On Wednesday, 28 August 2019 11:22:26 UTC+1, Jannie wrote:

Thank you again to everyone that has been contributing. I really appreciate it so much and I'm learning so much.


Can you trace out where the wires go? If it is a regulator, then the center lead will usually go directly to ground (nearly zero ohms to the negative output terminal). If it is a TL431, the left wire (looking at the flat surface with the wires down) will probably go to ground. And if it is a pass transistor, then probably none of the wires connect to the ground.

I have tried the above suggestion and the center lead goes to go to ground (nearly zero ohms to the negative output terminal, 1.6 ohms to be exact, but that's the same I get if I directly connect the 2 test leads together). Not sure if this means it's a regulator or if it's still more likely to be a TL431, but I hope this helps.


With mains stuff it's useful to test it by running it in series with a filament lamp if there's a possibility of shorting.

Thanks for this, I will definitely do the testing with this setup. I'm sure I still have a few old filament lamps in the garage somewhere.

I did some more detailed examination of the board on the front/top side with a magnifying glass for any possible damage and found another part with damage, the other IC (marked as U1 on the board).

The code on IC is:
H11552
OB2358AP
L d

Based on what I found, this seems to be a "Current Mode PWM Power Switch".. Not sure which one would be the correct replacement as there are many and not one matching the "OB2358" code (https://www.mantech.co.za/Stock.aspx...=PWM+DIL08and). Anyone have any suggestions?

Again, I hadn't noticed it before (probably because I assumed the problem was the clearly blown part). I have taken a picture and added it to the album (https://imgur.com/a/uBCXjwQ). You can see there is a crack in the center top side of the chip running down to the bottom right side.

I also added an image of the front/top of the board with an overlay of the circuit at the bottom. Don't know if that helps anyone, but it's definitely helped me starting to understand what goes where and the connection between the components.

@whit3rd: Thank you for confirming the TL431 part. Not sure if you meant to paste a link for the optoisolator part, but it is the same link as the one for the TL431 part. @Phil did however mentioned that any "xx817xx" opto-coupler part should work.

If the diodes are short-circuited, they'll look bad in-circuit; if a multimeter that does diode check is available, use that.

I have tested the diodes in circuit with a multimeter and some of them (D1, D2, D3, D4 & D6) looks bad (shorted), but as mentioned, they could be fine. I can also confirm that the few diodes where I could actually see the number, are in fact 1N4007 diodes. I reckon I just go buy a pack of 10 x 1N4007 diodes with the other parts (once confirmed) and then tackle the diodes tested as shorted in circuit 1 by 1, removing, testing and replacing if needed.

You asked for a suggestion, I have one. SMPSUs are not the simplest things for beginners, and this one sounds well knackered. I suggest picking up another piece of e-waste & fixing that. You're not gonna fix them all.


NT