I've been reading about this for a good 2-3 hrs now and it looks like the gx640 needs a cooling mod. Is the heat sink contact caused by the vram pads being too thick? If so cant i just apply a thick paste like mx-3 and tighten it? If I were to do a copper shim mod what thickness do you guys think is the best for the copper shim, how many would i need? Ive also got some spare thermal pads by swiftech but its 2.5mm thick, is that too much? Thanks guys help would be much appreciated![]()
-
GapItLykAMaori Notebook Evangelist
-
I applied MX-2 on the graphics memory; it improved my temperatures by a lot. I have a MS-1727 (GX740) though, I'm not sure if that would be the same.
-
yea, i would like to do this upgrade as well. i think i've read something about not needing to replace the thermal pads if you get the copper shim tho.
here is what you need according to this guy. would like to know what thickness it should be as well. and also the other dimensions of the copper shims.
http://forum.notebookreview.com/msi/473766-official-msi-gx640-owners-thread-63.html#post6502234 -
Where can i buy copper shims for msi gx640?
can some1 post a link? -
The thermal paste used by MSI is quite good already, so there won't be a significant change with the CPU. The GPU, however, really benefits from the improved contact with the heatsink. It used to idle at 63°C, so it's pretty dramatic.
As for the size of the shims, well, it's going to vary depending on the person you order them from. I ordered from this guy and the shims arrived in less than a week. He included a small baggie of thermal paste, so technically, you don't have to buy your own if you don't want to. The shims, measuring ¾" on each side, or 19.05mm in the civilized world, are a perfect fit for the heatsink. However, in order to fill the gap, you'll need to use two of them (2x .022" or 0.56mm). The guy has another auction for shims that are twice as thick (.044") if you just want to use one.
Alternatively, you can just remove the thermal pads completely and place tiny spots of non-conductive thermal paste on the memory modules. I haven't tried this one out, but apparently, it works. -
so place copper shim pad between gpu or heatsink?
-
-
GapItLykAMaori Notebook Evangelist
Thanks Dakins ur a very big help! I live in New Zealand and copper shims are hard to get hold of. I might just remove the thermal pads and place sum mx3 on the mem. I will report back once I get my laptop and done all this which should be within 2 wks cause i have exams :/
-
I also did the copper shim mod with tuniq TX-4. The temperature dropped around 10-15C. Some extreme case 20C drop. I also redone the pads for the rams. The stock thermal paste are just crap
-
You can try a crafts store to see if they sell copper sheets and cut it yourself. They sometimes sell rolls of copper but they'll be really thin and you'll probably need to double it. -
So yeah, I don't want to be responsible for destroying anyone's GPU, so as a word of caution, don't remove the thermal pads. Copper shims ship worldwide in a standard envelope, you have no excuse! -
GapItLykAMaori Notebook Evangelist
hmm ok then that worried me :/ O well ill be getting my laptop nxt wk ustoms held onto it
-
The dude who you suggested Dakins to get the shims off ... turned out to be dodgy, so I lost my payment through ebay...
Will look for another...
Cheers
Chumby -
a few question
do i need put thermal paste in the memories of the video card?
about the cooper shim, do i need buy a few more for the memory if i dont apply thermal paste?? and what should be the size of the cooper shim for memory -
SoundOf1HandClapping Was once a Forge
Don't put a copper shim on your memory. Just thermal pads.
-
but that thermal pads, where i can buy them?
http://cgi.ebay.ca/DV9000-DV6000-V6...738?pt=LH_DefaultDomain_0&hash=item3a5debe882
that thing is for the GPU or the Graphics Memery? -
SoundOf1HandClapping Was once a Forge
I use and recommend these.
-
Also, he dont ship worlwide.. im from chile -
I just replaced all the thermal pads I could remove with some better quality ones that don't fall to pieces when you try and remove them. noticed maybe 5-10*c drop, not sure though, it's a hot laptop after gaming lol. I use it on a coolermaster 19'' cooling stand (with 3 fans) I also put some copper heat sinks on the gpu plate.
-
niffcreature ex computer dyke
If we are to be using and replacing thermal pads, we need a much more reliable source, and description of the differences in quality.
Forge, your link was useless. At least after 5 days or so.
I don't see why RAM modules would be built in a fragile way. RAM is even more mass produced, used for much wider variety of electronics and has way more industrial standards.
2.5mm is way too thick for thermal pads. And .5mm CAN be too thick as well. Thermal pads come in so many different varieties which is why I don't really trust them Most people are used to the gel kind which are not as easy to find. These change their thickness a lot more than many pads on the market which do not have a consistency at all similar to what MSI intended. -
Meaker@Sager Company Representative
Because CPUs can take around 183psi of heatsink mounting pressure while the GPUs are covered in a plastic protective coating which is just designed for low pressure heatsinks.
-
Still, I would point out that some desktop graphic cards seem to apply the same pressure on the die and the memory chips which could mean they could withstand same pressure.
Anyway, I would like to suggest that people who plan on using shims check their heatsink first.
It seems there are several types of hetsinks that MSI might equip your laptop with.
For example, MSI GX640-260US that I use came with a heatsink which has an elevated surface in the center which makes better contact with the gpu die. And I've heard the MSI GX640-098US doesn't have this elevated surface(can someone confirm?)
There was another guy in this forum who had a totally different heatsink(something from older models?i don't really remember) unlike anyone else. -
Meaker@Sager Company Representative
CPUs and GPUs are made in a very similar process.
I know not of a single heatsink where the screws are set around the memory chips to also apply the same pressure.
They wont. -
So what you are saying is logical but its still a guess.
And I have not heard of anyone breaking their VRAM chips due to extreme pressure, although I know a couple of instances of GPU die failing under extreme pressure.
The pic is from a reference nvidia gpu cooler.Attached Files:
-
-
Meaker@Sager Company Representative
Do you really think those surround case assembly screws apply the same pressure as the ones surrounding the core there?
-
Can you give me a good reason or a data to show me it wouldn't?
And your missing my point really.
All I'm trying to say is, what you are saying maybe true but highly hypothetical since theres no data on how much pressure a GDDR chip can sustain. -
Meaker@Sager Company Representative
Thickness of the pads (paste on the GPU), the fact that the core is surrounded on all sides by screws close to it (the rest are spread around the case) the fact the core screws are likely spring loaded and the rest are normal screws, previous experience with graphics cards, the fact that memory does fine often without cooling, the fact the core is protected with a heatspreader protecting it from extreme pressures.
And the best reason of all:
Common sense. -
Again, you are citing subjective past experiences and assumptions as a proof that the GPU can sustain more pressure. Thats not very convincing.
Except for the pads, but personally I think they use pads because GDDR chip needs less cooling as you've said yourself. Manufacturers don't NEED to use a thermal paste.
Also the existence of heatspreader on the GPU die is really a disproof of the fact that GPU die can tolerate high pressure (more surface area=less pressure per area, I'm sure you would know).
Now, most importantly, all I'm trying to convey is that without any viable data, we shouldn't pass judgements on the solidity of GDDR chips. -
niffcreature ex computer dyke
Where did you see any viable data? Who brought that up? Is that what we need here? And what judgements have been "passed" so far, AND what exactly do you mean by that?
No, no he is not trying to prove anything. You read the damn posts. He is voicing his opinion backed by experience which is more important than a spec sheet which was probably written by someone who has never touched a heatsink. Just think for a second. And you are choosing to or not to listen to this opinion and take it for what its worth.
If you wanted to post a thread about finding spec sheets and pressure tolerances for some vague reason, then go for it. But this isn't your thread, and it isn't about that, its about the opinions and recommendations of users who have experience here AND more importantly improving the actual cooling capacity which again is NOT something anyone who is qualified to 'pass judgments' about the viability of this data cares about.
The disproof of pretty much anything you see a manufacturer doing is that it was NOT based on a specification, it was based on an incentive of lowering costs. This mere fact could be the entirety of the reason why thermal pads exist.
So, to simply say a heat spreader on a heatsink PROVES anything whatsoever about pressure is frankly ridiculous.
Anyway, back to business.
I'm pretty sure thermal pads are used mostly because of the discrepancies in thickness. The way many cards are manufactured, some of these MSI 5850 may have chips of different brands like Samsung and Hynix.
Since I was the one who brought this in the first place, I will say that I have put some pretty extreme pressure on vRAM modules in the past with no adverse effects.
I think the only reason we would have to think they may crack under pressure is actually for a similar reason to why I thought they wouldn't. Up until very recently, yes, GPU chips did not need active cooling at all, and RAM generally doesn't need it. But now gddr5 definitely needs it, and I don't know what RAM manufacturers know about heatsink interfacing etc at this time.
Maybe we should get a materials engineer in this discussion, no? Why exactly do things like this 'crack'.
Maybe its not even possible to crack a RAM chip in the same way that GPU dies have been cracked, thats going on what I said earlier... its a much simpler component. -
Niffcreature, you're quite right. i was the one who should have read carefully.
Meaker, I sincerely appolgize for my unnecessary and irrelevant criticism of your posts.
All I should have said was my own view on the matter and no more.
And I'm sorry if I have caused unpleasant feelings to any member of this site. -
Meaker@Sager Company Representative
It's fine, and if you find evidence otherwise I am quite happy to look at it.
I just think you want to be careful, one is I really don't think DDR chips will take the same pressure as the GPU, maybe a fair chunk but not as high.
But also consider that by putting more pressure on the memory chips you could well reduce the pressure on the core. -
Im bumping this old thread, since i need some support in this matter. Whats the general conclusion regarding the thermalpad-copper shim mod? I´ve had several heatissues with my gx640, and for now i only applied thermal compound on my vram directly, wich have lowered my temps VERY much. Im planning on upgrading my gpu though to this one DELL Alienware M17x R4 M18x R2 graphics card NVIDIA GeForce GTX 660M M3XJV | eBay , and i wonder what kind of cooling mods i have to do on that one.
Msi GX640 Thermal Pads and Copper Shim
Discussion in 'MSI' started by GapItLykAMaori, Aug 1, 2010.