***EVGA Precision X and Windows 7/8/8.1 and especially 10 bricking systems***

Guess mine was super-infected.

 

Nope...

1: I did not burn or cut any pins on my cable.
2: I booted with working lcd, then swapped to corrupted lcd.
3: Flashed it to a working lcd.
4: Booted into windows to check if the refresh rate was over clocked, but in this case it did not show the over clocked. Although at the time I was over clocked and using evga to monitor vitals. Then the gpu started major down clocking like I was gaming on battery. I rebooted and got the 8 beeps. I thought it was the card, but found out it was the screen. Anyway's... I rebooted into safe mode and uninstalled evga and the nvidia driver. And installed the 364 driver and everything seems fine.

I did bring everything with me just in case i need to re-flash on the go. LOL

 

Well the "Viking" breaths again. Turns out, the only pin damaged on the LVDS cable was the one supplying power to the back-light. Thank you, super bright LEDs!! Placed a few of them behind the original M15x LCD and could see everything... had to turn the room lights off, though. So like a bum trying to hack into CIA in the middle of the night I re-loaded the 18 and can now operate off the external monitor.

Can't wait till the new replacement LCD and LVDS cable comes in... my 18 looks like Frankenstein.

 

Howdy guys !
Back here after a 3 month break and traveling!
So this thing is STILL going on and no new fix besides the linux-flashdrive trick right?

 

Not from Nvidia, as evidenced by 32 pages of them not giving a duck: https://forums.geforce.com/default/...amsung-and-lg-notebook-lcd-display-panels/32/

Which is why I'm not buying any Dell or Nvidia product again. I highly suggest you don't as well.

 

So only AMD huh? a bit sad don't you think? Nvidia being (and having been for the last 5 years) the top GPU builder.... (For not buying DELL, that's ok, there are other better system builders)

 

I understand (and agree with) the reasoning behind it. If I thought it would make a difference I would do the same thing. It won't though. NVIDIA doesn't give a rat's butt about anyone except NVIDIA. They are not customer-centric in any shape, form or fashion. They are 100% profit driven. That being the case, I'm not going to do without, lower my standards, and settle for AMD garbage to make a statement. If AMD releases something on MXM that is better than NVIDIA's best, I won't hesitate to kick NVIDIA to the curb and drop a hot and steamy deep-dish colon loaf on them. I hate NVIDIA, but trying to get their attention is futile when they don't care. So, I make lemonade from life's lemons and choose whatever brand delivers the highest performance and best overclocking experience. Right now, and hopefully not forever, the only logical choice is NVIDIA garbage.

 

Its extremely sad to see this is still happening. I'm running a dual boot with 7 and 10 (man that took some work with messing with BIOS settings to figure out how to pull that one off) but I'm staying in 7 these days until Quantum Break gets here on Tuesday. Hot swapping LCDs and everything... what a hassle!

 

Its even more fun when you burn your backlight out and have to use LED lanterns behind said LCD in a dark room to see anything

 

Yeah that's not fun. At all.

 

Sure it is! It's like finding out the battle isn't lost. Just gotta go back to old school Gameboy with a twist Thank you EVGA!! **Satire**

 

I remember when choosing an AMD CPU and ATI GPU was the thing to do. Radeon, to me, meant something. I didn't really know nvidia were people until the first Xbox. Now, instead of gaming graphics, you have budget graphics. ATI may be gone but it'd be nice to see them make a comeback. Nvidia graphics at AMD prices? Yes please.

 

Actually, until Nvidia releases their next lineup AMD GPUs are in the lead for all things DX12. Granted they use even more power in DX12 than DX11, but that's OK for my main rig.

But that doesn't matter a whole lot atm, because I had just updated to a 980 Ti last fall and won't be due for an update for another year at the earliest. That and 1080P gaming leaves me a lot of wiggle room.

As for laptop gaming? I'm giving up on it - courtesy of Dell's lack of forsight, Nvidia's ducked drivers, and EVGA's shiptastic software.

 

DX12 only suffers on nVidia cards when async is used heavily. We need more DX12 titles to get a real picture of the performance difference.

 

Maybe that is the case, but in every single currently playable DX12 game Nvidia cards see a drop in FPS using DX12 when compared to DX11. I'd hazard a guess that Rise of the Tomb Raider isn't heavy into asynchronous compute either. http://www.hardocp.com/article/2016/03/28/rise_tomb_raider_dx11_vs_dx12_review/7#.VwOs03VMHqA

 

Look at the hit that the Fury X takes versus the 980 Ti and it becomes clear that besides the gimped 970, nVidia whipped AMD in DX12 performance. The Fury X lost a whopping 9-10 points on average while the 980 Ti lost 4-6 and had higher scores across the board.

I love how everyone is so shocked that there is a performance drop with DX12. There has never been a new DirectX release that didn't entail a performance drop, regardless of the promises Microsoft makes.

 

Well, to be fair the minimum frame rates are improved on systems with weak single thread process CPU's (cough, AMD).

Even Intel CPU's get a boost there.

 

So I have an Alienware 17 with Nvidia GTX 770M. I just re-installed Windows 7 to it. Should I upgrade to 10? I'm really not worried about my GPU temps as they never broke the 70's C when gaming. (Only when benchmarking). I use coretemp for my cpu though. Would I be ok on updating to 10? Or stay with 7?

 

Why such a big desire to install Win X? I can't understand this at all. Win X work with your AW17 as long as you stay faraway from EVGA Precision X.

 

I have a small tablet, laptop hybrid that has 10 on it and the store just has some apps on it I would like to run on my AW17. I don't use Precision X, I used to use Afterburner, but unsure of it. I just don't want to ruin my display. I actually don't mind Win X and running Ultimate I'll get the Pro Edition, which shouldn't have all the spy crap running and even so, I'm sure there is no amount of tutorials to disable that crap. I'm still not 100% certain on upgrading yet. I just got my Windows 7 interface themed out the way I want it and really like it. I'm just wondering primarily if I should be worried if I were to upgrade at anytime. I don't have the money to toss around on a new display. (My other computers where really cheap finds and luck.)

 

Ok. Everything will go well if you don't use Evga Precision and start to read what other says about new drivers from Nvidia. Newer be a Guinea pig(first man to install new drivers from Nvidia). Also read about how to tweak Win X in the forum guides. The same with stopping the spybox to gather data from your computer aka telemetry and stop win update from pushing you unwanted drivers.

 

≥$0.99 no-solder edid bulletproof mod​

Since it doesn't seem likely a permanent fix is forthcoming at the source-level (Nvidia), I've tested a method to prevent any writes to the lcd's edid eeprom, regardless brand of laptop or lcd. Doesn't even matter whether the eeprom can be write-protected or not and running Windows 10 and/or Precision X will be perfectly alright after finishing the mod. It would also ' revive' bricked panels, come to think of it ...

Basically, we're merely bypassing the lcd's edid entirely and substitute it with our own, one which we can control (via a hardware pin).

1.) Source a $0.99 I2C eeprom with pcb interface (inc. shipping ... ):



2.) Program it with the correct edid. Either use a simple programmer (plugs into the pcb's four pins), diy your own or send me a pm. If there's sufficient requests I could order a dozen of those and envelope-ship them pre-programmed to anyone interested.

3.) Snip the four edid wires on the lcd cable * and use four jumpers (90's leftovers!) to snag them securely to the corresponding VCC (voltage), SCL (clock), SDA (data) and GND (ground) pins on the I2C board.

4.) Swap the WP jumper one position and leave the rest as is (don't want floating pins).

5.) Install Windows 10, Precision X, overclock the display and worry no more; it's perfectly safe .

*) The thing is knowing which four pins to clip. This isn't very hard on the lcd-side of the cable; it's usually #4, #6 and #7 (plus any ground) and there's the panel's specification sheet to provide definite proof:

Spoiler: B173HW02 V0 - edid-only lvds cable (test platform)

Problem is that there's usually not enough room to spare inside the display lid and we need to tuck away that I2C board somewhere. It's pretty small, but still a 256 Kb eeprom, so 1,024 times larger than the panel's edid eeprom:

Spoiler: simple 256Kb vs. 256b eeprom

Well ... there's ample storage space near to the lcd cable's motherboard connector. No problem, except that we need to find the specific edid pins again. With a Clevo this isn't very hard; each model has a service manual, featuring simplified schematics that, amongst other info, list the purpose of each pin on the lcd connector:

Spoiler: P77xDM(-G) - LVDS connector

For Alienware and others there's several options:

• Hunt down schematics (quite an effort; look for the Quanta or Compal part nr.).
• Set multimeter to Ohms and test each pin until you've found the 4/6/7/10's corresponding pins at the other end.
• Unwrap the cable and simply trace each wire we need (wrapper is though!).
• Solder on four extra wires at the lcd end and route them back through the hinge, alongside the regular lcd cable.
• Free up space inside the lid.

Ps.
Here's what my test setup looks like (think of the adapter cable to the left as the lcd cable and its pcb as the motherboard).

Spoiler: Atmel AT24C256 - step 01 - setup

First thing was writing an edid to it and reading it right after (an AUO):

Spoiler: step 01 - write+read (AUO B173HW01 V4)

Code:

00 FF FF FF FF FF FF 00 06 AF 9D 14 00 00 00 00
01 13 01 03 80 26 15 78 0A 7D 45 AB 4F 38 A6 24
12 50 54 00 00 00 01 01 01 01 01 01 01 01 01 01
01 01 01 01 01 01 A0 37 80 B4 70 38 32 40 6C 30
AA 00 7D D6 10 00 00 18 00 00 00 0F 00 00 00 00
00 00 00 00 00 00 00 00 00 20 00 00 00 FE 00 41
55 4F 0A 20 20 20 20 20 20 20 20 20 00 00 00 FE
00 42 31 37 33 48 57 30 31 20 56 34 20 0A 00 F6

Then a different edid (a CMO this time):

Spoiler: step 02 - read+write (CMO N173O6-L02)

Code:

00 FF FF FF FF FF FF 00 0D AF 19 17 00 00 00 00
05 13 01 03 80 27 16 78 0A 08 05 9D 58 53 97 29
15 50 54 00 00 00 01 01 01 01 01 01 01 01 01 01
01 01 01 01 01 01 1C 2A 40 54 61 84 1A 30 30 20
35 00 7E D7 10 00 00 18 00 00 00 FE 00 4E 31 37
33 4F 36 2D 4C 30 32 0A 20 20 00 00 00 FE 00 43
4D 4F 0A 20 20 20 20 20 20 20 20 20 00 00 00 FE
00 4E 31 37 33 4F 36 2D 4C 30 32 0A 20 20 00 AD

Now we switch the WP jumper:

Spoiler: step 03 - write-protect ON

Writing yet another different edid (an LG):

Spoiler: step 04 - write (LG LP173WF4-SPD1)

Code:

00 FF FF FF FF FF FF 00 30 E4 6C 04 00 00 00 00
00 18 01 04 95 26 15 78 0A 0B B5 A3 59 55 A0 27
0C 50 54 00 00 00 01 01 01 01 01 01 01 01 01 01
01 01 01 01 01 01 B4 43 80 A0 70 38 1F 40 30 20
35 00 7E D7 10 00 00 1A 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 FE 00 4C
47 20 44 69 73 70 6C 61 79 0A 20 20 00 00 00 FE
00 4C 50 31 37 33 57 46 34 2D 53 50 44 31 00 3D

Then a final check is reading the eeprom's content:

Spoiler: step 05 - read (CMO N173O6-L02)

Code:

00 FF FF FF FF FF FF 00 0D AF 19 17 00 00 00 00
05 13 01 03 80 27 16 78 0A 08 05 9D 58 53 97 29
15 50 54 00 00 00 01 01 01 01 01 01 01 01 01 01
01 01 01 01 01 01 1C 2A 40 54 61 84 1A 30 30 20
35 00 7E D7 10 00 00 18 00 00 00 FE 00 4E 31 37
33 4F 36 2D 4C 30 32 0A 20 20 00 00 00 FE 00 43
4D 4F 0A 20 20 20 20 20 20 20 20 20 00 00 00 FE
00 4E 31 37 33 4F 36 2D 4C 30 32 0A 20 20 00 AD

So it's still the CMO edid, not the LG variety we've written last. Hence; not only is write-protect functioning (any data sent is disregarded), but we can also still read its contents, which would be impossible if we had just clipped the edid data wire and done nothing else besides. Benefit of the simple jumper is that we can always reverse the swap and flash our own, customised edid (for higher refresh rates, different colour calibration and/or timing frequencies).

Anyway, hopefully this is of some help to anyone and while it's not plug-and-play, it is a cheap, 100% fix.

Good luck !

 

Well my new LCD came in along with the new LVDS cable... Still no backlight. Motherboard faulted?


Sent from my iPhone using Tapatalk

 

I've already found out about the newer nvidia drivers being crap pretty much. I'm currently on the fence about updating though. I've customized my Windows 7 interface to the way I love right now. lol

 

Make an image while it is perfect and save it in a safe place so you can go back to it. The free version of Macrium Reflect works flawlessly for this. Then you can wipe away your Windows 10 regrets with only a few mouse clicks. The best way to find out how much Windows 10 sucks is to use it. For most folks it is one of those things you either love or hate, not much room in the middle for the undecided.

 

Before i buy a new Mobo, is there anything else that could be preventing the backlight from coming on? The output is present, just no light.


Sent from my iPhone using Tapatalk

 

I'm not an expert on where the screen draws power from, but if it comes from the motherboard the issue can be with the cable, connector (either end), or motherboard. If you've already replaced the cable then that leaves the connectors on the motherboard or LCD, or the motherboard itself.

 

Well, the LCD is new as well as the LVDS cable. I've read around that it could be a fuse on the mobo but I don't have the tools nor the part information to replace it. I could just flow solder over the top to bypass the fuse but I'd rather not

 

Hello everyone.

What's the general consensus on this matter now as I've come across a stroke of unbelievable luck.
Purchased a refurbished Alienware 15 (BGA boo, hiss - Just kidding, it's running fine and suits my needs) after selling on my Clevo.
Long story short. Paid for, and it was a solid deal nonetheless, for what I believed was just a Gtx965m equipped system and boom, turns out I'm running a 980m.

I've owned an Alienware 17 in the past and it burnt out (970m). May have been Windows 10 but I can't say for certain.

I'm concerned if I upgrade to 10 I'm going to go through the same problem.
Best to just completely avoid then?

 

Well ... why not? Can't hurt at this point; a replacement motherboard will be an expensive fix.

A few good photos of the mb should suffice to identify the fuses. Would help if you had a multimeter; there's several of them and you only want to bypass the broken one (or replace it, of course).

 

You're right. As for multimeters, no shortage there. I will post pics once i have the culprit or lack there of.

 

Hey!
I have a Clevo P775DM1-G laptop with GTX980 which arrived with LP173WF4-SPF1 screen G-Sync was working well, but unfortunately I had to send it back for screen change, because it had several dead pixels.
It came back yesterday and the screen was changed to LP173WF4-SPD1 and G-Sync option does not appear in the NVIdia Control panel. Windows and driver and BIOS update did not fix it.
The VGA has G-Sync because the option appears if I connect an external G-Sync monitor.

I am trying to flash the EDID of the SPF1 screen to teh SPD1 screen but the Linux with EDID tools linked in this thread does not work for me.
In BIOS I get black screen and nothing happens, in UEFI I get this:



And nothing happens...
I am able to boot Ubuntu 14, but that does not have the tools required to flash the EDID

 

The BGRT thing is a bug in 14.04, which the usb stick image is based on, but it shouldn't halt booting. The second message is apparently caused by the card reader, so switching to standard usb stick (if using an SD card) and/or disabling the card reader in the bios (or blacklisting its drivers in the image) should fix that. There's also a report that it may be caused by encrypted volumes; the live stick mounts all volumes it finds in order to access them for troubleshooting. A password line should pop up in order to proceed, so try hitting escape to force that. But, really, since you have 14 running anyway, it might be easier to just install the required tools on that setup instead.

And SPD1 and SPF1 should both be g-sync approved already, provided you have the latest vbios for your GTX 980. So try updating that first, if applicable. Can you also copy/paste that SPD1 edid? Only have a 75Hz+75Hz version for that model and perhaps your specific SPD1 merely has an outdated edid.

Spoiler: LGD046C (LP173WF4-SPD1) edid, 75Hz+75Hz

Code:

00 FF FF FF FF FF FF 00 30 E4 6C 04 00 00 00 00
00 18 01 04 95 26 15 78 0A 0B B5 A3 59 55 A0 27
0C 50 54 00 00 00 01 01 01 01 01 01 01 01 01 01
01 01 01 01 01 01 B4 43 80 A0 70 38 1F 40 30 20
35 00 7E D7 10 00 00 1A 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 FE 00 4C
47 20 44 69 73 70 6C 61 79 0A 20 20 00 00 00 FE
00 4C 50 31 37 33 57 46 34 2D 53 50 44 31 00 3D

Here is my SPD1 EDID:


Another strange thing is that when I use sudo i2cdump 12 0x50, I get the EDID:


however if I use sudo ./edid-rw 12 | edid-decode for the same bus, I get this:

 

WINDOWS 10 UPGRADE WARNING

You sre going to have to either boot with another screen attached to the LVDS port or manually reprogram the EEPROM EDID using a USB flash programming tool and another computer. The easiest solution is to get another screen to boot from, boot into Linux and hotswap the bricked LCD to flash it from the Linux desktop.

Windows 10 is not safe to use, but it may not not be 100% the sole cause of this issue. It appears to be a contributor or creates the situation where it can occur, either directly or indirectly, through changes NVIDIA made to their drivers accommodating Windows 10.

Anecdotally, I can tell you that all heck broke loose around Windows 10. Everything was just peachy, I installed Windows 10 RTM on my M18xR2 and that brought a screeching halt to the world I was happy in. To further confound my understanding, the simple act of installing Windows 10 on the Alienware 18 bricked the screen without EVGA Precision X being installed. EVGA Precision X being installed on the M18xR2 bricked the screen every day until I discovered it was contributing to, magnifying or exacerbating the issue. Removing it stopped the problem on the M18xR2.

I still do not understand why the experts decided EVGA Precision X was the offending product since there are reports of the problem, including my own with the 18, where the LCD corruption has taken place without the help or involvement of EVGA Precision X. But, I am past trying to figure that out now. I doubt we will ever get to the bottom of it or hold the responsible party legally liable for the damage they have done to so many Alienware and Clevo owners, and perhaps the owners of machines brands we have not identified.

Gabe Aul and Micro$haft ignored us as a group, and me personally as a customer. I sent several messages to Mr. Aul and found it pretty amazing that the dude did not have the courtesy to respond. He did not even have enough tact or diplomacy to delegate a one-time task to a flunky on his team to extend the common courtesy of a " we're sorry to hear of your situation, please have a nice life" sort of canned one-size-fits-all cameo response. I will never forget being treated this way by Micro$oft. They are incapable of and beyond any hope of regaining my trust or respect due to their lack of courtesy in relation to this issue. I know they don't care and don't need it.

Dell/Alienware was very responsive and I definitely appreciate that.

NVIDIA was responsive until I reported my findings with EVGA Precision X. Unfortunately, they were quick to blame EVGA and put them under the bus to deflect blame away from their brand.

 

Are you root?

And thanks for the edid. It is indeed a 60Hz+60Hz version, so probably from a non-validated batch. Flashing the LGD046C.bin from the archive should fix it. If it turns out the screen can't handle 75Hz (or more) then you'd best rma.

 

yes, same problem when I am root

it is now saying for ./edid-rw ? and ./write-edid that "Command not found"... ah its been years since I've used Linux, I feel like a total amateur ...

Meanwhile could you please upload the correct bin file for me ?

 

Just wanting to report on my AW 18. I installed Win 10 a few months ago and am on the latest Insider Preview and latest Nvidia drivers. I never had precision X installed. I have not experienced any issues so far and games are running great. I also do not overclock anything or do bios or vbios mods. I hope this contributes in some way.

 

Slowly, but I'm getting there, it did not work, because I was not inside it's folder -_-

Okay, next question: I am getting the same data for both bus 12 and 17...which one do i need? or I need to update both ?

 

Both, probably. Noticed this before on some systems; there was an edid-edid eeprom and a second one, that had more data as well as the edid. Do make sure the bus is still returning the expected edid right before flashing the updated version, don't want to risk writing to a bus nr. that's been re-assigned to something else.

The 75Hz edid can be found here: archive correct edids, v11.

 

Just ran write-edid.sh for both 12 and 17 buses. First used SPD1 bin and even tried SPF1 bin, but the EDID does not get overwritten, it remains the same...really starting to get on my nerve.

Result after using write-edid.sh with spf1 bin:


btw, I got the write-edid from here: https://github.com/ChalkElec/write-edid

 

That is not good; means it's write-protected and only de-soldering the chip + re-programming will work. An rma would be easier, really. On the positive side; no edid bricks are possible with this system ...

 

Shouldn't the script have given an error message if its write protected and cannot access it for writing ?

 

Nice! Sounds like they did it the right way. EDID writes should not be possible on any notebook. The fact that they are is why we have this problem. But, in fairness to the OEMs, before the Windows 10 cancer OS and the readiness updates for it, EDID writes were non-existent and I suspect they were blindsided by this. I think the goons at Micro$oft have taken liberties they had no business taking that created this sorry mess and placed so many systems at risk.

 

Yes, but the funny thing is that edid corruption did happen quite often long before Windows 10, but most reports concerned desktop monitors.

The weird thing in all of this is that the infection sticks to the gpu card, yet its vbios was bit-identical. So either one of their post-10 drivers wrote something beyond the vbios payload or there's a second firmware we haven't got access to. It would be nice to research an infected card and see what else is there ...

No, because it isn't aware of that; the write-protect pin is a hardware feature of the specific eeprom. There's no way for the i2c tool to know beforehand in which state it is without actually writing to it. That's probably why most eeproms also feature a software-flippable byte as this makes wp on/off possible without having to modify the hardware.

Nice job in making it this far, btw , only too bad it won't help with the issue. Also for anyone else trying to make their screen g-sync 'compatible'. The eeprom-bypass trick won't help since there's no edid pins to hijack, this being an eDP panel. Though you could solder them to the lcd's pcb, but then you may as well de-solder the chip and re-program that instead.

To make sure; you've checked the vbios?

 

Checked the VGA and according to the DeviceID it should be the one with G-Sync. To make sure I even tested it with an external G-Sync monitor and G-Sync option was displayed in the NVC for the external monitor.

btw, would it be possible to override the EDID in WIndows with an INF file made from an updated edid.bin to get G-Sync enabled? Or it wouldn't work because NV Driver checks the EEPROM edid no matter what?

If it would work, is there a guide somewhere so I could make a valid inf from an updated bin ?

Just tested it with the inf of an SPF1. Refresh rate changed to 75 Hz as expected, but G-Sync option remained hidden. Oh well, returning for RMA.

 

As far as I know, you need G-Sync capable monitor, motherboard, AND EDID or whatever for G-Sync to be usable. Someone with more knowledge of G-Sync should be able to chime in and say if I am correct or not, I'm personally not interested in G-Sync AT ALL.

 

The first two are of the LVDS jack and surrounding area, top and underside.

The last two pics include the only fuses that I could find on the board. Both of them are good.

Edit: Backlight is controlled by EC A replacement mobo is what I'm left with. Source = Viking 18 schematic. (Thanks google)

Sent from my iPhone using Tapatalk

 

My mb and VGA are G-Sync capable. (G-Sync is working with external monitor)

 

If you remove the discrete GPUs and boot only with the Intel HD Graphics does the backlight still refuse to come on? Unless you flashed the EC and that failed, I don't know how or why that would have been affected. I ask because some of the vBIOS testing I have been doing with @Prema, with vBIOSes that have problems (or if you have a soft-bricked GPU) the LCD backlight will not turn on.

If there is something wrong with the GPU(s) that could be why. If you have not already done so, see if the LCD works with no discrete GPUs installed.

 

I am actually glad that you mention this because I thought about that too. Especially after reading the mxm connector schematics. The IC that controls the backlight enable signal has a ton of other things on it too so I am having trouble believing that only that signal is bad and not knocking out a ton of other things.

I do have an unlocked vBios on the 780Ms. When the backlight was coming on (after EVGA/WIN10 corruption) the only image that would sometimes show was the pre-post vBios message from the 780Ms.

I'll give this a shot after work. In the event that it is the vBios, would a re-flash potentially fix this issue or are softbricks irreversible?