*** Windows 10 + NVIDIA WHQL Drivers are Killing Alienware and Clevo LCD Panels ***

Sure, just following @t456 instructions. See step #9. The only difference is the enumeration. On my M18xR2 it is bus 1 and in his example his was bus 6. So, for me the command is sudo i2cdump -r 0-127 1 0x50. That will show it on-screen. Or, you can dump it to text file using sudo i2cdump -r 0-127 1 0x50 > edidexport.txt (which is my bus number and if you open the terminal from the folder you are working in it will place the text file in that folder).

Try bus 0 and bus 1. If you get output with all X you know the select bus is the wrong one.

 

Thank you, that's some seriously complex low level stuff, at the moment everything is working ok. Therefor will I identify anything wrong in the EDID if I read from it? and can I ask, when you follow the tutorial and ;fingers crossed it fixes the LCD is it permanent or is what ever changed it in the first place do it again?

 

Yes, it runs two edid-extract methods. One might fail, but the other may have succeeded. If there's a 128 byte .bin file then you have the edid as currently on the eeprom (display chip).

Actually, it says 'FALSE':



Ctrl+C/Ctrl+V skills may have failed you there . Btw, the reader could have it wrong, but that chance is 1/255, so 0.3%; we can trust it with 99.7% certainty .

 

Nope, once Windows 10 cancer has infected the system BIOS it will never be fixed. I'm guessing motherboard replacement is the only fix. I've re-flashed the M18xR2 display probably 15 times now.

Hmm... strange. I copied and pasted it and it returned GOOD. Maybe the Linux clipboard had the wrong information and I just THOUGHT I copied and pasted the corrupted EDID. But, you're correct. Just did it again and got the same confirmation it is bad.

 

That's not the news I was hoping for, so is it the Nvidia drivers that will trigger the corruption? or will it happen regardless.. I've not installed any drivers yet and hasn't had any on since windows 8 but it's going to happen one way or another right?

 

Drivers that I have been using with excellent results since February (344.75 and 345.20) and never had any issues with before Windows 10 ruined the system BIOS with a root kit don't help this problem. Installing GeForce drivers seems to trigger the corruption regardless of the driver version. Although, sooner or later it's going to brick itself again without making any driver changes. Not sure why that is, but apparently the Windows 10 root kit checks periodically and when it sees the EDID is good it "fixes" the situation by writing corrupt code and bricking the LCD again.

 

Regarding moninfo the program is capable of pulling live EDID from the monitor(s) connected.If it doesn't that is a bad sign already that is something wrong as it has been mentioned in moninfo forums in the past.

 

Have an M18x R2 bios and searched that for the 'CompuTrace' rootkit:



So ... desolder bios and use programmer, that will clear the infection. Of course, any re-install of 10 will re-inject the code, unless you write-protect the bios as well. Check the autochk.exe (Windows 7) or Event Log (8.0/8.1/10) for traces of this thing.

 

That is actually in the BIOS on a menu. I will check and see if it is set to disabled.

I don't have the tools or the knowledge to do that. I also cannot identify which chip is the BIOS. There are many chips that look exactly the same and I don't know where on the motherboard to find it without a schematic to tell me where it is. Looking at the labels on the motherboard, it is not clear to me which one.

Look at this screen shot. I'm back in Windows now after flashing the display. Look at the registry for "Default Monitor" EDID. It's empty just like the dump with all zeros.

 

What about M18xR1? Isn't it old efi forced to work in legacy mode as BIOS?

 

Wow, I installed windows 10 this morning.. I so wish I come across this before hand :/

 

@GodlikeRU - yes it is

 

What abour pure BIOS systems as first M17x? Were they destroyed too?

 

Wow... just, wow. I wish you had as well. Sometimes life sucks. This is one of those times. So sorry, bro.

Look on the bright side. Now you have every good reason to hate what Micro$loth stands for. They used to be great, but they are one SORRY outfit now. I have nothing but contempt for them because of their Windows 10 cancer OS. Even apart from this problem, their filthy malware is an abortion and the scourge of the tech industry. Their executives deserve to be placed in front of a firing squad.

I don't know. I don't have one of those machines other than the M18xR1 and I'm not about to ruin it with Windows 10 cancer to find out.

It's safe to assume the M17xR4 and AW17 doom will be sealed in exactly the same way since @mariussx has had the same problem with flashing his LCD only to have it re-brick itself again.

 

Copied and pasted the EDID in the registry from the SEC5448 to the "Default Display" key. Maybe that will prolong the time until another flash is needed.

 

Man, £1100 down the drain they don't even sell motherboards over here.. with import charges and what not it would be extortionate. I checked my EUID in the registry and mine is 90 (where yours wrote 00) it such a shame.. I'm wondering whether to just install the nvidia drivers and get it over with.. does it corrupt straight away or can it take a while?

 

Longest Time-To-Die was 1 month

 

Usually when I reboot to complete the driver installation and it either fails then or the next system boot. If you SHUT DOWN instead of reboot, it seems you can POST at least once before the 8 beeps begin. If I reboot it gives short and faster 8 beeps, then power off and normal 8 beeps.

 

thanks, and have you been using an external monitor to reflash?

 

Don't know if it helps but the whole default_monitor key and its subkeys can be deleted if you run registry as SYSTEM

 

How considerate of Dell ... Asus hides the option in the menu.

Can walk you through that and tell which is the bios with a good photo of the pcb. And you have the tools; the edid programmer will do. Only need a soldering iron and a safety pin.

May try setting permissions, that would prevent a registry-overwrite, at least:

 

Yes, very. At least it's on a menu... I'll give them credit for it not being a secret. It was turned on. I just disabled it. There is an ominous warning that disabling it is permanent and cannot be reversed. Good, 'cause I don't want it to be reversed. I don't need that tracking device or Lojack for Laptops.

Spoiler: BIOS Menu: Computrace Disabled

I may take you up on that soon... as soon as I work up the nerve. I have a bad motherboard we can borrow the chip from. It has never been touched by Windoze OS X cancer, so we won't even need to program it.

I did the other day. Apparently, a change in display drivers replaces the registry keys with new ones with default permissions. My driver update experiment is probably what triggered the corruption and changed the permissions.

Yes, that worked. I deleted them under ControlSet001 and ControlSet002 as well as CurrentControlSet.

 

I'm going to install all the drivers and default nvidia drivers from dell.. I'll record the process, I'm absolutely dreading the BIOS beeps and a black screen the laptop will literally look dead.

 

I can confirm that even after trying a factory restore using the DVD's the system will have the default_monitor key with the BAD EDID value if it was "infected" in the past.
Just tried to a friend's AW17 with 780m. We are talking about a factory restore to Windows 8.1 and the built in very old versions of drivers.

Even so the problem will still be there so it doesn't matter which driver you will install or what OS.

 

I've just installed your... 344.75 Desktop Mod - refresh rate back down to 60hz from 64.. sli enabled and default_monitor seems fine.

 

Explain this:

Good EDID (Valid checkshum) via linux tools but when boot into windows the default_monitor key was still there. What the hell....

 

I deleted it running regedit as SYSTEM like you said. It came back. Leave it there and rename BAD_EDID to EDID. Go down in the registry to the real monitor, open the good EDID binary values, select all/copy, the go to the Default Monitor with the BAD_EDID key that you renamed EDID. Double-click EDID to open that and paste the content there. Then the Default Monitor will have an EDID that mirrors the real monitor instead of nothing.

Cannot explain it, but it is what I see as well. Good ol' Windows 10 rootkit... fresh from the Redmond Cesspool for Surface-loving panty-waisted nerds.

 

Is this what the rootkit has done?

 

So @Mr. Fox - the EDID corruption, does that ever happen in Linux or just in Windows?

I'm interested if this is related to Win10 writing things to the BIOS tables, and being triggered by an nvidia WINDOWS driver install. It will be very good to know if I have to take some companies to small claims court for damages.

 

I cannot say for sure. I'm fixing it in Linux. I don't know if it happens in Linux too because I only boot Linux to fix the problem for the most part, and to boot once the problem starts I have to use a different LCD to get past the 8 beeps. If I had to guess, probably not because the Windows 10 root kit appears to have been written to attack Windows machines, but given the malware appears to have embedded itself in the system BIOS there is a chance Linux might not be immune. But, we don't even know for sure if the EDID corruption is taking place outside of Windows. The rootkit may write to EEPROM and screw things up before Windows starts loading.

 

Where's the real monitor please? I did a registry search but nothing return true for those words.. just picks up monitor keywords

 

Look at:
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Enum\DISPLAY

...and see if you have more than one listed. I showed them in previous screen shots earlier today.

After correcting the BAD_EDID there, also check under:
HKEY_LOCAL_MACHINE\SYSTEM\ControlSet002\Enum\DISPLAY
HKEY_LOCAL_MACHINE\SYSTEM\ControlSet001\Enum\DISPLAY

If necessary, make the same changes there. ControlSet001 may correct itself automatically if you correct ControlSet002.

 

Any chance that you could mess around with Linux a bit to see if it does happen? I wouldn't blame you if you didn't want to due to the amount of time that you've already devoted to this problem.

 

Thanks again for your reply right default_monitor has the corrupt EDID so shall I copy the EUID from the SEC5449? there's 4 folders under SEC5448 and without opening each EUID they appear to look the same as each other.

 

I think soon enough Mr Fox and that other guy who come up with the EUID flash tutorial should put a donation link in your sig lmao.. you've probably done more work and provided more help to us than Microshite and nvidia already

 

Also.. just throwing it out their if anyone is bored. It looks like some have come to the conclusion that this come from Microshite cancer (Win 10), but given the evidence does this look like a root toolkit or could there be.. I don't know a program that was "intended" to do other things but mess up this particular hardware? I was just thinking.. if it was a deliberate attempt why target this small portion of users?

 

@Scerate

Since you once had 10 installed + display bricked with a non-Dell, would you run this from a usb stick:

Code:

afudos P77xZM.314 /N
afudos export.rom /O

Assuming 'P77xZM.314' is your current bios version, of course (and is on the stick ...). Mind that the '/N' command has the same effect as a cmos clear, so you may have to set raid/ahci again (and similar stuff).

That ' export.rom' would be mighty interesting, to be sure ...

 

@Scerate
@t456

ONLY run the /N command with the exact BIOS on the stick that is also flashed to the system or it'll brick it!
(My v2 Mod is not using the v14 base)

Also the /N command will WRITE the CMOS contained in the BIOS hosted on the stick to the system and NOT simply reset the CMOS.
So dumping it right after writing without a reboot will just match the one contained in the BIOS on the stick and not contain the system specific values you are looking for.

If you want to check the current NVRAM table just use Afuwin64 (in Windows) and simply hit the save button to create a backup.

 

Might wait till clevo releases laptops with desktop gpus...It is being done.

Sent from my HTC One using Tapatalk

 

Weirdly enough ... it doesn't; some process intervenes with that. Only using the programmer on the de-soldered chip was a perfect match.

Thanks, but something changes non-nvram as well (left is pre-flash, right is post-flash):

Spoiler: non-nvram change

Both .roms for your perusal: W230SS - pre-flash and post-flash.

 

I probably shouldn't mention this, but I have it installed on a couple of older Alienware laptops with faulty components for testing, so if it dmages them, so be it.
M11xR3 - 2 x Broken USB ports, scratches on LCD screen (Lots)
M14xR1 - Motherboard keyboard connection detacted, scratches on LCD screen (Lots)
M17xR3 - Faulty HDD connection, scratches on LCD screen (Lots)
3 weeks so far, powered on 24 hours a day with one reboot every 24 hours

What a kick in the guts, your's must be quite sore now.
Don't forget I have a spare system BIOS chip if your game!!!

You were one of the first, and one of the most active, you must be soo sick of this, thanks for sticking it out.

Might be worth a try, do we have any "creative" geniuses involved in this discussion

I got a couple of AMD8970's (R9 M290x)

Hahaha,lol - Now that's funny, Dell are absolutely hopeless at discontinued model support

Do you think it may be worth while trying a completely original with no modes or "extra's"

Dam, it must be really......................................annoying

Go on, I got every confidence in you, what have your really got to loose.

 

I’ve had something in the back of my mind since this thing started, but was not sure if it was a valid topic.
I’m the first to admit I have very large holes in my knowledge of computers, particularly when it comes to grass roots hardware and software engineering, but I do have a little experience with electronics.
Feel free to correct me if I’m wrong or add comments.
We’ll use Alienware as an example, but I’m pretty sure it applies to all other manufacturers as well.
Alienware components are made by contractors, not all components are made in the same factory or location.
For example, let say there were 100,000 nvidia GPU’s made for Alienware laptops.
10,000 would be made by one contractor.
50,000 by another.
40,000 by another.
This is most probably true for almost all components used.
Given that the parts are made by different factories there may be some miniscule difference in the final product.
If this is the case, is it possible that is why nvidia and Dell are having trouble re-creating the event.
There may need a certain “configuration” needed to allow this corruption.
I’ve seen it before when replacement components have been used on PLC’s I’ve worked on, sometimes the replacement is “compatible” but has unexpected side effects.
Just a thought I’ve had for a while, might also explain why things are being “re-corrupted”

 

So I was looking at my Registry and I had BAD_EDID in Default_Monitor under Current Control Set, but NOT under ControlSet001, it was normal there with the EDID and everything.
I opened up my wife's m14x that hasn't had 10's bad touch and to my surprise, Default_Monitor had no keys or folders under it... Im assuming this is the normal state?

 

This might be worth considering, as much as I dislike that product. It would be useful to see if AMD GPUs "cure" the problem.

Already did... makes no difference.

Extremely would be an understatement. It's downright maddening.

I'm seriously considering this. I have both of those. Let me get you some photos of the PCB so we can figure out which chip is the BIOS... probably tomorrow.

Nothing at this point. Having the flash the LCD daily basically makes the machine impractical for use.

---

BRICKED AGAIN THIS EVENING!!!​

---

I blind crisis recovery flashed the BIOS down to A03, which is a Legacy/EFI only BIOS, and no NVRAM. There are zero options for UEFI... good old fashioned awesomeness. My thought was maybe killing the UEFI and NVRAM might help. Several hours later, same old, same old...

But, this time the corruption is different... have a look.

Code:

Corrupted morning of 9/23/2015

00 ff ff ff ff ff ff 00 4c a3 48 54 00 00 00 00
00 14 01 04 00 29 17 78 0a c8 95 9e 57 54 92 26
0f 50 54 00 00 00 01 01 01 01 01 01 01 01 01 01
01 01 01 01 01 01 29 36 80 a0 70 38 1f 40 18 10
25 00 99 e6 10 00 00 1a 1c 24 80 a0 70 38 1f 40
18 10 25 00 99 e6 10 00 00 1a 00 00 00 fc 00 4c
54 4e 31 38 34 48 54 30 32 53 30 31 00 00 00 fe
00 4d 72 2e 20 46 6f 78 27 73 20 6c 63 64 00 b8

---

Code:

Corrupted evening of 9/23/2015

00 ff ff ff ff ff ff 00 4c a3 48 54 05 00 00 00
05 14 01 04 05 29 17 78 05 c8 95 9e 57 54 92 26
0f 50 54 00 00 00 01 01 01 01 01 01 01 01 01 01
01 01 01 01 01 01 29 36 80 a0 70 38 1f 40 18 10
25 00 99 e6 10 00 00 1a 1c 24 80 a0 70 38 1f 40
18 10 25 00 99 e6 10 00 00 1a 00 00 00 fc 00 4c
54 4e 31 38 34 48 54 30 32 53 30 31 00 00 00 fe
00 4d 72 2e 20 46 6f 78 27 73 20 6c 63 64 00 b8

Looks like the "05" code is back.


But, no "Default Monitor" or "DEL1821" monitors this time. Only my SEC5448 in the registry.

 

Thanx but since i deleted the default_monitor key it does not return after restarts. Also i don;t have the ControlSet002 as you do.I only have the

• ControlSet001
• CurrentControlSet

On anoter note i played some BF4 and checked the video settings. All resolutions i can select in game are at 60.01Hz. Sad thing is i do'nt rememeber how it used to be.I mean was it saying 60Hz back before Win 10 corruption?

 

Mine was 64hz before I installed the nvidia drivers, after installing i got 60hz then after running Assassins Creed the resolutions where crap, highest 1280 x 1024. Point being I think when the EDID gets corrupt which from my understanding contains information about the max refresh, lowest and resolution it's not surprising that the OS probably gives incorrect refresh rates and resolutions. Could be wrong but seems logical if the OS does reference settings from the EDID

 

​

If I purchased a new motherboard that never had Windows 10 on, would that solve the problem? I've found someone who has 2 to hand, so if/when this does break I may very well just do a quick fix if this is a way?

Also.. I come across someone saying what ever is changing the code may be stored in the eeprom? Hence why BIOS flashes wont even get rid if it?

 

Yes but mine shows 60hz in windows (windows settings and nvidia driver).

It is only in BF4 showing 60.01 Hz. Also the EDID is valid as it was checked from linux tools and moninfo as well.

I think i will create a custom resolution of 1920x1080p at 59Hz so it wont burn the display

 

Great, let me know how you get on. Have you installed Windows 10 as well?

 

I had on the past but i went back to 8.1 as soon as i found out about this thread. I went back to 8.1 using the factory restore DVD's.