PC boots to a black screen after windows logo. Hardware issue?

dh2311

Reputable
Mar 24, 2015
7
0
4,510
I’ll give you a short version because otherwise this would just be a wall of text. Other important information is below.

Specs:

Processor: i7 4790k OC 4.5ghz @1.25v
Ram: 16GB Dominator GT RAM
GPU: Powercooler R9 280x (Cooler upgraded to Arctic Cooling Accelero Extreme IV)
PSU: Be Quiet! Pure Power 650W
HDD: 240GB Sandisk Ultra SSD / 3TB Seagate Barracuda
OS: Windows 8.1 64-bit


Paused a game, left the room for 5 minutes, and came back to a grey and black stripy screen. Rebooted, after the windows logo had appeared. Black screen.
Rebooted a number of times, same error. Booted off the internal graphics, no errors at all. Uninstalled the AMD drivers using Display Driver Uninstaller, all works fine. Reinstalled the AMD drivers, and as it gets to installing the display drivers, the black screen returns.
This seemed a plan driver conflict issue. But after wiping the whole hard drive and starting again, the problem returns.
I am now at a loss, any suggestions as to what is wrong?

Important info

The powercooler cooler was removed and replaced by the Arctic cooler a few weeks before this error, but had only been running for a few days before the error occurred. However, I stress tested it for hours, and VRAM and core temps never exceeded 60 C

It was overclocked at the time – Afterburner 1150mhz core 1700mhz memory clock. Again, was stress tested and didn’t exceed 70 C. Power boost of 15%, and overvolted to 1.2v

The game was fifa 15 – which is hardly pushing the boundaries of the card.

It is fitted in the 2nd PCI slot because it wont fit in the 1st

I have another motherboard which I will fit in a few days to see if it’s the PCI slot (really clutching at straws here)
 

dh2311

Reputable
Mar 24, 2015
7
0
4,510
I was hoping for something a bit more constructive. It's not like it immediately failed after me changing the cooler. It has lasted perfectly fine for a week-ish before showing signs of issues.
 

ThomasLeong

Honorable
May 27, 2013
305
1
10,960
Sure seems to narrow the cause down to your graphics card. You did not state which motherboard (brand, model) as it is rare that it cannot fit into the first PCIe 3.0 slot - have only come across this due to the rear lip of a chassis requiring a hacksaw solution. Your motherboard's second slot may not be providing sufficient lanes/throughput for optimal use of your graphics card with the mods and over-clocking. Mobo manual should reveal.
 

dh2311

Reputable
Mar 24, 2015
7
0
4,510
Sorry, the motherboard is a GIGABYTE Z97M-D3H. Dont know how I missed that off.

It wont fit in the 1st slot as it is a MATX board and the rear heatsink of the Arctic cooler blocks off all 4 of the RAM slots.

The board claims to be able to support Crossfire, so I would presume that both PCIe slots are the same. Also the fact that it was initially working for a few days makes that look less likely. But thanks for the constructive help. I hope it's nothing major.
 
look from newegg specs.....
PCI Express 3.0 x16
1PCI Express 2.0 x161 (x4)

then your back to what I said at first
''It wont fit in the 1st slot as it is a MATX board and the rear heatsink of the Arctic cooler blocks off all 4 of the RAM slots''

so your improvements seem to be a issue ??? right ?? as in you installed all that at your own risk of things working / not working out ??
 

dh2311

Reputable
Mar 24, 2015
7
0
4,510


I didnt actually notice the slots were different standards. That hopefully could be part of the issue. Fortunately I have a new motherboard that I'm going to fit at the weekend.

I am not denying that what I did had an influence. it very clearly did, before this it had worked fine for a solid 8 months. Just I wanted advice on what could be wrong and how to fix it more than just being told I broke it.
 

dh2311

Reputable
Mar 24, 2015
7
0
4,510
If I boot into safe mode it all works fine. It also works fine with the drivers just uninstalled.

But I dont see it being driver related, because the problem came back on a complete clean install of windows
 
thing is when you do these things like gpu coolers overclocking what ever theres always that unknown risk as in mabye somthing did get damaged but it took this amount of time for it to show like the oc. it may of been hard on the memory chips , a cap, or the gpu its self and the ware and tare of the overclock took that amount of time to have it just say thats all folks and died on you ??
this is why manufactures may not honor warrentee if you take there card apart and have dissclaimers that overclocking is all at your own risk

i'm not trying to be smart but somethings work out some do not it may take 2 min. to fail or as like you 8 months or it may never fail at all ..

see it now maybe a certain part of your card needs some time to run and get heated up or just the right amount of stress put on it to fail

I got a card here that just blacked screened just sitting here like I am now all looked fine no pop no smoke nothing it just stopped displaying .. slaped a old back up card in and all works good .. only got 14 months out of it and its done for ..

if you did what you can with the software part of things its down to the card just went bad or something on the motherboard - with out a back up card to test things out no telling there or another computer to test the card in to see if it replicates the issue

 

ThomasLeong

Honorable
May 27, 2013
305
1
10,960


Being a Z97, the second slot is not a x16 lane even though it may be a x16 slot.
According to the specs for your mobo:

1 x PCI Express x16 slot, running at x16 (PCIEX16)
* For optimum performance, if only one PCI Express graphics card is to be installed, be sure to install it in the PCIEX16 slot.
(The PCIEX16 slot conforms to PCI Express 3.0 standard.)
1 x PCI Express x16 slot, running at x4 (PCIEX4)
(The PCIEX4 slot conforms to PCI Express 2.0 standard.)


So your gfx card presently in the 2nd slot has only 4 PCIe 2.0 lanes to work with. I dare say for the overclocking you have applied to the gpu, the combo may be the cause of your problem -
a) Try without overclocking the gpu if it has to be in the 2nd slot.
b) Use another cpu cooler that does not block off the install of your gfx card in the first x16 slot with 16 PCIe3.0 lanes!! These PCIe 3.0 lanes are direct from the i7 4790K cpu. The PCIe 2.0 lanes of the 2nd slot are from the PCH (aka Z97) - slightly slower and with less throughput capacity.

Edit: Just realised the Arctic Cooler you used is on the gpu, not cpu. In this case, as you have mentioned, another mobo may be the answer. But mATX will always be a tight fit. I'm using a MSI Z97M Gaming myself with the i7 4790K (previously a MSI Z9M-G43) but non-modded gpu.
 

dh2311

Reputable
Mar 24, 2015
7
0
4,510


No, I totally get that, its just me trying to diagnose what went wrong. The temps never got dangerously high enough to do much damage (I know thats not the only cause) and I'm just confused as to what actually broke in the process and why. Everything looked and functioned perfectly fine.

I also (perhaps wrongly) presumed that GPU's had the same sort of failsafes as CPU's when it comes to overclocking. You push too hard and they just shut off before damage gets done, which I had a bit of when tweaking the settings. But what it was set at OC wise had been running unigine heaven for over an hour with 0 signs of any issue and all temps staying lower than they used to be.

I'm about 95% sure the card is just dead and burried, but I'd love to know the exact root cause so I know what to do differently. I.e. Was it the cooler mis fitted, did the OC push too hard, or is it just a pure coincidence that the card failed when it did and its just a wear and tear thing.



I presume by a complete reinstall of windows the overclock settings were lost since they were set in afterburner, I think they just triggered the issue.

As for the new mobo, I have now upgraded my tower and motherboard to full ATX, so with the new board I will have both slots availible to me due to the better spacings. I'm hoping that is somehow the issue, however I am beginning to doubt that.
 
bottom line is once you go outside the manufactures specs your on your own with electronics any thing no matter how slight can cause a failure - you tampered with the card and all and this is what can happen.. unless you got a back up card or computer to do tests on your failed parts you just guessing .
was it over-volted on a weak part and it failed ? undervolted and caused that when you removed the stock cooler and put on the aftermarket did something get out of whack cracked, shorted misaligned ?? I mean the list of things can go on and on -- for all you know its a psu issue as well ????

this is why these things are at your own risk and parts are binned to go into a higher end factory overclocked card and such..
I wish I had the golden answer for you but anything can happen at any time .. now a days its hard enough to get out of the box stuff to work correctly or last as long as you should feel as it is ...
 

ThomasLeong

Honorable
May 27, 2013
305
1
10,960
With the new ATX mobo, if you want to try discovering a cause, try -
1. Install in the 1st PCIe 3.0 x16 slot the gpu modded as is, but without overclocking. See if it works well for ?? days/a week or two.
2. Then try with the overclock settings you had before.

If the gpu is done, one of the 2 above should allow some conclusion. If all is fine, then that PCIe 2.0 x4 lane was the likely cause.

Good luck!