Older Samsung TV SMPS repair: What is shorting my MOSFETs?

[iamdarkyoshi]

Hello again everyone. I picked up a free 55in Samsung TV. When I got it, the board (BN44-00342B) had a blown fuse in the middle of the board (FS802S) and two shorted MOSFETs.

It originally had two 15N50 (15A 500V) MOSFETs and a 3.15A fuse. I replaced the fuse with a 2A one, and the MOSFETs with 12N50's (12A 500V) just to see if it would power on, and it tried but it wouldn't stay on. Unplugging the backlight, it stayed on (albeit with no backlight) so I figured the MOSFETs were being overloaded.

This was good news though, as it was completely dead before.

So I found some 20N60 MOSFETs (20A 600v, higher specs than the original ones) and it worked! I had lines on the screen, but thats a different issue, the BGA LCD driver on the video processor board needs to be reballed, it works when I press on it.

I left it running for the day at work while I worked on other TVs, planing to bring this one home.

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

So I took it home, set it up in my bedroom, plugged it in, everything was working. 2A fuse and 20N60's in place.

I was going through the menus, tweaking settings, and I eventually turned up the backlight brightness. As I was continuing to tweak the image, about 60 seconds later, the TV turned off and I heard a "tink"

I was hoping the underrated fuse just blew from the increased backlight draw, as it was 1.15A lower than the original one, but nope, the damn MOSFETs shorted again, which of course popped the fuse.

Why do these MOSFETs keep shorting? Is there an issue with the snubber network? The resistors and tiny ass capacitors near the MOSFETs look fine, so I'm really not sure what to do next.

I'd rather not just buy a replacement board, as it would be infinitely more expensive than the TV cost me, and there's no learning involved 

[T3sl4co1l]

Lots of possible reasons.  "Too old" is a prime suspect -- in turn, anything that ages, electrolytic capacitors being the quickest aging of common components.

Presumably it's either: 1. a design "defect"*, 2. a latent production defect, or 3. a freak incident (caused by individual environmental stresses -- electrical surges, chemical corrosion, condensation, etc.).

*You may not like it, but part of design is satisfying the lifetime and cost requirements of the product.  Hence, "too old" is "beyond design lifetime".  If you want a "younger" TV, you should've bought a younger one.....but then it wouldn't have been free, would it?  Alas, that's just how exchanging money for goods works.

In any case, whatever failed, the damage has been done.  It's amazing that it worked at all by replacing just the transistors -- when a transistor dies, it goes three-way-shorted, dumping high voltage onto the chip driving it.

In a power supply, normally the controller, gate resistor, transistor and maybe a couple of other components (current shunt?) fail, when the culprit might be the secondary side rectifier overheating.  The fuse is the last thing that blows, ~10 milliseconds after the transistor and all has long since passed.  The only way to be sure you've fixed the circuit is to trace the entire sequence of events, and replace them all.

Tim

[iamdarkyoshi]

Lots of possible reasons.  "Too old" is a prime suspect -- in turn, anything that ages, electrolytic capacitors being the quickest aging of common components.

Presumably it's either: 1. a design "defect"*, 2. a latent production defect, or 3. a freak incident (caused by individual environmental stresses -- electrical surges, chemical corrosion, condensation, etc.).

*You may not like it, but part of design is satisfying the lifetime and cost requirements of the product.  Hence, "too old" is "beyond design lifetime".  If you want a "younger" TV, you should've bought a younger one.....but then it wouldn't have been free, would it?  Alas, that's just how exchanging money for goods works.

In any case, whatever failed, the damage has been done.  It's amazing that it worked at all by replacing just the transistors -- when a transistor dies, it goes three-way-shorted, dumping high voltage onto the chip driving it.

In a power supply, normally the controller, gate resistor, transistor and maybe a couple of other components (current shunt?) fail, when the culprit might be the secondary side rectifier overheating.  The fuse is the last thing that blows, ~10 milliseconds after the transistor and all has long since passed.  The only way to be sure you've fixed the circuit is to trace the entire sequence of events, and replace them all.

Tim

What confuses me is that the thing worked for a straight 6 hours... I'm not an expert in SMPS repair, but maybe next I'll try replacing the caps in the snubber network, since they have caused issues for me in the past