# LED failure 240V

Status
Not open for further replies.

#### debe

##### Active Member
This is one of 8, 240V LEDs I purchased about 9 months ago. It started flickering & eventualy would come on, but eventualy failed. I decided I needed to know why, but it had to be destroyed to get it apart. The diffuser part is glass so it had to be broken, on dismantling I find the constant current source was set in a soft grey rubbery stuff which I removed. I discover 2 dry joints on the circuit board, on re soldering it worked again but is not abled to be reassembled. The LEDs are on an aluminium backed heat sink. Have drawn the circuit just for interest. You will be lucky if you can find any brand name LEDs that aren't made in China..

Last edited:

#### Nigel Goodwin

##### Super Moderator
I discover 2 dry joints on the circuit board, on re soldering it worked again but is not abled to be reassembled.
That's annoying!

Nice job finding the dry joints.

#### Diver300

##### Well-Known Member
I had something very similar happen to some LED lamps that I had. I also destroyed the first one to fail by opening it, but I found that the main rectifier capacitor (the equivalent of the 1.8 μF 400 V one in your circuit) had failed. I also found that if I ran the light from rectified and smoothed mains voltage, it didn't matter if that capacitor had failed.

When the next one failed, I put a rectifier and capacitor from a redundant switch mode power supply in the fitting, and fed the lamp from that, and it worked without flickering and without having to be opened up.

Your LED lamps would certainly work from rectified, smoothed DC.

On the circuit, I don't see why the makers used such a large capacitor, in either voltage or capacity, for the 2.2μF 400 V. It shouldn't really need more than about 150 V for 100 V of LEDs, and the current in the LEDs is maintained by the inductor, and the frequency will be 100 kHz or so. On the photo, that capacitor looks smaller, so have you got them the right way round?

Also, the resistor are marked with two digits meaning the value, followed by a digit showing the number of zeros. So the ones labelled 514 are 510000 Ω or 510 kΩ

#### Tony Stewart

##### Well-Known Member
This appears to be a popular chip for large strings with any current range up to 150 mA or less depending LED string voltage but is rated for a wide range using offline SMPS regulator chip.

BP9831 not not widely available.

#### Attachments

• 64.6 KB Views: 92

#### RODALCO

##### Well-Known Member
Surprising to see a proper label on the lamp. Often the cheap led Chinese lamps have no data on them at all.

#### RODALCO

##### Well-Known Member
Just noticed looking at the picture again B22 lamp while it has an E27 screw thread. Talking about QC.

#### debe

Sorry about the picture but the one I pulled apart was a B22. The E27 just happened to be another one same brand that is still good. These bulbs were $10 each Trade price. Not from Ebay either. #### Tony Stewart ##### Well-Known Member Most Helpful Member The This appears to be a popular chip for large strings with any current range up to 150 mA or less depending LED string voltage but is rated for a wide range using offline SMPS regulator chip. BP9831 not not widely available. View attachment 100948 Your schematic shows some improvement to input transient suppression current limiting of a 10us pulse with 10R>1.8uF then L1 to main storage cap 4.7uF to buck regulator with Effective load (ESR) of 30 series LEDs of 93V typ @100mA is around 10W with 1/3W each or 3.3 ΩESR*30~100Ω ESR string, giving an RC decay T=470us so switching cycle is <10%'of this. Appears to be a reasonable design for <$1USD/W. It will not too hot for Tj. with Density of packages, on Alumclad board.

I agree with other comments.

Status
Not open for further replies.