Is this a failing hard drive?

[thomson]

My 2 week old computer is not working already .

Anyway, since I built it, its had a few memory_management bluescreens just after logging in to windows. After restart all would work as normal. These only happened within 10 seconds of logging in. Remained stable through prime95, furmark, and the everest stress test.

Today, after logging into windows it had a system_service_exception bluescreen. After restarting, it told me some files were corrupted, then it bluescreened again. After restarting, the same again, more corruption messages and a bluescreen. By the 3rd or 4th try, it just booted into a black screen with a cursor. Bluescreens were a mixutre of the system_service_exception and memory_management ones.

Installed windows onto the computer's second hard drive, and it hasn't died yet. Ran windows memory diagnostic and it passed. It's only been half an hour though, and it hasn't been turned off yet except the reboot to run the memory diagnostics, so not getting my hopes up yet .

So anyway, I know nothing about what different bluescreens mean or anything, so anyone fancy helping? I probably should have wrote down the codes, but I didn't.

I'm guessing its probably either a dying hard drive or dying ram. And as it passed windows memory diagnostics, I'm guessing that the ram's fine... But then again I've heard windows memory diagnostics is crap . I'll try memtest when I can get my hands on some blank cds.

I suppose if it doesn't die again I can just blame it on the hard drive, and if it does, I can rule out the hard drive. But other than that, what else can I do to figure out what is wrong?

[Sputnik]

you could always try running the hard drive diagnostic program for your hard drive. What hard drive is it ? Have you made sure there is no known issues with this drive and your motherboard?

[jim]

Get a few tools to test the hard disk, seems the easiest thing.

Western Digital Lifeguard, HDTune, Seagate Seatools are all worth a try. Run some scans on the possibly broken drive from the working one and see what happens.

[thomson]

It's a Seagate hard drive, model number ST3500418AS. Motherboard is GA-870A-UD3.

Doing a Seatools test now. Which one should I be doing? Doing the long DST atm.

Here's my stuff copied and pasted from the CPU-Z validator . In case it makes any of the below slightly different, I haven't installed all my motherboard drivers yet for my new windows install...

CPU Arch : 1 CPU - 4 Cores - 4 Threads
CPU PSN : AMD Phenom II X4 955 Processor
CPU EXT : MMX(+), 3DNow!(+), SSE (1, 2, 3, 4A), x86-64, AMD-V
CPUID : F.4.3 / Extended : 10.4
CPU Cache : L1 : 4 x 64 / 4 x 64 KB - L2 : 4 x 512 KB
CPU Cache : L3 : 6144 KB
Core : Deneb (45 nm) / Stepping : RB-C3
Freq : 803.66 MHz (200.91 * 4)
MB Brand : Gigabyte
MB Model : GA-870A-UD3
NB : ATI 870 rev 00
SB : AMD SB850 rev 40
GPU Type : NVIDIA GeForce GTX 260
GPU Clocks : Core 300 MHz / RAM 100 MHz
DirectX Version : 11.0
RAM : 4096 MB DDR3 Dual Channel
RAM Speed : 669.7 MHz (3:10) @ 9-9-9-24
Slot 1 : 2048MB (10700)
Slot 1 Manufacturer : Corsair
Slot 2 : 2048MB (10700)
Slot 2 Manufacturer : Corsair

It's a shame I didn't get DDR2 ram, I could have stole some from my other computer...

[thomson]

It passed the Seatools test, so unless I did the wrong test I take it the HDD is perfectly fine and it's something else.

Memtest, and if it passes that its the motherboard?

[jim]

Possibly. Could be the CPU for instance.

Have you tried removing one stick of RAM at a time to see if the performance improves/worsens?

2GB is perfectly adequate for most uses, so I'd try running it on one stick at a time for a while. Then swap over to the other stick. Might show one of them to be a bit flakey.

Also, did you buy all of the components from the same retailer at the same time?

[smargh]

PSU.

Check voltages with something like Speedfan or HWMonitor.

[thomson]

How should I be testing the voltages? Check them when idle and under load or what? And what would I be looking for?

How would the CPU cause this problem? Or was it just a random part that I hadn't asked about yet? It just seems to me that as the problem only happened while everything was loading in windows, it would be mostly the ram/hdd/motherboard that are doing things. Not saying you're wrong btw, I would just like educating .

RAM passed 8 passes of memtest btw. I'll try again later from a cold boot though, thats the only time I ever had an issue. It's never had a BSOD from a reboot after its already been turned on a while.

Not tried with 1 stick of ram yet, I'll give it a go after I try this second memtest. Which will probably be tomorrow morning after I leave my computer off overnight.

As I have no idea what I'm looking at, can someone interpret for me? I'm assuming the lower VCORE value is Cool'n'Quiet, I can turn that off if it would be better.

This was done with a few minutes idle before a prime95 blend test. I'll do it with furmark too once I install my graphics card drivers, that uses more power .

I realise I cut off the headings... columns are current/min/max



But again, I have no idea what I'm looking for. And as I've only ever had a BSOD on startup, will showing my power supply is ok after startup show anything?

I bought all my components at the same time, but from different retailers. I got everything from either Scan or Ebuyer, depending where was cheapest.

[smargh]

Those voltages look fine, but checking voltages with software often gives incorrect results - the -5V, as an example from your screenshots. Generally a multimeter is best for checking voltages.

How old is your power supply and what brand/model? I realise you said 'new computer' but perhaps you re-used one or two old components.

Did you do the SeaTools long test, not just the short test?

Run this and show us the screenshot: http://www.nirsoft.net/utils/blue_screen_view.html

Also do a chkdsk for a read-only filesystem check, and perhaps also do a chkdsk /R then reboot and view the results of it in the event log ("Application" - look for 'winlogon' entries) after you've rebooted.

At the moment I'm thinking that you either have a broken driver, inappropriate drivers, or just a dying motherboard, memory or PSU etc with an intermittent fault.

[thomson]

Everythings new, I'm selling my old one to my parents .

PSU is a 2 week old Be Quiet.

Would I be right in thinking I need to run that from within the OS that died? I can't actually get into that OS anymore, I just get a black screen with a cursor.

[jim]

How would the CPU cause this problem? Or was it just a random part that I hadn't asked about yet? It just seems to me that as the problem only happened while everything was loading in windows, it would be mostly the ram/hdd/motherboard that are doing things. Not saying you're wrong btw, I would just like educating .

RAM passed 8 passes of memtest btw. I'll try again later from a cold boot though, thats the only time I ever had an issue. It's never had a BSOD from a reboot after its already been turned on a while.

To be honest, anything out of the CPU, memory, motherboard can cause all sorts of problems if one of them is playing up. I don't particularly think that it is, just pointing out that it could be.

A few thoughts:

If it is a driver issue, as smargh is suggesting, then try booting into safe mode. If that works, it implies driver problems.

If the RAM (both sticks) did 8 passes successfully, then it would seem that the RAM is okay.

Your mention of a cold boot is interesting though, could mean power supply issues.

The reason I asked about retailers is that if all your components were bought from Scan, you should be able to send the whole lot back and ask why it isn't working. Since they're from multiple retailers, if you send back to Scan only the parts that you bought from them, it might transpire that a part you bought from eBuyer was actually the faulty one. And then Scan will charge you £30 or so as a "Not found faulty" fee. So that's still an option, but it's more of a pain that it would've been otherwise.

[thomson]

Ok I got it .

So I was wrong about half the bluescreens lol, there were others too .

If its important there is more info further along that wouldnt fit on my screen... It doesn't look important though...



Safe mode doesn't work either, but I think thats because I have loads of corrupted files now...

The last time I had multiple bluescreens, booting into safe mode worked, but then booting into normal mode straight after worked too...

Oh, and I did bother the long and short tests, it passed both.

[smargh]

I might have seen those kinds of errors recently when a driver was having problems with various power-saving modes.

Disable every power-saving, sleep mode & any "allow the computer to turn of this device" option in Device Manager. Run "powercfg -energy" to get hints at other options.

I still lean towards some kind of hardware thing though. The only thing you could do is to go through a process of elimination by trying each bit one at a time in another PC.

[thomson]

I think I already have all the power saving things turned off. Or I did anyway...

I'll disable them all again and run that when I get back on . I'll have to wait until it breaks again to see if its worked though right?

I can't really do much with trying parts in a different computer. This computer has an AMD processor with DDR3 ram, my other has an intel processor with DDR2 ram. The power supply in that one doesn't have the 2 PCI 6-pin cables I need to power my graphics card in this one, and I don't have onboard graphics. The only thing I could do is stick this power supply in that computer, and see if it starts breaking.

Oh, and swapping parts on a pre-built computer would void it's warranty wouldn't it?

I wish something would just fail a test so I could replace it and be done with it...

[thomson]

Looks like this person here has just about the same problem. They think one of their ram sticks is dodgey, so I'll try running on one stick at a time if I get another BSOD.

http://social.answers.microsoft.com/...3-1caceae2c16f

[Terbinator]

I'm not an expert on BSOD codes, but "Memory Management" is the most frequent error on there: It could quite possibly be a RAM issue.