Grab copper and a drill, now!

Kondor removed the entire grill and said he got better GPU temperature with little to no effect on the other components. With my minor drilling I can confirm that my temperatures look better.

I'm getting annoyed with my temperatures. Soviet, you having problems with your card clocking up while browsing these forums?

 

Switch to the "old nbr style" down at the bottom left of the site. it removes the ads, and stops the GPU from clocking up.

 

Complete coincidence - I just received some of those short copper heatsinks. Here's my experience with them ... http://s65.photobucket.com/albums/h216/Ray_GTI-R/PC/BGA MOSFET RAM heatsinks

Sekisui #5760 thermal tape (in the UK, try Ebayer 1jupiterway) I'm not sure how it works in laptops that are prone to flex.

 

Great minds think alike, brother. Welcome to NBR.

 

Now it's time to test it in practice...

 

Did another furmark run.

1366x768, Xtreme Burning Mode, 8x AA.

Hardware in " Crysis mode", meaning this is what I would normally play Crysis with. CPU at full 11.5 multiplier, 3.06 GHz, undervolted to 1.250v. GPU at 580/1410/870 (1::2.5::1.5 ratio).

On top of modified Zalman. Meaning that the Zalman was off and the G51 propped up in the corners with roughly quarter-inch high object (.22LR bullet casings. I know I will be ridiculed for this).

Ambient 74F.

Max of 87C, stabilization of 86C


I'm still wondering why my max is higher than usual. Usually I would hang around 86C for a bit, and then the fans would kick in, bringing the GPU temperature down. I started noticing this after installing a 0.0216''/0.55mm copper shim on the GPU core and Vista SP2.

Could be a bad shim job. While I was sure to hammer it flat, I didn't lap it mirror-shiny, which might have a detrimental effect.

So, if the ramp up temperature of the fan is 86C, and the fan gradually spins up--as opposed to going to speed at 86C exactly--it could mean a bad shim wouldn't transfer heat optimally, and the GPU would continue heating up to 87C before the fans kicked in.

Gah.

 

So was I supposed to grab the copper before the drill? Cuz... I grabbed the drill first and well...

Does anyone know first aid?

Seriously though... Forge you would think even if the copper wasn't that thoroughly polished, that the thermal paste would have hit enough heating cycles to fill those pits...?

 

I have no clue, really. It should. I might just remove the GPU shim entirely and just do a thick layer of MX-3 like I was using before.

 

That's what I'm thinking as well. It might be best to revert back to using a thicker layer of thermal paste.

 

Will do, especially since you reported that MX-3 tends to work better in thicker layers. I'll keep the northbridge shim, though.

By the way, ordered an XPad.


Edit: Removed the shim and redid the paste. Did a quick Furmark run to see how hot it would hit. Still hitting 87 max. Granted, it's 80F, as opposed to my usual 74F testing range, so I'll wait a bit.

 

Those are pretty damn good numbers for 80*F. Just out of curiosity, do you push down a bit on the heatpipe just above of the area where the GPU die isn't making much contact with the copper? Maybe that might help.

Other than creating more area for air to traverse underneath the notebook, the second advantage that the Xpad bears is less more clearance for the fan. Small notebook fans usually need ~2cm of ground clearance to be relatively free of turbulence when sucking in air. Most notebooks have feet that only prop up the notebook so much that there is just enough space for the fan to suck in air. However, it won't be sucking in that air efficiently. The amount of clearance that the feet give is definately less than 1cm, probably averaging ~6-8mm for most notebooks, which will generate that vacuum turbulence when you have the fan engine so confined against solid ground. The Xpad, compared to the NC2000 and Cryo LX, gives the notebook fans the extra clearance it needs to be able to suck more efficiently. More clearance means less turbulence, less dust sucked in, and better passive cooling overall under the entire face of the notebook underside. Along with low cost, portability, and versatility, I would never go back to active notebook coolers ever again.

Nope. My GPU stays at 160/320/80 whenever I browse NBR when logged in. I always use the old NBR skin. I'm telling you man, those JavaScript ads really pack a wallop.

 

I'll try that. Temperature right now is 79F and dropping, so I hope to see some improvement. If I can get my usual 86C max and 83 stabilization, that'll work out nicely.

What if Edova put a fan in the XPad, and your current pad broke. Reality meltdown.

Seriously. I've switched to simple mode and I'm not having the issues anymore.

And you keep pretty low 2D clocks. Those aren't standard, right?

 

You will be getting much nicer temperatures once your room settles at 74*F. I can tell you've already hit a new record.

If Edova put a fan in the Xpad, then I would just make my own. I already have the design for the Xpad so I'll just pick up a slate of Torlon and make my own Xpad tailored just for my notebook.

My factory idle clocks are 200/400/100 underclocked down to 160/320/80, which is 20% across the board.

 

Hope so. Just a few more hours, I guess.

And since the GTX260m/280m can handle pretty much any 2D environment... hmm. Come here, NiBiTor.

 

Just be careful when going under 160/320/80. I know that 150/300/75 makes my notebook cranky when waking up from sleep mode.

 

Ah, that's for later. Time to give that MX-3 some thermal cycling in Left 4 Dead.


EDIT: After finishing a game, I shut if off and did what you suggested. In the middle of putting the bottom back on, I said "what the hell" and disassembled the heatsink again. After cleaning I slathered on a bit more MX-3 than before. Actually, I would put on a layer, smoosh the GPU heatsink down, lift it, examine where I was getting gaps, and dab/spread as I saw fit. After that I put the shim back on the northbridge and everything was happy.

Did a Furmark run at ambient 77F. Noticed that my GPU would "hover" at 86C longer than it did before... yet the fan would not kick in. At the exact moment I hit 87C did the fans kick in, lowering me to 83C (This isn't a quality result, since I only ran the test for about five minutes; will do better tests later).

I'm wondering if software (namely a MS update, like SP2) could somehow, in one way or another, effect the threshold at which the fan speeds up. I'm thinking it's a likely no, but I'm really puzzled. Software-wise, the only significant change between now and a week ago was installing SP2. Hardware wise I have more copper on the heatsinks and a swiss-cheesed bottom panel. And the RAM heatsink on the top rack.


And since you still appear to be online, Soviet, quick question: When you replaced your MX-2 with MX-3 (I'm assuming you had your MX-2 on for at least a month or so) was it hardended and less "pasty" when pushed from the tube? I'm thinking that, instead of magically evil software interference, I'm expecting results when I should allow it some cure time.

And to bother you again, did you lower your throttle and 3D clocks as well?

 

Software updates have absolutely no effect on fan thresholds. It is probably the additional modding at the hardware level that is benefiting your temps.

MX-3 squeezed out of the tube similar to MX-2. It wasn't too pasty until it was exposed to oxygen. Carbon is less inert than aluminium oxide, and the composition of MX-3 seems to use a bit more carbon with the silicone base compared to MX-2, which is the reason why MX-3 is chalkier prior to mounting.

 

Sorry. What I meant to ask was if your paste had "hardened"/changed/what have you after being on the computer for, say, a month. I'm thinking the reason I'm getting "worse" max temps is because when I first ran my tests, the MX-3 had had a month to cure, and in my current tests its freshy applied.

And I might have edited after you typed. Did you underclock throttle and 3D?

 

Oh, that's what you meant. Yes, MX-2 hardens up after a good week. MX-3 does take longer to cure than MX-2, but the results are pretty decent after a few days of use. A month of curing time for any non-silver composition paste is unnecessary. A week should be more than enough providing that you break it in with fun.

My throttle and 3D clocks are mildly overclocked.

 

All right, thanks. I'll stick with stock throttle and 3D, and try a 20% underclock on 2D. Much appreciated.

EDIT: Woah. With .75 2D voltage, 160/320/96 clocks (stock was 200/400/120), and ambient 75F, my GPU is idling at a nice even 48C with occasional dips in 47C. Zalman off.

Let me turn it on and see what happens.

EDIT 2: Slept. Woke up and turned on computer. From a cold boot and ambient 70F, I hit the usual 87C max, but then after the fans kicked in, it dropped to 81C and stayed that way for a minute or so. I'll run a better test later. Time for work.

 

I got a package attempt delivery on my door today, maybe it's the copper?

 

I won't be a package. It'll be a slightly padded envelope.

 

Basically a package. http://www.google.com/search?hl=en&source=hp&q=define:package&btnG=Google+Search&aq=f&oq=&aqi=

 

Btw what paste do you recommed, i ordered some ICD7, but i think not silver based, you think would perform better than MX3?

 

http://forum.notebookreview.com/showpost.php?p=5185488&postcount=47

 

Silver-based thermal pastes are so 2003.

I will say this, though: Arctic Silver 5 was by far the easier of the two pastes to spread. Granted, if you messed up applying AS5 you could short out your components or not get any thermal improvement.

 

Well i ordered one ICD7 but i yet to see how hard it will be to apply, but if its good ill take my time into it, hoping to never do it again.

My temps went down some, seems the AS5 is getting a little better, back to idling at 50C (at least at night), and in wow max i saw was 79C. I do think that 186.81 drivers were not allowing the GPU to downcloak at 2D, and got also 300points extra on 3dmark with 186.42, so im sticking with them.

Btw Forge did u get the last batch of copper sinks, and if so how many of each should i order and how much tape to cover all stuff?

 

I have 23 inside my computer (I somehow lost one of them). Depending on how I'm doing right now, I might add some more near the fan and most likely some on the bottom RAM stick once I get my final heatspreader.

If you know what you're doing, two strips of tape is more than enough.

 

Alright thanks for the info, ill order 24 then. Im assuming its all the small ones no large ones.

@Soviet
Thanks for the info, after reading your post went to browse reviews and stuff over the net, and although there are different results, ICD7 seems a good paste, although already outperform by Indigo Xtreme and Shin-Etsu X23-7783D. I also ordered an MX3 yesterday will compare both compunds when i get them, the other 2 i cant buy since they are from resellers that dont take international cards so ICD7 or MX3 will have to do.

 

I've applied it, an it was not hard. Didn't put it in a cup of hot water as some people suggested. and it was still fine to apply.
I applied it by making a pea (maybe bit smaller) size blob in the middle of a core and then placed the heatsink onto it (with no spreading).

 

Indigo Xtreme is made for a desktop CPU with a large IHS. It won't work with mobile CPU's. Shin-Etsu X23-7783D also works better on a large IHS and not so much on an exposed die. ICD7 or MX-3 will do the job just as well. But if you want to waste you money, be my guest.

 

Alright, thx for the explanation. Btw with MX3 i was checking more reviews directly to it, and seems to be really good, just a lot of comments that its way to thick, and that artic cooling recommends the old bead style let it spread by preasure of the heatsink, but with my GPU not touching fully do u think this is the best way to apply or should i just do the old style of plastic bag/finger spreading the paste so its all around the DIE and more on the farthest not so good contact part?

 

Anyway, pictures of the bottom plate:

Angle 1

And Soviet will not catch them all

 

You drilled from the top didn't you. I can see the sandpaper scuffs leftover when you were sanding off the leftover plastic hanging off the holes.

Nevertheless, nice job on the holes. You can expect good numbers to come out of that setup now.

In more fateful news. http://forum.notebookreview.com/showpost.php?p=5333993&postcount=746

 

I drilled from the inside going out. Had to, since there were ridges and things on the inside.

And I'm still scratching my head at the temperatures. I cannot seem to stabilize like I did before.

 

Is it possible to remove that fake vent hole? I don't care if there's a non filtered hole there as long as I can remove it. It wouldn't be hard to replace it with a regular fan grill + filter.

 

A Dremel and a router bit can take it off nice and easy.

 

In that case I'll wait on the spare part to come back into stock and take a dremel to that.

 

Hopefully it does. Took a while for the bracket to come back, so I hope the sait for this isn't as long.

I'm going to replace the copper shim, this time with slightly more MX-3 than I used last time, and check out the thermal pads.


EDIT: Replaced shim on the GPU. Going to play a bit of Crysis Warhead to heat up the GPU and entertain myself a bit. Going to run Furmark after.

 

Quick question: Is it the fins on the copper blocks that help reduce the temps, or just the added heat capacity on the heat sink??

 

Both. The fins give more surface area to dissipate heat, and the mass of the copper block sucks away heat.

 

So what if instead of using copper blocks, i was a sheet of copper and shape it to suit the heatsink for extra capacity.. Should that work as good as the blocks?

 

the whole point of the fins on the copper blocks is the increased surface area across which cooling occurs (i.e. the surface area of the fins). Putting a sheet of copper on there may work but it will not yield the results you may be expecting. In my opinion there isn't any point of adding copper sheets. I added some copper blocks (and will be adding more). My gpu temp sits at about 51C and maxs out at 87-88 under full load with furmark.

 

ok cool, I have a G50v-A2..the 9700m GT one...My temps would hit 93 degrees depending on room temp..89 degrees would be average...

But then i dusted out the fan with compressed Air and Cotton swabs and got a Zalman Nc 2000, now im maxed out at 76 Degrees...Would i see any improvement with copper blocks or is 76 as sweet as it gets?

 

What is the temp when you are idling? IMO, 76C is excellent if you are playing video games. Did you try running Furmark to see what it would max out at?

 

i dont know off hand what my idle temps are...i'm in Uni atm so dont have my G50v with me, will post them later (approx 2hours time)..

 

Forge,

Wow, with all those holes you can now play a game a battleship while on the go.....B4.

 

Have you guys given any thought to using a Y cable fan splitter + an additional low profile fan bolted to the fake grill on the inside (after it's been dremeled) to add active air intake? The main fan header can power both fans and this should in theory dramatically cool the system down. Once those G50 back panels are back in stock and I go home in a few weeks, I'm going to order a fan, Y cable + panel and give it a shot.

edit: if there isn't room near the fake air intake, is there enough clearance near the GPU? Based on photos I've seen it seems like there would be enough room for a low profile fan to fit in there and be bolted to the back plate and a hole could be cut out to allow the air intake to occur.

 

I tried plugging the Y cable splitter but its too big for the notebook 4 pin power (made for desktop). I also tried adding a 10mm high fan just above the GPU, but it didn't work without major modding tot he backpanel. But I think that a 10x10 fan above the gpu + copper blocks, should significanlty lower the GPU temps. We just need to wait for ASUS to restock the backpanels as 5150Joker said.

 

I'm so incompetent with hardware that I wouldn't dare do that.

More like Grill-4.

Anyway, the aim of the many holes was to simulate the removal of the backcover without actually removing the backcover. I get impact protection and less dust, but still a really sexy airflow.