Not sure whats going on here....(Artifacts M18x R3)

Had my AW 18 running for a few days in a row. Simply sitting in Windows (10 Pro 64), only thing running was a dedicated (local) ARK Server.


Woke up this morning to "artifacts" (I put in quotes as it isn't artifacting like I am used to, this is just a distortion across the entire screen). Well damn. Rebooted into the BIOS, still have the same issue (so its not an OS problem). So I began with a full power down then back up, no change. Pulled power and battery, held power for 15 secs or so. Reconnected and powered up, no change.


Then I got more in depth. I assumed that I had a bad GPU so I tore down the entire laptop and removed the 2nd GPU. Powered back up, no change. Figured the 1st GPU must be the issue so I removed the HSF from both GPUs and swapped (re-pasted) GPUs with the 2nd GPU in the slot for GPU 1. Still, no change. So now I am a bit confused as to what the issue could be. Has anyone ever had this before? Any suggestions?


A few pics can be seen- http://www.tgupdate.com/M18xR3.html


The system has always been VERY well taken care of. No parts have ever been OC’d and until today, the system has never been taken apart at all (beyond a simple RAM upgrade about a year ago).



M18x R3

Windows 10 64bit

I7 4700MQ

24GB RAM (8x8x4x4)

GTX 770m x2 SLi

256GB Samsung Evo



Edit- Have pulled all but a single stick of RAM with no change (Swapped that stick with another from another laptop as well with no change).


EDIT- I connected to a second screen and, the display is just fine on the second screen which means, the built in screen spontaneously went bad?!?!?!??

Edit- Continued the tear down all the way to the LCD. Removed the LCD from the bezel. Re-seated the LCD cable at the laptop and on the screen itself. No change.

 

Could be as simple as a damaged cable between your LCD and motherboard. There are really thin wires in those cable assemblies which sometimes break slightly, making poor contact, or sever entirely.

 

Agreed, the cable is most likely damaged and is causing this.. OP get a replacement cable and see if it fixes this.. Its not too expensive anyways!

Sent from my LG-H811 using Tapatalk

 

I had the EXACT same problem with my brother's M18x-R1. I don't want to scare you, but it most likely is the MoBo. Feel free to try the cable though (I tried cable and display prior the MoBo). Good luck!

 

I have the similar type of issue as OP, here's my thread on it. Could you just have a look, thanks !
http://forum.notebookreview.com/threads/gpu-render-related-or-panel-w-r-to-gpu.789391/

 

Mobo, that, would SUCK.

I got a replacement screen. Wasn't that because the issue remains. Last two possibilities I see are cable and mobo.

The case for it not being mobo: My only thought pointing to it not being the mobo is aside from the weird artifacting, the system runs fine. No "sparks" or other power issues. Games still run without any issues. This is no fun at all!

The case for it not being cable: This happened spontaneously. Left laptop open and running at night, found it the next morning (without having been touched) with the screen issue.

Another question, I assume the cable for the screen is different in an M18 vs an M17x R2? I happen to have an older M17x R2 that I could always.....sigh....take apart if the cable match (and I mean connectors, length of the cable may very well be different due to screen size)....for testing.

 

So then, what is the important info on the cable label? As in, what am I searching for to find one?

 

Posted a short vid of the issue:



Note- This vid was taken after over an hour of Witcher 3 playing with no issues aside from the screen "artifacts".

 

From my experience , it might be the gpu getting too hot, and stuff will start to happen .." artifacts ". I also noticed on benchmark , if the card is clock too high , it gets hot , ram gets hot ..and usually i see artifacts.

 

I would tend to agree but in my case A- all parts are running stock speeds no modifications of any kind. And B- The issue occurs instantly as soon as the system is powered on, long before the GPU(s) could get to the point of a thermal failure.

 

I would agree it is an overheating issue of some component in the video pipeline.

The point of thermal failure can happen in less than a second on a defective part.

 

Agreed (though you are stretching it pretty thin there as indeed in all but the most severe cases, you will be able to get past the initial boot logo LONG before a thermal failure would manifest..... ). However I have swapped the GPU now three times. Used each of mine (770m) separately as a single GPU and because I still wasn't sure, I grabbed a 770m from a totally different laptop. Same issue. It's not the GPU(s).

Further, in my experience the symptoms will get worse as time passes in almost every case. But here, I can push the system pretty hard (ran Witcher maxed for an hour+, ran some Heaven for a while and then ran 30 mins of Furmark) and the issue never gets any worse, never changes.

 

Never said it had to be the GPU (just in the image pipeline...).

 

Are you padding the VRAM and VRM modules (chokes, mosfets, ICs) on your video card? The GPU die is not the only component that can get hot. I had artifacts on my 7970m before properly padding said components, and not again since padding them properly. Even when these artifacts were happening, the GPU core temp was hovering around 67-73C under load.

These are all the components that should be padded. Gap sizes here may be different for your heatsink though:

 

I've made sure multiple times that everything is cooled properly and connected properly (and hated the M18x for crappy/tedious disassembly procedure). Tried a few GPUs, tried a new display and a cable. Tried a new MoBo in the end, with the old display and old cable. That was my case, but I'm VERY inclined to think that it is the same here.

A bonus - after leaving it on for about 10-15 minutes (even idle) and rebooting (not shutting-down) everything was OK and would be so until it's turned off. At least till the new MoBo came. Haven't tried if it would go worse, because of the mentioned crappy/tedious disassembly procedure.

 

Not necessarily; display size has little to with interface (main thing is resolution and eDP vs. LVDS). Which panels are they, specifically? Easy enough to check cable compatibility on the lcd-end, though the motherboard side could be different. Without schematics that is difficult to tell, but could always strip the wrapper and check where each wire goes. Length of the cable is irrelevant; no need to affix the display and with the keyboard 'loose' there'll be plenty of shortcuts, sufficient for the M17x's cable to reach the 18's mb.

Part nr. is DC02001PC00.

 

That's a great illustration you have (I will assume you have done a guide at some point, and if not you should because that could be quite helpful for many folks!).

Yes I did re-pad the card(s) (Still running with just a single GPU atm to eliminate any SLi related issues). Used all new pads.

My gut tells me this isn't mobo (that and my distinct lack of any "sparks lol). LVDS cable on the way. If that doesn't work then I'll replace the board, then the CPU, then.....the laptop!

 

Well, I got the cable, tried er out, same result. So, if its not HDD, RAM, GPU, Cable, PSU, or LCD Screen related. That pretty much leaves the Mobo (or the CPU I suppose). Right?

So, where on the board will I find its model information? (Dont want to go after the wrong board).

Still seems so odd. System runs great, no issues at all aside from the graphical artifacts.

 

Have you tried the warm restart, which I suggested in my previous post? That's how I "fixed" things until the new MoBo arrived. As for the MoBo itself, it's underneath the left side black MXM tape... or it was the right one. It was quite some time ago. Anyway, your part numbers are GRP9C or LA-8321P. Try both. Good luck!

 

Yep, tried it. Same result.

 

Well DAMNIT!

Decided oh well, time to try the mobo. I Ordered another motherboard (used, online). Pics of the board were identical to mine but it seems the board I have received is different and now I am getting 7 beeps upon start (CPU related).

My Board (Printed on mobo itself)- VAS10 LA-9332P REV 1.0 (A00) 2013-05-29
My Board (Sticker on RAM slot)- CN-04703X-12961-37C-024E REV A00
My Board (Sticker Under RAM)- VAS 10 02

The Board I Purchased (Printer on Mobo)- VAS10- LA-9332P REV 1.0 (A00) 2013-05-29
The Board I Purchased (Sticker on RAM)- CN-04703X-12963-45F-0135 REV A01
The Board I Purchased (Sticker Under RAM)- VAS 10 01F

Re-assembled, powers up but gives 7 beep's and no POST. Not the right board huh....I'm buggered aren't I?


My CPU- i7-4700MQ

Pics of the boards- http://www.tgupdate.com/sitebuilder/images/20160409_140113-3040x1811.jpg

 

This is the right board, I made a mistake in my previous post thinking that this is an M18x-R2. I'm used the R3 to be just AW18, since it's no longer M18x. Anyway, sorry for the misread.

As for the beeps, have you tried to hold down the "fn" key while booting?

 

Yeah, when i posted, i dunno where my mind grabbed the "x r3" portion of the Title as the system is just an Alienware 18.

Have tried many methods of booting. Holding FN, pulling all power and holding power to drain the juice before trying again. Doing the same with no CMOS battery. I am wagering the key is the REV00 vs REV01. Man, this is like, maximum frustration level here.

Suppose I will....sigh....return the board and go after a board more exactly matching my previous board.

 

Pics with the differences I mentioned circled: http://www.tgupdate.com/sitebuilder/images/mobodiff-3100x1847.jpg

 

I went through two boards, before getting a perfectly working one, even though the seller swore to be "100% tested". Have you tried to return the parts in your old board, just to rule out the CPU? In your case I think it's a defective board. Even though there are differences (like missing DMC slot, and probably a few less obvious stuff), it is supposed to work, even if newer, since this is a replacement part, and it is not supposed to break compatibility. I hope that there is a return policy, ask for exchange, or better yet, return it, get a refund and shop elsewhere, that's what I did.

 

Managed to get the seller to take the board back.

Now, the hunt begins again to find this board.....

 

Well, got the correct, identical board....SAME...DARN...PROBLEM!!

So. Having replaced every single component in play cept for one, I decided it HAD to be the one thing I really didn't think it could be. The CPU. Got an identical CPU (4700mq), plugged er in. Problems resolved. Indeed.

HDD, SSD, LCD, LCD Cable, Mobo, GPU(s), PSU(s), Ribbon Cables, SLi Cables, Thermal Paste, Thermal Pads...............all swapped out. And it was the CPU that went bad. Incidentally, of the 50+ desktops/laptops I have owned in the last 5-10 years, the very FIRST CPU I have ever had go bad. And just might be the first CPU I have ever seen fail that was 100% stock and never altered, never OC'd, never not cooled properly. So. Odd. But the good news is, I am no longer timid to pull GPU's out of these things. So.....

My M17xR2 with Dual 5870's.............whats the max I can put in there (thatd be a plug-n-play card)? (And where do I go for the thermal pads needed as I am fresh out)?

 

A bad CPU is indeed rare. Glad you got it sorted out though. I'd head over to the M17X forum for video card suggestions. Dual 7970Ms seem to be a popular upgrade on the R2.

http://forum.notebookreview.com/threads/m17x-r2-revisited.778180/page-2

 

So, the only possible explanation are failing PCIe lanes, who feed the dGPUs and from there on corruption occurs. This is what I thought when it happened to me as well, but there's no way you can "catch" such a problem, even less to fix it without new parts. Sorry, that your part bin ended-up even bigger than mine, but as I said, it's a nasty problem to locate and fix.