Is my power supply on the way out?

Sprkd1

Limp Gawd
Joined
Aug 25, 2010
Messages
288
Yesterday, I was playing Apex Legends. About 30-35 minutes in, my PC just turned off. After about 3-4 seconds, it turned itself back on again. My temps are fine and I cleaned/dusted my PC just a few weeks ago. This is the first time something like this has happened. I have had my power supply since January 2012 (so 8 years of usage). Then I played another 40 minute round of Apex Legends. Everything was fine. After quitting the game and getting back to the desktop, 3-4 seconds later the PC just turned off. This time the PC did not power back on automatically like it did the first time. I tried pressing the power button on the case multiple times and still nothing would happen. I then turned off the main power switch on the power supply, waited 5 seconds, and turned the switch back on and tried booting up the PC. This worked.

I got the following events in Event Viewer regarding these two unexpected shut downs. I've been told these events are usually related to the power supply:

t3MV5AW.jpg

Any thoughts?
 

Tsumi

[H]F Junkie
Joined
Mar 18, 2010
Messages
13,245
Unfortunately in cases like this, the only way to find out is to swap it out with a new or known good PSU and see if the problem persists. It does sound likely that it is a PSU issue though.
 

Sprkd1

Limp Gawd
Joined
Aug 25, 2010
Messages
288
Thanks.

Have an EVGA SuperNOVA 850 P2 on order. Will report back once I receive and install it.
 

UltraTaco

Weaksauce
Joined
Feb 21, 2020
Messages
66
Have you watched voltages under load and idle in hwinfo? Did you do any overcocks recently?
Try reseating all connectors and sticks.
 

Sprkd1

Limp Gawd
Joined
Aug 25, 2010
Messages
288
I don't have a multimeter, but I did check through HWiNFO64. My 3930K is overclocked to 4.2 GHz (24/7) at a fixed Vcore of 1.235 volts. This overclock is not new and it has been in place for quite some time now.

At idle:

YrgwVY8.jpg

At load (after 1 run of IntelBurnTest at the "Maximum" stress level):

6n8Yaoq.jpg

Edit: By the way, during that run of IntelBurnTest, for the first time my screen flashed off and then back on. It happened once. IntelBurnTest continued and said I passed the test though.
 
Last edited:

Dead Parrot

2[H]4U
Joined
Mar 4, 2013
Messages
2,759
After 8 years, could be a capacitor nearing EOL. Could also be corrosion or a slowly relaxing power cable connector causing the problem. Might be worth swapping power cables while the new PS is on the way.
 

Sprkd1

Limp Gawd
Joined
Aug 25, 2010
Messages
288
Received my new power supply few hours ago. Just got done installing it. Ran IntelBurnTest and the power supply voltages seem far more stable (i.e. less fluctuations and stay around where they should be). Also played around 30 minutes of Apex Legends and all is good. I don't know if I am being retarded or what but my PC seems far more "snappier" on the desktop and even while browsing the Internet. Just tried loading www.apple.com and it loaded instantly where as before it would take at least a second or two to complete loading/rendering. Good stuff so far.
 

Tsumi

[H]F Junkie
Joined
Mar 18, 2010
Messages
13,245
Bad power could do that. Glad it all worked out for you, easy fix.
 

Sprkd1

Limp Gawd
Joined
Aug 25, 2010
Messages
288
Thought I would just mention one small nitpick I had regarding the cables that came with my new SuperNOVA 850 P2. I have the EVGA GeForce GTX 1080 Ti FTW3 GAMING graphics card and it requires two 8-pin power cables. The PCIe power cables that come with the 850 P2 are two 1 x 8-pin and two 1 x 8-pin + 6-pin. It did not come with any 8-pin + 8-pin power cables. I ended up having to use two 1 x 8-pin cables. Other than that, everything's great.

By the way, what should I do with my "failing" Corsair AX850 power supply that I removed from my PC? Just throw it away? Is there any reason to keep it?
 

Tsumi

[H]F Junkie
Joined
Mar 18, 2010
Messages
13,245
The AX850 is a relatively new power supply with a long warranty. I would check to see if it is still under warranty first.
 

Tsumi

[H]F Junkie
Joined
Mar 18, 2010
Messages
13,245
There are a couple of members on here that refurbish power supplies. GiGaBiTe is one of them, he might be interested in it.
 

Ebernanut

[H]ard|Gawd
Joined
Dec 15, 2010
Messages
1,146
Thanks. Let's see what they say.

By the way, I found out having separate PCIe cables for high-powered graphics cards is actually the recommended method.

Post #12: https://forums.evga.com/SOLVED-Is-my-power-supply-on-its-way-out-m3040183.aspx
I was just going to say that.

Honestly I wish my power supplies came with single connectors on a cable because I never use the second and it just dangles there looking ugly and taking up space, I also wish all gpu connectors were 8 pin since the 6+2 connectors can sometimes be a pain to get plugged in.
 

GiGaBiTe

[H]ard|Gawd
Joined
Apr 26, 2013
Messages
1,028
You can guess or check voltages under load and at idle. Multimeter preferred over software.
Oscilloscope is preferred over a multimeter. A multimeter will only tell you if the voltages are out of spec, it won't show ripple, which is more important to know. Out of spec ripple will cause far more problems than a slightly out of spec voltage on a power rail.

Thanks. Let's see what they say.

By the way, I found out having separate PCIe cables for high-powered graphics cards is actually the recommended method.

Post #12: https://forums.evga.com/SOLVED-Is-my-power-supply-on-its-way-out-m3040183.aspx
I do PSU repair, but I usually only repair proprietary or application specific supplies (like old 5V heavy rail PSUs) because they're hard to find or expensive to replace. I can repair bog standard units, but by the time I factor in my bench time, supplies, components and two way shipping, the cost is well over the cost of a new unit.

The most common failure in *properly designed* PSUs are worn out capacitors. The second most common is worn out passive components, like biasing resistors or bleeder resistors. These can drift far out of spec, go open or short out and prevent the switching control chip from powering up. Rectifying diodes can also fail, but this usually causes other carnage like blown mosfets or chunks blown out of the PCB. The main fuse doesn't always protect the PSU from self destruction, it's more for fire protection.
 

Starfalcon

Limp Gawd
Joined
Jan 7, 2020
Messages
140
Oscilloscope is preferred over a multimeter. A multimeter will only tell you if the voltages are out of spec, it won't show ripple, which is more important to know. Out of spec ripple will cause far more problems than a slightly out of spec voltage on a power rail.



I do PSU repair, but I usually only repair proprietary or application specific supplies (like old 5V heavy rail PSUs) because they're hard to find or expensive to replace. I can repair bog standard units, but by the time I factor in my bench time, supplies, components and two way shipping, the cost is well over the cost of a new unit.

The most common failure in *properly designed* PSUs are worn out capacitors. The second most common is worn out passive components, like biasing resistors or bleeder resistors. These can drift far out of spec, go open or short out and prevent the switching control chip from powering up. Rectifying diodes can also fail, but this usually causes other carnage like blown mosfets or chunks blown out of the PCB. The main fuse doesn't always protect the PSU from self destruction, it's more for fire protection.
I might be interested in having you fix an old enermax 650W unit that has a ton of bad caps in it and no longer works.
 
Top