Problems with Alienware Aurora R1 Desktop, any ideas?

[drnick5]

Hey folks, I have an Alienware Aurora R1 desktop computer that I'm working on, wondering if any of you guys have seen a problem like this before. It has the 875 w power supply, i7 920 cpu, 6GB RAM. (3 x 2GB). 2 hard drives running in RAID 0. On the latest A11 bios. Running Windows 10.

It was originally having issues on bootup (sometimes it would boot to a black screen and the alienware lights cycling colors). If I restarted the computer (which required me to pull the power plug, then plug it back in) it would boot to Windows just fine.

When in Windows, it would hard lock up occasionally. (everything on the screen freezing). This was seemingly random, not just when playing games. It had 2 x ATI hd5780's in crossfire, when I removed 1 card, the crashes seemed to stop. I ran it for hours with prime 95 and furmark stress tests with no issues. My suggestion was to run it on 1 card, and I recommended an upgrade to a GTX 1060.

I installed the GTX 1060 (first removed all ATI drivers, shut down, swapped the cards, installed latest Nvidia drivers) I booted it up, I see the alienware splash screen, then it went to a black screen again with the alienware lights cycling. But this time, if i let it sit for a few seconds, it would restart itself, I'd see the splash screen again, and it would go right into windows.

While it was up, I ran a furmark test overnight to stress it out. It ran perfectly fine over night and didn't crash or lock up. CPU temps are fine, GPU temps are topping out at 80C under full stress test. (a little higher than I'd like, but still ok as the max is 90C). So I'd think if it was the power supply, it would have crashed during the stress tests. (voltages in HWinfo all seem fine)

I restarted, and it went to the black screen again with the lights changing colors. After a few seconds in this state, it restarted itself and got right to windows.

At this point, I'm not sure what the issue could be. I ran a memtest 86 on the ram, which passed 15 passes overnight, so I don't think its the RAM. the CPU didn't fail any of the prime 95 tests, so my guess is thats fine as well. I'd think if the PSU was dieing, it would fail during the stress tests. visual of inspection of the board doesn't show any bulging or burst caps.

I tried swapping the 6 pin connector in the gtx 1060 from P14 to P16, but still had the same occasional problem with the black screen on bootup. Again letting it sit, it would reboot itself and go right into windows.

Any ideas on what else I could try? I've searched around and found several problems with black screens on alienware desktops, but nothing quite like this where it goes to the black screen, reboots and goes right into windows.

 

[HCHTech]

RAID0, ugh. I'd say back up the data, then test the disks. An R1 has to be 4 years old by now, yes? It could certainly be the motherboard (maybe the controller). If the disks test ok, then it might be time to call it. You could start throwing parts at it until it works, but most customers aren't going to agree to that. Otherwise, you could be just rearranging the deck chairs on the titanic, so to speak.

 

[GTP]

It could certainly be the motherboard (maybe the controller)

I would also suspect the motherboard.

 

[sapphirescales]

Clean it out REALLY well, and bench it. It's possible the board or case could have flexed a bit and something's getting shorted out. Take the board out and test it outside the case on an ESD safe mat. Try updating the BIOS. Switch the SATA port on the motherboard. Try resetting the CMOS jumper. Run memtest. There's lots of stuff you can do to get to the bottom of the issue. The lazy option is just to pronounce the board toast, but if that's not really the problem you're going to get a lot of heat from the client. Intermittent issues like this are a real pain to diagnose. That's why for custom systems and Alienwares and such, My diagnostic rate is DOUBLE my normal rate.

 

[GTP]

Clean it out REALLY well, and bench it. It's possible the board or case could have flexed a bit and something's getting shorted out. Take the board out and test it outside the case on an ESD safe mat. Try updating the BIOS. Switch the SATA port on the motherboard. Try resetting the CMOS jumper. Run memtest. There's lots of stuff you can do to get to the bottom of the issue. The lazy option is just to pronounce the board toast, but if that's not really the problem you're going to get a lot of heat from the client. Intermittent issues like this are a real pain to diagnose. That's why for custom systems and Alienwares and such, My diagnostic rate is DOUBLE my normal rate.

And after the customer has waited way too long and you've done all the diagnostic tests you can think of you scratch your head and pronounce the board deceased you replace it anyway!

 

[Altster]

As far as what you are experiencing here, my way of diagnosing issues like this is to strip it down to its bare necessities (min. ram, etc.) just to make it work (boot into whatever O/S or diagnostics program you wish to use) and run the heck out of it. If it passes a basic test, then start adding more memory - just one at a time - and then run the heck out of it again. Take your time.

And always make sure the internals are clean - no clogged fans, etc.. Make sure the heat sink(s) are clean and clear of dust / debris. You might even want to check the heat sink compound between the CPU and heat sink. I clean those all of the time and re-apply fresh compound.

I learned how to do much of my type of repairing from trial and error in my early days. I bet I've forgotten more that I've ever learned.

 

[GTP]

Time is money though.....
Is your client ok to pay for 5-6 hours of "diagnosis" only to be told "Yep - It's your motherboard..!"

 

[glennd]

. I bet I've forgotten more that I've ever learned.

I've pretty much forgotten everything I've learned.

 

[ComputerRepairTech]

How long did you run prime95 for? also what 2 drives are in the raid0 normal hard drives or (hybrid or ssd drives)?