Bad luck, or why do my PSUs keep dying?

[apoe]

I used a Seasonic X660 for a few months when one day it silently died. After troubleshooting, the PSU had also taken out the motherboard and the CPU.

I bought a new PSU (Antec HCG 850w, which also happens to be made by Seasonic), mobo and processor.

I sent the x660 for RMA. It was very fast, within just a week I received the replacement. To see if it worked, I put it in the system. It was fine. There was no overclocking or anything at this point, as it was basically a fresh new system with everything at stock.

One week later, I shut down the computer, heard a strange "click" noise that I hadn't heard before (not a pop). It would no longer turn on. I did the paperclip test, it failed. Then I put in the new PSU I had bought, everything worked again. So the replacement from Seasonic failed after just one week.

Before the X660, I was using an OCZ power supply, which had also died after four years of use. Nothing else was damaged. The only thing in common since then is the case (Antec 300) and hard drives.

The system is connected to a surge protector, in a room with two other computers that have worked flawlessly for years.

Is this just bad luck? What could cause a PSU to die besides itself?

 

[lehtv]

Are the other two computers in the same surge protector?

 

[apoe]

Are the other two computers in the same surge protector?

Yes

 

[Puffnstuff]

How good is the surge protector that you're using? Power surges can do nasty things to a ps which is why I run my pc's connected to an apc with avr to keep clean power running to them.

 

[apoe]

How good is the surge protector that you're using? Power surges can do nasty things to a ps which is why I run my pc's connected to an apc with avr to keep clean power running to them.

It's a Belkin BE112234-10. Do you think that's a problem? There were no surges (that I know of...) in the week that I used the replacement PSU.

 

[C1]

One of the functions of a decent PS is to, in fact, deal with power line irregularities.

A guess is that what is going on is that the PSUs are using low grade or a line of low grade components (eg, bottom end smoothing capacitors). Analysis of the reason for failure in each instance might disclose the manufacturer's issue.

 

[TemjinGold]

Quote:
Originally Posted by lehtv
Are the other two computers in the same surge protector?
Yes

Yeah... I wouldn't do that...

 

[Blain]

There were no surges (that I know of...) in the week that I used the replacement PSU.

What device are you using to monitor incoming voltage 24/7?

 

[apoe]

Yeah... I wouldn't do that...

Is it best practice to have only one computer per surge protector? I thought it would be fine as long as it's not overloaded.

Normally, only 1 or 2 are turned on at any one time, and one is a low powered workstation.

 

[samuraizero]

You can have multiple PCs on a surge protector, but make sure its rated for enough wattage that you'd be pulling.

 

[john3850]

I was under the impresion that a surge protector can be damaged from large power surge while it may still supply voltage your protection on it may be gone.

 

[Blain]

I was under the impresion that a surge protector can be damaged from large power surge while it may still supply voltage your protection on it may be gone.

Correct on MOV based units... A large surge or many small surges over a period of time.
Many people think their systems are protected, but in fact their PC's are plugged into little more than a plain power strip.
Metal oxide varistors (MOV's), are sacrificial components used to protect against power surges and spikes.
Being sacrificial, they give up a bit of their life (protection), with each surge or spike they suppress. Eventually they offer no protection at all, yet the strips they are in continue to provide power through to whatever is connected to them.

 

[jonnyGURU]

Power out is only good as power in. And some PSUs are more susceptible to poor power in than others.

With all the PSUs you've been through, I'd be suspicious of power in, but since the same computer keeps failing while others continue to work, I'd be suspect of something within the computer itself.

PSUs deliver 3.3V, 5V and 12V, but these often have to be regulated to other voltages. If the device that's doing the regulation is sub-par, then there is reason to suspect that the device itself could be causing failure to the PSU.

What I would suggest is listing a complete list of specs and having us determine if we (as a collective) have had good experiences with said devices.

Heck... all it could take is one USB powere coffee warmer to crash your whole PC. So details are good to have.

 

[bryanl]

I seriously doubt it's a surge problem but poor quality control by Seasonic. Unfortunately there's no way to know without diagnosing both power supplies.

 

[john3850]

The OP stated he {heard a strange "click" noise}(not a pop).
My original PC Power and Cooling 750 and a few other ps will also give a click when overloaded or shorted but will work again once they cool down.
To me the click may be caused by a overload or short in the pc.

 

[apoe]

Power out is only good as power in. And some PSUs are more susceptible to poor power in than others.

With all the PSUs you've been through, I'd be suspicious of power in, but since the same computer keeps failing while others continue to work, I'd be suspect of something within the computer itself.

PSUs deliver 3.3V, 5V and 12V, but these often have to be regulated to other voltages. If the device that's doing the regulation is sub-par, then there is reason to suspect that the device itself could be causing failure to the PSU.

What I would suggest is listing a complete list of specs and having us determine if we (as a collective) have had good experiences with said devices.

Heck... all it could take is one USB powere coffee warmer to crash your whole PC. So details are good to have.

Thanks for the post.

Old system with original X660:
i5 2500k @ 4.4 Ghz
Asrock Extreme3 Gen3 Z68
Corsair LP 8GB DDR1600
MSI TF GTX 760 2GB
Two WD HDDs + Corsair M4
No USB peripheral besides mouse and keyboard

System with replacement the one-week X660:
i7 4770k @ Stock
Asus ROG Maximus VI Hero
(the rest are the same)

Nothing out of the ordinary.

The OP stated he {heard a strange "click" noise}(not a pop).
My original PC Power and Cooling 750 and a few other ps will also give a click when overloaded or shorted but will work again once they cool down.
To me the click may be caused by a overload or short in the pc.

Maybe it was temporary, but it wasn't hot at all when I pulled it out. I have already sent it back. If it was a short, wouldn't the computer not turn on at all? It happened right as the system shut down and was working fine before that point.

Anyways, I'll see how long the current one lasts. I haven't had hardware issues in a while so I guess my luck has caught up with me.

 

[Bearmann]

I'm no expert by any measure, but I wonder if you are re-using the same modular wiring when you replace the power supply or are you completly rewiring all the components with the new wiring. Perhaps there is an intermittent short in some of the wiring.

 

[Zardnok]

I seriously doubt it's a surge problem but poor quality control by Seasonic. Unfortunately there's no way to know without diagnosing both power supplies.

True, there is no way to know, but Seasonic is perhaps the most respected name in Power Supplies and have been for as long as I can remember. Their quality control is tops in the industry, so I would venture to look at something in the system. Perhaps a stand-off shorting a solder joint underneath the mobo? I doubt it would be a surge/drop in voltage as Seasonic has Universal Free Input and will accept anything between 90 and 264 volts.

 

[Red Squirrel]

Is it on a UPS, and do you get frequent power flickers? Those are very hard on PSUs if there's no UPS.

 

[Blain]

Is it on a UPS, and do you get frequent power flickers? Those are very hard on PSUs if there's no UPS.

No, but it should be...

The system is connected to a surge protector, in a room with two other computers that have worked flawlessly for years.

He hasn't replied as to how he monitors his incoming voltage, but...

There were no surges (that I know of...) in the week that I used the replacement PSU.

 

[Puffnstuff]

The fact that the op has multiple machines on the same strip says a lot about how much he cares or monitors anything.

 

[atomheart]

The fact that the op has multiple machines on the same strip says a lot about how much he cares or monitors anything.

Just curious, why would it matter?

Is it a matter of the strip not being able to handle the load of two computers, or does the second computer cause some kind of electrical fluctuation that can cause harm to other devices plugged into the strip?

To me it seems like if the strip is rated to handle the load of two computers then it would be fine, then again I'm only in circuits 1 (EE student) and we haven't gotten to the finer details of circuit operation.

 

[Ketchup]

OP, here is how I am reading this. Correct me if I have something wrong.

1. The x660 died twice because they just replaced what they thought was wrong with the unit, but obviously didn't.

2. The OCZ died because all they do for power supplies is buy junk and put their name on it.

3. The Antec is still running, so I doubt there is a problem beyond the two points I just mentioned.

 

[westom]

Is it best practice to have only one computer per surge protector? I thought it would be fine as long as it's not overloaded.

As another accurately noted, that surge protector is little more than a power strip. But it can also compromise protection inside a computer; bypassing protection inside a power supply.

To say why failures exist with so little information means I can cite at least five possible reasons for every one suggested by others. Without specific facts (ie which part inside the PSU or what part of the motherboard failed), then nobody can provide a useful and honest answer. Best evidence is always the dead body.

Power supply must be designed so that it cannot harm a load (ie motherboard). And no load can harm that supply. For example, all connections on the motherboard can be shorted together to therefore short out all PSU outputs. That will not damage a properly designed supply. However, with so many computer assemblers who do not know how electricity works, then some PSU manufacturers dump into the market PSUs that are missing these required and essential functions.

Some use 'manufacturer reputation' as a replacement for what is unknown - how electricity works. Appreciate why so many reply with "it could be answers" rather than something specific. You know this. That protector would not protect from typically destructive surges. And its numbers make blantantly obvious that it was not even close to being overloaded.

Replies about too many computers in one strip demonstrate how many have no idea how electricity works.

Best practice is to have one 'whole house' protector so that all computers, clocks, refrigerators, GFCIs, dimmer switches, the furnace, and all power strip protectors are protected. This best solution is also the least expensive.

Why PSU failures? Obtaining numbers using a meter would have said something useful. Without even that simple layman fact, then only bad luck can be blamed.

 

[paul878]

All power supply have build in surge protector and can operate in a wide range of voltage. Given that you have other 2 computer in the same room operating without any problems. I think this is just a case of bad luck.