BSOD only on cold boot - troubleshooting help

[tfboy]

My main PC has been suffering from a BSOD on cold boot for the last year or so, pretty much since I built it. I've been "lazy" and never really sorted it out, but it's about time I did.

Hardware is:
Gigabyte GA-Z77-D3H motherboard
Intel i5 3570k CPU at stock speeds
16GB RAM in 4x4GB running at stock SPD speeds
Gigabyte nVidia GTX 560Ti GPU
Corsair RM 650W PSU
W7 64bit Ultimate

Issue it a BSOD pretty systematically when cold booting. It will appear at the end of the Windows 7 bootup before logging in or after logging in. Sometimes, it will only appear a minute or two after boot. I haven't been able to identify a trigger.

What I've tried so far in trying to identify the cause:

• If I reboot the system before the BSOD, I will get it on the reboot (or shortly afterwards).
• Once I've had it, it will be stable and no matter how hard I push the system on CPU, RAM or graphics, it never misses a beat.
• I've run individual burn-in tests but can't get it to crash.
• I've run memory tests for hours, not a fault found.
• I've taken out the discrete nvidia card and run it on the intel iGFX, makes no difference.
• I've tried multiple different BIOS revisions. Makes no difference.
• I thought at one point it might have been the PSU (I did some reading and a lot of people with Z77 boards have pointed the problem at the PSU which was a Novatech high-end modular 600W affair. However, it's still doing it with the new Corsair PSU which I bought just last week.
• Once I've had the BSOD, an immediate shutdown, new-cold-boot is fine. I wondered if it was a "component heating up" that was the issue, but if I put the system in sleep for hours and then resume, it's fine.
• I have just done a fresh reinstall of Windows7 on a brand new SSD, it's still doing it. Whether the system is set up in intel RAID or AHCI makes no difference: I get the BSOD on both setups.

Now the cause of the BSOD varies. Sometimes, it's page fault, sometimes irq, sometimes a device driver (intel storage, nvidia graphics).

I've now run out of ideas on tests that can isolate the issue. The only thing I haven't tried is a safe-mode boot (from a cold system) to see if it still happens.

It's one of those things that isn't a huge head-ache, I can just boot the system from cold in the morning, wait a minute for it to BSOD and then know I'll be fine on the reboot, but it shouldn't be happening in the first place!

Any ideas where to go from here?

Thanks.

[kalniel]

A few things I'd try that you don't mention:
1) Putting a delay on the HDD boot up in BIOS, just a few seconds, in case there's an initialisation problem
2) Having a look at voltages - it could be that the motherboard is clever enough to boost a voltage or disable a power saving mode after a crash. Setting some manually rather than in auto mode might help in the case where the default voltage chosen by the motherboard isn't enough first time.
3) Disabling some power saving modes - specifically C3/6/7 advanced states. It could be that the CPU has a problem going through these states on cold boot.

[tfboy]

Thanks, I'll give them a go and report back. It will probably need a day or two for me to be sure. Much appreciated

[tfboy]

Well I've disabled quick boot and it seems better. Surprisingly, there's no HDD delay option in the EFI BIOS.
Anyway, will keep monitoring over the next few days.

[Obraxis]

Some diagnostics to try:

Seagate Seatools on any mechanical HDDs (not SSDs).
Memtest+

If all is fine, try a re-install making sure ahci sata is on before you install windows NOT IDE emulation. Re-Seat all your RAM.

[tfboy]

A long time update, this is till happening.
Either on startup, within a few minutes of starting up, or when shutting down.
When it happened this morning, the system rebooted and then I didn't have any network connection (crossed out network in system tray) even though the link was up and the lights were flashing!
I've taken out the PCIe graphics and have just the mainboard with no extra cards. I will try another SSD just in case: it's currently running on an OCZ Vector and after having a horror story on another drive where it just vanished overnight, I'm wondering if this one is also causing the issues. I really want it to fail to replace it with something else!

[DBarber]

A long time update, this is till happening.
Either on startup, within a few minutes of starting up, or when shutting down.
When it happened this morning, the system rebooted and then I didn't have any network connection (crossed out network in system tray) even though the link was up and the lights were flashing!
I've taken out the PCIe graphics and have just the mainboard with no extra cards. I will try another SSD just in case: it's currently running on an OCZ Vector and after having a horror story on another drive where it just vanished overnight, I'm wondering if this one is also causing the issues. I really want it to fail to replace it with something else!

maybe it is network related, many updates try to run on cold boot, maybe try different drivers for network card or disable and put in a cheap nic with different chipset to test.

[tfboy]

Thanks, but I don't think it's NIC related. This is the first time it's done it, and I reckon it's just "another symptom" of a more fundamental issue.
I have my money on the SSD (which I will change / try something else) or the motherboard, but narrowing it down is proving difficult.

[tfboy]

Ahh forgot to mention that when it makes it past the initial boot and appears to be running fine, often, plugging in a USB device (scanner, flash drive), etc, will cause it to BSOD.

This morning, booted OK, waited at the login screen (I was typing this post!), then about 2 minutes later, it's just rebooted without any BSOD or anything.

Next thing I'm going to try is individual RAM sticks on a cold boot.

[Millennium]

Hey there

Not a fun debugging on this problem, well done living with it

I would like to suggest trying a fresh windows 8.1 install on a spare hard drive, also (or alternatively, well perhaps even first) try the (latest z77) intel chipset driver installation with the "setup.exe -overide -overall" switches. One or the other may fix that, otherwise do let Gigabyte support know they'll be interested though they may not be able to fix it unless you can or feel like asking them to look at is as a possible BIOS issue.

I don't really think that has more than a 50/50 chance of working (60/40 and 30/70 ? ) though that's what I'd try. Ok, maybe also, alternative SATA cables or run them straighter in your case.