GA-Z97X-UD7 TH - Weird cold start / resume issue

[NessPJ]

Hello there,

Ever since i built my system it has been working fine. Over the past few months a weird issue has developed out of the blue.
The motherboard will no longer cold boot or resume from sleep mode.
The system powers on and midway through boot *poof* it turns off again.

I tried:
Doing a "clean" cold boot, by unhooking the power supply from the mains and from the motherboard. Hitting the power buttons on the case (for the PSU) and on the motherboard to discharge for 30 seconds. Connect everything again and power on. This doesn't work.

Here's where it gets funny (and i've reproduced this!):
It seems like when i change the 'load' on the motherboard once by unhooking any random component...discharge, and connect it back after...the mothboard will boot again, just fine!
I tried this on seperate occasions with: My 2 SATA Harddrives, My PCI-e NVMe storage (Intel 750 series) and my graphics card!
So each time, the PC will not POST...i disconnect it from the mains...take out one of the components mentioned above, wait for 30 seconds. Connect the component again and the motherboard will POST without problems!?

Does anyone have a clue what might be going on here?

My specs:
GA-Z97X-UD7-TH
Intel Core i7 4770K (no overclock)
2 x Crucial Ballistix Tactical tracer 8GB 1600MHZ BLT8G31608DT2TXOBC
Gigabyte Windforce GTX980TI
Seasonic X-850
Intel 750 Series NVMe SSD 400GB
2 x Seagate ST4000VN000 4TB (RAID1)

[dmdilks]

Have you tried another power supply?

[Sylvester]

I don't know what is happening to yours OP but the same thing started happening to my GA Z97MX - Gaming 5 only yesterday.

The way it happened I was playing Forged Alliance in Win 10 with Win 7 compatibility set (probably should have been XP SpP3). When the game ended, exit from the client caused the PC to crash and reboot, not even a BSOD.

It rebooted once OK but it provided a boot failure notice and I entered BIOS and exited without saving and it booted with existing settings. It then ran normally for a session but next boot did the same, providing a boot failure notice so I powered down with the on off button and shut it down and it wont boot properly ever since and will not even show the boot  fail screen, just powes on and then the CPU fan stops rotating and it does nothing or it does the same repeatedly in a fail boot cycle. The only way I can get it to boot is to move a stick of RAM.

This is a dual boot machine with one UEFI SSD for Win10 and one standard GPT SSD for Win7, both installed in isolation so they can be managed independently. I have no idea what is wrong. Worked fine for 18 months.

PSU Seasonic 860W Pt : SS-860XP2 modular
CPU 4790K + Noctua NH-L12 (LianLi PC V354)
Gigabyte Z97MX Gaming 5, BIOS F4.
RAM GSKILL 2400 F3-2400C11Q-32GXM 4x8GB
GFX nVidia Palit GTX 1070 Jetstream
SSDs Crucial BX100 and M550

At first I thought it was CMOS so I changed the battery but the old one (2 yrs) was a good 3v and the new one at 3.2v didnt fix it.

So then I wondered if UEFI foul up from Win10 so I stripped it down to CPU mobo and 1 stick of RAM and cleared CMOS with the jumper then it would boot.  So I added the SSDs and it provided the boot failure on the first boot, then it would boot on exit BIOS after that but would not reboot again from OS and entered the CPU fan stop mode again, until I shifted one RAM stick, then it would boot to the boot failure notice and I could get through the BIOS to the OS again.

It appears to be a log jam in the CMOS but I have tried clearing it and it doesnt help. When its powered up in the OS it works fine. Only thing which works reliably to get it to boot is moving the single stick of RAM to a different slot. I have also tried using different sticks in case. Same situation.

I am at my wits end, just noticed the similarity about moving a connected item with this thread so I thought I would chime in and hope for some kind of correlation to help make sense of it.

[Sylvester]

OK I have had a breakthrough with my problem and got my PC back, it looks like it was windows boot manager listing in the BIOS boot priorities table coupled with UEFI interface imho. Fingers crossed, its working as I type and will now reboot.

I think WBM was messed up somehow by the crash in Win 10 from a UEFI SSD. I have had problems with UEFI + WBM and this mobo before. Could that be anything to do with your issue @NessPJ ?

So more by luck and vague intuition than judgement I left it running in the Win7 OS (GPT) while disconnecting the UEFI drive with Windows boot manager on it. The BIOS somehow got used to not having Windows boot manager around and it was no longer listed as a boot option the next time I tried a reboot a couple of hours later; the reboot worked for the first time in 24 hours.

So I then reconnected the UEFI drive and it booted straight to the UEFI drive completely ignoring the BIOS settings I made for the other drive, which is the kind of thing WBM does and is why I suspect it is not very compatible with this iteration of the tech. But BIOS erased whatever tangle it got into by booting with a GPT drive only connected and leaving it for a bit.

Anyway its working again but I lost all BIOS profiles, so something broke. Its saving fresh profiles but it makes me nervous of running Win 10 since I am pretty sure it was the Win 10 crash which borked the WBM and started the whole problem, if that makes any sense at all.

I know that is weird but its what happened. FYI hope it helps.

[NessPJ]

@ Sylvester: Thanks for posting your findings. My problem however, does not seem to be Software of UEFI related however.
It really seems like a hardware failure.
The strange thing is that; if i use my 'trick' to make the system POST/Boot the system will run perfectly fine afterwards.
The issue solely comes down too not being able to boot properly.

I measured the Seasonic PSU both under idle and full load on the system and all voltages stay well within their margins.

I really think this has something to do with a certain capacitor or power FET on the motherboard, that makes the motherboard fail POST on a cold boot.

During my many years as an enthusiast computer/hardware user i have seen this failure happen on a lot of motherboards. I was wondering if someone from Gigabyte perhaps would be able to clarify what this issue is usually related too (it's a very typical motherboard issue).

[Sylvester]

Checking your PSU makes sense but it looks like you ruled that out re: power, unless its a quirky interaction thing.

To be honest I have completely disabled sleep mode for my PC as I have always found the suspension states very unreliable in a desktop. Workaround might be to disable sleep mode.

re UEFI, based on fixing my PC today, I would still say UEFI related settings in the BIOS seem able to block the POST, which surprises me, perhaps it is related to detection. I had to open up all the UEFI related options to get boot to work with my UEFI SSD, before doing this the boot failed before the POST screen when the UEFI drive was connected and since enabling UEFI stuff it has been working normally.  I dont know if its a red herring in your case but it might be worth a look if you have nothing else to try.

[Sylvester]

OK strike all the above, update, it was a corrupting BIOS, thought I had it fixed but it slowly degenerated until I could not even boot by shifting RAM around. So I thought about updating the BIOS to F6 (which cannot flash back to F4) read around and read a guy saying GA have a habit of updating BIOS files without changing the # so instead of of using F6 or trying to trigger the backup BIOS which seemed reluctant to operate I redownloaded F4 and Q flashed it in BIOS from USB stick and loaded defaults before setting the RAM etc.

BIOS was the last access I had, like a shrinking window into my PC, as it would just blue screen and bug out with a SSD this afternoon when for the last two days and this morning it was booting OK with a RAM shift. So I was lucky I flashed it before the BIOS bugged out completely but it seems to be back to 100% functionality for now at least.

I am not sure if there is a hardware fault behind it, time will tell. If F4 goes pear shaped I will shift to F6.

But it might be worth considering updating your own BIOS NessPJ for the sleep function, either get the latest or reflash the existing one if its already the latest and cross your fingers.

There are also beta BIOS out there, though not recommended, there is a big thread on TweakTown forum.

https://forums.tweaktown.com/gigabyte/