m17x r3 (3d screen) 680m serious gpu issues

Hey guys,


Ive been having some serious issues with my m17x r3 680m GPU lately---but first let me start from the beginning of this whole mess. Ok about 2 months back nvidia released new drivers, and naturally I wanted to try them out---as you may know, the only way to get ANY drivers to install on a m17x r3 with a aftermarket upgrade, especially the 680m, you have to hack the INF file. So I hacked it, although hacking it was a bit different than normally( I think it was in a different file this time with limited values to hack) afterwards I installed the drivers using clean install and my screen went black. so something went wrong with the installation. I left my computer on for over 30 mins (in case it was still installing) and then cut it off. When I rebooted, still no video, just black screen. so I tried to go into bios and change the display, I went in and changed it wrong and rebooted and I got the beeps of death. so now I got to the point where my pc wont load. So I had to re-flash my bios all over again( it took multiple attempts, because I had no idea i had to first remove GPU in order for it to work) after i re-flashed bios from A10 unlocked to A12 unlocked, my computer started working again, I went into safe mode and fixed the driver issue doing a re-install. SO i finally got everything working again :thumbsup:, except one thing--------my 680m GPU ----

for some odd reason, almost every game I try to play, the frame rates are ridiculously low and the GPU temp scale up high upon first load of game (temp rise to 90c within 10 secs, when not gaming it idles at 60c) even on games like deus ex human revolution the frame rates are horrible, sleeping dogs frame rates are super low and extremly unplayable. all these games i normally play maxed out on 1080p----SO i did everything I could think of the try and fix the issue, I re-installed different drivers using clean install like 3 times, still no luck. I pulled gpu out and cleaned it and heatsink, still no luck. I re-installed bios A08 and A10 and that didnt solve the issue (im still on A10 now)---- I ran a diagnostic test and found out my laptop battery was going out, thought this was the issue, ordered a new battery and replaced the old one----still no luck...

the dianostic test said eveything else on my computer was good to go, no hardware issues----so im stumped -- I have no idea whats wrong with my computer----I even did a full wipe using ccleaner and driver sweeper in safe mode to make sure there were no old nvidia drivers left over---still no luck

if you guys have any suggestions or comments on this or how to fix this, PLEASE let me know........ lol

thanks

 

is the card showing up in the bios and device manager? If so, That would rule out a Hardware issue for the most part.

 

yes, it shows up as NVIDIA GeForce GTX 680M in device manager and bios as nvidia

 

have you tried whitelisting the games you're trying to play? If you don't know how to do that there is a link in my sig.

 

Ill check out that link, never tried or heard of that technique, what is it?

 

basically forces the discrete gpu to run if optimus doesn't recognize a game. Used to have to use it a lot more often than now.

 

there isnt an option under my nvidia control panel to select "High Performance NVIDIA Processor"

 

also, i have the 3d m17x so i dont think i have optimus

 

Temps indicate that the card is not properly padded or poor die/heatsink contact that's the reason for throttling which prevented the card from being fried; you're lucky.

 

so if a gpu is poorly padded the card will autimatically detect this and throttle ant any temp? as my temps reach 90c almost instantly when i crank up any game (except dishonored) which is kinda rediculous

 

90C makes the card throttle, probably no gpu die/Htsnk contact at all.

 

now that I think about it, those thermal pads looked pretty shushed in when I removed the heatsink to check the gpu---also----the heatsink I have is the original one that came with my 460m, before i upgraded to the 680m---so I probally dont even have the right heatsink anymore

 

Re-paste it and check if the gpu fan is spinning correctly.

 

ok, i i will definately do that, and thanks alot for the help guys!! ill post the results!!

 

Man thanks for chiming in man, I didn't think of thermal throttling at all.

 

Clean up gpu fan/vents.

580M's legend, I mean Posttraumatic stress disorder

 

ok guys i replaced padding and i did get a slightly better temp its now around 82c-84c, the performance of the games are still bad, but slighlty better--(I guess dell throttles at 78c??)also, Im beginning to think I need a new heatsink--I checked the heatsink and one of the 3 copper tubes, at the very end of one tube, it looks likes its pressed down, almost as if it wore in, it looks quite awkward. the rest are flat all the way across. what do you guys think?

 

Have you checked gpu fan/vents ?

 

the fan and vents seem to be fine, the fans are spinning and the vents are clean

 

although now it seems to be idle at 73c which is too high---in HW monitor it shows 2500 rpm next to gpu

 

You must ensure that the gpu die makes proper contact with htsnk, M18x example GUIDE: Installing Clevo GTX 680M in Alienware M18x

 

Hmmmm.... Thanks for the guide-- my thermal pads are thin so I may need thicker thermal pads then -- mines are super thin -- which means it may not have full contact-- although it sure feels like it does when I screw the heatsink in--

 

do you have more thermal pads? You could try doubling them and seeing if that helps with contact.

 

The issue here is the gpu temps reaching 90C real fast, that indicates no proper htsnk die contact, try peeling off the black isolating htsnk tape then place the thicker pads mosfet (border) area, the rest should be thinner ones if not you will end increasing the gap between the gpu die and htsnk http://forum.notebookreview.com/ali...reen-680m-serious-gpu-issues.html#post9331732

 

Yes, you should see around 65c to 70c under normal clocks. For mine the full slow down is hardcoded at 78C and can only be overidden with some tricky stuff, and not reccomended. Get it to run cooler.

Also stay away from any NV driver after 314.22. Overvolting and bad things have been happening including fried cards. Get your modded inf's from here: LaptopVideo2Go: Drivers

Just follow the link for the driver, then the inf, and manual save when it opens. Easy way, and the inf name changes a lot but they have always worked when I have needed them.

Look at j95's pic and take yours apart, and make sure you are using the x-bracket for a 680m. This is a common overheat cause since the old one will fit but the post height is very core specific, so a thinner core on the 680m will result in poor contact!

HTH

Edit, take a pic of the card core and core area on the heatsink as soon as you pull them apart. Seeing this will help identify a poor core contact

 

x-bracket GPU Spreader x Bracket for MXM 3 0B NVIDIA GTX and Quadro with Matching Screws | eBay US $39.00 stand up.....

 

hi guys,


sorry it took so long for me to get back to you guys. first I would like to say thanks to you all!! you have been a huge help~ today after I got home I solved the problem---the problem was rather simple as well----I had to do a re-paste on the gpu--I searched around the house and managed to find my IC Diamaond paste (luckly, as I thought I had lost it...) and got to work with cleaning my GPU with alcohol. the paste on the gpu looked worn down and it was really tough to get it off---Once I repaste and put everything back together--bingo! everything in my pc gaming universe is restored!!---- I cranked up crysis 3 on max and it ran smooth like butter!! temps stayed under 73c!!

 

That's great news. I'd still be a bit concerned with the heatsink clearance, there should have been barely any there even if it had gone hard. But it works OK so just keep an eye on it

 

Glad you got it fixed!!

 

I definitely will bro! ive had the 680m for about a year now so it was probably time to redo the pasting anyways, lol

 

Im back in business!, now all that is left is to try and get the HDMI in to work, so i can hook my ps3 to my laptop--I hacked the regedit, by adding DisableHDPFakeMonitor, and I got it working, and after the first day it stopeed again---any ideas anyone?

 

Yeah, one year is when it is said to start to go downhill. This one died on me a few weeks ago, but that wasn't a heat issue. I think that the better paste tends to leach quicker than the thicker, less efficient, stuff.

 

Remember to reapply the thermal paste every time you reassemble the heatsink.

 

that was one of my mistakes right there--I remember removing the heatsink to check GPU a while back--when i removed it I never did re-apply new paste...

 

j95, have you been having problems with your hdmi in as well?

also, hows that 780m working for you? how much better is in than the 680m?

 

780m is faster than the 680m, but not enough to warrant an upgrade or anything like the jump from the 580m to the 680m.

Sent from my Samsung Galaxy S2 running Resurrection Remix

 

Doh, well spotted j95, reading through all the thread this information IS there....