Had an LED go out in the bathroom over the vanity. Decided to do some testing on it. After determining that it wasn't a poor electrical connection I broke it open. Found a very cheap set of electronics inside. Nothing like the better made PWM type lights.



This is the second cheapest construction of a driver that I've seen so far. The only worse things I've seen have been those cheap Chinesium made which have hardly more than a capacitor for voltage drop.

Anyway, the next two pictures focus on R13 (F1). When examined, it looked like a broken solder joint. Only it wasn't. The other side of R13 (F1) looked good but IT was the side that was open. Now, given that these are heat sink mounted it's not easy to solder on. But I managed to set the iron at its hottest and let the iron dwell on the joint for quite a few seconds. Maybe on the order of 20 to 30 seconds. But eventually solder began to bridge from R13 to the pad. You can clearly see the end that I resoldered. And yes, that's WAY too much solder. But that's what it took to transfer sufficient thermal energy to get the pad to flow.





You may also note the pinch type wire junctions of both the Hot/Neutral lines and Capacitor (mounted through the board). If this is the reason why these things are coming down in price then we can begin to expect more failures in the future. After the butcher job - I mean solder job - it's working. Now, what to use it for?

 

Question??? Does the bulb work now?? You should be able to test it. Just be careful because that is mains voltage present. But please let us know if that was the problem.

 

I’m not disagreeing but I’m wondering what you expected to see besides “cheap“ electronics. Poor construction probably led to the failure and the electronics might otherwise last as long as the LEDs themselves.

 

The Bright Power chip is a modern linear regulator chip that works quite well. You might want to watch of of BigClivedotcom's videos on YouTube where he's analyzed a large number of LED bulbs. On some of these chips, 2 resistors in parallel set the current through the LEDs. In many designs, it's like it was made for hacking. Often, the resistors are about a 2:1 ratio. Removing the higher value resistor reduces the current through the LEDs by about ⅓, resulting in a slightly dimmer bulb that will run cooler with a much longer life. Removing the lower value resistor instead reduces the current by maybe ⅔, resulting in even a dimmer but far longer life bulb.

Those push-in contacts for line power and the cap are quite effective. In fact, I've never seen one fail. They provide some isolation from heat for the cap as well.

 

From a Facebook post I made.....

Dollar Tree has a good selection of LED bulbs that are decent quality of the latest technology. Two 40-watt-equivalent bulbs for a buck and a quarter is a great deal, and they even have a 3-way LED bulb for that price.

My subject of today is a 3" diameter globe light. I got one to test last night, in warm white. The color is pleasing, towards the yellow side of warm white. I have a 3 lamp bar above my sink that uses this type of bulb, that I've had daylight-white bulbs in. I actually prefer the warm white, so I'm going to give these a try. But three 40 watt bulbs is a bit bright for my tastes....

Watchers of Big Clive 's YouTube channel will know that the latest LED bulb design typically has two resistors in parallel to set the brightness. In this case, those resistors are 47 and 62 ohms, giving an effective resistance of ~27 ohms. If the 62 ohm resistor is removed, leaving only the 47 ohm resistor, only half the current will flow, and the lamp will be roughly half as bright (your eyes may not think it's that much different because of their logarithmic response, but it will be noticeably dimmer). Less current means less heating, so the bulb life will be increased too. Removing the 47 ohm resistor instead of the 62 ohm resistor would make the bulb about ⅓ of original. It's almost like they designed it for us!

The modification isn't difficult to make. The plastic globe has to be pried off, which takes a bit of effort – it's glued on with silicone glue. Press in near the bottom of the plastic globe to get a gap between the the globe and the base, then work a spudger or butter knife around to break the glue line and pop it off. Then remove the 62 ohm resistor – the easy way it to crush it with small needle-nose or diagonal pliers. Push the globe back on, and you're good to go. No glue is required, it will stay on.

CAUTION: if you power it up without the globe keep your fingers away the the circuit board. It's at line voltage and it WILL HURT YOU.

 

 

Question??? Does the bulb work now??

Yes, but in my learning about this failure I ended up tearing the base apart. Prying the globe off the bulb resulted in a globe that can not be put back. Oh well. It was more of a learning experience than anything else.

I’m not disagreeing but I’m wondering what you expected to see besides “cheap“ electronics. Poor construction probably led to the failure and the electronics might otherwise last as long as the LEDs themselves.

Those bulbs re evolving fairly quickly. From comments above it appears the bulb I disassembled (and can't reassemble) what I've found is not all so bad a design.

Big Clive - I've seen some of his videos. Not many. Probably because I haven't had a desire to learn more about LED bulbs. Will consider watching more video's. Thanks for the suggestions.

Don't know what I'll do with this carcass.

 

Put it in the trash where it belongs, I just don't understand why people spend so much energy trying to use the junk they accumulate.

You learned something, now move on.

 

Put it in the trash where it belongs, I just don't understand why people spend so much energy trying to use the junk they accumulate.

Can't find the trash can for the junk I've collected. I'm GOING to do something with it. Just like all the rest of the junk on my work bench. Guess this is the definition of "Pack Rat". I've got an Xenon Flash Tube, HID lamps and driver modules, transformers, relays, LOTS of LED's, caps, heat sinks, Wall Warts and so much more. Even have an "Engine Degrees Advance" meter. Just the meter alone. No shunts, nothing directly related to it. Just an unholy pile of poop. And a bottle of marbles. But I keep those so no one can say I've lost them.

What a mess. Yeah, I gotta start tossing stuff. I've had lots of this stuff over 10 years and still haven't done anything with it.

Spring time is coming.

 

Tracing out the circuit could be a good learning experience. Seeing what a circuit optimized for production is always educational. AND you might be able to use that series string of LEDs for some application. But was the light bulb filled with a gas that conducted heat better than air?? there might be a cooling issue if that was the case.
If you find that all of the LEDs are in a single series string, and know the voltage, and if there is a single series resistor that you can know the value of, then you can figure the current and understand the operating point of those LEDs. THAT will be very interesting.

 

was the light bulb filled with a gas that conducted heat better than air??

No.

If you find that all of the LEDs are in a single series string, and know the voltage, and if there is a single series resistor that you can know the value of, then you can figure the current and understand the operating point of those LEDs. THAT will be very interesting.

Took quite a while just to trace out as much as I could. The IC shows two wires entering underneath from the left and the right. I have no idea where they go. They are shown as blue dashed lines. I suspect they go to these pins but I don't know that for certain.
And now - this:

Click on the drawing to get a full screen view.

And I have no idea how the LED's are wired internally.

Just read 142VDC across the capacitor.

 

Each "LED" has multiple LED elements in it, to get the Vf of the string near the powerline voltage.

Once again, I recommend Big Clive on YouTube. He has dissected many LED bulbs, discovered the datasheets for the linear regulator chips and drawn the schematic for each bulb he has dissected.

 

OK!! You can measure the voltage across that 270 ohm resistor (271) to determine the current and measure the voltage across the 47mFD capacitor to know the string voltage. Then you will know what voltage 8 LEDs in series use.But probably both volts and amps will be average values because why filter if it is not needed.

 

You can measure the voltage across that 270 ohm resistor

When measuring the voltage on one of the two (in series) (R2 & R3) resistors I'd be blinded by the light.

 

Ok, didn't think about that part. So solder a couple 2foot extensions on it #24 pair from a scrap of cat5 or cat6 twisted pair, from a cable scrap. Skin and tin about 1/8 inch before soldering to the resistor.

 

Ok, didn't think about that part. So solder a couple 2foot extensions on it #24 pair from a scrap of cat5 or cat6 twisted pair, from a cable scrap. Skin and tin about 1/8 inch before soldering to the resistor.

Didn't think of that. But given how difficult it was to get the solder to flow to R13[F1] it might not be so easy to tack a couple wires on. I'll consider it tomorrow. Don't be surprised if I don't.

 

One of what I'd call a cheaper design now appears to be more along the lines of the first one I posted. I'm now looking for information on a chip: SM2196E. It appears to be a current controller from what I can read of Chinese (which I can't). But this schematic seems to appear as such.

From
https://datasheet4u.com/pdfhtml/0362/1330504/page-000001.png

 

Couldn't blow the fibers off. Forgive please.

 

It appears to be a current controller from what I can read of Chinese (which I can't).

But Google Lens can! It sometimes makes a bit of a mess, but pertinent points can be found.

Again, I will recommend Big Clive's videos. He has sleuthed out many of these linear LED regulator chips. Many of them seem to be essentially the same with different manufacturers and part numbers. Even if the pinouts aren't the same, the theory of operation is.

 

I've seen and have been watching them. Interesting for sure. Thanks.