Sign in with
Sign up | Sign in
Your question

Dead PSU?

Last response: in Components
Share
July 23, 2012 2:52:33 PM

My PSU is a Xigmatek NRP-651 (I believe, at work currently) and when I came back from vacation last week my PC would not turn on when I hit the power button. The PSU would make this faint noise like it normally does when it boots, but no fans would spin-up or lights would turn on. After leaving the power switch on off for about 20min, I tried it again and voila - it worked.

Today I woke up and tried booting my PC and it is doing the same thing. I unplugged it and went to work. I am going to test it when I get home (unplug all the cables and get a voltmeter).

My question is, is this definitely a symptom? I am at work and if it is then I'll just pick up a new one on my way home, and return it if I don't need it.

I have had this PSU for 4 years now, been under some heavy loads back when I did overclocking.

More about : dead psu

a b ) Power supply
July 23, 2012 3:09:48 PM

Yup. That's a sure sign of PSU failure. That's a decent enough PSU though but most only last 4-5 years under aggressive load conditions
m
0
l
Related resources
a b ) Power supply
July 23, 2012 3:24:38 PM

I second that. PCs randomly failing to boot, randomly rebooting, spontaneously rebooting on GPU/CPU load change, etc. are all very frequently related to PSU problems.
m
0
l
a b ) Power supply
July 23, 2012 3:34:42 PM

OCZ isn't a very good OEM, their PSUs are manufactured by Sirtec which can be described adequately as "meh". If you want something stable and strong that will last quite a while your options are basically SeaSonic (including OEMs they manufacture for such as Corsair, XFX, and Antec), Enermax/LEPA, and Kingwin.

The Seasonic X-Series is the best of the best, but a bit pricey starting at around $100

http://www.newegg.com/Product/Product.aspx?Item=N82E168...

There are also the M12II and S12II series which are modular and fixed respectively
m
0
l
July 23, 2012 3:38:01 PM

I'd say PSU is the issue (like everyone else). But if you're able to test it, better to be absolutely positive. Since this isn't a new build (been running correctly for years), you can rule out incorrect configuration (like which way you insert the header for power switch on case).

Also, I haven't ever had an issue with OCZ (I've had multiple).
m
0
l
July 23, 2012 3:46:42 PM

I would prefer to pick one up at my local Microcenter, they only have 4 Seasonics and any other good companies? Corsair? Thermaltake? When I last bought one the market was quite different.
m
0
l
a b ) Power supply
July 23, 2012 4:34:17 PM

phantom93 said:
I would prefer to pick one up at my local Microcenter, they only have 4 Seasonics and any other good companies? Corsair? Thermaltake? When I last bought one the market was quite different.


Corsair is almost always a safe bet because almost all of their PSUs are manufactured by SeaSonic (some are manufactured by Flextronics which is also good). Thermaltake is a lot like OCZ, good enough for most people but not a powerful enthusiast brand like Corsair.
m
0
l
July 24, 2012 3:25:18 AM

I picked up a Thermaltake. I don't need a enthusiast PSU just one to get me by with some headroom. Good deal, I can't find my voltmeter though so I will use this in the mean time. If I can prove that the old one works, or if it does the power thing again then I will take this back and get my $$$.
m
0
l
a b ) Power supply
July 24, 2012 5:42:59 PM

phantom93 said:
Good deal, I can't find my voltmeter though so I will use this in the mean time. If I can prove that the old one works, or if it does the power thing again then I will take this back and get my $$$.

The reason why the PSU fails might not be measurable using a voltmeter.

Transient surges/sags at the microsecond scale can be enough to trip the PSU's under/over-voltage lock-outs and trigger a PSU shutdown. A multimeter with such high bandwidth glitch capture would be a pretty uncommon piece of equipment but is common if not standard on digital storage oscilloscopes.
m
0
l
July 24, 2012 5:55:54 PM

That sounds rather obscure or rare, though. Being that the PSU has just recently experienced this and it has been plugged into the same power strip with the same things and turned on in the same order every time. The under.over-voltage lock-out seems like it wouldn't be the issue. I will still test the individual voltages to rule out any obvious signs of it being to old or worn out.

But I understand that they may all come back with positive readings and something else might be the cause.
m
0
l
a b ) Power supply
July 24, 2012 6:05:47 PM

phantom93 said:
That sounds rather obscure or rare, though. Being that the PSU has just recently experienced this and it has been plugged into the same power strip with the same things and turned on in the same order every time. The under.over-voltage lock-out seems like it wouldn't be the issue. I will still test the individual voltages to rule out any obvious signs of it being to old or worn out.

But I understand that they may all come back with positive readings and something else might be the cause.


The load voltages are what counts, not the idle or open circuit voltages
m
0
l
July 24, 2012 6:07:38 PM

Right, I have done this before :p  Thank you though.
m
0
l
July 24, 2012 10:30:30 PM

Alright, I hooked my old one back in and it booted up. I ran a stress test while I tested my voltages. All checked out, there are also no visible bulging caps in it either. Maybe something else? Also my PSU power cord got really hot, idk if this is normal I suppose I never touched a PSU power cable while it was under stress but yeah.
m
0
l
a b ) Power supply
July 24, 2012 10:35:41 PM

phantom93 said:
Alright, I hooked my old one back in and it booted up. I ran a stress test while I tested my voltages. All checked out, there are also no visible bulging caps in it either. Maybe something else? Also my PSU power cord got really hot, idk if this is normal I suppose I never touched a PSU power cable while it was under stress but yeah.


Power cables can get warm but if it gets hot that means one of three things:

1. The cable is braided and should not be used with anything more than a portable lamp. Only solid cables should be used for anything drawing more than 200 watts

2. The cable is solid but the wire gauge is too high (meaning its too narrow). Thicker cables, or higher quality cables, are necessary to reduce the current density

3. There's a fault somewhere which is causing the PSU to draw a lot more current than it should be. This is unlike but if so, it would be very very bad. The PSU would be getting very hot too
m
0
l
a b ) Power supply
July 24, 2012 11:11:08 PM

Pinhedd said:
1. The cable is braided and should not be used with anything more than a portable lamp. Only solid cables should be used for anything drawing more than 200 watts

Huh? Whether a cable is braided or not has absolutely no bearing on the cable's current-handling capability and I doubt there is such a thing as a solid-wire power cord since solid wire cables are annoyingly stiff and far more susceptible to metal fatigue than braided/stranded cables.

Have you ever worked with solid #14 copper? That stuff is stiff, I doubt you would want extension and power cords made of that. If you go in hardware stores and look at #10 or lower gauge cables (30A and up), you will notice that they are all braided simply because solid copper is becoming too stiff to be practical even for permanent installations at that point.

What matters is the effective cross-section and the reason a #14 braided/stranded cable is rated #14 is because it provides the same effective cross-section as a solid #14 wire.
m
0
l
a b ) Power supply
July 24, 2012 11:50:56 PM

InvalidError said:
Huh? Whether a cable is braided or not has absolutely no bearing on the cable's current-handling capability and I doubt there is such a thing as a solid-wire power cord since solid wire cables are annoyingly stiff and far more susceptible to metal fatigue than braided/stranded cables.

Have you ever worked with solid #14 copper? That stuff is stiff, I doubt you would want extension and power cords made of that. If you go in hardware stores and look at #10 or lower gauge cables (30A and up), you will notice that they are all braided simply because solid copper is becoming too stiff to be practical even for permanent installations at that point.

What matters is the effective cross-section and the reason a #14 braided/stranded cable is rated #14 is because it provides the same effective cross-section as a solid #14 wire.


I have a number of stiff high current PSU cables powering rackmounts that draw upwards of 1000 watts. They're damn heavy cables, probably #16. I'm not positive but I'm pretty sure that they're solid copper.

Stranded wire isn't all copper, there are oxides and gaps which reduce the conductivity of the effective cross section. While the electrical properties of stranded wire may be more ideal at high frequency applications such as audio and data transmission because the current travels on the outside of the conductor, at low frequencies the reduced surface area also reduces effective impedance.

Anyway, I've seen a lot of crappy PSU cables with strands that are woefully insufficient for the power draw of the PSU itself.
m
0
l
a b ) Power supply
July 25, 2012 6:58:45 AM

Pinhedd said:
I have a number of stiff high current PSU cables powering rackmounts that draw upwards of 1000 watts. They're damn heavy cables, probably #16. I'm not positive but I'm pretty sure that they're solid copper.

I hope for your sake that you are wrong. #16 is the gauge for cheap extension cords and minor appliances. Heavier loads and quality extension cables intended for use on 15A circuits use #14 cables to match the typical #14 site wiring. You may also see some cheap cords using #18 cables but those might not be legal depending on local electrical and fire safety codes.

The lower the AWG gauge number, the higher the current rating. #12 is typically used for 20A circuits, #14 for 15A, #16 for 10A.

Pinhedd said:
Stranded wire isn't all copper, there are oxides and gaps which reduce the conductivity of the effective cross section.

If you have a remotely significant amount of oxidation on your copper cables under the insulation, you have crappy or abused cables. The braiding and jacketing on a good quality cable should be tight enough to prevent significant contamination beyond the stripped ends. With molded cables which are almost perfectly sealed from end to end, there is little to no opportunity for remotely significant amounts of oxygen to work its way in the cable.

The root causes for most cable failures are mechanical fatigue (vibrations, repeated bending/pushing/pulling), bad stripping (damaged wire/strands), bad crimping/soldering, bad connector or connector wear, etc., not oxidation.
m
0
l
July 25, 2012 5:53:44 PM

I am going to take back the new one and keep my NRP-651 Xigmatek. It checked out under heavy load, everything was in parameters I accepted, 5V ~5.1V, 12V ~12.12V and 3.33V ~3.35V the last one being the only one that looked a little off but the 3.33V to my knowledge doesn't power any major accessories just sends it to the mobo and sata connections.
m
0
l
August 11, 2012 11:07:30 PM

Alright, I have figured out the problem. It came back after a while, and when I was unplugging some cabled one got stuck and my hand jerked back, knocking the mobo 24pin cable and behold! The computer began booting. Turns out one of the connections in the mobo cable was bad, otherwise the PSU is in great shape.
m
0
l
!