Alienware 17 880m screen artifacts

As @MogRules said it doesn't reset after reboot, that's the whole point of a MUX switch. It completely removes the need to reroute the signal through the iGPU.

 

UPDATE

Alright - I decided to start using the PC as normal again, and I fired up a game. I noticed as I play that the artifacts were kinda returning. Note: At this point the PC had not been restarted since it started working again, so nothing was changed or touched - it came back during use.

Just to clarify, this is nothing compared to what it was before. At this point it looked more like a few dead pixels rather than anything else. Annoying but not hindering the use of the machine in any way.

Anyways, if it wasn't fully working there was no reason to wait any longer. So I did it - I restarted the machine.

Good news is that it doesn't make it worse. The few small artifacts that came back during gaming are still there as well though.

Went straight to NBR to type this, and about halfway through the post I tab'ed to check on it and what do you know, they are not 100% gone but definitely diminishing.


Any thoughts on this guys? I'm testing right now to see if I can provoke it by running some stress tests as I think it might be related to heat, whatever that means. Will return back with more asap.

 

This is just weird.

I ran prime95 and it made no difference at all, which I was kinda hoping it would, since CPU is what most theories are pointing towards.

Next one was FurMark, and within a few minutes the artifacts started happening.

Left it to cool off and they disappeared again.



Note: If you havent read the thread, the artifacts are still there with the 880m removed from the laptop. Right now the iGPU are completely disabled and I am able to provoke it with usage/heat from the GPU (which is not the cause of the problem).

Am I just SOL at this point?

 

Well ... Your options now MXM port, Mobo screen port or maybe PCH (Even though I don't believe PCH can cause such an issue)

 

We've blown out some Dell Precisions (M6400 M6500) like that over time and never really figured out if it was the GPU or the motherboards (or something else). I did swap out the nVidia Quadro card in two cases, but the symptoms persisted. Also, when it did produce the symptoms with the transient blooming, it would also do it via video out (either on the computer or through the docking station ports). So it wasn't the LCD panel. It seemed to triangulate to the motherboard or chassis, if anything.

In our cases it was where the computers were 3-4 years old and beat-to-war-and-back from Asia to Europe to North America by frequent travelers. So after I had no diagnosed repair, the business replaced the computers with new ones without a second thought.

Now here's the strange part. Yes, it gets strange....

In the last case, for some reason, I decided to remove the built-in keyboard and just have the user use an external USB keyboard 24/7. The problem never returned. He ran fine for over a month that way. We put the builtin keyboard back in, and the blooming returned via external monitor about 45 minutes later. I pulled the keyboard, again, and he was OK. He ran that way without incident for 7 days until the new Dell Precision arrived him.

Explain THAT one.

 

DELL happened

 

TL;DR (to my above post) Try pulling the keyboard and run without it, assuming you have a USB one. See if the symptom follows the keyboard. YMMV. This is a hard one. Good luck!

@MahmoudDewy, LOL! I cannot agree more!

 

Can you elaborate a bit? Not sure what exactly these things are, especially the PCH i have no idea.

@Aroc

Now THAT sounds stupid

I do travel alot with the machine but obviously always taking good care of it. Not going to do it now but over the course of the weekend I'll give your keyboard-theory a shot for sure!

 

open HWinfo64 settings and report PCH temperature and see if the temperature affects artifacts

 

Looking through HWinfo64 now and I can't find anything called PCH.

 

You have to enable PCH reporting in HWinfo64 safety settings

 

@MahmoudDewy

Thanks!

I think you are on to something here. I still don't know what it is but there are some interesting tendencies.

The PCH idle temp is around 60C.

It heats up to 75c under use.


BUT! It cool downs alot slower than any other component in the PC.

When the artifacts appear under use they don't go away immediately when exiting or the GPU/CPU cools down, it takes a couple of minutes longer and looking at the artifacts and the PCH temperature these things follow each other perfectly.


What is a PCH and what does this mean? I think this is very interesting.

 

Oh, trust me, it is stupid. but your latest symptoms reminded me of my repressed memories of those Dells.

 

@Aroc

At this point I am listening to any and all suggestions. I appreciate your input and will follow up on it this weekend.

 

Actually what @Aroc said made me think of the PCH as it is passively cooled and when you use external keyboard it doesn't heat up as much as it is passively cooled and it is right under the keyboard. For PCH to reach 75 you must have insane load on it. In my case it reaches 75-80 when I am intensively pushing my machine with GPU pulling 100W and CPU pulling 47W.

What kind of load heats up your PCH to 75 ?

 

It was just a 5 minute run of Overwatch. The CPU and GPU doesnt skip a beat or run hot at all, there are many, many other game that requires alot more.

Let me give FurMark a try and see what happens.


Edit: 5 minutes of FurMark brings the PCH temp up to 89C


Not good I guess.

 

@ MahmoudDewy

That's awesome diagnostic skills you have there. This reminds me of the old days when people used to mod the cooling of Northbridges and Southbridges (add a fan, or a better heat sink), back before we had the PCH.

 

There is still something different about using the Fn sequence and fully disabling the iGPU via a bios hack. If it was that easy nobody would have needed to set PEG for the machine to use an upgraded new GPU? Unless it's for some other reason, anybody know?

And I don't recommend furmark on a suspect GPU. It has been known to trash them. It stresses them far more than ever in use. Stick to something like Heaven 4, ( https://unigine.com/products/benchmarks/heaven/) you can leave it running to see any temp creep. More realistic test.

No help I'm afraid with the root cause...

 

@MickyD1234

Agree with your comment on FurMark. But when the sole purpose of the test was to produce heat, it seems like a decent choice.It ran for 5 minutes tops.

 

Dude that's flaming hot ... That's either the reason behind your problem or soon enough it will be a reason for a bigger problem ... especially that Furmark stresses only the GPU not CPU & GPU

Gratitude ... I actually changed the thermal pads on mine and it helped a little bit

 

PCH hitting 85ish is pretty normal TBH. The PCH also sits smack dab right beside the CPU. Iooked into everything I could when my system was overheating and that was a pretty common number under load on these machines. Unfortunately at this point OP your probably not going to solve the problem without swapping at least some of the hardware to figure out what the problem is.

Assuming you have already tried a clean install of Windows then there isn't really much else you can do. It's pretty safe to say at this point there is a hardware problem somewhere but nailing it down to the GPU/CPU or MB is going to be expensive and a PITA.

Usually artifacting points to a GPU problem but to have it persist across two GPU's and to have it disapear completely on an external monitor would suggest otherwise or more likely a failing MB.

 

@MogRules

Thanks alot for your very thorough response.

I am going to be taking apart the machine completely this weekend (again).

I'll try to locate the PCH and see if there is anything visible wrong with it.

The problem is, I don't really know what it is or what to look for. Is it possible to replace just that or is it soldered on? Can you repaste it? Questions like that, just anything I can go by when I try to locate it.

 

The PCH is on the top side of the mobo and it's under the CPU heatsink backplate. PCH can get hot but honestly under the most intense of loads on both CPU and GPU I have never seen mine break 80c. It's soldered on the motherboard and you will find a thermal pad on it. What you can do is change this pad to a better quality thermal pad.

 

PCH temperature is hardly ever an issue as most are rated to handle over 100°C. The 17's Lynx Point Mobile for example has a tmax of 108°C. Passive cooling alone is often sufficient to reach such temperatures, hence most systems even lack a thermal pad for this chip.

 

@t456

Thanks for your feedback. The interesting thing is that the PCH temperature follow the artifacts on the screen perfectly and it's the only thing I have to go by right now.

Even if it is the problem I'm not sure how feasible a replacement will be. Waiting for a quote from Dell now.

 

That is only possible with a bga rework station, so you'd need a repair shop that has one.

More likely is that the issue is with some other component that is struggling. When the system is under load so will the pch and, for instance, the adapter and/or battery. To conclude that these items are the culprit would be a bit of a long shot. Still ... it is a possibility, so you could test the pch-theory by adding some cooling option to it, even as simple as blowing cold air from a hair dryer directly at the chip.

 

@t456

I was not aware of that, but I had requested a price from Dell for them to replace it.

The interesting thing here is, that being under load itself doesn't trigger the artifacts, they only show up gradually as the system heats up, and then stay there after during no load until the PCH(and probably other components) have cooled down.

Edit: I will deffinitely try cooling it, but I am not sure how I will go about running the machine while everything is dissasembled. I'll give it a shot for sure.

 

@t456 @MahmoudDewy @Aroc

Sorry for spamming the tags but it was a bit urgent.

PCH - where is it?

Picture of what I am looking at atm:

https://ibb.co/caAvka

 

It's under the aluminum small plate that is part of the CPU backplate

 

Thanks for the quick response!

Unfortunately I didn't make any progress, but it was worth a shot.

 

I wish we could help you seems like nothing is working

 

Ah well, atleast we have tried. The amount of help I have gotten is astonishing and thats really more than I could have ever expected.

I got a quote from Dell that says 500ish Euro (550 USD) for a on-site motherboard replacement.

Honestly it's cheaper than I feared, but it's still an easy 20% of a brand new system with far better specs, not to mention that replacement is not even guaranteed to fix my problem.

Decisions decisions...

 

It would be better to get a quote on fixing the problem, not changing the motherboard. As changing the motherboard might not solve the problem. Let "them" diagnose and propose a way out.

 

--
It could be that the PCH is just "left holding the bag" so to speak. Or maybe the term "lighnting rod" is more fitting. One possible of it overheating like that is 1. a short someplace or 2 power delivery problems.

For 1. you might re-seat each and ever connector on the motherboard. A pin or line crossed somewhere could cause that, but typically, we see this more on desktop motherboards. It is worth checking. Look for a youtube complete teardown video to familiarize yourself.

For 2., you might try running load tests, but changing the power source. I don't remember if the Alienware will run without a battery, but you might try pulling the battery, and running with just the AC Adapter. Then try your second AC Adapter, if you have one. Then try the test on batter only, without the AC Adapter. See if the symptom follows one power source but not the other. It could be the power source or the cabling/connector (a Short).
--
The PCH looks like a smaller miniaturized version of a CPU die. It is probably under the metal plate like @ MahmoudDewy said. I didn't see anything else in your pic that looked like it. Given that it does not have a huge heatsink on the PCH itself, Dell must have designed it to be passively cooled and for it to not need to dissipate that much BTU or it has a high thermal tolerance. It must, if that is all the cooling the OEM built into it.
--
I might be tempted to get a used, tested, known good motherboard from another forum member or from ebay. Like $30 or under $100 or something as a test. Then if that works, either use that board or try to find a new one then. I agree with @ MahmoudDewy that you need a quote or a dollar figure on fixing it, if you are going to sink $500 into it.

Another option is to send it back to Dell and let them diagnosis it and quote a repair. Typically they will refund your money if they cannot fix it in the end. They are good that way.

Another option, is Eurocom will take an Alienware (dead or alive) in as a trade for some of their computers. 25% off. Depending on how much you still like your Alienware and how much it meets your needs. But this is the option I might lean towards even though it is the most expensive.
--
You've spent a lot of time and energy on this. You need a reliable computer at the end of all of this. So I agree that should be your goal. I feel your pain.

 

Hello brutus,

Still I'll ask you to provide the following:
- Screenshot (not photo) of your pixelation
- All Temperatures of your PC (HWInfo or Speedfan will suffice)
- GPU-Z Screenshot of your GTX 880m and Intel
- Photo of your Mainboard (Preferbly as detailed as possible)
- How long does it take from start to "pixelated"?

Once u've answered those questions I think I'll be able to tell you the issue you have.

 

@Aroc

Thanks!

I tried both of your suggestions, didnt make a difference. I'll ask Dell what it will cost to have the diagnose it.

Looked at Eurocom and couldn't find the trade-in option for Alienwares, only Eurocom PC's.

@Danishblunt

Let me get back to you asap.



Note: I had an interesting find. During normal windows operations there would be very small artifacts sometimes. I changed the refresh rate to 50hz from 60hz and these seem to have now dissapeared.
This does however not work in games, as the artifacts still show up there no matter what I do.

 

@Danishblunt

1: I cannot show you a screenshot of the artifacts, they do not appear on screenshots.

2: See the links, couldn't get the logging feature to work in HWinfo so I just took screenshots of everything. Min. is Idle and Max. is during medium load (playing Overwatch).
Edit: The paste job right now is far from perfect since I have been taking the machine apart many times without repasting, the CPU and GPU temps are a little higher than they would normally be but thats about it. Im not using the machine under load for extended time now anyways.

3: See the links

4: See the links.

5: Pretty much immediately when it starts heating up during load (Following the PCH temps it happens after 60c-ish)


Pictures:


http://ibb.co/kqMMQa
http://ibb.co/eJoeBF
http://ibb.co/kNzVJv
http://ibb.co/fNvsWF
http://ibb.co/jNt1Qa
http://ibb.co/cv7T5a

https://ibb.co/caAvka - Photo of mobo

 

From all you've posted I can gather the following:

The GPU's are probably fine, both the intel and GTX 880m, if u're not able to screenshot the artifacts then the VRAM and GPU are probably fine, meaning likely have an issue on the mainboard itself. The fact that your screen runs normally and only gains artifacts when you're playing games and slowly fall of when you're not doing anything (cooling down) makes me believe that the components on the board itself all are working, but once you strain the board it will start giving artifacts. The type of artifacts look much like the artifacts you would get when your AGP bus is about to die or needs resoldering. Now what I would suggest before u start going crazy is to take the board of, get some isopropyl alcoh ol and clean the entire mainboard and your connections.

If youre lucky you might just have dirt, dust or something fluid that simply makes a weak short on your board causing some chip that's connected with the screen to not work properly by overheating, hence producing the artifacts. Also check for all chips that should be cooled by cooling pads if the pads are intact or missing entirely. All the important chips like GPU, CPU and so on are well withing the norm tempwise, so they are probably fine as well, so your mission is to look for other smaller chips that might be damaged, dirty, not getting cooled properly etc.

 

@Danishblunt

Thanks for your suggestions. I am a little scared to pull out the entire mobo, I am well in over my head just taking apart the PC. I want to give it a try though, is there something I need to know before doing this?

 

If you don't feel comfortable then look for a dissasembly guide. Don't use to much pressure when opening the parts, since you don't want to break the casing. Also when opening the upper case try to grab some hard plastic (like an old credit card ) which you don't care if it gets ruffed up, the process can leave marks and make the card unusable. Just slide it through the upper case when you unscrewed all screws. If the card is stuck somewhere and the case won't pop open, you know that you missed a screw or 2, when it klicks open you know all screws are removed and u're good to go(will probably the "hardest" part)

Also when opening the case I suggest you sort the screws. There are often only 3-5 types of screws, so make sure when taking the notebook appart that you sort the types in maybe 3-5 cups or figure a way to place them on a piece of paper or similair, just don't mix them up.

Also don't ever take something off fast, like the mainboard for instance, you don't want to accidently rip off a ribbon cable or other smaller calbles, since they are rather fragile. I would also suggest you make a picture of the other side of the board and post it here (once you removed it from the caseofc), in case you might have a short caused by some kind of corrosion.

Also I suggest when putting the notebook together that you look 2 or 3 times if everything is connected or screwed in properly, there is nothing more annoying to notice that a touchpad doesn't work because you didn't put it in properly and have to open the whole thing again

 

@Danishblunt

Thanks! I have taken it apart quite a few times now (you saw the picture of the mobo) so that part I'm kinda comfortable with now, but never taken the actual mobo off.

Do I make a 'bath' with isopropyl alcohol and soak it or do I take a q-tip and poke around all over the mobo?

 

q-tip is probbaly 2 weak. Bath might be overkill but you can opt for it. I usually just use makeup cleaning accessories in combination with the alcohol (pads and stuff). So you can probably go for something like that. If you have a lot of the alcohol and some time, go for the bath, but remember to clean let it stand for a while, to be 100% sure that the alc. is all gone, you don't want to risk anything even tho chances are very very very low that something might short.

 

@Danishblunt

So I tried your suggestion, pulled the mobo out and cleaned it with isopropylalcohol.

It didn't work, and whats even worse is I think I managed to screw something else up.

The PC works, sometimes. But other times it doesn't and is giving me following problems:

A: Refuses to post at all

B: Gives me 4 beeps during post


Any idea what I've messed up here?

Edit: I know 4 beeps is ram problems, but there is obviously nothing wrong with the ram so I'm wondering what i have destroyed in the process

 

This only supports my theory that there is something wrong with the board.

Still try this:
Take battery out, take AC out, press the power up button for 20 seconds, release the button, put AC back in and try to start it. See if it changed anything on the beeping

Do you have a multimeter around?

 

Thanks for the tip - although right now it's running fine with only 1 ram stick inserted.

I don't have one at hand but there might be one at work I can borrow.

What do you want me to do with it?

 

You have to get yourself a scematic and then measure the power rails.

Since you say it works perfecly fine with only 1 RAM slot occupied you might already have a great lead from where the issue might come from.
http://laptop-schematics.com/

Find the proper scematic and measure all your rails until you find a chip or something that causes your notebook to act weird and has improper output.