Alienware 17 R3 - Suddenly screen goes black / system appears to lock after months of stability

Now I know what you are thinking. What did you install? What did you change? That's the funny thing. Besides installing NBA 2K18, I hadn't done anything but Windows 10 64-bit updates in MONTHS. I added a 27" Samsung 4K display maybe 3 months ago, when a buddy traded it to me. It's a slightly older model without HDMI 2.0, so I'm using a DisplayPort cable to run it at 4K res and 60 Hz. It works fine, as does the internal 4K display, at least until I try to play any games.

I was playing NBA 2K18, which is to say, managing the team in MyGM mode. After the first four days or so, I went to sim a few more games into the season, and the first black screen occurred. From game to nothing, unplugging the DisplayPort cable didn't do anything. The system appears to lock up, and I can't do anything but hold the power button and turn off the notebook.

I immediately started looking at video drivers, but I'm on the newest Nvidia that works with my built-in 970M and the ASUS 1080 GTX in my AGA. I'll mention I AM running ASUS GPU TweakII , but I'm running it in the standard "Gaming" mode with no-overclocking. Driver is 385.41, Asus tool is v1.3.6.0.

I then looked at drivers. There was a Thermal one on Dell's / Alienware's site that I didn't have so I loaded it after months without. Didn't seem to make any difference, so I uninstalled it again.

I then looked at the BIOS. I was on 1.3.6, which I'd been told in the past was the best for my AGA, so I'd never gone past that. But with the black screens being increasingly frustrating I downloaded Rufus, put 1.3.10 and 12 on a USB stick, changed my BIOS settings from SecureBoot and UEFI to Legacy, and first booted using F12 and applying 1.3.10, putting the BIOS back for regular Windows 10 use, and testing the OS. Once that tested okay, I updated to 1.3.12. That's what I've got the system using now.

Thinking because this had worked for months with Halo, Witcher III, and a few other games... maybe it was just NBA 2K18. But I loaded Halo 2, and while it took a while, it eventually locked the system as well.

Then, as I've been sitting here typing, the keyboard seemed to "slow down" and wasn't working very well. Thinking the system might be overheating, because I don't know what's performing fan control under Windows, but the CPU fans never speed up, I went again and used HWinfo64 to run the fans at about 4000 RPM, both CPU and GPU. (Since I'm using the AGA, I didn't notice the GPU setting make any difference.) The laptop fans are audible, but I've been able to type better with them turned up like this.

I know there is an option in the BIOS that seems to run the fans at full ALL the time. Forums like these said that was unnecessary and don't use it. But something seems to be affecting my system stability. And it all started maybe 3 days ago, after MONTHS of use without issue.

Any thoughts are appreciated. Thanks so much for your time.

 

Okay. So I uninstalled ASUS GPU TweakII, and installed the 3 remaining drivers I hadn't used previously from Dell's site, hoping if any of them affect the fans and/or heat management, worth a try. The sytem rebooted and seems good, but the fans STILL weren't spinning up at all. Figuring it was worth taking a look in HWiNFO64 again, I opened it up to similar info as before. The CPU is sometimes as low as 38c, but as gotten as high as 69c. I went to the fan page, wondering if it might remember the settings I used last time, but it hadn't. Both the CPU and GPU fans *seemed* to be set to AUTO. I say seemed, because clicking Auto on CPU didn't seem to make a difference. But when I clicked Auto on GPU, I heard fans audibly spin up. I wondered if this might be my 1080 card, but no... the noise is definitely coming from the laptop chassis. I've left it set that way since I started typing this post, and the keyboard is working well, and things seem a bit more stable than before.

Think I'm gonna try uninstalling and reinstalling the Nvidia driver. Despite it supposedly not updating recently, I wonder if TRYING to update to the latest might have screwed something up. Hoping to find that ONE THNG that might resolve the chaos my primary computer is somehow causing now.

 

Okay. Wasn't exactly sure which video card I had, so I searched for it on ASUS website and followed a link for their "Tweak" tool that controls fans and such. They released an updated version to fix issues with the old one at the end of August. No chance that's what I had, even though I'd used the old tool to look for updates (supposedly), so I uninstalled, rebooted, and installed clean. I've also followed the advice from another thread here about using HWiNFO64 to manage the CPU fan, so hopefully that will help keep this system cooler and more stable. For the time being I still haven't uninstalled and reinstalled the video driver. That will be next on the agenda if it comes to that. Gonna try and game a little and see how it goes.

 

Read what you typed but I haven't seen anything yet other than software troubleshooting. Hard locks and black screens from my experience are almost always power related. So first thing I would do is narrow it down to the laptop or the AGA. Disconnect the AGA, is the issue repeatable? Yes then it is something with the laptop. Does it happen while it is on the charger or off? If it happens with both now we're looking at either a driver or hardware issue.

My guess is either the PSU in the AGA is giving out or your charger is. That's my first knee-jerk reaction. I'll give an example of an issue I just had that I fixed:

All of a sudden I was having issues with flickering and graphics corruption in the game Prey. This I first noticed while on my laptop away from my AGA so again I did the same checklist you did. Ok it works on the AGA and I didn't notice it until I disconnected it so it must be a driver issue or my GPU is dying. I tried a different game, it was OK. I ran futuremark, it was OK so I know for a fact the GPU is OK so it must be a driver issue. I rolled back the driver I had currently installed. No change. I tried changing the resolution and it went away but later came back. Slowly tried every setting I could think of and thought it might be related to the display panel. In this case I started to troubleshoot the panel and the first thing I disabled was vsync (vertical sync). Immediately everything was fixed. I tried it again over a couple of hours to see if the issue would reoccur. Then I re-enabled vsync and the issue immediately came back. So the solution was a driver bug related to vsync specific to NVidia.

That's how you do it. You've got to narrow down the issue before you move to software

 

Have you checked the stability of undervolts on your CPU? Which one are you using 6700hq or 6820hk? Try updating to Intel latest drivers for hd 530. Use DDU to clear out older intel drivers in safe mode and install the newer drivers.
You could turn on CPU Performance mode in BIOS to speed up the fans, but whenever you're doing BIOS upgrade revert it back to Disabled.

 

Intel Core i7-6700HQ CPU. Undervolts? How do you mean? Watching the voltage in HWiNFO64 for a while, I see the voltage moving between 1.08 and 1.13, just keeping an eye on it for about a minute. It moves around a lot on the System Summary screen.

Intel Drivers - Yeah, I thought about that. Those probably aren't up to date, so I'll do that now. I'm not familiar with DDU, but I'll look it up.

As for the CPU Performance mode - That's the funny thing. I've NEVER had to "mess" (bad terminology, but I assume you'll understand my meaning) with the fans to keep the system stable before. Now all of the sudden, I have to upgrade the ASUS GPU Tweak II app, and run HWiNFO64 with a custom profile to cool the laptop. It was fine for months, what changed? Wasn't like I upgraded the BIOS and THEN the stability problem happened, or I'd think THAT did something.

Definitely appreciate the suggestions guys. After messing with it for FAR too long yesterday, as long as I make sure both the new GPU app AND HWiNFO64 are running and the fans are set to keep things cool. (I've got HWiNFO64 set to start at boot, then minimize, but I'm not sure if the fans are running right without opening the sensors and looking at the fans at least once specifically. I saw someone had a line to add to the ini for that, but it wouldn't let me edit the file. And while I can force it, I was surprised it wouldn't just let me do it with the app closed, so I haven't done it yet.)

 

May I know the rpm set in HWINFO?

 

Sure thing. 2500 RPM for 0 to 50c. 5200 for 60 to 80c.

I've downloaded the latest Intel driver, and it wouldn't install (installed driver is dated 3/2017 and 21.20.16.4627 version) with the EXE or by pointing to the extracted ZIP folder. Gonna reboot into Safe Mode now and use DDU, then install the new one.

 

Obviously, it's no guarantee things are "fixed", but my system booted and loaded all the "extra stuff" down in the taskbar a LOT faster after updating that driver. And having the latest isn't a bad thing, so thanks for those suggestions.

Was a little worried the "default" driver from Intel might not "Integrate" well with the AGA and all, but seems to be ok thus far.

Gonna try running NBA 2K18, since that's what started this whole mess. Course it running isn't prove of anything, but it'll give me an idea if it's STILL working. If I get brave, I might even bump the resolution back up to 4K and see if it locks the system or not. Running the game at 1080p seems fine, but resolutions that worked for 4 days have brought pain the last several times I tried them.

Any idea if I did Safe Mode, DDU, and then tried installing the latest Nvidia driver that came out on the 20th, if that works on a Alienware 17 R3? Currently it gives an error about not supporting the laptop's integrated gpu (970M). I know I can hack it to work, but I haven't tried that yet either.

Thanks again for the time. That's why I love these forums. When I have no one else to ask, folks here are VERY helpful. Notebook Review and it's followers ROCK!!!

 

One more thing I'd like to say is, disable driver updates from windows update using winaero tweaker. That'll prevent WU from breaking drivers. If you want to cleanup older drivers to trim Windows use Driver Store explorer.

 

Ironic you would suggest that. I thought for sure because the new drivers were from August, they'd be fine for a while without Windows Update messing with them. But sure enough, I DID see a flash, and while the game ran for over an hour, it did have a little trouble when I exited it. Instead of exiting and going to the windows desktop, things looked hung up. But when I hit the power button it logged me out of Windows and I was able to log back in. When I checked the Intel driver, it was from Jan., not August.

I hate to kill ALL of Windows Update because I do believe MS does a better job at updating our drivers correctly than many users do. But my system SHOULD have most of the correct drivers and need few updates besides video drivers. Guess at some point I'll go back to safe mode, disable WU, redo the Intel driver, and see how it goes.

Thanks again!!!

 

So I reinstalled the newest Intel driver, made sure it "stuck", and tried running NBA 2K18 at 4K resolution. It actually took it, I got in and fooled around for a few minutes, seeing the video flash 1, and then maybe 20 seconds later the system locked.
Held power, rebooted, got back into the OS, went to look at the temp settings, but before I could get there the system locked again.
Rebooted, played with the ASUS Tweak Settings, customizing the fan to come on a little sooner, bumping the power to the video card just one slider over, hoping if it's getting under powered, that might help, and then came back here to post. Gonna go watch football at a cousin's house, then I'll come back and see how it looks. I'll power the system down so if it IS a heat issue it doesn't get worse while I'm gone.
Certainly tempted to wipe the Nvidia driver, figure out which one was installed for the Destiny Beta if it's not what I'm using, and go back to that. That driver was rock solid before this whole mess began. If IT changed, well maybe that's a big part of the problem. Or maybe it IS hardware like another poster noted. I didn't really look at power info before this mess happened, so I have no way to compare and know.
The troubleshooting continues...

 

I think ASUS tweaking app is the culprit. Not sure though. Grab a beta version of ASUS tweak utility, I think it has a conflict with driver and creators update.

 

We got a 2 17 R3's here one with an AGA and the one without that has issues as well it will lock up randomly but has gone 4 days without a problem and sometimes when starting or restarting it would come up with blank screens so after reading this thread and doing a search I found that the Intel graphics have had common problems with the blank screen so I removed both the Intel and Nvidia graphics drivers with DDU and updated the Nvidia and reinstalled the intel but when checking windows update it started downloading two different version intel drivers(I didn't like the idea of windows doing this) so I stopped it updating and just got the latest driver from Dell support site installed it and did a windows update check and it was up to date.

Will need to a few days to see if this fixes it as it is totally random.

I also use HWinfo to control fans and have to bring up the fan control panel for it to control the fans.

 

Sorry your having trouble lunatic, but it's nice to have some help with the situation. I did the same with my Intel, blocking Windows Update from letting it mess with things as Vesudev suggested. I'll go look for a beta to the ASUS Tweak II app and see what I find.

Update: Didn't find a Beta, but the link to all the versions is here. https://www.asus.com/supportonly/GPUTweak II/HelpDesk_Download/

Didn't see the skin'd version before. I need that Destiny 2 one for my 1080 GTX!

 

Can't help with the black screen lock, but if you mentioned higher temps with same ambient as before, that's always due, assuming fans are not damaged, to degrading thermal paste, or dust in the fans (if the fans are spinning at proper RPM). Again probably nothing to do with your posts but it's still something useful for the future. A repaste with IC7 (for badly fitting tripod heatsinks), Phobya Nanogrease or Kryonaut (for fixed heatsinks that are properly mounted for solid pressure and no gaps) and repairing brushless DC Fans (several videos on youtube for disassembling brushless laptop fans) can work well also.

 

I don't think I'm running higher temps. I think stability has simply gone to ****. I dunno if NBA 2K18 is causing ALL of this or what, but it's damn frustrating to have 2x-3x as many processes running, controlling fans and other things via user profiles that were TOTALLY unnecessary before, AND I can't run the game at the same resolution as before. WTF!

 

Are there any apps that can help me control the CPU / Motherboard fans on the Alienware, without be locked to 2500 and 5200 RPM? Certainly seems like heat is locking up my laptop. I was in the middle of moving files around, after playing a video for a while. Out of the blue, the computer got flakey, apps quit responding, and I couldn't even get to Task Manager. Rebooting, it took HWiNFO64 5.57 (Beta) a LONG time to open after the sensors screen popped up. Maybe 30 seconds later, I could control the laptop again. Once I could, I opened the sensor screen, went to fans, and set 50c and above to run the fans at 5200. But having to run them THAT high creates more noise than I want, and then the fans turn off and on again more often than they really need to. With a little better speed adjustment, I think I could keep the laptop cool, without having such an "off and on again" scenario.

Thoughts?

 

You need to enable EC support and PCH Thermal device in hwinfo preferences.

 

Ahhh. Much better. Thanks!

Before changing that, I set my BIOS to Performance CPU Fan. Seems that's overriding the HWiNFO64 settings and speeding things up as I go. Gonna reboot, turn that off in the BIOS, and see how it goes. First with the fan setting of System Auto, and then changing back to MY fan profile as necessary.

I was trying to rip a disc, and play NBA 2K18 earlier. The game worked for about a minute, maybe less... and crashed. Then when I rebooted, I noticed the BT driver wasn't loading correctly. Uninstalled, reinstalled, and got that fixed. But then for several attempts in a row, every time I started the laptop it said the video card in my AGA wasn't valid, and I needed to install an Nvidia or/and AMD. Ironically, once the BT driver was fixed, I got a message to reboot to finishing installing the AGA, which I hadn't made any changes I know of, but then it started working again. So odd.

Anyway, I'll reboot and check my fans again.

 

So NBA 2K18 got an update today. After it did, I switched the graphics back to 4K resolution, and it seemed to be pretty stable for 10 minutes or so. I was looking at trades and various things, and then the game black screened.

It was at this point I decided I don't know enough about the settings for my ASUS 1080 GTX Founder's Edition, so I google searched the card, and came to realize I was providing NO additional voltage to the card, but the clock frequency in GPU Tweak II was set to the BOOST mode of 1700 something. I have now dropped the card's speed down to 1616 in GPU Tweak, set the memory clock to 10010 Mhz, and left the power target set to 100% and the GPU Temp Target set to 83c. (Though the fans are set to kick in well before that of course.)

Gonna try again and see if these settings work for me. I suppose I should have asked long before here, but does anyone else have the ASUS Founder's Edition, or another "First Edition" 1080 GTX with the earlier "reference" cooling system? If so, what clock speeds, voltages, and such are you using?

 

Just check your RAM for any errors or buy another stick of RAM. Reduce Page file to 1024MB.

 

Downloaded the latest memtest and created a USB stick. Ran the test through 1 and a 1/2 passes, not a single error. My experience has been that if you have a bad stick, or a problem with a socket, the 3rd or 4th test in a pass throw errors right away and often, so I don't think that's it. Though I didn't realize I had Corsair before. DDR4 running at 2134. Not sure if I can run faster in this device.

I also changed the page file. I had already moved it from one SSD, to another SSD, and now my SATA drive. Last change before this one it was set to about 4GB, but as you suggest I reduced it to 1GB.

However, when I ran NBA 2K18, and made sure I hadn't gone back to 4K mode which seems like creeping death to that game, it STILL crashed after a couple minutes. It MAY be the video driver isn't compatible with the game. I've read several reports of Nvidia users like myself having issues. And I haven't unplugged the laptop from the AGA and tried it without it. But since the Internal video card is a Nvidia too, I don't have high hopes that's a proper fix.

It's been about two weeks since I bought the game. I don't suppose 2K or steam would give me a refund at this point, but they ought to give me a medal for not *****ing sooner. What a mess.

 

One more step of troubleshooting, try changing Physx from Auto select to CPU.

 

I went more radical. At one point, the black screen occurred after playing Halo 2 on the system for a time. But instead of fully locking up the system, the external LCD went black, then the laptop LCD turned on. And while it wasn't possible to make changes in apps with the mouse, I was able to use the keyboard to bring up task manager and kill the game. Once that was done, I had full Windows 10 control again.

It was at this point I shut down the system, unplugged the Thunderbolt cable and anything else that was connected to the AGA, and plugged the necessary cables into the laptop so I wasn't using the AGA or the external, just everything plugged directly into the laptop. I just got down passing a particularly long level in Halo 2, and the system never black screened, never glitched, and I have a feeling I MAY have been right, and the power supply on the AGA might be the issue. After all, I opened it, the cables are all plugged in, the "motherboard" or whatever you call the board the video card plugs into looks fine, the fans are clean and NOT dirty or impeded, and I can't explain what else would be causing the instability.

Also of note, when the system booted for the first time after the AGA was removed, the onboard Nvidia (GTX 970M) had no driver loaded and wasn't functioning. Since I had the latest non-beta GeForce Experience loaded, I used that to download the latest Beta driver. That driver had NOT loaded while the AGA was connected, I THOUGHT because the driver didn't support the GTX 970M. But when I ran the express install without the AGA and my 1080 GTX, it installed and worked fine. THAT is the driver I just got done playing Halo 2 with, and everything worked great.

I have not tested NBA 2K18, because I had to downloaded and reinstall it for the *5th* time. Sometimes after the black screen, I go to run the game, but the OS doesn't seem to remember it's installed. The only thing I can do is delete the former install, and reinstall it where it was, or install it on a different drive. Comcast is probably gonna charge me $1000 for the over 250GB of bandwidth I've used JUST on NBA 2K18 in the last two weeks. Not even considering my Destiny 2 gameplay, etc. Uggh.

So yeah, if anyone has any AGA troubleshooting tips, let me know. I'm leaning toward getting another power supply and trying that at this point. I think I saw a post recently where someone used a Corsair 650W power supply that fit inside the case and it could close. Sounds like a good option to me, as my last couple Enermax weren't as good as older units were, and Corsair's have been bulletproof in my experience.

 

So for giggles I hooked back up the Alienware graphics amp, but I left the lid open. This way, I could make sure the front fan was spinning (it is), etc. What's interesting is there is NO airflow through the power supply, afaik. One more thing that has me leaning toward trying to replace the PS with a Corsair 650 like the gentleman in the AGA thread said it got to work. First post of page 301. Haven't 100% decided to open it and swap the PS yet, but I'm getting closer all the time.

NBA 2K18 was stable playing in lower resolutions on the 970M. I couldn't get reasonable speed even at 1080p. Just moving through menus was painful. So I decreased the res to something like 1300 x 768 or something. But I'm not running the ASUS GPU Tweak software, or the HWiNFO64 software right now. I tend to think it's power or heat, but maybe those apps add to the problem?

 

Okay. So I realized that the game was playing WAY too slow. Just navigating menus wasn't work how it had before. So I went to the desktop, right-clicked, chose the Nvidia menu, created a "profile" for NBA 2K18, and told it to use the 1080 GTX to play the game, not the Intel integrated graphics. Performance immediately went back up, but the game crashed to the desktop with the AGA lid still open. I'm still leaning toward trying swapping the power supply and seeing if the game plays better. Otherwise, I must have a setting that's wrong in the driver, and it's impacting the ability to play the game in the 4K resolution with all the usual settings enabled, etc.

 

Okay. So this morning I couldn't restore the Nvidia Control Panel defaults. If I did, the app crashed. Figured with my uninstalling and reinstalling, something must be amiss. I also noticed my Intel video driver was doing it's own thing again. So I disabled Windows from changing drivers, rebooted to safe mode, wiped the Intel driver and rebooted, safe mode, installed the Intel driver from Dell's website and rebooted, safe mode, wiped the nvidia driver and rebooted, safe mode, wiped the geforce experience, loaded Nvidia 385.41 using custom and telling it to do a clean install and rebooted.

At this point, I was back to being able to reset Nvidia settings to their defaults. I tried running NBA 2K18, and it was slow. Knowing this meant the Nvidia card wasn't running the game, I went into Nvidia control panel, setup a profile for NBA 2K18, changed the settings to be Nvidia with most functionality set to program managed or on, clicked apply, and ran NBA 2K18. Froze my system.

Rebooted, deleted the NBA 2K18 profile in Nvidia Control Panel. Then I set the whole system to use the 1080 GTX for all programs. Ran NBA 2K18, and it locked up. Rebooted, ran Halo, it locked up. Shut down the system and came over here to type what I had learned.

Seems to me, while I don't want to install the Corsair 650 power supply JUST to do it and find out NOTHING has changed... if I AM having some sort of a power problem in the AGA, that should fix it. And that driver has NOT had problems running other things previously. Granted, it seems Halo 2 may have been using the Intel 530 driver all this time, as the game plays fine with it. I could do some testing with Witcher 3 and other games, and see if THEY lock up when using the Nvidia card, but unless NBA 2K18 has driver issues... and again, it DID run for 4 days straight without issues before all trouble started, a power problem from a power supply who's fan never seems to run, might be an issue.

I presume the Corsair fan will actually pull some air in from the side, (when mounted in the AGA,) and expel some air out the back if installed. I don't think it's designed to just sit there and not seem to move ANY air like this Dell PS is.

Thoughts before I pull the Dell PS and try the Corsair? Got knows I've run out of patience and not sure what else to try on my own.

P.S. All my testing last night and this morning was with the AGA, and the 28" Samsung connected via DisplayPort cable, but turned off. So the only monitor I've used during this time is the Internal LCD. Because the Samsung was off, Windows 10 doesn't even see it for me to select it, etc. But all my issues from the last two weeks are happening on the internal LCD just the same as they do on the Samsung external. Just FYI.