Blown fuse and massive spark and crack from PC

TegGhola

Distinguished
Dec 9, 2009
68
0
18,640
Apologies if this has been covered elsewhere (in which case, do point me in the right direction), but . . .

My rig had been running fine for at least eighteen months. Then, yesterday, in a cool (17°C), dry room I reached behind the PC case to hit the power switch on the mains socket. In that position, I was nicely placed to hear a loud crack and see a big electrical-arc flash inside the PC case. I couldn't tell exactly from where in the case the spark had come, but it quickly became apparent that the PC would no longer boot, and the power LED on the motherboard itself was not lit.

I initially wrote the whole thing off as being fried, when I realised that the 5A fuse in the kettle lead had blown. Switch to another kettle lead (13A fuse this time) and everything now seems fine. Ran Memtest86+ for three hours (four passes) and it found zero errors. Boot up, and the disks are fine, graphics output seems fine.

So I want to know whether anyone else has seen this in their rig? And ask whether there are any electrical / electronic engineering types who can tell me whether the loud and bright arc was caused by the blown fuse, or was the cause of the fuse blowing. Also, how can the inside of a PC generate such a noise and bright light without anything seeming to be damaged? Could static electricity (caused by the cool air) explain the occurrence? Any other damage I should be checking for?

Oh, and what fuse should a PC kettle lead need? Seeing as the power supply is only 650W, I would have thought that 5A ought to cover it (650W/240V = 2.7A).
 
(1) Kettle lead ???, Goolged it, is this a in-line fuse on the AC line to the PSU?

(2) If so get rid of it and plug your computer into a switche power strip.

(3) the Arc and loud "crack" was not ESD. Althoug static build up on YOU can be very high (ie seveal KVs) the arc would be barley visible and not very loud (your ouch may be).

(4) Inspect your MB very closely and any leads that are not terminated (plugged in) for Black marks.

(5) Now for my quess and that is all it is is a guess!!!. (1) your on/off PSU switch is faulty and when you turned it off you had what is called contact bounce and the surge (Inrush current) triggered a overvoltage or over current protection circuit inside the PSU. Dougtfull, but a ESD discharge from you into the PSU could also trigger it. If this is the case, then you now have a PSU without a protective circuit, and although it "seams" to work, I WOULD replace it.

PS: Have to leave, so forgive spelling, also would have elaberated more.
 

TegGhola

Distinguished
Dec 9, 2009
68
0
18,640
1) Looking it up on Wikipedia quickly, by "kettle lead" it appears that I actually mean a mains lead with a BS 1363 on one end and a C13 connector on the other.

2) Is that what you mean by "in-line fuse on the AC line to the PSU"? The mains lead was already plugged into a switched power strip (Belkin, "surge protected"), and it was the power switch on the Belkin strip that I was switching on. I made no contact with the PC or PSU, I was just near it.

4) I did take a good look at the motherboard and the cards, but I was looking for burst capacitors. I'll take another look in daylight for any burn marks.

5) As I didn't touch the switch on the PSU, would that mean the switch on the Belkin strip would be at fault instead? Oh, and I was turning it on rather than off, if that makes a difference.

What does the protective circuit do, and what might happen without it being intact?
 

False_Dmitry_II

Distinguished
If the same thing happens again it will not have a protective burst and will leave your components unprotected from a repeat performance.

You can't get a ESD from not touching things.

I'll have to defer to someone more specifically electrically experienced for how there would be a huge pow with no power in there.
 
OK looked at your ref, wiki. The term just refers to the type of plug.
Where was the fuse you replaced located.

Did I read you correctly - You switched your belkin strip Off, not the rocker switch located on your PSU.

Yes the Belkin Switch could be defective. Two things with switches. When you open them, they do not instantly open as they start to open they can have a short time where an arc will occur until the gap between the contacts are far enough apart. This is in micro to milliseconds. The second problem that can develop is called contact bounce. This is where the contacts close and open a couple of times before remaining closed. Both conditions are undesirable. At work we have had to pitch some brand/models of power strips due to know problems. In both of the above cases the problem is the power is on/off/on or visa versa. This relates to what is referred to as "In-Rush" current.

In Rush current is the current that is drawn when you first turn something on and is very short in duration. Example: I have to measure the In Rush cuurent for three PSU suppling power to a satellite instrument. Normal current is Very low, 1 Amp on one and .05 Amps on the second. I measure (Clamp on Amp probe to O'scope) about 4 to 6 Amps for several Millisec. A regular light bulb has a very low resistance when AC is applied, But as the wire heats up the resistance goes up and the current goes down if not you would need Very large fuses for table lamps.

As you noted 650/240 = 2.7 Amps But the fuse used is a 5 Amp. If a 3 Amp fuse were used, Normal current would not blow it, But depending on the initial surge duration it might blow. This is also why "Slow Blow fuse where invented, Takes a High surge current and will only blow if the rated current is longer than X time.

Sermom over - IF Arch was inside your computer, and it is currently working I'm betting it was the Input protection circuit in the PSU. This is simular to what is inside a surge protector - They can make a loud bang and blow open - Basily they are a one shot protection device - Last long enough to protect one time.

Put 2 to 2 together - The plug end got its name because it was used on the electric tea kettle popular in England - O'well blew that one.

Good luck - Swap out the Belkin power strip (if that is what you turned on/off). If NO burn marks on MB and Computer components, See if you can open up the PSU and look around in it. If nothing is found I would not take a chance, I'd swap out the PSU.

 

TegGhola

Distinguished
Dec 9, 2009
68
0
18,640
Yes, it was the Belkin strip switch, but I was switching it on rather than off.

I gave the board and loose connectors a good examination, but I see no scorching. I've not pulled the PSU out yet, but I'll see if I can build up enough steam to tear all the parts out of the (less than spacious) chassis and replace the unit.

Is the PSU dangerous (fire risk) if the circuit has gone, or it more a risk of frazzling my motherboard et al?

As for kettle lead, I'm obviously not that clued up either, as Wikipedia states that kettles actually use a slightly different connector. (Same shape, but the kettle connector seems to have a notch in the bottom.)
 

bgd73

Distinguished
Feb 18, 2008
201
0
18,690
its the psu. check for circuits close to each other, heatsinks etc.
the wind down to high amp low volt can make an interesting spark.
 

Paperdoc

Polypheme
Ambassador
As a simpler explanation of protection circuits (and some surely are built into your PSU, maybe some also on your mobo in its own voltage regulator section), they are often connected across the output + and - (Ground) of a power supply section. Under normal circumstances they have a very low leakage current and hence have a negligible impact on output from the supply. But under specified designed conditions (usually, in the event of voltage higher than some limit) they suddenly become almost a dead short across the output. This completely consumes all the current output capacity of the supply section, reducing the output voltage to very low and thus protecting the connected normal load from the voltage surge. BUT in doing that job the protection circuit itself usually is destroyed by carrying that heavy current even briefly. In the aftermath we find that the whole system appears to be working OK. But in reality that very-low-leakage protection circuit is now gone completely and the rest of the system is simply unaware that the minor load of that circuit is missing. However, when the next high-voltage condition happens, there is absolutely nothing to short it out and the overvoltage is passed on the the sensitive normal load, potentially causing damage to it instead of to the protection circuit that is gone.
 

Paperdoc

Polypheme
Ambassador
A really good UPS will have its own protection circuits built in to protect both itself and its load. However, if you have that, you don't start ripping into your PSU to disable its protection circuits. They stay there. They are just much less likely to be forced to deal with a power surge that come down the line because the UPS may have reduced it substantially.