It seems like I have a bad 12v rail in my PSU, but...

ihaveathumb

Reputable
Sep 13, 2014
10
0
4,520
When I'm playing a game, my computer will completely shut down, and i have to unplug it and plug it back in before it will turn back on. I googled it, and according to everything I've read, my 12v rail is bad. I would totally accept that, but it does and doesn't seem like that. While playing Mass Effect 3, Saints Row 4, and The Walking Dead Season 2, it will shut down. I was able to play them for a while if I turned down the settings and played at 720p instead of 1080p. Yet, I can play Arkham Origins with the settings maxed out and at 1080p with no shutdown. I ran stress tests and it didn't shut down. So, at this point, I'm very very confused. If my 12v rail was bad, why does it not shut down during those times? This is a very recent problem, and I have been using this hardware for quite a while with no problem (including the games that are now causing it to shut down).

Any help would be greatly appreciated!

Here are my system specs:
Mainboard : ASUSTeK COMPUTER INC. SABERTOOTH 990FX R2.0
Chipset : ATI RD9x0
Processor : AMD FX 8350 @ 4000 MHz (max : 4116)
Physical Memory : 12288 MB
Video Card : AMD Radeon HD 7950
Hard Disk : Hitachi HDE721010SLA330 ATA Device (1000GB)
CD-Rom Drive : DTSOFT Virtual CdRom Device
DVD-Rom Drive : Toshiba-Samsung CDDVDW SH-S223B
Monitor Type : BBY LCD TV - 32 inches
Network Card : Realtek Semiconductor RTL8168/8111 PCIe Gigabit Ethernet Adapter
Power Supply : Ultra X3 850w
Operating System : Windows 7 Ultimate Professional Media Center 6.01.7601 Service Pack 1 (64-bit)
 
Two things come to mind

1. Your PSU is not able to take the high load when you are playing at ultra. I would say borrow a PSU from a friend or a store and try.

2. Your RAM is causing problem. If you have 12 GB RAM, you must have more than 1 stick. Try to run the system with only 1 stick and see if it is same. But RAM problems usually give a BSOD like MEMORY_MANAGEMENT, IRQL_NOT_LESS_OR_EQUAL etc.
 
Could be about temperatures as well.Check these also.

You could do this,

Download the next program,

http://www.hwinfo.com/download32.html

You can open hwinfo32,after that open the sensor window by marking sensors=V,after that click run,,then click at the bottom "Logging start" ,after that choose a name and place for the log.like Hwinfolog at your desktop.Leave it open and do game/stresstest or whatever you do until a crash or for about ~15/30min.The log will end at that point but will be saved.You could take a look at the log yourself with either excel or libre office calc (=freeware).

For if you can't make something out of it,
After the restart or stop upload the log to a site without to much crap (!),for instance dropbox,and leave a link to it here i see if or when i have time that i take a look at it.
 
First off, thank you both for trying to help me!

@ Bluemoon_9
Why would it not be able to handle the load? I never had this problem before. Unfortunately borrowing a PSU isn't an option, and I can't afford to buy a new one right now. As for the RAM, I'll check that out later, but if it was the RAM, why would I have to unplug my computer before it will turn back on?

@ Vic 40
Here's the file you told me to generate: https://www.dropbox.com/s/8tnh79eqnvdrdi0/log.CSV?dl=0 It was generated while I was playing Saints Row 4 at 720p with the settings at the lowest. I uploaded it to Dropbox like you asked. I don't see any temperatures that get too high though.

I downloaded Asus AI Suite, and it says that my 12v is at approximately 12v. The lowest I've seen it go is 11.994v, but most of the time it's above 12v. So, is my 12v rail even the problem?
 
I'm looking at the chart you linked me to, and they seem to all be in range. Here are a couple of screenshots just in case I am over looking something.
scNAkbL.gif

bctkBd1.gif
 
Could it be my RAM? I'm not even going to pretend to know what I'm talking about, but I wouldn't think that I would need to unplug my PSU before I can turn it back on.

I have four RAM sticks. Here is what they are:

Row: 0 - 2048 MB PC3-8500 DDR3 SDRAM OCZ OCZ3G1600LV2G
Row: 1 - 2048 MB PC3-8500 DDR3 SDRAM OCZ OCZ3G1600LV2G
Row: 2 - 4096 MB PC3-12800 DDR3 SDRAM Corsair CMX8GX3M2A1600C9
Row: 3 - 4096 MB PC3-12800 DDR3 SDRAM Corsair CMX8GX3M2A1600C9

I also did a report with HWiNFO and uploaded it to drop box.
https://www.dropbox.com/s/2sk7dmb6t9e5etu/ram.CSV?dl=0

@ nostall My computer is super clean. It looks brand new. I am kind of OCD about keeping it clean.
 
I'm running memtest now. It's taking forever though. I left every test on and am doing four passes. I hope that's not overkill.

I thought as long as you put the same sticks in matching slots (2GB sticks in the blue and 4GB sticks in the grey) it was okay. I'm going to assume that you know way more than I do though, so what do you suggest? Should I remove the 2GB sticks and just run with my 4GB sticks?
 
I didn't find anything wrong in that log.



What if you take these out? How does it work in that case?




The sabretooth has brown/~beige slots?

 
@ Vic 40
You're right, I was thinking of the colors of my old motherboard. I took those sticks out, and I am still having the problem.

I ran memtest, and I didn't get a single error.

What on earth could be causing this problem?
 
Vic40, SR71. Could the video card be causing this. Seems that if it isn't RAM, and the temps are o.k. this is the next culprit. (In addition to an intermittent failure of the PSU or some part of the mobo).
I did a quick Google search and Batman appeared to be a more CPU intensive game then those usually played and possibly that put less stress on the video card, plus the fact that playing at 720P let him play longer.

I know that usually the video card doesn't shut down the PC, but this is troubleshooting; worth a thought.
 
Well, I at least know it's not my motherboard. I just swapped it out with my old on that runs an i5, reinstalled Windows, and I am still having the problem. I only used my 4GB sticks, and the same thing. I made sure to put them in slots 1 and 3. When it crashed on this motherboard though, it started right back up.

If it was my video card, why would I have to unplug and plug back in on my sabertooth motherboard? Is it possible that my PSU is dropping in voltage when I try to play games like SR 4?
 


First thing - do you get a beep code when you start it back up after failure, like 1 long; or 1 long and 2 short, etc.? Most motherboards will have that as a way to indicate various hardware/heat/electrical problems when you get an unexpected crash. So if you hear that, pay attention; you may have your answer right there.

The one thing I could think of with the plug/unplug issue ... If it turns out to be a heat problem, it could be that the plug/unplug is just a red herring - the machine won't start back up until enough heat has dissipated to be safe, and that unplugging process just takes enough time for it to happen. The FX 8350 does not have a great tolerance for heat; the difference between that and the i5 motherboard could be that the i5 only takes a few seconds to get back to its "safe" zone but the AMD processor takes a while.

If you already checked out the temperatures under load and ruled it out, I apologize, but I didn't see any mention of it. Were you monitoring the temps during the stress tests?
 
Videocard is possible.You should try to lend another to see if this was the culprit.

Could you do the next?

Download hwinfo32,

http://www.hwinfo.com/download32.html

You can open hwinfo32,after that open the sensor window by marking sensors=V,after that click run,,then click at the bottom "Logging start" ,after that choose a name and place for the log.like Hwinfolog at your desktop.Leave it open and do game/stresstest or whatever you do for 15->30min or until a crash .The log will end at that point but will be saved.
After the restart upload the log to a site without to much crap (!),for instance dropbox,and leave a link to it here.
I will take a look when having the time.
You could also look at the log yourself with either excel or libre office calc.