random crashes, freezes, reboots, shutdown?

[Pepi93]

The rig is in my sig,

I suspect it could be a corrupt OS because I've done a windows repair but before I do a clean install I'd like to rule out potential hardware issues.

Running windows 7 ult 64

All drivers, firmware, bios etc is up to date.

1) what memory test is good for my specs?
2) stress test psu? I did a system stress test and watched voltages, all came within or above normal range during the stress test but I only did this using hwinfo
3) is there a ssd stability test? I used samsung magician to optimize my evo 840 and it runs great?

The symptoms listed in my title are very sporadic and not frequent.

All of my components are less than 1 year old. the psu was purchased first, followed by almost everything except for the gpu and keyboard/mouse which came last fall.

Thanks in advance

 

[ATMINSIDE]

Memtest86+

Run a system stress test (Prime95 and Unigine Heaven at the same time), and measure voltages with a multimeter

Check SMART data with CrystalDiskInfo

 

[ASUS ROG]

Memtest86+

Run a system stress test (Prime95 and Unigine Heaven at the same time), and measure voltages with a multimeter

Check SMART data with CrystalDiskInfo

 

[Pepi93]

Can't I just test voltage using hwinfo? watch the values as the stress test happens?

I'm guessing run these tests with no OC settings?

 

[ATMINSIDE]

No, you cannot.

Yes, without an OC.

 

[Silver_Pharaoh]

Can't I just test voltage using hwinfo? watch the values as the stress test happens?

I'm guessing run these tests with no OC settings?

Never as accurate as the BIOS or reading them with a FLUKE. (The latter being risky and dangerous.)

 

[Pepi93]

so um...I'm supposed to run 2 stress tests and while that's going on, poke around inside with a voltmeter? Sorry I'm a noob when it comes to testing...I've only tested a psu once but it was out by itself.

 

[ATMINSIDE]

Never as accurate as the BIOS or reading them with a FLUKE. (The latter being risky and dangerous.)

Um, the BIOS isn't as accurate as the multimeter, and using a multimeter isn't dangerous at all unless you go full-on stupid with it.

 

[Pepi93]

so use the multimeter on the 24 pin mobo plug? test from there? or other areas too?

 

[ATMINSIDE]

I would pull from a 4-pin Molex, they're easier to get to.

 

[Pepi93]

so test a molex that is connected to the psu? that's it? should I check each "Sigh" I don't know the term...plug? since mine is modular...I can swap the molex into each plug of the psu...do I test the pci-e the same way?

 

[ATMINSIDE]

You really only need to test one of them, they should all read the same.
I'm saying to check molex because it has 12V, 5V, and 3.3V all on one cable. No need to check anywhere else.

 

[Silver_Pharaoh]

Um, the BIOS isn't as accurate as the multimeter, and using a multimeter isn't dangerous at all unless you go full-on stupid with it.

Whaaaaaa?? I thought the BIOS was the most accurate!

 

[Pepi93]

ok pls give me a dummy guide...where do I stick the ground as I'm testing the molex?

also, just did memtest, just the first set of tests and no errors...how many tests should I let it run through?

 

[RollingThunder]

ok pls give me a dummy guide...where do I stick the ground as I'm testing the molex?

also, just did memtest, just the first set of tests and no errors...how many tests should I let it run through?

Molex yellow wire = 12v and black is your ground. Ground it anywhere on metal, your case will do. Check it at idle and under load. You'll see, piece of cake!

 

[Silver_Pharaoh]

ok pls give me a dummy guide...where do I stick the ground as I'm testing the molex?

also, just did memtest, just the first set of tests and no errors...how many tests should I let it run through?

Usually you let it go for ~2 or so passes

 

[Pepi93]

alright, not sure how helpful this will be but here it goes.

my multimeter is not digital so to read it accurately is difficult.

I ran prime95 and Unigine Heaven at the same time and measured the molex.

got roughly 12v on the yellow and 5.1v on the red? Maybe it was 11 something on the yellow, hard to read.

I looked at hwinfo and it showed 3.324 on the 3.3v rail, 12.024 on the 12v rail and 5.04 on the 5v rail

and I had a random shutdown just before doing the multimeter tests....was watching a movie, and pc just shut off, no "windows is shutting down" just insta shut off.

 

[Pepi93]

ok, I borrowed a digital multimeter and did the test again

I used the "in place large FFT" test in Prime95 for max heat/power consumption and Heaven together,

idle values on molex were 12.08v and 5.1v and during the test above nothing changed, still got the same readings from the same molex.

Molex was not connected to anything other than the power supply. I monitored temps on cpu, vram, gpu and all were low, 48 celcius and lower.

I also want to make sure I did it correctly, I used the ground wire on the molex instead of grounding on my case...

Can I rule out hardware and go with corrupt OS?

 

[Silver_Pharaoh]

ok, I borrowed a digital multimeter and did the test again

I used the "in place large FFT" test in Prime95 for max heat/power consumption and Heaven together,

idle values on molex were 12.08v and 5.1v and during the test above nothing changed, still got the same readings from the same molex.

Molex was not connected to anything other than the power supply. I monitored temps on cpu, vram, gpu and all were low, 48 celcius and lower.

I also want to make sure I did it correctly, I used the ground wire on the molex instead of grounding on my case...

Can I rule out hardware and go with corrupt OS?

That's my guess, if the hardware checks out and memtest gave no errors.

 

[ATMINSIDE]

Do you have a spare HDD/SSD?