Clevo 7970m crash issue

hey all,


recently i bought a clevo 7970m. installing was easy, and it seemed to work with 13.8 drivers. while playing light games such as nba 2k13, it easily pumps out 60 fps. however, while playing NFS MW 2012 or FC3 on high/highest setting, i encounter a problem. it gives me 60 fps for around 5mins, then laptop shuts off. i thought it was fan was not quick enough, so it overheats.

i increased the fan rpm incrementally, while trying to play NFS. it still kept on shutting down. then finally, it crashed horribly. screen turned black, then appeared again, with artefacts. i had to hard shut down. now it wont even boot up or post, giving me the flashing caps/scroll light, with numlock solid. that means video card is preventing post. back to stock 260m. i dont want to end up with expensive paper weight, any help?

thanks

ps. added log from hwinfo (view with wordwrap off). note the GPU thermal diode going up, even though GPU fan is increasing. crash with artefacts happened when log ends

 

as the card is artifacting, I would test the card with a underclock on both core and mem to be on the safe side.
I take it you have the correct x-bracket on and applied new thermal pads and thermal paste ?

when playing a game I would ramp the fans to max or close to it personally..

as I do not have a clevo card im unsure if they have auto fan controls or not, I guess with you using hwinfo/mon ?
does the card have a dell bios ?

 

I used stock 260m heatsink, with screws taken out and replaced with ones I got with card. it fit the x bracked perfect.

the thing is, it doesn't boot up anymore, gives me that flashing signal thing I said in first post. I read somewhere it might be dead/damaged vbios chip. to reprogram it I would need a programming platform, which I don't have.


the card didn't have native fan control. used hwinfo for that. and in the 2 days the card worked, I didn't realise to check if its dell bios or not. likely it wasn't, as it did say clevo card when I bought it

 

have you tried a bios reset ?

or have you tried to remove the card and retry your older card to see if systems boots ?

 

that is what I'm using right now, I have tried to take the CMOS battery out, but nothing changed

 

It sounds very much like the heatsink wasn't apply proper pressure to the die. I experienced this with a 6990m which would shoot up to 91 degrees within 3 minutes of playing Tomb Raider.
With a 7970M it shutdowns down before you get to those temperatures. Before your card stopped posting did you monitor temperatures at all? That might have been a clue.

I'm pretty sure the x-bracket from the 260M doesn't fit the 7970M at all as some central transistors stick out and stop it making good contact. That x-bracket may also be shorting some components in this central area.
You really need the right x-bracket. Without having one, I wouldn't try the card again. It might not be a total write-off but you may have done some damage by continually allowing it to get to very high temperatures.

X-brackets can be hugely expensive for what they are but perhaps ask if you could borrow one from someone or try to find one cheaply in the marketplace here.

 

x bracket came with the card. the screws didn't fit from 260m heatsink, so I took them out and used the ones supplied.

the log attached in post 1 has everything logged up until the final crash.

also does anyone know where the temp diode is located? because my heatsink had pads in location for 260m, and I didn't add new ones.

 

I wouldn't run the card without thermal pads. They are important for keeping VRAM cool. Here is a pic of the optimum thermal pad placement on the 7970M
View attachment 100759

 

the thing is, shouldn't it lag/throttle once it gets too hot? I was getting 60 fps on NFS then crash.

 

Its good that you have the correct screws. I don't experience thermal throttling on my Dell 7970M. Only power-based throttling when I try to overclock cpu too much.
I also get thermal shutdowns @ 85 degrees and above without any signs of throttling before that.

I suggest you invest in some 1.0mm 5w/mk phobya thermal pads immediately and try to run your card again.

 

now theres the thing. shouldnt something, like post or other gibberish, at least appear on screen even when its just plugged in and no heatsink is attached? i know gpu will crash and burn if booting up with no thermal stuff attached, but it should work, if only for a second. no amount of thermal pads would revive a burned out card

 

I can't say for certain if the card will boot or not with thermal pads but they aren't expensive and it might fix the problem. I have read of people that can't boot when heatsink isn't attached properly. You should also try re-seating the card just to be safe.
If that doesn't solve it and you're out of options you should google how to bake the card. Some people have success baking a broken card. That should be a last option though.

Also I just remembered that I thought my 7970M had died as the fans would spool up to maximum but the screen would be totally black. It also wouldn't shut down with removing the battery and AC. It also wouldn't boot up. I fixed it by removing the card and placing it in a non-static bag overnight. Installed the next morning and everything was perfect.

 

the card has been sitting in the box in a bag since it broke down. will try to resit it with proper pads again later. im running out of thermal paste with all this re seating