580m gtx repaste form

i am attempting to use AS5 thermal paste and would like to know what the right way to apply it. should i spread using a something flat or should i go ahead and just let the heatsink do its job?

btw. thanks widzeu69 for all your help

 

Nah just a small line in the middle and let the heatsink spread it flat

 

this is what i did for my 6990m and it has shown dramatic results.

Curiously, if you go to the AS5 website, they recommend using the credit card method for laptops...i.e., using a credit card to spred a thin lower over the whole die. I didn't do that though

 

I don't recommend it. What happens if the spread when bad and the middle was thinner than the outside and then it creates an air bubble which is the worst thing for components.

 

ahh i c...i need to do this asap

 

I always spread it myself using a very thin layer,tried using the pea dot or line method but it never seems to work for me.Recently i replaced my GPU so i tried again the pea dot method,idle temp at 59-63,max at 76 degrees.Going back to spreading it myself,my final temps were 50-54 idle,max 71 degrees.You should try out various methods to get the best results,that's the key.

 

i finally did he line repaste then the idle temps went down as low as 45C but the furmark tests hit up to 75C

 

That's decent. Furmark usually stresses the GPU more than any game.

 

I had my screen and speakers changed over this morning.
Asked the engineer 'nicely' to repaste my GPU.

It was actually quite tidy and even (from manufacture) but since I had the opportunity I thought I would let him to do it.
He did it quite quickly and not in the tidiest way either.
I had to ask him to take the old paste off.
Said it doesn't make much difference.

Should I be worried about the tardiness +/- overspill etc..
How can I check if it has made any difference?

 

That usually isn't that great. Techinicians are usually ok at servicing and really bad at repastes and stuff. I would do it myself again. You can check using furmark brun-in. Post your temps after about 10 mins and we'll let you know if the are good or not.

 

which test/setting should I carry out? I've got latest furmark

 

OK this is the main 'BURN-IN' test for 10 mins

FurMark v1.9.1 - Burni-in test, 1920x180 - 15 FPS

FPS - min: 13, max: 22, avg: 15
Frames: 9156 - time 00:10:00
GPU 1 - Geforce GTX 580M
core: 476MHz, mem: 1500MHz, temp: 77'C
VDDC: 0.850V, fan speed: 0%
GPU load: 99%, mem load: 6%

How does this look.
I have nothing to compare with pre-repasting

 

Ok the temps are good but you GPU has throttled. Have you tried the throttling fix? I should stay at 620Mhz for the whole time and not got down to 476Mhz.

 

Thats good.
Any chance I can compare temps to yours? (or where I can compare with others)
Regards throttling; I'm a bit of a novice to try all that flashing, blind or not
I'll have to wait until Dell release something official or something thats is REALLY easy and reversible.

Maybe this throttling is why my 3D Vantage score, about 13.3k for GPU, and P14560 overall, is said to be 'low' compared to similar systems when I submit its scores.

 

Welcome to the thermal throttle issue. As of right now as soon as the GPU temps reach 78 degrees the core clock with throttle back. To avoid this use a Laptop cooler, move to Antartica, do what ever you can to keep the temp down and it will not throttle. I have tried everything in my power to keep my temps down but I still hit that thermal throttle so I am actually going back to a 6990m for my needs.

The Blind flash only addresses the Power throttle issues where the GPU with throttle back if it is using too much power. The Thermal throttle is a issue for everyone and no fix I have seen to correct it.

Good Luck

 

Those numbers are normal and not low at all. The other score that have been submitted are higher because people overclock. It seems you have a temperature throttle. My advise, use the laptop with the rear propped up. You will have no throttling if you are running on all stock clocks

 

since your gpu is overclocked... when you run furmark or games like battlefield 3 do you use a laptop cooler or just simple prop your computer up?

 

I am not finding this to be accurate on a system with the P0 and P1 power states set to identical core/shader/memory clock speeds. While the power state may change, throttling of clock speeds does not happen because changing power states under load or temperature becomes irrelevant.

Is there something different about the MXM module in the M17xR3? The SLI setup in the M18x is totally unaffected by throttling if the proper steps are taken to correct it. As you can see in this video, I ran a burn-in test well beyond 78°C (up to 86°C, in fact,) with no effects from throttling apparent. And, flashing the vBIOS is not necessary for those that find doing so intimidating.

<param name="movie" value="http://www.youtube.com/v/XEGWWT8yhAU?version=3&amp;hl=en_US"></param><param name="allowFullScreen" value="true"></param><param name="allowscriptaccess" value="always"></param><embed src="http://www.youtube.com/v/XEGWWT8yhAU?version=3&amp;hl=en_US" type="application/x-shockwave-flash" width='480' height="360" allowscriptaccess="always" allowfullscreen="true"></embed></object>

 

What is a good furnark temp on the 580m? I keep hitting the 79C throttle, occasionally I can get 10 minutes until it hits, but it does it every time eventually.

 

I get it too no matter what I do. It's fine You shouldn't hit it when gaming.

 

I think that temps may depend somewhat on your clock speeds. The video above is 800/1600/1800. Furmark is a pretty brutal test... more so than any game. Are your P0 and P1 clock speeds set exactly the same?

 

I get a max of 70c after 10mins of furmark, that's with a coolmaster under the LT.

 

Are you using the stock/throttled vbios or Widezus full 620mhz vbios?

 

Hello,

I am wondering if this is normal but I am hitting the 77/78 throttle after like 30s of FurMark with no overclock what so ever. Also after this happens FurMark reads core:73MHz, mem:324MHz. I also noticed under nvidia inspector that there are P8 and P12 powerstates that have low settings(P8 =73/324, P12=50/135). My idle temp is around 55. Is this normal?

What should I do about this? Aside from SC2/Skyrim every game I play throttles (shogun 2/dead island in matter of minutes). I tried the various fixes offered on this forum but nothing solved this issue. Do you guys think something is wrong hardware wise?

Thank you,

timurStas

 

I am using Widezus BIOS, with the 620mhz P1? state, so furmark does not cause a downclock to 475.

 

@timur nothing is wrong hardware wise. Just have to try and get the temps under control. A repaste, elevate the back, notebook cooler (I suggest U3) and move to somewhere with a cool ambient.

 

Sorry to kick a dead horse guys but I have a few questions. let me first start with I own a computer repair and servicing business so I do in fact know my way around hardware.

Now that is out of the way, After dell confirmed that my GTX 580m was in fact dead, they sent out a repair technician with a refurbished motherboard and GTX 580M. I watched

the so called "technician" nervously disassemble my precious lappy and then admitted to me that this was the first M17x R3 he has ever worked on... . So, As he disassembled and placed parts on carpet with no grounding, I realized that he was gonna reuse the old thermal paste and I was trying to inform him that thermal paste is a one time use thing, of which he disagreed. So after putting the new motherboard he put the cpu in with barely any thermal paste on it becasue it rubbed off on the carpet ( this guy was clueless ). He then proceeded to rip the old thermal pads of tearing and stretching them and placing them on the new GPU. After closing everything up he just left.

I gave your guys the long story because I am still occasionaly get the Driver State Power Failure, IRQL NOT LESS OR EQUAL, and lastly DRIVER IRQL NOT LESS OR EQUAL. Not to mention temps skyrocketing ( to be expected since the pads and pate were reused). Thinking that maybe just a fresh OS install would assit in stopping the regular occurecs of the BSOD, I reinstalled Windows 7 and put in all of the most recent drivers, and i still get these. I want to know everyones opinion as to what i should do. Im really tempted to call up dell and them out.

I want to repaste the the GPU and CPU now, but the only thing that has me a little hesitant from going forward with this is I dont know the thicknesses and amount of thermal pads I need in order to properly do the entire GPU. I also didnt get a close look at the back of the card to see if there was a heat spreader like the 6990. Any help in the this area would be greatly appreciated.

 

I'm not really an expert on the 580m, but I get the same BSOD if I don't have windows 7 sp1 installed when i reformat my machine. I would check and make sure you have that installed and it should fix the issue.

 

First thing needed is to get paste (and pads) on that 580M. It's going to be hard to troubleshoot the BSOD's until temps are stabilized. They are very likely related.

Most people use 1.0mm replacement pads. These are easily compressible and quite effective. Fujipoly Premium System Builder Thermal Pad - 1/4 Sheet - 150 x 100 x 1.0 - Thermal Conductivity 6.0 W/mK - FrozenCPU.com.

Here's the back side of the 580M. Just re-use the large pad that came with the original card.

 

^^ I never understood why they put a large pad on the back/bottom side of the card liek that. Makes no sense to me

 

^^ So the GPU components don't touch the motherboard! If they do touch your system will short circuit and cause beeps, cause weird artefacts, or just destory your machine completely.

They probably won't, but you can see in the previous picture where the components have made a little dent in the pad. I guess it could help with the 580M vram chips on the back also.

 

Ya, that makes sense +

 

Based upon the impressions in the pad, I'd guess it is primarily to avoid contact between the GPU and whatever is directly underneath (which I have forgotten ATM), but also to improve heat dispersion from the back of the VRAM chips, which extend through on the 580M, by increasing the surface area.

 

Thanks bro, Geo took my last rep for 24hrs...will get you later

 

Ok so heres the thing tho, that thick pad is what the so called technician tore up and used as thermal pads on the vram.... and he put to little squares on the back to separate it from the motherboard. So I guess i need to replace that thermal pad with a 2mm thick pad and replace the vram pads with 0.5mm thick pads right?

So my shopping list to repaste / repad my laptop includes:

0.5mm pad - for Vram (thermal conductivity is 7W/m.K)

2mm pad - for back (thermal conductivity is 3.2W/m.K)

Arctic Silver 5 - CPU & GPU


Does that look right? I already own Arctic Silver 5, but if anyone knows of a paste that would substantially help cooling performance please feel free to mention it.

 

Use PK-1. In my tests it was second best after IC Dianond but I wouldn't recommend that stuff.

Also 0.5mm pads are too thin. Get some Fujipoly 1mm ones. They are good.

 

Those phobya 7w/mk pads are apparently too tough and hard to work with...i'd go with the 5w/mk phobya pads instead. And like Widezu mentioned, go with 1mm pads

AS5 is old. Get PK-1 if you can...if not, something like MX-4 is readily available

 

Ok, I read a review that the 6w/mk worked well on the 6990 and he saw a 5-8 C difference so I'll go with those. And i thought the stock thermal pads were 0.5mm? So are we going with 0.5mm or 1mm with the Vram?

Revised list:

0.5mm pad - Vram

1mm pad - for the back

PK-1 - for CPU and GPU

Look good?

 

Nope, 1.0mm for the VRAM. The .5mm pads are too thin and flimsy; very hard to avoid tearing.

It may not matter for the back, but the stock one was at least 2.0mm, maybe more, so I'd tend to stay in that range, mainly for durability.

If the tech used the 2.0mm pads for VRAM, they probably are holding up the heatsink and preventing tight contact with the die. Those pads are the likely cause of the heat problems.

 

SpecialK, ^ ^ ^ ^ This ^^^^...so just get 1mm pads, and make sure they are the 5w/mk if you're gonna buy phobya...you can double up on them to make 2mm for the back.

PK-1 is one of the best pastes on the market today, so that definitely looks good

 

Awesome guys, thank you so much for bearing with me, I just wanted to make sure that i was gonna have everything that I needed before I cracked this thing open and realize I dont have what i need to do the project properly.

 

Just realized something, the MXM 3B card is 105mmx82mm and the thermal pads are 150x100mm, so i guess i need to buy two sheets?

 

Woah dude, the thermal pads are not for covering the whole card. You only need like 120mm x 15mm in total for the vrams.

 

lol I understand that, but I need to replace the pad on the back of the card because there isnt one there... the technician just tore that pad and put 2 little squares on the back to make sure it doesnt touch, and put the rest on the vram. So basically I'm cleaning up the snafu the technician created.

 

Oh right in that case the single pad should be enough.