into question. Now you would think that
there cant be any mechanical problems with a pc board with
LED's mounted right? Me too! But somehow mechanical
problems seem to creep their way into seemingly nice board
layouts :-( and cause all kinds of problems.

What im talking about is the hole spacing drilled for
the mounting of the LED's. If the hole spacing isnt just
right, when the LED leads are pushed into the holes
the leads either spread farther apart or get pulled in
toward each other just a tiny bit more than they should
be...which causes a fracture in the seal between the
plastic housing and the lead...which allows contaminants
over time to seep in and destroy the LED.

The other possibility is that the same thing can
happen even if the holes are drilled perfectly but
the soldering was done with too much heat or for
too long. Since you had said the soldering looked
awful, this could be what happened. It may be that
it's so hard to solder them it also becomes very hard
to avoid this kind of problem.

What worries me is if we go through all this to fix
what amounts to 16 percent of the LED's are more going
to 'burn out' next week?

One way to find out is to run the LED's for another
three months to find out if any more are going to
go bad. If so, you may wish to look into another solution.
If not, it may last forever once those two LED's are
replaced.

If you dont want to wait that long, perhaps you can perform
an accelerated moisture test...which involves running the
light in a small chamber with a humidifier for maybe a
week...to find out if more go out.

Remember also that you're only losing 16 percent of
the light right now, and that isnt too bad. If you
wait another 3 months you can be more certain that the
design isnt totally crap perhaps.

Think it over and let me know what you want to do.

Take care,
Al