M15x & 6970M / 6990M - What you need to know - upgrading, problems, solutions

91C is too hot for the 6990M.

Firstly make sure you have applied decent themal paste properly. I recommend ICD 24. Secondly use ati tray tools to undervolt said card to 1V and run stock clocks. Thirdly Elevate the back of the machine (the 9cell battery is excellent at this). Fourthly take the back panel off and strip out all the plastic wire mesh that cover the intake vents. This helps airflow hugely. Now more dust will enter but simply clean with compressed air once every couple of months and you are golden.

Temps should remain around 85C max. This is even when playing BF3 at max settings and 1080p (except AA which I had turned down) and the 920xm running at 3ghz on all threads.

Happy gaming.

P.S. Use HWinfo to force GPU fan to max speed when gaming. Although I found full speed wasn't necessary. Full speed should keep temps in check around 80C at full load in games if you follow steps I listed above. I use the 9 cell AND use inch thick gel feet at the back and 0.5 inch gel feet at the front. This way there is about 2 inches of clearance under my intake vents Do this if you want ultimate cooling. As I type my 7970M is idling at 38C in ambient 25C Fans idling too.

 

Greetings.

I have a Alienware 15x with a Dell Radeon HD 6970m, using the heatsink from a 260m. The screw pegs for the backplate are bigger than that of the 260m, however, and I don't think the heatsink is making proper contact. I'm not a "shop" type person, so I'm at a loss as to how to grind/file the screw pegs down.

Help, please? I've searched and searched and know to file them, but I don't know how.

 

Don't grind anything down. Simply don't use the backplate from the 260M it isn't compatible. Chuck it away!

Instead purchase a backplate that fits: GPU Spreader / X-Bracket for MXM 3.0b AMD HD69x0M/HD79xxM VGA Modules | eBay

also compatible with the 7970M card if you ever upgrade.

If you are using the correct backplate already then it definitely does NOT need modifying. The only way the heatsink might not be making contact properly is if you are using too thick thermal pads over the memory or haven't screwed the heatsink onto the card properly.

Either way temps should speak for themselves. Run some tests.

 

UPDATE:

My card is now working, using a combination of the latest drivers, the latest bios (A09), downvolting with ATIFlash, less thermal pads, and the 260m backplate.

Alright, folks. I'm going to post this to help any future upgraders, since my path has been filled with bad information and frustration.

As previously said, I bought a Dell 6970m card from eBay that was pulled from a 17x R3. It came with a heatsink. I repasted the GPU with Icy Diamond 7. I initially experienced the thermal shutdowns around 90c as others have mentioned.

To get the drivers working, you -must- use the 17x R3 drivers first, OR manually "find a driver" in control panel/device manager and link it to the drivers extracted when attempting Catalyst installation. Both of these methods work. I installed driver 12.8.

I flashed my bios to A09. No more thermal shutdowns -- why? Because I didn't let it run very long. My temperatures quickly shot up to 110C even with low settings on Guild Wars 2, causing the card to throttle to a pitiful 5fps. Fans were clean and functional. Even running fans at max RPM on HWINFO had little effect on the temperature. Next step was downvolting.

When it comes to downvolting, ATI TRAY TOOLS DO NOT WORK WITH MOST NEW CATALYST DRIVERS IF YOU INSTALL CLEAN. A VITAL DLL IS MISSING. You can download this DLL separately, but it seemed like a waste when you could just use ATIFlash. I downvolted to 1.0 using ATIFlash.

Temperatures still shot up, though somewhat lower. The card still throttled and games were unplayable. Idle temperatures were around 70c, with game temperatures about 100c.

Next step was taking it apart and trying it again. I tried every combination of thermal paste you can imagine -- too little, too much. I then tried different thermal pads.

On the verge of giving up, I noticed that the screw "poles" on the backplate/GPU spreader of the Radeon 6970m and my old GTX260m were different. THE SCREWS OF THE HEATSINK WILL NOT TIGHTEN ADEQUATELY INTO THE RADEON 6970M BACKPLATE FROM DELL. Judging by the eBay link for the ridiculously priced 40 dollar backplate (after shipping), that one probably won't work either if your screws are the same size as mine.


I tried grinding down the screw poles (AFTER TAKING THE BACKPLATE OFF) of the Radeon 6970m backplate with a grinder drill-bit, but they're too strong. If you feel more confident or have better tools, you too can try. Instead, I used the backplate of my old GTX260m on the Radeon 6970m. No grinding necessary. It's a little difficult to line the screws of the heatsink up, but it's doable.

The MXM upgrade guide did the same, using the old GTX backplate.

Alienware m15x upgrade guide HD6970

YOU NEED TO MAKE SURE THAT THE BLACK PLASTIC IS STILL ON THE BACKPLATE. If not, use some THIN masking tape. Why? Because the backplate, as previously mentioned, is NOT a match with the Radeon 6970m. The holes still line up, but the actual plate WILL BE TOUCHING RESISTORS. The plastic will prevent any short circuiting. BE VERY CAREFUL WHEN TAKING EITHER BACKPLATE OFF. The metal on the 260m backplate is much sturdier, and the paste connecting it to the card is much weaker. You -will- bend the crap out of your Radeon 6970m backplate, just don't break it and be -very- slow as the paste is very strong. I took ALL thermal pads off with the exception of the two longest strips.

You COULD theoretically do this without replacing backplates/modifying the backplate by using different screws on the heatsink, but I couldn't figure out how to get them out -- if you even can.

I did not bother with the heatpipe mod.

Results?

Idle temperatures are at 40c. Game temperatures with Guild Wars 2 on high (everything except shadows and LOD) absent a laptop cooler, with previously mentioned downvolting and new bios, WITHOUT HWINFO are at 65c running at a smooth 40-60fps (I also upgraded my RAM to 16GB 1333mhz). With HWINFO manual control on the fans, it runs about five degrees cooler. A laptop cooler from Thermaltake doesn't seem to have much effect on the actual GPU, but it keeps the bottom cooler.

No more shutdowns. No apparent short circuiting. Finally, upgrade complete.

Happy trails.

PS: In the process of this, I stripped a screw on the heatsink which caused me to be unable to take it off. These screws are VERY soft -- it only took a couple of turns to strip it. IF this happens to you, don't listen to people recommending a Dremel to cut the screw off unless you're an expert. Don't bother with the rubber band trick, flatheads, or left-handed drill bits -- you will only make it worse. Instead, get a Grabit Pro from ACE Hardware or Amazon. They're 15-20 dollars. One side drills in to expose threads (it will leave a big hole in your screw as even the #1 doesn't actually fit into the screw, BE CAREFUL). The other side takes the screw out -- again, be careful as it takes about 4 seconds. After you're done, make sure you clean up all the metal shavings. I didn't bother trying to replace or retighten the stripped screw, as it was finally tight enough. My temperatures speak for themselves.

 

Thanks for the advice -- genuinely, as you were the only one to reply -- but other people in this thread have mentioned that the poles don't fit adequately (even judging from that picture, the poles wouldn't fit my heatsink's screws) and that they've had to mod their backplates. I was using the heatsink that was present for the 260m, and the backplate present on a Dell/Alienware 6970m. The card actually finally worked when I used the 260m backplate. MXM also uses the 260m backplate. ( Alienware m15x upgrade guide HD6970)

 

No problem mate. The only problem I ever found was that yes the screws might not fit between backplates however getting screws that do fit is an easy fix.

I use the orginal heasink plus 7970M and it's own backplate. An absolute cinch to install. The screws on my 6990M backplate didn't fit it actually but I was able to find fitting screws easily and install the card.

 

Hey guys!

So I just bought a Clevo 6970m and I got in my m15x just fine, thermal paste, thermal pads, booted up first time, downloaded the 12.10 drivers, worked like a charm, fixed the multi-display bug that had it running at 400/900 idle.

Started up WoW thinking I'm gonna get some awesome frames... and nope.. recommended settings i'm getting like 14fps..
Started up GW2.. 5fps on recommended settings

temps haven't gone above 65c in HWinfo64
tried several different drivers and NOTHING is working...

also stealth mode is NOT on, i've made sure.
and Battery is set to High Performance

i dont know what i'm doing wrong.

i7 720qm
6gb Ram
Clevo 6970m
150w Power Adapter

 

I need you to download GPU-Z and monitor the sensor tab while you're gaming. If it doesn't say 680/900 which is the default for core and memory, then I guess that's the issue.

One solution I recommends is to try and do a power drain trick - power off, remove the battery and AC, press and hold the power button for 20 seconds, connect everything back up and see if it works.

 

Okay did the power drain

opened up GPU-Z, did a log and ran wow for 15 minutes.

Didn't go above 100/150... at all..
2nd time around, I deleted CCC, and got it to go up to 680/900
BUT, my frames didn't move.. still ridiculously low

also I noticed this

 

a couple of questions first are you on bios A09? also hit the "?" mark that you have by the bus interface to run a render test, switch over to your sensor tab and check your gpus clocks along with its temps and see what its doing. you might also want to download this and try forcing the stock clocks. Im unsure if you need to have a dell vbios for the clevo gpu?

 

I figured it out. I googled it and it turned out that just some dust or something got into the actually bus interface slot, and so I took the card out, cleaned off the connection and blew into the slot and re-seated and boom. It works now. ;running at the proper x16 2.0

HOWEVER. now i'm having overheating issues <.<
wow runs 40-50 fps on high aaannnndd I hit 80c in under an hour and shutdown

 

Try running HWinfo to run the fans at 100% and see if that helps.

 

Okay I had the fan running at 100%
also took my card out and re-pasted, and put more thermal pads, also put some underneath the card.

After an hour of wow, I got no higher than 75c with no shutdown.

I'm gonna try a few other games and see what I can come up with.

 

Sounds better than before. Do the 6970m cards really get that hot? Then again my 260M can get hot as 73C and that's after a re-paste. Only really demanding games get my gpu above around 64C.

How's the temps on your processor?

 

Played Black Ops 2 last night, shot up to 80c and shutdown withing 20 min..

 

Are you on the latest A09 system bios? If not patch up. It will allow higher gpu temps before shutdown.

 

yeah I am I flashed that awhile ago.

I am noticing however that with the 3 different temps i'm getting for the GPU in GPU-Z, one is getting up to like 90c, while the other 2 hover around 80-82c

I did downvolt/clock to 500/600 .9v and after an hour of black ops 2, it didn't shutdown and highest temps were 79c

so hopefully that works. fingers crossed X(>.<)X

 

Try undervolting to 1v at stock clocks. Also use hwinfo to force fans to max. Then report back temps. Maybe another re-paste is required

 

Shutdowns can be due to sheer power draw and not temps, so undervolting should help.

 

sigh... i just went back to play bf3 after a 2 months break from the game and the shutdown is back. haven't experience in other games i played recently like sleeping dogs. i just did a full laptop repaste and pad. the gpu-z temp logged 69C before shutdown.

as for dota2, i also experience shutdowns, but later solved the issue by turning off specular bloom in the options while leaving the other maxed. it's weird but it does fix the issue. butter smooth fps 60+ and gameplay by the way, no hiccups whatsoever when bloom is off.

 

Wanna upgrade my 5850M, so i'm locking for 6970M or 6990M.
And i'm planning to do some overclock for every day use. At 5850M i had 840/1200/1.1V, this equal to ~2600 3DMark11 score.
Which card will be better for M15x, and what clock's & temperatures i should expect?

 

Why not a 7970M. It is alot faster than either card and runs cooler too!

At stock clocks you get nearly 7000 3Dmark11 score

You haven't got alot of OC headroom on these high end cards because the m15x is fussy where power is concerned. Therefore running a 7970M at stock clocks still destroy any OC on the 6990M that would actually be stable in the M15x.

I owned the 6990M inside this laptop before the 7970M. Also ran the 6970m before that in a different laptop.

 

@King of Interns Thanks for reply. 7970M - is really beast (especially with 900p screen ).
7970M atleast 2-3x more expensive than 6970M, and i have not big budget & only 150W PSU.

 

It will work fine with 150W. 240W doesn't make any difference any than cooler brick.

Agreed for 900p the 6970M should be plenty powerful for the lower res 900p for a few years yet.

 

where can i find M15x & 6970M / 6990M - What you need to know - upgrading, problems, solutions info at??? I forgot the dos command for flashing the video card.

 

The original post is here.

Why are you flashing?

 

Laptop started shutting off while playing Dota2, and on guildwars2 it shuts off even faster. I have an X940 and 6970 card. Thinking of down clocking video card and lowering voltage. Also i'm wondering if it's the ac adapter 150 watts is causing the shut downs.It gets so hot you can fry an egg on it, I hear the ac adapter has a safety feature that makes it shut down when adapter gets to hot or laptop trying to pull to much current. I'm thinking of buying a new adapter but a 240 watts adapter.

 

Thx for link.

 

I'm also thinking of getting the 240w adapter but I haven't experienced those problems. Try undervolting the card to 1.0v.

Sent from my cm_tenderloin using Tapatalk HD

 

I'm at 1.0v and still having shut down issues. How do you keep your laptop cool did you use thermal pads?

 

Most importantly, did you grind down the x-bracket and use thermal pads of appropriate sizes? I play Dota 2 everyday without a single problem.

 

With my card and x-bracket. I didn't need to grind. But since you mentioned it. I'm gonna check out the x-bracket and see if it needs some grinding, also i don't have thermal pads in my laptop i.m gonna order some thermal pads today.

 

Ordered the pads this morning, as soon as I get the pads. I'll check to see if the bracket needs grinding. I'll keep you posted.

 

I have a clevo 6990m (which I think might be non-reference) and when I try to edit BIOS in RBE the gpu registers are greyed out.
Is there any way to edit voltages on this card as doing it in the main window doesn't work apparently (and not using software).

If anyone has an undervolted clevo 6990m BIOS could they send it to me.

Thanks

 

Adding to the list.

My 3rd shutdown since Thursday, March 21, 2013, the day I got a 6990m.

Shutdown #1 - stock bios from clevo. Temps hit 90C at some point playing... Diablo 3 on maximum settings 60 fps and whatnot.. I live in an oven of a country [Philippines]. Normal game play. idle temps are at 49-50C [Good, right?]

Shutdown #2 - undervolted to 1.0V from 1.1V stock clocks. Temps did *not* hit 90C at all but still it shut down. Was playing Assassin's creed 3. Also, it happened at the part where you run to the city from the frontier area to take down some ships. I got through the whole visual load on max settings without it shutting down. It was when I got the ships that it died on me.

Shutdown #3 - undervolted to 1.0V from 1.1V and still stock clocks. about 2 hours ago.. or was it 3. I was playing diablo 3 with 2 friends. we were in act 4. Barbarian [me], barbarian [friend], and monk [friend]. We were going at it for about 2 hours I think. The first hour I was in the office so airconditioning was there. Temps did not go beyond 85C. For the Second hour, I went home [computer was shut down properly so temps are cooled down], booted up and continued the game. At the second hour mark or thereabouts, laptop shut down. Temps were... 86/85/88 or something like that. Also worth noting is that all my shutdowns have only occurred in an room without air conditioning.



Currently: I reflashed bios to dell's from svl's thread over at TI and still maintained the undervolt at 1.0V from the 1.1V and downvolted the 1.05V ones to 1.0V. Clocks were left at stock so that.s 680/900.

If I get another shut down after this doing whatever, I may have to reduce clocks to 650/900. My primary concern is that... is the overall gaming performance hit going to be obvious? These things I'm not entirely familiar with. I just know bigger = better BUT I do not know how other factors such as bandwidth/stream processors may be affected. maybe bottle-necked? not fully utilized?

Specs:
150W stock PSU
740QC stock
ati 6990m 13.1 drivers, 680/900, 1.0V
500gb sata
6gb 2+4 stock ram sticks.
900p

 

What is the thickness of your thermal pads? Which X-bracket did you use?

 

.5mm pads on the heatsink
clevo xplate. This is what they looked like when I got them and they fit nicely.

 

Your paste looks good. Did you grind the x-bracket to the appropriate height?

 

yep. The heatsink and the die make excellent contact and I didn't need to do the heatsink mod where you... hammer a pipe because of something that gets in the way of the heatpipe.

 

this is my set up fyi

 

That's a LOT of pads all over the card. I'll consider it if I get just one more shutdown.
Also why is it that your.. memory bandwidth and fill rates are faster than mine? Are those an effect of the OC?
mine reads at 21.8 Gpixel/s, 38.1 Gtexel/s, and 115.2 GB/s.

 

its the oc