LED worklight SMPS failure - Can you help?

Thread Starter


Joined Nov 11, 2009
Hi guys,
I am trying to repair an LED worklight that suddenly stopped working after a period of use. The unit is an LED driver but also has a Bluetooth speaker incorporated with it.
I am relatively new to electronics and have a decent understanding of Switch mode power supply theory but troubleshooting them is a whole different thing.

On opening up the unit I found that 3x resistors in parallel with the Source of the mosfet were blown (See picture – new resistors under black heatshrink bottom right). Also the main 500ma board fuse had blown. So that was 2 obvious failures, but I knew that it was likely something else had caused the resistors to blow. I removed and tested the MOSFET and it was conducting both ways and essentially shorted. I also tested the Diodes on the rectifier bridge and on the secondary side. Secondary Schottky diodes were also conducting both directions.

So I replaced all components I found faulty – Diodes (x2), MOSFET, 3x resistors (1.3 Ohms each – I used regular bead type resistors instead of smd type to test) and the 500ma fuse.
I powered up the unit and nothing happened. However I did smell a slight burning smell which I now believe to be the 500ma fuse blowing again. There is another short somewhere…

But now I am stumped. I have no idea what to look for next? I assumed it might be the PWM chip to drive the FET but the number "5673A" does not match anything and I cannot get a data sheet to test the pinouts. Could this PWM be shorting causing the fuse to blow? They usually have overcurrent protection but seeing as the resistors all burnt out could this be the culprit?

Capacitors look ok, and test fine with an ESR meter. Would a bad cap cause a dead short though?

As you can probably tell I am pretty new to this so please feel free to correct me if I am talking nonsense… but at the same time go easy. :0)
Any info you guys could give to steer me in the right direction would be much appreciated.




Joined Jun 4, 2014
This is the problem with SMPS: when they blow, if you don't replace all the faulty bits then they are very likely to blow again - and repeat.
Without a schematic, so you can make some reasonable guesses at which components should be checked the task is nigh impossible. I speak as someone who spent quite a number of years repairing these things.

Thread Starter


Joined Nov 11, 2009
Hi Albert, thanks for the reply. I was thinking that would be the case. Is there anywhere I can obtain schematics?
Failing that I was going to fit a basic LED driver and forget about the Bluetooth side of it. The Lamp was rated at 20w. Any advice on how to select the correct driver for this? Do I need a larger powered driver for this or does it have to be exactly 20w output?

Thanks again...


Joined Jun 4, 2014
You would need to know more about the LED specification - voltage and current.
I assume it isn't a single LED and multiple LEDs could be connected in series or some combination of series and parallel.


Joined Feb 20, 2016
These can be hard to fix as AlbertHall says above.
I did notice the spark gap near R1/R2 looks to have flashed over.
Have you measured continuity of the filter (yellow "transformer") above the spark gaps?


Joined Feb 8, 2018
There is a chance the control IC has been fried. The original failure was probably the FET short circuiting. The three resistors in parallel are probably for sensing the current through the FET (between source of FET and negative of input supply). Normally the peak voltage across them would be something in the 1 volt range, give or take a volt. If the FET fails short circuit, then then almost the full rectified line voltage will be applied to the resistors. Unless there is a fairly high value resistor between the "high" end of the sense resistors and the sense input on the control IC, the latter will likely be toast. The fuse eventually gets around to blowing.

I suspect the output diodes were probably actually OK. Did you check them out of circuit?

ROHM makes a BD7673AG SMPS controller in a 6 pin package. Perhaps they have a 5673. Diodes Inc and ON Semi have some 6 pin controllers.