Unpredictable Black Screen... 8 beeps... no boot... panic

Hey Team,

I have an Alienware m17x R4 with nVidia 675m.

In the beginning:
Over the last few months every so often my screen suddenly went black, but still backlit, and I was usually forced to hold the power button to shutdown and reboot. Usually it would not work again straight away but after a wait of around 20 mins it would then boot and run as normal. Naturally I thought this to be an overheating issue as it would work as normal after a 'rest' period, but it is more likely to be something else, read on....

Then:
Only a couple of months ago it started to occur more frequently this time giving the lovely '8 Beeps' after an immediate restart (and subsequent 5 restarts). Now I think it's the LCD, so to test this I removed the nVidia card and booted with just the Integrated Intel 4000 card. Booting with only the Integrated card functions as normal and the LCD has never once 'gone black'. NB: Dell Diagnostics gives no errors at all!

Now:
Of course now I think it's totally related to the nVidia Graphics card again as the issue does not occur when it is not installed. The GPUz logs indicate that there are no big temperature rises (highest ever has been 68'C) when the issue occurs and I am still able to use the PC "blind" (ie, I can still pause/unpause music and execute a proper shutdown without seeing what I'm doing). I also had an external monitor attached at the time the issue occured which also went 'black'.

Today: TL;DR
1 - With nVidia card attached it will function anywhere from 5-180 mins before screen goes 'black'. (reboot gives 8 beeps)
2 - Removal of nVidia card is then necessary to boot again with only Integrated graphics.
3 - There seems to be minor (major?) physical damage to a few pins on the motherboard where the nVidia card is inserted but this seems odd seeing as the issue began months ago with the nVidia card still inserted. (have photo if needed).

Pretty sure I will either need a replacement mother board and/or nVidia card... Question is finding out which and not throwing away too many $$$ in the process - any/all advice welcome!

 

Hi there, sorry to hear of the problem. Sounds just like the way a couple of my GPU failures went. Intermittent and gradually getting worse.

The 675m is a really poor card and is just a rebadge of the 580m. Check out how many of those failed here! Luckily for many R4 owners the 680m was released a couple of months into the R4 run and it replaced the 675m.

It sounds like the card was failing and maybe that damage occurred when you replaced it after removal? Post the pic .

I have been through two 580m's and one 675m (all under warranty) before I purchased the 680m and never looked back. It is of a next generation over the 675m and is a real performance increase. With it being so reliable and cool running it also has a huge capacity for overclocking! Best of all it will be plug and play, no bios or driver messing!

Good luck

 

Since you have bent pins that would be the first indication for the issue. Next try to borrow a GPU from a friend or a shop and do some testing.

Sent from my SAMSUNG-SM-N915A using Tapatalk

 

Great info and advice on the 680m, I think that will have t be an option regardless as it is a better card, thanks. Is it 'normal' that the diagnostics show no errors at all for the video card? Because that is very confusing indeed!

Here is the photo of the damaged pins on the motherboard that has me a little more worried:


I brushed the pins with a tiny flat-blade screwdriver and brush so that they now appear to not be touching as much, not sure if it's enough to run normally or not with a card inserted.

Is it likely that this damage has been caused by me taking the card out and putting it back in several times over the last few months during my troubleshooting? That would be most excellent if I am the cause of this and now need a new mobo too...

 

Hmm, doesn't look great . The socket (IMO) is not of the best quality anyway. I see so many people have their machine 'go bad' only to find a reseat gets them running again. Both my R3 and R4 have at times refused to start (eight beeps) and just removing and refitting the card fixed it. BUT with the 580m/675m you can be fairly sure it's gone to meet it's maker!

If the damage was always there then you might be OK? I've never heard of anyone damaging the socket but it's difficult to say.

If you do hunt down a 680m then there are two models available - 2gb or 4gb memory. The 2gb versions are dell branded and the 4gb clevo. I am told that both work fine without any mods needed but I went for a dell one as nobody could tell me if the 3D would work with the Clevo card? Chances are that a different ID would fail the licensing check, well that's what I was thinking...

 

Those pins don't look good.....

 

The plastic insulator looks damaged, the pins themselves appear intact. If it works, then I wouldn't worry about it. Your card is probably failing though as previously suggested.

 

That's right the pins are right and the plastic is damaged which might be ok. But you need to look into a little bit on what caused the damage to the plastic.

Sent from my SAMSUNG-SM-N915A using Tapatalk

 

I will look for a replacement card that I can test with and will then most likely end up grabbing a 680m as mentioned. Are there any suggestions as to where/who I might find a good deal to purchase such a card?

Also, here is another photo of the pins/socket after I brushed and cleaned them a little. I think it has come up alright and I'm guessing that I am most likely the only that damaged them during one of the many card removals and/or insertions.

 

Any hints/tips on which of these options is the better or okay choice? I guess i'm not against another 100 bucks but would like some extra input on whether it would be worth it for a new card as opposed to used...

The Aliexpress one, USED and cheaper:
http://www.aliexpress.com/item/NVID..._1&btsid=3c55d39c-ac1d-4089-8592-88d3a1db8318

or the Ebay one, UNUSED and more expensive:
http://www.ebay.com/itm/301936684370

thanks again team

after this i'll order one up and do the install, then give feedback as to the result! (finally)

 

Ended up getting the new/non-used option and after a re-seat of the Heatsink and a second layer of thermal pads in 2 places to make up the distance gap I am up and running. GPU Temp is a steady 65ish while gaming and around 50 when idle. In the beginning there was a few restarts and freezes but I assume that was due to the change-over from the Integrated card to the newly install dedicated one.
As mentioned previously by MickyD1234 it was totally plug and play - no hassle with drivers, BIOS or anything else. It was totally the option I was after.

Big thanks to all involved and for the many hint/tips/suggestions from all. I will keep a close eye on the system and will report anything unusual but for now.... back to 'normal'. 680m all the way until a new rig in the future ready for proper VR.

 

Great to hear.