Major Power Supply Failure

[Terabyte2007]

This gives a whole new meaning to letting the magic smoke out!

Have not seen a bridge fail quite like this before. It had a whole burned through it...

[dr.diesel]

Caps are bad as well.  How old is the supply, manufacturer?

[Terabyte2007]

Caps are bad as well.  How old is the supply, manufacturer?

Oh yeah, the caps are definitely bad also. I will post the MFG and Model when I get back to the lab later today.

[NiHaoMike]

Was it set to 120V and plugged into 240V?

[N2IXK]

From the look of the solder joints, the bridge was running hot enough to melt the solder before it finally shorted and presumably blew the input fuse.

Was the supply running at or above rated current output?

[Terabyte2007]

Was it set to 120V and plugged into 240V?

The switch was set to 120 and this power supply was in the US so I don't think it was ever exposed to 240.

[Terabyte2007]

From the look of the solder joints, the bridge was running hot enough to melt the solder before it finally shorted and presumably blew the input fuse.

Was the supply running at or above rated current output?

All I know was a tech in the field swapped it out. No other info was given.

[NiHaoMike]

The switch was set to 120 and this power supply was in the US so I don't think it was ever exposed to 240.

Some server rooms are 208V and there have been instances of uninformed techs accidentally breaking hardware with it. They assumed the IEC plugs are all 120V.

[Terabyte2007]

The switch was set to 120 and this power supply was in the US so I don't think it was ever exposed to 240.

Some server rooms are 208V and there have been instances of uninformed techs accidentally breaking hardware with it. They assumed the IEC plugs are all 120V.

I guess that's a possibility.

[Tinkerer]

The last time I saw something similiar(but no where near as bad) was when arcing occured.

[N2IXK]

I see that the rectifier bridge has no heatsinking.  The larger hole appears to be factory issue, intended for a holddown screw. Are the charred remains of a part number visible?

My SWAG would be that the bridge failed (in short circuit) because it ran too hot for a long time. Melted solder, baked look of PCB around AC input tracks both signs of a long term thermal issue. When the diode finally failed, it went dead short across the mains input until the fuse blew or the branch breaker tripped. The rectifier package (probably nice and brittle/crispy at this point) burst from the fault current, blowing lots of black soot and metal vapor around.

Of course, a massive overvoltage spike on the input could take out the bridge as well, but the other signs scream "lousy thermal design/underrated components" to me.

[Terabyte2007]

I see that the rectifier bridge has no heatsinking.  The larger hole appears to be factory issue, intended for a holddown screw. Are the charred remains of a part number visible?

Of course, a massive overvoltage spike on the input could take out the bridge as well, but the other signs scream "lousy thermal design/underrated components" to me.

The part number is no longer visible. A long term thermal issue may be the reason, I am not in my lab right now but I believe it was a 250W supply and it may have been pushed to it's max potential for a while. The bridge is brittle and crispy, any real pressure, and it will crumble. The worst PS failure I have seen to date.

[AmmoJammo]

The capacitors are across the rectified mains.

So if the caps were bad, they could have overloaded the bridge rectifier

[N2IXK]

My guess is that if the caps were drawing enough leakage current to present a significant extra load on the bridge (an amp or so, maybe?), they would be vented more violently than the mild bulging we can see. That much heat burned up in the caps would be obvious with shrinkwrap melted off, etc.

No doubt the caps are bad. Probably sky-high ESR. Typical Chinese capacitor failure stuff. Outgassed enough over the years to bulge the can, and vent pressure, drying out the cap.  I don't think it really contributed to the catastrophic failure here, though.

What is that melted-looking yellow wire connected to? Does it look melted from excessive current through it, or was it melted by touching that hot bridge?

[mariush]

yellow wires are most likely for 120v/230v switch.

[Terabyte2007]

No doubt the caps are bad. Probably sky-high ESR. Typical Chinese capacitor failure stuff. Outgassed enough over the years to bulge the can, and vent pressure, drying out the cap.  I don't think it really contributed to the catastrophic failure here, though.

I think I am going to pull one and check the ESR for grins!

What is that melted-looking yellow wire connected to? Does it look melted from excessive current through it, or was it melted by touching that hot bridge?

It looks melted from the hot bridge.

[codeboy2k]

This gives a whole new meaning to letting the magic smoke out!

Have not seen a bridge fail quite like this before. It had a whole burned through it...

Awww... that ain't so bad.. I'm sure you can just collect up all the little black dust scattered about and stuff them back into the hole