System BSOD, even from the repair disc.

[Lewis_Hamilton_]

So my problems started last night. I was enjoying a Tomb Raider 2013 session when my PC suddenly reset itself, and never again would anything appear on the screen. After the usual troubleshooting problems, I swapped in my old GPU and voila, I could see my BIOS posting. So my newer GPU died I figure.

Ok, so I can send that in and get it replaced as it is still under warranty, however, whatever happened here has corrupted or deleted the ntfs.sys file. I had this once before in the past and repaired it using a recovery disc. This time though, not even the recovery disc will start. When selecting to boot from CD-ROM, I sometimes get a BSOD within seconds, again, regarding the ntfs.sys problem, and others about a change in hardware and to insert the repair disc (Good Lord this error message makes my blood boil, it is the fricken repair disc you daft bugger).

Sometimes I get to see the solid white "Loading Windows Files" bar, and sometimes the green "Microsoft Corporation" loading bar, but here is yet another problem. Even if these bar begin to load, they slow to a crawl in no time and the disc drive stops spinning completely within 30 seconds. I know the repair process can take a long time, but 6 hours of a slow moving loading bar that has no physical measurement and no disc activity from the CD-ROM for this whole time tells me that nothing is going to happen.

I have tried every available boot option and all of the above happens regardless, it just swaps between BSODs or stuck on loading. Anyone with any suggestions? How the heck does a dead GPU cause this much mess?

 

[Grayfox]

Download Memtest 86+ and let it run.

Also go into the BIOS and make sure the hard drive controller is not set to RAID unless you are using RAID.

Also set it to IDE and try it and see if the BSOD is related to the sata controller.

The Windows is loading files is just the needed files being loaded to memory.

In theory it should be quite fast, doing the automatic repair can take time.

 

[Lewis_Hamilton_]

Download Memtest 86+ and let it run.

Well, I used another computer to download memtest86+ and put it on to USB, went into my BIOS and checked the settings that you mentioned above, they are as they should be and haven't been changed for 3 years. I selected boot from USB to run memtest86+, and nothing happens, I'm just left staring at DMI pool verified message, there is no further activity.

I didn't think it was a RAM problem as the BSOD were to do with a required HDD file as far as I know, but went and tried the RAM troubleshooting method of taking out each stick and trying each one in each slot. Now though, I'm back to a blank screen, I can't even see the BIOS again, no matter which RAM sticks are in which slots.

So now what do I do? This is only going backwards.

 

[Grayfox]

Well, I used another computer to download memtest86+ and put it on to USB, went into my BIOS and checked the settings that you mentioned above, they are as they should be and haven't been changed for 3 years. I selected boot from USB to run memtest86+, and nothing happens, I'm just left staring at DMI pool verified message, there is no further activity.

I didn't think it was a RAM problem as the BSOD were to do with a required HDD file as far as I know, but went and tried the RAM troubleshooting method of taking out each stick and trying each one in each slot. Now though, I'm back to a blank screen, I can't even see the BIOS again, no matter which RAM sticks are in which slots.

So now what do I do? This is only going backwards.

Well you are crashing so it is either a chipset driver issue, SATA controller issue, or RAM.

What kind of BSOD

What are the stop codes.

Also it is a bit weird when memtest cant run.

 

[Infinital-NG]

Couple questions:

1. What kind of computer (specs & age)
2. Have you tried resetting your BIOS to factory defaults?
3. Take the hard drive out of the equation, just pull the power from it and try and boot memtest and/or recovery disc. The recovery disc should at least stop and say "No hard drives found" or something.
4. Does everything look ok in the computer? (ie, CPU Heatsink/Fan is in place and hasn't come loose)
5. Knowing what the BSOD error is will help out a lot (as Grayfox mentioned)

 

[Lewis_Hamilton_]

Couple questions:

Quad core i7 920 D0 stepping CPU
3x2GB Corsair XMS3 RAM
OCZ 700W PSU
Radeon HD4830 (old) and XFX (DD edition) HD7850 GPUs - Obviously was running the 7850 alone and kept the 4830 just incase something like this happened.
Gigabyte GA-EX58-UD5 motherboard

I've tried the clear CMOS/BIOS button on the motherboard, tried without any RAM at all, tried with the OS hardrive disconnected and checked that there is no short circuiting and all components are fitted correctly.

Since taking out the RAM sticks from the first attempt at fixing things earlier, the motherboard has been displaying the b2 memory code error on the LED display. Since then, I can't even get to the BIOS and the same error is there whether there is RAM fitted or not, and it doesn't matter which stick is in which slot, or whether I have 1, 2 or all 3 sticks fitted.

The BSOD I was getting before was 0x50 win32k.sys and NTFS.sys file missing or corrupt. I don't think this is to do with the RAM personally, when the computer originally shut itself down, it wouldn't display anything upon reboot, but would if I fitted the 4830 back in. Put the 7850 in again, the fans start up on the GPU but no display or BIOS, back to the 4830 again and the BIOS would boot.

Now I'm in the situation where I have no BIOS even with the 4830 and no matter what I do with the RAM or HDDs. Bad times.

 

[Danoff]

Something is broken... probably on the motherboard but possibly the CPU. If you can get a CPU stress test to run, do that and see if your CPU passes it. It's not the GPU, you swapped that already and it didn't fix the problem. It's not the ram, you swapped that out (if you have two sticks, you can try running with one and then try running with the other). Pretty much you're left with the HDD (unlikely since it fails while trying to boot a CD) and the motherboard. My money is on a dead motherboard.

 

[DQuaN]

What happens if you use the onboard gpu?

 

[Lewis_Hamilton_]

There is no onboard GPU.

I went to a repair shop and described everything above, they think the PSU is the problem. I wouldn't be suprised as this is my 3rd PSU in the 5 years I've had this build while everything else has worked great, despite always going with >200W above the recommended power output and sticking with respectable brand names.

The shop has offered to diagnose my PC for free tomorrow so I'll see what happens then. Thank you all for your input, can't believe this all happened the day before I was going to set my G27 wheel up, so bummed about it. I was nearly finished with Tomb Raider, that game was better than I expected and man it looked beautiful on my Plasma TV.

 

[Danoff]

There is no onboard GPU.

I went to a repair shop and described everything above, they think the PSU is the problem. I wouldn't be suprised as this is my 3rd PSU in the 5 years I've had this build while everything else has worked great, despite always going with >200W above the recommended power output and sticking with respectable brand names.

That would be an interesting one. I've never seen a PSU work halfway - when they've died for me they die 100% dead. But it's an interesting datapoint if that's what it is.

 

[Omnis]

You might be overloading a psu rail. Might be bad memory.

 

[dice1998]

I'm assuming that if it was bad memory, POST would beep like no tomorrow... or at least spew out an error code on the MoBo LED stating that it was a bad memory, which for Gigabyte boards (at least with my UD3H) should be 53 or something.

I'd go with bad rail on the PSU.

 

[Grayfox]

I'm assuming that if it was bad memory, POST would beep like no tomorrow... or at least spew out an error code on the MoBo LED stating that it was a bad memory, which for Gigabyte boards (at least with my UD3H) should be 53 or something.

I'd go with bad rail on the PSU.

Not really.

It will do a RAM ERROR beep code if the ram was very bad.

A few bad memory blocks will still cause a PC to post fine, but once something attempts to write data or read data from those bad blocks you wont know.

Now your saying the BSOD is mentioning the HDD.
So it would mention a driver file or string that caused the issue.

any of these mentioned?
These are common intel sata controller drivers
atapi.sys
ataport.sys
msahci.sys
pciidex.sys
iaahci.sys
iastor.sys

 

[BobK]

Just out of curiosity, how does one get a Blue Screen of Death aka BSOD before Windows even loads?
Or is BSOD now a generic term for any system failure?

 

[Grayfox]

BSOD is a Blue Screen of Death.

It happens when the system does something it shouldn't, has issues with something or software.

Things it shouldn't do.
Attempt to load and use the wrong driver either by a user forcing it you it thinking it is the right driver.

Has issues.
Happens when the RAM is corrupted, has issues loading a file from HDD or RAM, PCs hardware settings are incorrect.

Software.
Malware can cause bluescreens as they will try to mess with system files, Software being loaded is corrupted.

A common HDD one is 0x0000007E or 0x7E which is UNMOUNTABLE_BOOT_VOLUME
0x7E can mean other things but I have mainly seen it as UNMOUNTABLE_BOOT_VOLUME when trying to load to windows, and it can be fixed by running CHKDSK X: /r or CHKDSK X: /b
X: is the drive in question,
It can be C: or D:

Just depends on how many drives installed, which partition windows is installed on, in which port the drive in question is connected into and when windows mounted it.

 

[Danoff]

Just out of curiosity, how does one get a Blue Screen of Death aka BSOD before Windows even loads?
Or is BSOD now a generic term for any system failure?

This is occurring as windows loads or during a windows repair disc - both of which are fundamentally windows software failing. BSOD doesn't happen for non-windows operating systems.

I still think a CPU stress test is in order. RAM can totally do this (but has been checked). Motherboard failures are so hard to diagnose.

 

[BobK]

Yes, I'm quite familiar with BSODs from my Windows 3.0, 3.1, 3.11 etc days, and their myriad of causes. They are a Windows-specific error and it appeared to me that the OP said he was getting BSOD's even before Windows loaded. This is why I asked if it had become a generic system-failure term.

I'd agree with @Danoff that a CPU stress test is in order, and definitely a couple hours of running memtest86. Overnight would be good. Memtest86 is a much more thorough test of memory than the quick test the POST does.

 

[Infinital-NG]

I've seen PSU's only work partially. Usually the system isn't very stable, and can be a hard one to figure out.

Any idea what temperatures you were running?

I'm sure the repair shop would do the same, but at my work I'd probably do the following:

1. Try different PSU with your original GFX card
If that didn't work, I'd try another GFX that hadn't been in your system.

2. Then I'd try a new RAM

3. Then CPU

4. Then Motherboard with original components.

More I think about it, sounds like your PSU went Ape and messed up something else. (Cpu/Mobo/RAM)

 

[Infinital-NG]

Any news? Always curious to find out what the problem was.

 

[Lewis_Hamilton_]

Any news? Always curious to find out what the problem was.

So it's been over a month since I had this problem. I gave up on it for a while as I had other things to do.

This week, I took another crack at repairing my PC with some success. It's all up and running now, everything works... except for my newer graphics card.

In the end, the motherboard's BIOS needed to be flashed. That was done and from there, I could finally(!) repair both Windows 7 and 8 (both ended up corrupted). Suprisingly, the latter was actually a quick process, taking just a few minutes until I was back to a fully running PC.

Only issue I have now is that whenever I swap the HD7850 in, the motherboard's logo flashes up for a split second, and from there on, nothing. However, the motherboard's LED display shows that it is posting (after having to press F1, a hint I read online) and I can hear the computer/HDDs fully booting up, which is more than it used to do. Previously the motherboard's LED display would throw up some errors while it was posting and would grind to a halt.

I've tried the different graphics card priorities and settings in the BIOS settings to no avail. Anytime I throw in the older HD4830, everything works; put the 7850 in, everything works with no display. At least the PC works now, even if I can't run all my newer games on the ancient card.

I was sure all along that somehow my graphics card having problems was the cause of everything, and now it's the only thing that's not working properly. I'm ruling ot the PSU as both the old and new graphics card draw the exact same amount of power and use the same 6-pin connector. I'll send the 7850 back as it is still under warrenty, and perhaps try and get it upgraded to a 7870 while I'm at it.