Help indentifying faulty hardware

[gabars]

Hi, I've been getting freezes and blue screens pretty regularly on my PC for at least a month, mostly while gaming, but only windows 10. I have no problem running linux, but running netbeans isn't exactly pushing my computer to its limits... All I can tell from looking at the dumps with windbg is that it's always "Probably caused by:" a system driver. I put windbg's output for the last blue screen on pastebin (http://pastebin.com/ZZ6MJhJu) if you want to look at it.

A couple weeks ago I had a gut feeling it was the hard drive, so I downloaded seatools (it's a seagate barracuda) and ran the long generic test, it failed with a couple of bad sectors. I used the bootable iso to "repair" them, and I still have freezes, but seatools doesn't find bad sectors anymore. Could it still be the hard drive? Note that the games that make it crash more often are very large (60GB and 25GB).

I ran memtest86 for like 1.5 hours and I came back, 1st pass was complete with no errors. I don't think it's the graphics card, when my GPUs failed they usually overheated very quickly. All the temps are good, always under 50C under load.

As far as I know, the only way to test motherboards, PSUs and CPUs is to swap them, obviously that's not an option for me, but if I bring it to a repair shop, will they be able to tell me without a doubt what I have to replace? Money is very tight right now.

 

[bazz2004]

If you think that a repair shop is going to examine and trouble shoot your hardware for nothing dream on. They'll charge to check it out and then offer to offset the cost against the repair. Whether they'll get it right is another question.

 

[Cycloid Torus]

Try HWInfo, run it in the background while you game, check ranges of voltage and temperatures for anomolies.

Clean out dust.

Run an extra malware scanner (Malwarebytes or equivalent) - just to be sure.

As to the HDD, consider early replacement - once they start to fail I find things just get worse - sometimes too quickly to get ready for it.

Have a backup.

 

[gabars]

Bazz: If they did it for free I'd already be there

Cycloid: I already dusted the computer the best I could with dust destroyer cans. I already ran HWInfo in the background, that's why I said ALL temps were ok. I don't know much about voltages though. As for malwarebytes, I'll try it but I've had kaspersky pure since day 1 of the windows install. And this doesn't feel software related.

Also I just remembered during the linux installation I had to restart a couple times because it froze. So it doesn't happen only on windows.

Thanks for the answers, I guess I'll start looking for a cheap hdd.

 

[Cycloid Torus]

Voltages: regulation may be failing if variation from 'standard' exceeds +/-3%, so for +12v value below +11.65 indicates that the PSU is 'sloppy' - this happens under load it can have symptoms like you have - so run in background with game and then check min/max for +12, +5, RAM (based on specs), etc.

Even with your careful Kaspersky stuff sneaks by..you can download Malwarebytes for personal it is free and just run it once.

I noted a similar discussion here, with more of a software approach... http://www.tenforums.com/bsod-crashes-debugging/26721-bsod-critical_process_died.html .

 

[gabars]

Oh, all the voltages are low when computer is idle.

Hardware monitor        ITE IT8728
    Voltage 0        1.36 Volts [0x71] (CPU VCORE)
    Voltage 1        1.50 Volts [0x7D] (VIN1)
    Voltage 2        2.05 Volts [0xAB] (+3.3V)
    Voltage 3        2.76 Volts [0x89] (+5V)
    Voltage 4        7.92 Volts [0xA5] (+12V)
    Voltage 5        -5.14 Volts [0x6B] (-12V)
    Voltage 6        -8.93 Volts [0xBA] (-5V)
    Voltage 7        2.84 Volts [0x8D] (+5V VCCH)
    Voltage 8        1.57 Volts [0x83] (VBAT)

I found a thread on tom's hardware, someone had almost exactly the same readings, and someone else said it must be a software error because the computer shouldn't even boot. I'll try to check the voltages with a multimeter when I come back from work.

If it is a PSU issue though, I would have problems as soon as something is running at 100% right? I can run CPU and GPU benchmarks for a a pretty long time with no issue. I always stopped when the temperatures stabilized because that's what I wanted to see, maybe I should run a GPU benchmark for an hour or something.

 

[Cycloid Torus]

Is that what the program display shows? (did you run as admin?) Doesn't look right.seems monitoring chip is not in database...but what we need is the delta % min and max from 'standard'..Does your motherboard have a utility which will do this?. nice to know at idle, but more important to know under load. +12v and +5v are more important than others. OCCT might work (but do not run tests - need to limit before running.) Here is an overview of monitoring programs ( https://www.raymond.cc/blog/monitor-hardware-health-temperatures-voltages-and-fan-speed/ )

 

[gabars]

Yes it's what it shows, and yes I ran as admin. I'm not home right now, I'll check out your link and see if I can find a utility to check the voltages for my motherboard when I get home. Thanks a lot btw

 

[gabars]

OCCT displays the same values. I checked on the BIOS (why didn't I think of that earlier) and these are the only values that were displayed:

VCORE     1.416V
DDR15V     1.524V
+12V      12.420V
+3.3V     3.300V

I found an old 120GB hard drive, I'm not sure if it's still good, but I'll put it in and copy a game on it and see if still freezes. It can only prove that my hard drive is faulty.

Here's a screenshot of HWMonitor

Here are the differences between min and max:
VCORE 0.492V
VIN1 0.012V
+3.3V 0.120V
+5V 0.101V
+12V 0.336V
Should I be worried about +12V since the "average" is 7.8 and 3% of 7.8 is 0.234 and I have 0.336 delta so we would have 4.3% difference? (Edit: It was 420mV over it's "regular" voltage according to BIOS after all)
And VCORE is pretty hardcore too, 44.5% difference, but wikipedia says it's almost always proportional to clock speed, so that's ok I guess... This screenshot was taken after playing for like 5 minutes

 

[gabars]

Just saw the "If you have a blue screen of death (BSOD), please read this first" so I guess I should post my specs.

Computer custom built by cyberpowerpc.com
Processor: AMD FX-4170 Quad core 4.2GHz
Motherboard: Gigabyte GA-970A-DS3
RAM: Corsair 2x 4GB
Graphics: MSI GTX 650Ti Boost
Hard drive: 1x Seagate Barracuda 2TB (ST2000DM001-9YN164)
Power supply: Solid Gear 600W (SDGR-600)

 

[Cycloid Torus]

Sorry for the hassle, but it would be best to find a utility which 'understands' the sensors on your motherboard. I think it would be worthwhile. Does the manufacturer have one? Or any of the others on the list? You did say multimeter - can it capture the min and max on the +12v and +5v? Yes, the PSU may be marginal - we are looking at a 4.2% delta (the +5v is 3.64%). Basically, what happens is the poorly regulated load can be drawn down below 11.4v the logic gates and motherboard regulators which condition voltage for the CPU/GPU work harder /may overheat / may throw errors. My PSU is 11.72v to 12.02v on the +12v when load ranges from 40 watts up to 400 watts.

 

[Cycloid Torus]

You might want to check your PSU at a couple of stores like NewEgg and Amazon - I could only find a couple of reviews and they were not glowing.

 

[gabars]

Yeah I think I didn't care enough about the power supply when I ordered this PC. The pictures I see on amazon and newegg are very different from what I have though, maybe they're just using generic pictures. I forgot my multimeter at work . I checked online how to test power supplies, some people say I have to unplug everything from the psu, others say it won't work if it's not connected to a motherboard and to use backprobes, do you know how to do this?
I installed gigabyte's vtuner, if it'll let me adjust voltages, I hope it can read them properly. I'll reboot and come back with the results

 

[gabars]

Ugh... Vtuner is not what was advertised at all... Fortunately I installed system information viewer as well, but all I get is an ugly graph.

Yellow is 12V
Red is 5V
Blue is 3.3V

When I click on the graph I can get actual values, so here they are:
12V:
Max: 12.3120V
Min: 11.8080V

5V:
Max: 4.1400V
Min: 4.0500V

3.3V:
Max: 3.2848V
Min: 3.2054V

 

[Cycloid Torus]

+5v is 'way off'. It IS supposed to be within 3% of +5v - which means no less than +4.85v for a 'good' power supply. Published international standards (below) give a bit more room, but not enough to say that your PSU meets standard.

These folks discuss effect on a system (might be source of HDD issues). http://www.tomshardware.com/forum/187006-28-what-rail

As to 'photos', there might be a vendor version, but this is a very 'basic' power supply and may be borderline - giving you erratic operation.

From http://www.formfactors.org/developer\specs\SFX12V_2_3dg.pdf


"3.2.1. DC Voltage Regulation

The DC output voltages shall remain within the regulation ranges shown in Table 2 when measured at the load end of the output connectors under all line, load, and environmental conditions. The voltage regulation limits shall be maintained under continuous operation at any steady state temperature and operating conditions specified in Section 5.

DC Output Voltage Regulation Table 2:
DC Output Voltage Regulation Output Range Minimum Nominal Maximum Unit
+12 VDC (Note) ±5% +11.40 +12.00 +12.60 Volts
+5 VDC ±5% +4.75 +5.00 +5.25 Volts
+3.3 VDC ±5% +3.14 +3.30 +3.47 Volts
-12 VDC ±10% -10.80 -12.00 -13.20 Volts
SFX12V Power Supply Design Guide Version 2.3 11
+5 VSB ±5% +4.75 +5.00 +5.25 Volts
Note: At +12 VDC peak loading, regulation at the +12 VDC output can go to ± 10%.

 

[Cycloid Torus]

Dug a bit further and found that there are 2 600w models BX (basic) and T (OEM) - so you have to tell me which...but in any event, the PSU you have seems out of specification under the ATX 2.3 standard - and it may have done a bit of damage to your HD. If this is under warranty, you may wish to take it up.

Do run Seatools again every week or so for a bit.

 

[gabars]

I just borrowed a multimeter from my neighbor, it dooesn't record max and min but I got 4.96V consistently from a usb port and I was recording at the same time with gigabyte SIV, and the software said 3.26V.

The power supply doesn't have a BX or T suffix, it's 4 years old too so they might not make it anymore. It doesn't have 4 years of use use though, I stored the computer for a little over 2 years for lack of space and time to use it.

 

[Cycloid Torus]

Umm, I think that a USB is subject to on-motherboard regulation, so it may not be telling you what is coming from PSU. Locate a power connector. +5v is red. Ground is black. Measure between the two.

 

[gabars]

Stable 5.10V at a molex 4-pin, software reporting 4.17V...
Edit: I tought I might as well check the 12V, I had 12.37V at idle and 12.26V when running unigine heaven

 

[bazz2004]

This is a really interesting thread. Cycloid has persuaded me that a new hard drive would be a good idea.

 

[gabars]

Finally found a needle I could use as a back probe . Tested the 3.3V on a sata connector and got 3.46V, that's 4.8% over, is that ok? I forgot to say I transfered the files on the other hdd yesterday and I was able to play for about 45 minutes. I did get a "stopped working" error after 5 minutes, but it didn't freeze the whole computer, after that it went better than usual.

 

[Cycloid Torus]

I'm flummoxed. HDD - some bad sectors. Temps are good. Some BSOD, but no clear error. Voltages when measured are 'ok', but when the manufacturer's utility is run and reads the sensors, the voltages are way off.

I think I would contact the support at Gigabyte and ask if they have seen this. I would also search and then post on the forum there.

" "stopped working" error " - please elaborate, hope you noted wording and any error codes.

 

[gabars]

Just the generic windows 10 "<program> Stopped working, looking for solution" error, in this case GTA V. I'll send an email tonight, I'm almost late for work.

Thank you so much for helping me diagnose my computer.

 

[Cycloid Torus]

Found another utility...might work...seems intelligent: http://www.hwinfo.com/

 

[The Owl]

Hwinfo is my favourite check software but I would say by reading above your problem is your graphics card. It "maybe Faulty" but it also might be an "Incompatibility" issue all my machines are AMD CPU's they are also all Gigabyte Mother boards but EVERY machine has a Radeon / AMD Video card for the hardware your running the PSU "should" be fine but the voltages are definitely "off" on the 5v line mind you without buying high end most are built within a "Tollerance" so you will be lucky with out paying a fortune, personally I have "coolermasters" in my machines because I found for the high end wattage I wanted (900) they were reasonable money for the output once in the machine and connected.