That acrid electrical burning smell

The one that catches in your throat and sticks in your sinuses for a day or so.

It started last night, and I immediately put the PC into sleep mode, so it and the monitor power down. The top of the monitor was quite hot, and easy to reach out and touch, so that was the first suspect. I never pay attention to monitor heat though.

I also have a new router, and it's the only piece of kit that's new here, so it too is a suspect. It sits on top/front of the tower.

So I switch off the monitor completely, and power up the PC. And there's that burning smell again. Not the monitor and probably not the router. Power off quick, as the smell is burning metal and plastic now, and it's getting quite thick and worrying.

Open up the case, and unfortunately the mosfets on my mobo are covered with a heatsink (joined to the Northbridge heatsink by a heatpipe). I can't see the mosfets without popping the spring-loaded gromets on both.

[censored]

Is it the PSU? The mosfets are two inches below the PSU, so I can't tell. OK...assume it's the PSU, pull it and inspect it. No obvious damage, leakage or anything that looks untoward, but it and every other piece of kit in the tower is reeking of the burning smell.

[censored] (it's probably the mosfets)

So given the choices of:

1. Probably the mosfets and any more power to them will blow the mobo and lose the system

2. Probably not the PSU, but it might be, so replace it and see

[censored] Option 2 it is.

Thank God! New PSU and no smell. phew!

So that's an OCZ Stealthxstream 700w been running since March 2009. Probably on 90% of the time, and not in sleep mode or off. So at 3.5 years of say 24/7 times 90%, that maybe amounts to 2-4 times the usage of business or some home users. Can we estimate 7-14 years equivalent?

Also bear in mind that it's a 20% overclock system, I'm not too down-hearted, especially as it's not the mobo. Bought a 750w OCZ today from PC World, or would have to wait until the 3rd or 4th to find out if it was the PSU. I had an old 480w one, but didn't dare.

So that's my little bit of panic/excitement for 2012 over. (I hope...)

We had a similar issue not too long ago with a burning/fish smell coming from the landing area of upstairs. It turned out to be an electrolytic capacitor inside of a CFL light-bulb that had gone bad and was giving off the smell.

Looking at the inside of that power supply the main filter capacitor for the rectified mains appears to be located right next to two large heatsinks, so it could have been slowly cooked over those years of heavy use ( :

I often wonder how many realise that in the bases of the CFLs, there are those HF Oscillator circuits, to drive the actual Compact Fluorescent Tubes.

Much has been said about the trace of Mercury that is required to produce the UV light, which is converted to the Visible Spectrum by the Rare Earth white powders coating the interior surface of all Fluorescent Tubes, both CFLs and traditional, whether straight or curved as seen in signs.

The CFL HF circuits contain the mix of components needed to form an Oscillator circuit, eg Capacitors as mentioned (both Electrolytic and others), Toroids and Transformers on Ferrite Cores, Semi-Conductors, PCB etc.

The Electrolytic Capacitors in turn contain Alkalis, which can cause chemical burns.

Yes, most of those hazards apply to Electronic Equipment generally; but CFLs are frequently in locations where children can readily access them.

In most CFL, HF Oscillators, there is a total of 26 discrete components, as well as the hefty plastic moulding forming the Base.

I wonder how much of that circuitry is recyclable?

In reply to a post by eckiedoo:

I often wonder how many realise that in the bases of the CFLs, there are those HF Oscillator circuits, to drive the actual Compact Fluorescent Tubes.

Much has been said about the trace of Mercury that is required to produce the UV light, which is converted to the Visible Spectrum by the Rare Earth white powders coating the interior surface of all Fluorescent Tubes, both CFLs and traditional, whether straight or curved as seen in signs.

The CFL HF circuits contain the mix of components needed to form an Oscillator circuit, eg Capacitors as mentioned (both Electrolytic and others), Toroids and Transformers on Ferrite Cores, Semi-Conductors, PCB etc.

The Electrolytic Capacitors in turn contain Alkalis, which can cause chemical burns.

Yes, most of those hazards apply to Electronic Equipment generally; but CFLs are frequently in locations where children can readily access them.

In most CFL, HF Oscillators, there is a total of 26 discrete components, as well as the hefty plastic moulding forming the Base.

I wonder how much of that circuitry is recyclable?

Personally if I come across an electrical device that has a problem I like to reverse engineer the thing and see if I can fix it. Its always a treat if I need to use my oscilloscope to track down a fault on a misbehaving circuit.

If it turns out the device is not economical to fix then I usually strip the board for the various components (capacitors I will check with an ESR meter first).

I once was designing a switched mode power supply for the education of it and made the silly mistake of connecting the rectified mains filter capacitor in reverse. I observed it puffing out before the top blow off in a vicious spark filled explosion.

Luckily I was not looking directly at it as I had my eyes fixed on the oscilloscope to check the waveforms of the MOSFET (it was using flyback topology so controlling the turn off spike from the leakage inductance of the primary coil was important).

A colleague many years back, also reversed the connections to an electrolytic capacitor of the period, probably 16 to 32 MFds, 300V DC.

Probably bout 1.5 inches diameter by six inches tall, say 37.5 mm diameter by 150 mm tall.

Large enough to be potentially lethal.

It exploded directly in front of him, as he switched mains power on.

The contents, rolled aluminium foil, separators, corrosive alkaline electrolyte etc, headed upwards at high speed to splatter on the laboratory roof, about 25 feet (8 metres) above


The resulting damage, splatter etc was there for many years, being pointed out as a salutary warning, "tae mak siccer, afore switching on".


By sheer coincidence, his name was highly appropriate to the event!

Well I've been over the old PSU with several toothcombs and can't see an obvious issue.

There's a little wire melting in pic 4 (black wire behind the top coil), but said wire has been against the copper coil for 3.5 years, so that's pretty likely I would say. The whole thing was pretty dusty inside. Maybe the heat built up beyond the dissipation efficiency (or whatever the correct terminology is).

Pic 1
Pic 2
Pic 3
Pic 4
Pic 5
Pic 6
Pic 7
Pic 8

I would rather see a big blackened mess, so I could be sure.

That inductors core looks like it has been getting very hot as it looks pretty discoloured.

Edit I just googled "OCZ Stealthxstream 700w smell" and there are lots of forum posts relating to this issue. I will keep digging ( :

Just a thought but perhaps it could well be related to some other electronics, as inductors just don't start to saturate and get hot after running fine. But then again if was related to the primary drive circuitry then the main switching transformer would exhibit problems too.

Its a coupled choke that carrys the high current +12, +5 and +3.3v rails and is normally used to help smooth out the supply.


Hmm...Perhaps its related to the mixture the inductor was made out of.

Another Edit: Please be careful when poking around inside of the PSU. Its unlikely but there is a small chance that there is still a charge stored inside of capacitors on the rectified mains side, and those heatsinks can be electronically connected to various points in the circuit.

Plus the metal case is referenced to earth internally so try and keep it to using only one hand unless you are sure that everything is discharged. Take it from me, 325v across your chest really sucks.

Edited by Alex1M6 (Mon 31-Dec-12 14:28:48)

In reply to a post by camieabz:

There's a little wire melting in pic 4 (black wire behind the top coil)...

Nice pictures!

That coil appears to have received a good 'burning'. It's probably in the 3V3, 5V or 12V output (whichever output has the highest current rating).
If you move the output cabling from the PSU chassis, you may find the culprit underneath the cabling - look for the smoke trail/two legs (with charcoal or an air gap between them) or a damaged electrolytic capacitor. Also check the underside of the circuit board.

EDIT: Picture 6 appears to have captured the damage - zoom in to the base of the coil/inductor - the windings appear to be scorched. The damage should be visible on the underside of the PCB.

Edited by JimmyBoy (Mon 31-Dec-12 14:12:29)

Read this, it pretty much sums it up LINK

According to this the inductor gets very hot and slowly cooks/begins to melt everything close to it.

Edited by Alex1M6 (Mon 31-Dec-12 14:30:26)

More than half-a-century back, I was asked to repair a neighbour's radio.

Typical 1930's model, valves/tubes, mains dropper etc, in a relatively large cabinet to allow cooling air to circulate.

On taking the fretted back off, I was confronted by a mass of loose felt or fragments of thread that had gathered inside, rather like the contents of a vacuum cleaner bag, used mainly on carpets, apart from where it had been singed by the mains dropper.

Several handfuls of felt later, I started to expose the valves etc, rather like an archaeological dig.

Yes, I did manage to repair the radio.