Help! LED modules with certain strands going dim??

Thread Starter

Chikan

Joined Jan 18, 2008
24
Hey guys! Been a while since I posted, but thought I'd check here first since you guys are the pros.. :p

I've got a bunch of LED modules taken from our vehicles at work. My boss said that certain sections of each one were 'burned out'.. (Seems they were going out in groups of 3 or 4.. sounded like maybe a bad solder joint to me)

Brought them home and started taking them apart, and was surprised to find (when I tested them) that the entire thing IS lighting up, but certain sections are very dim.. Seem to be getting only partial power. Keep in mind they were all working fine at first, this just started happening after they've been in service for a few months. The pictures kinda suck, so I'm not sure if you'll be able to see the circuit layout well enough to figure it out, but I've included pics of the front/back of the board. The section that's dim is colored in red.

If anybody could give me a tip or tell me what to do I'd appreciate it greatly!




EDIT: Oh and forgot to mention. I've got 4 of these modules here. 3 are identical like the one pictured, the 4th is a longer version.. (Instead of 4x5 LEDs, it's 4x11) Each one has a DIFFERENT group that has gone dim.
 
Last edited:

beenthere

Joined Apr 20, 2004
15,819
How are you with a soldering iron? If nothing looks like it has darkened from heat, then the problem is likely to be poor solder joints. The technique is to use fine solder (the old lead kind is best) and heat each solder joint until you can deposit a small amount of fresh solder on it.

For the record, 'fine' is solder that is less than .030" diameter. The soldering iron should be 30 watts or less, and the tip no larger than the solder pads on the PCB.
 

Thread Starter

Chikan

Joined Jan 18, 2008
24
How are you with a soldering iron? If nothing looks like it has darkened from heat, then the problem is likely to be poor solder joints. The technique is to use fine solder (the old lead kind is best) and heat each solder joint until you can deposit a small amount of fresh solder on it.

For the record, 'fine' is solder that is less than .030" diameter. The soldering iron should be 30 watts or less, and the tip no larger than the solder pads on the PCB.
Meh.. I've got an iron and some fine solder... And I know HOW to do the work.. Unfortunately I don't have surgeons' hands.. :D I don't know how well they show up in those pics, but there are 3 TINY components labeled D3, D6, and D7 on the board.. They look like resistors, but they're absolutely tiny... They appear to be kindof a copper color, but one end is black on each one.. The thing is they're so damn tiny I can't tell if they're supposed to be that way or if they're charred.. LOL
 

SgtWookie

Joined Jul 17, 2007
22,230
Here's a guess for you:
The LED boards have an onboard current or voltage regulator, and they also have series current limiters in the form of resistors.

In order to get the product to market more cheaply, they used resistors that were marginally rated for the power requirements. After they have been in use for a while, the resistors change in value due to the excessive heat, and the string associated with the resistor that has changed the most becomes noticeably dimmer.

Another possibility is that a resistor has become so hot that it has caused a "cold solder joint", or resistive joint. Re-flowing the joint may provide a temporary fix. Replacing the resistors with higher wattage rating would provide a much more permanent fix.
 

SgtWookie

Joined Jul 17, 2007
22,230
I don't know how well they show up in those pics, but there are 3 TINY components labeled D3, D6, and D7 on the board.. They look like resistors, but they're absolutely tiny... They appear to be kindof a copper color, but one end is black on each one.. The thing is they're so damn tiny I can't tell if they're supposed to be that way or if they're charred.. LOL
Those are most likely 1N914 or 1N4148 diodes. They could be Zener diodes, too - you would need a magnifier like a loupe or microscope to read the numbers on them. The black end is the cathode. The "D" in the reference designator is a giveaway that they are diodes.

Zener diodes usually have a reference designator beginning with "Z" or "ZD", but I've occasionally seen them inappropriately referenced as "D".
 

SgtWookie

Joined Jul 17, 2007
22,230
It would help a great deal if you could take better photographs.

You need to be just a bit further away (so that the camera will focus) and have a lot more lighting - diffused lighting is greatly preferred. Harsh direct lighting would not be very helpful; too much glare.
 

Thread Starter

Chikan

Joined Jan 18, 2008
24
Heya Sgt! LTNS!

Here's a guess for you:
The LED boards have an onboard current or voltage regulator, and they also have series current limiters in the form of resistors.

In order to get the product to market more cheaply, they used resistors that were marginally rated for the power requirements. After they have been in use for a while, the resistors change in value due to the excessive heat, and the string associated with the resistor that has changed the most becomes noticeably dimmer.

Another possibility is that a resistor has become so hot that it has caused a "cold solder joint", or resistive joint. Re-flowing the joint may provide a temporary fix. Replacing the resistors with higher wattage rating would provide a much more permanent fix.
Ok that's kinda along the lines I was thinking, that either some solder joints were bad, or maybe a resistor had gone out. The thing is, their layout is so screwy I can't tell for sure which resistor is associated with that string of LEDs.. I can say that straight off the incoming + line is a 1Kohm resistor, (Brown/Black/Red/Gold).. And each strand of LEDs appears to have a 6.2ohm resistor (Blue/Red/Gold/Gold) nearby.. The thing is these 6.2's are scattered around and seem to connect to more than one group at a time, or may even daisy-chain their way around the board.. Like I said, it's all screwy.. LOL

Those are most likely 1N914 or 1N4148 diodes. They could be Zener diodes, too - you would need a magnifier like a loupe or microscope to read the numbers on them. The black end is the cathode. The "D" in the reference designator is a giveaway that they are diodes.

Zener diodes usually have a reference designator beginning with "Z" or "ZD", but I've occasionally seen them inappropriately referenced as "D".
Ok cool, so those are probably fine then. Now I know. :D
 

Thread Starter

Chikan

Joined Jan 18, 2008
24
It would help a great deal if you could take better photographs.

You need to be just a bit further away (so that the camera will focus) and have a lot more lighting - diffused lighting is greatly preferred. Harsh direct lighting would not be very helpful; too much glare.
Let's see if this is any better.. As I mentioned, I don't have surgeon's hands.. LOL (Again, the dim group is marked in red)


 

SgtWookie

Joined Jul 17, 2007
22,230
So, what do those five LEDs have in common that the others don't?

That's what you need to find out.

Those new photos aren't much better I'm afraid. Wait until daylight. Bright overcast days are best for taking such photos. If you don't have an overcast, take the photos in the shade. Direct sunlight will create terrible glare.
 

Thread Starter

Chikan

Joined Jan 18, 2008
24
Unfortunately all I have is my phone to take pics with. Don't have a decent digicam at the moment. :(

Looking at the back of the board, I'll # them from left to right.. LED1 to LED5...

6.2ohm Resistor >> LED1 >> LED2 >> LED3 >> LED5 (the path also continues PAST it to another row of LEDs) >> LED4 >> 6.2ohm resistor

From there the path goes to Q2 and Q1 on the board. Going the other direction, that first resistor goes on to connect to another row of LEDs...

You guys think that first resistor might be the problem child?
 

SgtWookie

Joined Jul 17, 2007
22,230
Try replacing it.

One thing you might try doing is connecting the board to a variable power supply. Bring the voltage up slowly, from say 7v to 14.5v. See if that string gets brighter before it gets dimmer. If that's what happens, you may have a shorted LED.

Try powering the LEDs individually, using a supply with a current limiting resistor. The supply might be a few batteries in series.

I don't know what color your LEDs are. Red and green LEDs may be 1.8 to 2.2v. Yellows somewhat higher. Blues and whites are usually 3.4 to 4v.
 

Thread Starter

Chikan

Joined Jan 18, 2008
24
Oops, forgot to mention, these are amber/yellow LEDs... of the 5mm variety.. No clue on the actual ratings unfortunately..

Since the pictures aren't helping, I'm drawing up a diagram of the board to see if maybe that can help. I'm thinking it's that first resistor since that seems to be the only one that group of 5 has in common, but I'm no expert either.. lol

EDIT: Ok, Got it drawn up... Don't mind the crummy MS Paint styling.. LOL
Everything is marked to help.. The only one I'm not sure on is the resistor that's straight off of the ground.. It's a 5-band, but it's one of those crummy blue ones so it's hard to tell the colors for sure.. It LOOKS like it's Green/Brown/Black/Silver/Brown, which comes up as a 5.1... But I can't say for sure... Let me know if this helps you guys narrow down which resistor I should go for first... The other chains should be easy to deduce on the other modules that are having this problem.. lol

 
Last edited:

SgtWookie

Joined Jul 17, 2007
22,230
OK. Look in the corner OPPOSITE of Q1. In your schematic, it's in the upper left corner; look at the lower right corner.

Look at where the LED is soldered in that corner. The PCB may be cracked at that point. Try re-flowing the solder. Use isopropyl alcohol to clean it. Rosin flux if you have it.

It looks like D1 is where +12v is sourced to all of the LED strings. D1 prevents damage in case the board is connected in reverse polarity.

D3, D6 and D7 are there to "make up" for a "short string" of just two LEDs. With 20mA of current flow, an 1N4148 will drop about 0.8V across itself; therefore the three diodes will drop a total of 2.4v. Your "problem string" is not a five LED string, it's two separate strings; one with three LEDs and one with two. They are both being supplied with 12v via the trace that goes nearly 2/3 of the way around the edge of the board.

Q1 and Q2 are wired as a constant current sink or a voltage regulator (one of the two). The thing is, now you only have 5 out of 7 strings working properly. The 5 strings that are working properly are getting too much current. The one problem 3-LED string and one problem 2-LED string being the same intensity tells me that the problem should lie with that positive trace around the right edge of your schematic; as it's the only thing they have in common that would have little effect on the rest of the board.

I'm off to bed; it's late.
 
Last edited:

hgmjr

Joined Jan 28, 2005
9,027
Did you remove the attached pictures on purpose? There doesn't seem to be any attachments at the moment.

hgmjr
 

SgtWookie

Joined Jul 17, 2007
22,230
hgmjr,
I still see them - perhaps a router is down between you and his hosting site.

I'll attach the schematic, and a marked-up schematic.

The marked up schematic has red circles around the five problem LEDs (top row) and an aquamarine box around the area which I believe contains the problem.
 

Attachments

Thread Starter

Chikan

Joined Jan 18, 2008
24
OK. Look in the corner OPPOSITE of Q1. In your schematic, it's in the upper left corner; look at the lower right corner.

Look at where the LED is soldered in that corner. The PCB may be cracked at that point. Try re-flowing the solder. Use isopropyl alcohol to clean it. Rosin flux if you have it.

It looks like D1 is where +12v is sourced to all of the LED strings. D1 prevents damage in case the board is connected in reverse polarity.

D3, D6 and D7 are there to "make up" for a "short string" of just two LEDs. With 20mA of current flow, an 1N4148 will drop about 0.8V across itself; therefore the three diodes will drop a total of 2.4v. Your "problem string" is not a five LED string, it's two separate strings; one with three LEDs and one with two. They are both being supplied with 12v via the trace that goes nearly 2/3 of the way around the edge of the board.

Q1 and Q2 are wired as a constant current sink or a voltage regulator (one of the two). The thing is, now you only have 5 out of 7 strings working properly. The 5 strings that are working properly are getting too much current. The one problem 3-LED string and one problem 2-LED string being the same intensity tells me that the problem should lie with that positive trace around the right edge of your schematic; as it's the only thing they have in common that would have little effect on the rest of the board.

I'm off to bed; it's late.
Thanks for all the help here Sgt.. :D

I'm not seeing any cracks in the PCB or any bad joints in that area, but I'll try reflowing it tonight after work and see what happens. And thanks for the explanation of the layout.. I was kinda confused by that little loop of diodes and couldn't quite figure it out.. LOL Your explanation makes sense, and will help greatly when I take apart the other modules to fix them as well.. (They all have different chains of LEDs that are doing this.. But I believe the others are all vertical groups of 3, so those should be pretty easy to figure out)
 

SgtWookie

Joined Jul 17, 2007
22,230
If re-flowing doesn't work, try scratching down through the coating to the traces on the top and bottom of the board, and solder a jumper wire from that lower 12V trace to the upper one. Insulated, of course.
 

hgmjr

Joined Jan 28, 2005
9,027
hgmjr,
I still see them - perhaps a router is down between you and his hosting site.

I'll attach the schematic, and a marked-up schematic.

The marked up schematic has red circles around the five problem LEDs (top row) and an aquamarine box around the area which I believe contains the problem.
I can see it now so whatever prevented me from seeing the images has passed.

hgmjr
 

Thread Starter

Chikan

Joined Jan 18, 2008
24
Well... Now then... Yes, I'm still at it.. lol :(

Ok, reflowing did nothing. I added a jumper wire as you recommended Sgt from point A to point B (as shown in the first attachment).... Again, nothing..

So I tried the direct route, with a jumper wire from that bottom LED directly to the LEDs that start the 2 upper chains, bypassing the trace completely (attachment 2)... STILL nothing...

I'm at a loss here... I was thinking the other modules would be easier to figure out, since it's only single chains (3 vertical LEDs) that are doing this... but now I'm not sure... I'm thinking I may just tell my boss sorry, but these are junk and you're just gonna have to keep buying replacement modules everytime this happens.... (Guess he shouldn't have bought the cheap imported lights just cause they were horrendously low-priced.. lol)

Any other ideas guys?
 

Attachments

Top