PSU just went pop, what do I do?

[westom]

With regards to the plug, I thought, that a surge could come through the AC side to the PSU and this is what triggered the motherboard warning.

Again, assumptions by using word association. You have no idea what the word 'surge' means until defined by specific current, voltage, frequency, duration, etc numbers. Those completely different disturbances, each called a 'surge', are from different anomalies involving different (and unrelated) circuits.

Same mistake was made with the word 'load'. Long before recommending what causes a load, first define what is accomplished by a maximum load. And where that load must exist. Where a load must be significant, Prime95 creates a tiny load.

Will connecting or disconnecting USB devices cause a surge? Unknown until those changes are first defined with parameters. Numbers that define which word 'surge' applies.

Also assumed, current on the 12 volt line is relevant to what happens on the 3.3 or 5 volt rails. Each voltage is defined uniquely by its own current and other unique electrical parameters. Each rail must be loaded.

Defined are how the OP obtains numbers so that a reply is without doubt or speculation. To either define or exonerate so many suspects. If the meter and load confirm those BIOS numbers, then move on to other suspects.

 

[jackocurly0074]

Will connecting or disconnecting USB devices cause a surge? Unknown until those changes are first defined with parameters. Numbers that define which word 'surge' applies.

Ok, so what numbers define this surge? (other than voltage/current and frequency)

Also assumed, current on the 12 volt line is relevant to what happens on the 3.3 or 5 volt rails. Each voltage is defined uniquely by its own current and other unique electrical parameters. Each rail must be loaded.

It should be "loaded" which was what I said no? Without proper testing equipment there is a limit to how much current/wattage you can draw from the power supply.

So could you explain why Prime95 creates a "tiny load" with numbers/parameters as I am unsure what you mean?

By running Prime95 and a GPU stress test creating 100% processor usage (both CPU and GPU) power draw will come close to the max CPU (95W TDP), GPU (170W TDP) and motherboard (around 140W TDP depending on configuration). This will be pulling a large amount of current from the 12V rails and the motherboard uses the 3.3V and 5V rails, add to this the hard drive searching/optical drives usage and this is likely to be the highest current draw/power draw possible without purchasing more components, or a proper PSU load tester. Then whilst the PSU is under this "load" or high power draw, or more specifically high current draw, the voltage can be measured to see if the current draw is pulling the voltage down, indicating a problem.

 

[westom]

Ok, so what numbers define this surge? (other than voltage/current and frequency)

USB reports a surge. What is that surge? Something similar to what blows a fuse. How is that surge 'solved' by a box on a computer's AC power cord? It's not. That power strip recommendation demonstrates insufficient electrical knowledge. And demonstrates knowledge using word association.

Normal is for a defective supply to boot a computer. "If it boots ok etc then there is likely nothing at all to worry about" is false. That computer may work for months. Then fail after a warranty expires. The defective supply might have been identified on day one by multimeter numbers. But many computer techs do not know how electricity works. Do not know how to use a meter. Then assume a defective supply must be good only because the computer boots. Another example of why knowledge requires numbers.

Purpose of this thread and the OP's question is not to teach basic electrical concepts. The OP's surge is a completely different electrical disturbance.

The OP may have a useful reply by performing the recommended procedure with a meter. Then posting those numbers.

 

[Mei]

why does it have to be worded so confusing
are you saying ppl should use a multimeter on their psu's and if the voltages are not perfect to send it back? and argue with the shop to get it replaced?

seems a little extreme

 

[Bacon?]

Hardly Mei, if I had a defunct component it wouldnt be in my house long.

 

[Mei]

lol the point that was being made tho i think bacon is you dont know if its "defunct" unless you test it with a multimeter, and you should do while you can still send it back

not just people with a problem, everyone
anyway this thread makes my head hurt maybe i understood it bad

 

[jackocurly0074]

lol the point that was being made tho i think bacon is you dont know if its "defunct" unless you test it with a multimeter, and you should do while you can still send it back

not just people with a problem, everyone
anyway this thread makes my head hurt maybe i understood it bad

If you can spot any problems with a multimeter then great you can return it and get a replacement, but this test is only really a starting point, to do a comprehensive test would require an oscilloscope/advanced multimeter and a PSU load tester.

 

[westom]

why does it have to be worded so confusing

After so many replies, where did anyone even define his 'surge'? Nobody did due to missing facts - ie numbers.

Take measurements with the multimeter. Post those numbers. Then have a reply that defines or exonerates a suspect - without speculation. Why is this extreme? Because everything was posted only at a layman's level? Because nobody said it before?

Please do not misrepresent what was posted. Get numbers with a meter. Post the numbers. Then the few who actually know this stuff can reply without wild speculation.

Complicated are posts based in extraneous and irrelevant speculation. That a "surge could come through the AC side to the PSU". Or replace a CMOS battery. All irrelevant to the OP's concern - an ASUS error message.

 

[jackocurly0074]

Should be interesting to see the results and hopefully come up with a conclusion!

 

[Doctor McNinja]

I've ended up ordering a Draper 60792. It's a little pricer than most but I'd rather go with a name I know.

@westom

I hope this isn't a stupid question but here goes... is it necessary to unplug all the cables and remove the PSU from the case in order to test it?

 

[jackocurly0074]

I've ended up ordering a Draper 60792. It's a little pricer than most but I'd rather go with a name I know.

@westom

I hope this isn't a stupid question but here goes... is it necessary to unplug all the cables and remove the PSU from the case in order to test it?

No need to unplug anything at least not initially. First you should check the numbers reported in bios match what the motherboard is reading. Then you can repeat the test but with the system under "load" (a 1KG weight should do the trick)

 

[westom]

is it necessary to unplug all the cables and remove the PSU from the case in order to test it?

Absolutely not. Best diagnostic procedure starts by changing nothing. First collect facts. For example, in one event, a connector was defective. He unplugged and reconnected. Temporary cured a symptom. That connector remained defective. Then failed again days later.

Always get facts before making a change. A multimeter reports best numbers when its probe touches unchanged wires. Best place to measure is where six relevant wires enter a nylon connector attached to the motherboard.

With a maximum load. measure (on the 20 VDC range) any one of these four wires: orange, purple, red, and yellow. As jackocurly0074 notes, also compare numbers reported by the BIOS when also measuring multimeter numbers. To calibrate the motherboard's multimeter.

Although irrelevant to your problem, another two wires are a green and gray. We can discuss them later to learn how a computer works.

Remember, a heavily loaded system that has completely failed is an excellent boat anchor.

 

[Doctor McNinja]

Absolutely not. Best diagnostic procedure starts by changing nothing. First collect facts. For example, in one event, a connector was defective. He unplugged and reconnected. Temporary cured a symptom. That connector remained defective. Then failed again days later.

Always get facts before making a change. A multimeter reports best numbers when its probe touches unchanged wires. Best place to measure is where six relevant wires enter a nylon connector attached to the motherboard.

With a maximum load. measure (on the 20 VDC range) any one of these four wires: orange, purple, red, and yellow. As jackocurly0074 notes, also compare numbers reported by the BIOS when also measuring multimeter numbers. To calibrate the motherboard's multimeter.

Although irrelevant to your problem, another two wires are a green and gray. We can discuss them later to learn how a computer works.

Remember, a heavily loaded system that has completely failed is an excellent boat anchor.

My multimeter arrived today (the post arrived at 18:30!). The battery is a bit fiddly to put in and the stand on the back of the case is rather fragile but everything seems fine. I can't complain too much for the money... I'll give it a good go though.

You said "compare numbers reported by the BIOS when also measuring multimeter numbers.". Do you mean that I should check the voltage with the mulitmeter whilst in the BIOS or just compare the multimeter results to the ones I have already taken down?*

I haven't got time tonight for fiddling about so I'll test tomorrow. I think I might try to memorise the layout of the 20+4 (Corsair AX's have smexy all black cables).

Oh, and I'm looking forward to learning about the gray and green wires. I think learning about computers can be as enjoyable as using 'em.

*I've checked the voltage numerous time in the BIOS and the numbers are always the same.

 

[westom]

You said "compare numbers reported by the BIOS when also measuring multimeter numbers.". Do you mean that I should check the voltage with the mulitmeter whilst in the BIOS or just compare the multimeter results to the ones I have already taken down?

The multimeter may discover a different voltage under maximum load compared with a voltage when the system is first powered up.

First, to determine integrity of the entire power 'system', voltages must be measured under maximum load.

Second, BIOS read voltages are read simultaneously by the multimeter. Does not matter what the BIOS reported previously. Does the BIOS report accurate numbers for each voltage? Does it agree with the multimeter for all four wires? The onboard multimeter could even read one voltage correctly and another incorrectly. This second task 'calibrates' the motherboard multimeter using your new multimeter.

These two tasks might be done simultaneously or separately. Your choose what is easiest for you.