Start a Conversation

Unsolved

This post is more than 5 years old

6698

August 24th, 2017 19:00

Aurora R4 booting directly to black screen with blinking cursor

Okay I'm posting on behalf of my brother, he has an Alienware aurora R4 desktop computer that's been having problems about 6 weeks now.

started with the BSOD popping up and forcing the computer to reboot, I tracked that problem to a dying hard drive so I ordered a new one and installed it a few weeks ago.

had about 3 weeks where the computer was perfect aside from an issue had while trying to restore from an external that lead the thing to say there was two operating systems but it didn't seem to effect the system so I'm not sure it's relevant.

night before last it blue screened again.

according to the crash info I had someone look over for me, it appeared the bios was outdated, there was issue with the amd drivers and evidently the new hard drive was throwing errors.

on top of this when he tried to boot it, it would come to a screen that said previous over clocking had failed (it's never been overclocked) and asked if you wanted to reset original values but if you select yes the computer would freeze.

managed to get to Windows once but it froze solid as soon as I attempted to load the Internet.

i currently have the side off the tower, there's no beeping or Amber light upon start up, sometimes the colourful side lights come on, sometimes they don't and currently the alien head is not lit up.

when I attempt to reboot now the fans kick on, all the lights come on for a moment and it goes straight to a blank screen with a little blinking dash in the upper left hand corner.

no splash screen or anything information is being showed.

i was told earlier that I should reset cmos so I looked up how to do that but try as I might I can NOT find any jumpers, cables or the damn battery that would accomplish this reset.

at an absolute loss here.

3 Apprentice

 • 

4.4K Posts

August 25th, 2017 09:00

Hi misspaperwait‌,

How many video cards does the system have?

For the CMOS reset you can see a picture here or you can follow the steps on the online manual.

8 Wizard

 • 

17K Posts

August 26th, 2017 14:00

What version of Windows?

- Did you clean install it or just a clone? 

What AMD video card?

Does machine pass ePSA Diags?

Just troubleshoot like a normal computer:

 

 

August 27th, 2017 00:00

Only one graphics card, clean install of windows 10.

I reset CMOS yesterday and now there's a splash screen and all that good stuff, it passed ePSA diags with no errors reported at all.

I've been troubleshooting all day, currently running Memtest86+ on the individual sticks of ram, no errors there thus far either.

Since resetting CMOS it no longer shows an overclocking error so thats an improvement.

8 Wizard

 • 

17K Posts

August 27th, 2017 12:00

Also, Aurora-R4 must be running latest BIOS.

 

The final BIOS fixed PCIe-slots/lanes to cure BSOD problem for Nvidia cards. However, some users reported it then broke the nice AMD cards of the time (causing them to start random BSOD-ing). 

 

To verify your motherboard is good, I would not only try a different video-card, but might be a good idea to make it a small Nvidia one.

8 Wizard

 • 

17K Posts

August 27th, 2017 12:00

misspaperwait wrote:

1a. Only one graphics card,

1b. clean install of windows 10.

 

2. I reset CMOS yesterday and now there's a splash screen and all that good stuff, Since resetting CMOS it no longer shows an overclocking error so thats an improvement.

3. it passed ePSA diags with no errors reported at all.

 

4. I've been troubleshooting all day,

5. currently running Memtest86+ on the individual sticks of ram, no errors there thus far either.

 

1a. Good. Like that thread shows, maybe try a different one.

1b. Good

2. Good. I would pop-in a new CMOS battery.

 

3. Good. Not only for those checks, but maybe that checks Liquid Cooler in a slight way outside of Windows. Too bad LQ is so hard to check on Aurora R3-R4 (unlike most Alienware desktops). If ePSA is pushing CPU hard-enough, it should over-heat in 10 minutes if LQ is bad.

 

4. Been there, done that. Systematic problems you are describing is sounding-like like a bad motherboard, video-card, power-supply etc. If you can't run Windows, or get through a Windows install, all hardware must be tested outside Windows. Again, too bad no true DOS/Linux 32bit Dell Diagnostics for this machine.

 

5. Good. But why not test it all at once (does it fail ?). 

 

The re-occurring HDD problem is also troubling. Do those drives pass SMART checks in other machines? I would pop-in a new SATA-3/600 rated cable for C: regardless.  A SSD would have been nice (old spinner HDD becomes D:)

 

Power-Supply must also be checked. Be sure machine is on good UPS like APC.

August 29th, 2017 00:00

I've been testing the individual sticks and slots of ram because when I originally ran memtest86 (first time I got memtest86 version 7.4 free not realizing it wasn't the same thing as memtest86+) I had all four sticks in and it started throwing errors almost immediately.

i downloaded the proper software and pulled all the sticks but one, tested each stick and each slot on their own and had no errors, I put the 4 sticks back in and ran memtest86+, left it going over night it completed 12 passes and found no errors.

a fellow that's trying to help me out on another website told me to start the testing over again with the first version of memtest I was using so I once again did each stick and it came back with no errors, I'm currently working on different combos with 2 sticks and have no errors but as soon as I put them all in it starts throwing errors again so I have no idea what the deal is there and I'm going to keep trying combos to see if anything else makes it throw errors.

nothing has ever needed to be replaced on this thing up until a month ago when it started having BSOD and the crash logs pointed directly to a dying hard drive. 

It's possible that the errors shown in the hard drive are caused by a partial restore that happened right after the new hard drive was installed; when the old one was dying I did a full back up on an external so once the new drive was installed I tried to restore from the back up, I tossed it on and went out with a friend for a few hours, when I got home I discovered that the percentage had apparently stopped moving for an hour or so and my brother decided it was frozen and forced the computer to shut down. When it booted back up there was two windows operating systems showed, only one working of course but I'm wondering if the partial, clearly corrupted os could be to blame for the hard drive errors.

after I reset CMOS when I booted the computer it would attempt to load windows and then after a few moments of spinning it goes into "preparing to automatically repair windows" mode and eventually comes up with "windows was unable to start" type thing and gives the different repair options and such so it's possible I may be able to do a new clean windows install at this point, I haven't tried at this point as I've been up to eyeballs in hardware testing

August 29th, 2017 01:00

Oh and to answer your question, right before it started getting the black screen thing (which was resolved by resetting cmos) it was pointed out to me that the bios version is from 2011 but I haven't had the chance to attempt and update it as I've been testing hardware the last few days. 

8 Wizard

 • 

17K Posts

August 29th, 2017 10:00

Ok, good post.

YOU MUST get final BIOS on there eventually (before you clean-install Windows), but only do it after ram-problem is figured-out and machine is stable enough to complete it without crashing. Otherwise, you might brick/kill the motherboard. I still say doing it outside of Windows is safest.

Machine on good UPS like APC. This first step is NOT optional. These tests must run for hours at a time. I don't care how good you think your power is, there is no way it will make it thru them successfully without a single power-glitch. 

 

Must pass ePSA Long-Test first. Under the circumstances, we will say "at least twice".

 

You want to use PassMark's MemTest86.com, and do the Long Test. On a newer UEFI machine like that, the newer v7 version should run automatically. You should not have even one RAM error. Two complete (error free) passes would be nice.

 

If you get this far, then update BIOS.

 

YOU MUST get final BIOS on there eventually (before you clean-install Windows), but only do it after ram-problem is figured-out and machine is stable enough to complete it without crashing. Otherwise, you might brick/kill the motherboard. I still say doing it outside of Windows is safest.

 

Finally, Nuke-and-Pave system and clean install Windows-10 64bit. All that "two windows operating systems showed" stuff needs to be formatted away. Start with a completely blank and uninitialized FDisked C: drive (DiskPart). See this post:

fixed 

 

Give main-user a spare computer to use while you work on this Aurora-R4. Don't let owner start using system for real until you are sure machine is 100% working. You should be allowed to erase it at a moments notice (until it's fixed).

 

Install completely repaired Aurora-R4 at final site, with good UPS (like APC). Should be on "safe-side" of router's firewall. Windows Defender should be updated and running.

August 29th, 2017 11:00

I've been running memtest86+ for literally days trying out different combos and such and getting zero errors (each test is,left to run for 8 passes, the final test with all four stick in went for 17hrs and completed 12 passes without a single error) 

the he only time I've seen an error is when I run memtest86 version 7.4 free with all four sticks in for whatever reason.

i haven't experienced any freeze ups or error screens since resetting CMOS so I'm calling that progress lol 

i have a few more combos of two to try out with the ram to see if it triggers any errors in duel channel mode so if I figure out why I was being thrown errors I'll let you know.

how does one update bios? Can you put the update on a thumb drive and install it?

oh the computer makes it past POST every boot up since I reset CMOS, no more auto reboots.

ePSA ran the short version first and found nothing and then asked if we wanted to run more extensive tests and I said yes, it  ran for an hourish I'd say and again spit back zero errors so in theory that would be a good thing right?

there is a single, quick beep when you first hit the power button but I'm not sure if that was always there or not, never payed that much attention to be completely honest lol

8 Wizard

 • 

17K Posts

August 29th, 2017 15:00

misspaperwait wrote:

1. I've been running memtest86+ for literally days trying out different combos and such and getting zero errors (each test is,left to run for 8 passes, the final test with all four stick in went for 17hrs and completed 12 passes without a single error) 

 

2. the he only time I've seen an error is when I run memtest86 version 7.4 free with all four sticks in for whatever reason.

i have a few more combos of two to try out with the ram to see if it triggers any errors in duel channel mode so if I figure out why I was being thrown errors I'll let you know.

 

3. i haven't experienced any freeze ups or error screens since resetting CMOS so I'm calling that progress lol 

oh the computer makes it past POST every boot up since I reset CMOS, no more auto reboots.

 

4. how does one update bios? Can you put the update on a thumb drive and install it?

 

5. ePSA ran the short version first and found nothing and then asked if we wanted to run more extensive tests and I said yes, it  ran for an hourish I'd say and again spit back zero errors so in theory that would be a good thing right?

 

6. there is a single, quick beep when you first hit the power button but I'm not sure if that was always there or not, never payed that much attention to be completely honest lol

1. Good, but I don't think old MemTest86+ is not even maintained any more. I don't think it supports newer ram memory technologies, so it's likely just testing in a lesser compatibility mode.

 

2. PassMark's MemTest86.com v7.x should work I think. Might be best way to test at full speed and in dual-channel. I think it failing might point to a problem.

 

Aurora-R4 has 2 memory banks of 2 DIMMs each. Each bank must have IDENTICAL DIMMs installed (a matched set). See PDF Service Manual because slots might not be assigned to Banks as you would think.

 

3. Good. 

 

4. See my Featured Content

5. ePSA is not "all telling" but it's an initial hurdle you must pass.

 

6. I forget if that's normal or if its trying to alert you to something. Someone with an Aurora-R4 will have to answer.

August 29th, 2017 17:00

I've been testing with memtest86 version 7.4 since yesterday, did each stick individually and I'm currently running different combos with two sticks at a time as per the other guys request, so far absolutely nothing has shown any errors.

i have one combo left to run but so far absolutely no errors with version 7.4 unless all 4 sticks are in.

the sticks are the ones that shipped with the computer day one, never been touched up until now so it shouldn't be a compatibility issue I wouldn't think

August 29th, 2017 19:00

So I finished the list of ram to slot combos the other guy requested I do and there was no errors in any of the tests so I don't know what the deal is at this point because every slot and stick has been tested individually, they've now been tested in pairs and no errors any where unless you test with all the sticks in at once.

false positive?

8 Wizard

 • 

17K Posts

August 29th, 2017 20:00

misspaperwait wrote:

1. So I finished the list of ram to slot combos the other guy requested I do and there was no errors in any of the tests so I don't know what the deal is at this point because every slot and stick has been tested individually, they've now been tested in pairs and no errors any where

2. unless you test with all the sticks in at once ... false positive?

1. With PassMark's MemTest86 ?

 

2. Possible but doubtful.

 

What memory is this? Is it the original Dell supplied memory? Is it all the same or at least matched pairs?

August 29th, 2017 21:00

All the two stick combos I tested passed PassMarks memtest86 with zero errors.

the only time I've ever gotten errors during this testing is if all 4 sticks are in.

someone recommended adjusting the voltage settings and trying again?

the ram is dell supplied, it's never been changed so it's what they put in while building the machine 

August 29th, 2017 21:00

I have idea if I was having ram issues or not to start with because the crash dumps didn't mention anything about memory at all.

what they did mention was the fact the bios was from 2011

thw amd display driver needed to be removed and reinstalled 

and finally that the hard drive was throwing some errors but I'm thinking if I clean installed Windows that those errors might go away as I know there's some corrupt files on there anyways.

the only reason the ram was tested at all was because someone recommended it as part of the basic troubleshooting steps

No Events found!

Top