A friend got one of those garage/workshop lights that has folding wings with LED panels, and that you can screw into a standard bulb socket. It has since failed and she knows I'd rather futz with it than send it to the dump. It'd be fun to fix this but at this point I'm mostly playing around. It's not worth enough to justify any real work.

I though it might be something simple like the PSU rectifier bridge or the output caps, but now I'm stumped. Any ideas?

Here's what I found inside:

One of the eight LED panels that are in parallel and share the same power supply.
The voltage at the power pins is ~90±10VDC and moves around. Only one of the eight panels lights at all, only dimly, and happens to have one dead LED on it. The capacitors alone kept the panel lit dimly for minutes, down to 80V and below. (The capacitors were an attempt to replace those on the PSU board. Didn't help.)


The PSU:
Two caps are removed, C5 and C6, both are 100µF, 100V rated and mounted in parallel. Four jumpers on the right power the panels, 4 "wings" with 2 LED panels each, all in parallel.
AC power on the left. Voltage on the rectifier (behind the orange cap) was >160VDC.



PSU from the other side.



Backside of PSU. Shows the parallel wiring of the four jumpers and two caps on the right side as well as the two inductors.

 

If one damaged panel partially works and the others do nothing, I would first suspect the wiring to the dead panels, then the LED arrays.. Do you have another power source of around100VDC to test the panels with?

 

If one damaged panel partially works and the others do nothing, I would first suspect the wiring to the dead panels, then the LED arrays.. Do you have another power source of around100VDC to test the panels with?

No I don't, and that's the problem. Maybe I'll try to rig up an alternate supply and check the other panels.

I'm perplexed that the output caps are rated to 100V and I see the voltage there hit 95V or more and yet the panels remain dark. How much more could they need?

The panels are all wired in parallel and while I may not have checked every single one, I did find continuity everywhere I looked. I'm pretty confident the wiring is OK.

 

Big Clive on YouTube is the hands-down expert on LED lamps. He has dissected many types of LED bulbs and reverse-engineered them.

Yours may not be exactly the same type as this one, but perhaps they are similar enough that this video will give you some ideas.

 

Does the power supply provide the 100V you mentioned to the LEDs? Another case in which everything seems ok but nothing works.

 

As there are 24 LEDs and if they are connected in series I would only expect each panel to only require about 72 to 80 volts unless there are some series resistors also on the panels. For the panel with the faulty LED I assume it must have failed short circuit for the rest of the LEDs to still work. This would clamp the voltage to the other panels to about 3 to 4 volts less than there would be with 24 LEDs in series. It would be interesting to disconnect this panel to see if the remaining panels then light.

Les.

 

Probably a camera artifact, but there's what looks like a solder whisker between the two pins at the top right corner of the fourth image?

 

Does the power supply provide the 100V you mentioned to the LEDs? Another case in which everything seems ok but nothing works.

Roughly, yes, although it moves up and down when a panel is attached. It's more steady at ~95V with no load at all.

 

Probably a camera artifact, but there's what looks like a solder whisker between the two pins at the top right corner of the fourth image?

I noticed that and I'm pretty sure it's an artifact. I'll verify though.

 

As there are 24 LEDs and if they are connected in series I would only expect each panel to only require about 72 to 80 volts unless there are some series resistors also on the panels. For the panel with the faulty LED I assume it must have failed short circuit for the rest of the LEDs to still work. This would clamp the voltage to the other panels to about 3 to 4 volts less than there would be with 24 LEDs in series. It would be interesting to disconnect this panel to see if the remaining panels then light.

Les.

They do not light even when tested one at a time. Only the one glowing in the photo at ~80V lights at all. I'm thinking of placing a battery in series with the panel to see if the other panels will light with a little more voltage.

 

They do not light even when tested one at a time. Only the one glowing in the photo at ~80V lights at all. I'm thinking of placing a battery in series with the panel to see if the other panels will light with a little more voltage.

If you use more voltage you had better replace the two capacitors with a higher voltage rating or something will go pop!

 

The SMD LEDs found in light bulbs sometimes have multiple LEDs in series. I bought a cheap 120V bulb thinking I could cut traces to run the LEDs from 12V, but it took much more than 12V across each LED to make them light up. If your panel runs off less than 100V, then the individual SMD LEDs might have two internal LEDs in series, and need 6V to light up. You could use a 9V battery and a 1K resistor to check the individual SMD LEDs.

 

It's the LEDs. I've just done some probing with a battery and resistor much as mentioned by @bassbindevil. The LEDs are in parallel pairs and when you probe one, both light up (if they're working). Each panel is a series string of 12 pairs. I didn't probe all of the 8 panels but on the ones I looked at, there were occasional pairs that would not light and so the whole series string is dark. The single panel that lit a little got very bright (it was connected to the PSU) when I probed the pair where one LED was dark. This all tells me that the PSU may be fine but the panels are kaput. Oh well.

 

Further testing showed that each white panel (like the one pictured) in the lamp had a single failed LED. The panel would light fully if that one bad LED was bypassed with, say, a 1K resistor. In the course of this testing I observed that the power supply voltage was occasionally much higher. I'm sorry I can't recall the value but 180V sticks in my mind.

I quit working on this thing because I speculated that the sporadic high-voltage condition caused one LED to fail like a fuse on each panel, breaking the circuit. I don't bother fixing SMPSs (even if I could).