Clevo Overclocker's Lounge

Nope nope, just be very very careful when cleaning the CLU on the heatsink/IHS, its a pain to clean it.

 

I have a special LiquidUltra Cleaning kit that was made specifically to clean CLU. Bought it from Coollaboratory

 

Dang didn't know that even existed.

 

that is like heaven man. I used it before, it melts the living heck out of LCU and makes it very easy to remove

 

If the heatsink fit well on the IHS, I would still use Liquid ultra. I think @Mr. Fox Have tested this well enough to take the right conclusion. Higher Oc and higher temperature will demand a better paste than a conventional paste is. I think the heatsink is the culprit for some with this monster.

 

Yeap, would love his input on this.
Look at this :
http://i.imgur.com/IskYG28.jpg

Do you see how the paste layer is way thicker towards the bottom (Towards the MOSFET/VRM side), and this was a test i did without any thermal pads applied. This gap gets larger with the thermal pads on. CLU barely makes contact towards the bottom side (side of the VRM's/MOSFT's).

I remember from @Mr. Fox 's video that his heatsink was making perfect contact.

@Phoenix has unusually higher temps for CLU on the IHS. Its worth a shot.

 

I have a new extra heatsink that I bought from HIDevolution as a spare. I will use that in my next re-paste just to be on the safe side.

 

@bloodhawk

can you please remind me how to apply IC Diamond? Pea sized dot? Cross Pattern? Or Line?

And how much

I wish you can show me a picture of how much I need.

Sometimes I feel that a pea sized dot isn't enough to cover the entire IHS

 

@Phoenix See one of my latest post in this thread. I'm on phone now. You will find it And bookmark the post.

 

your latest post doesn't have any picture of how much paste I need to put.

Please finish your phone call this is more important

 

I would give a small pea a shot, then screw it down and take it off again, to see if it is making thin even contact.

 

Use a small ball of very soft modeling clay or fresh (soft) Play Doh with a square of plastic wrap on top and bottom, then you can easily see if it is being squashed evenly. Just be sure whatever you use is extra soft and easy to squish. Even fresh cookie dough would work as long as you don't leave any chocolate chips in the way, LOL.

Assuming you use a very soft material, if it is making proper contact it should be almost entirely smashed out the sides with little or nothing left between the layers of plastic wrap.

 

Haha never thought of doing this. But i did try multiple mounts with the small tube that comes with the system, and for me there is barely any contact towards the bottom. (MOSFET/VRM side). Same thing with CLU, there was a very thin line of contact towards the top.

 

Please read this guide and click on all the images https://www.pugetsystems.com/labs/articles/Thermal-Paste-Application-Techniques-170/

 

Yeah i have seen that. But it is applicable in cases where the mounting pressure is much higher and you know that it is even. Also now a days the application doesn't really matter when it comes to desktop coolers, and the difference is barely worth even considering.
For laptops like ours however, since the mounting pressure isnt as much and some people are not getting even mounts it might matter to a certain extent.

 

@bloodhawk See the image from the guide. With small pea in the middle. And compare with your application.


Your application

Can you see the similarity with how the paste behave? Same problem as you. Use X application!!!

 

Edit: If you have the same problem with Liquid Ultra... Use a little more Clu where you see it is lesser of paste in the image. Aka a little more Clu on the place on the IHS with lesser pressure.

 

Nope, not the same. The IHS is making good contact with the Heatsink in the image above ,right in the middle. Our problem is that the contact region is skewed.



If the contact was happening properly in the middle, then the application methods you linked to would apply, at least in my case. Those are being rendered useless, because of the gap towards the bottom.

Also in my case the thermal is spread pretty well towards the top, where the heatsink is making proper contact with the IHS, almost towards the edges. Nice thin layer.

 

I did try more CLU, but only the top portion is making proper contact. Heck i actually used that to confirm that there isnt even contact taking place. Because the bottom part would not even touch the heatsink. Only the top part.

 

Read

Jeah. The sinner is bad fit. Same problem as @Mr. Fox had with a warped heasink on his Panther. Unfortunately so can some get a heasink with inferior contact against IHS.

 

I actually tried my processor on a desktop board with a Corsair H110i mount. And the contact was perfectly even. Im sure it varies from user to user, but at least my IHS is perfectly flat.
Not sure if my seller will cover this under warranty and give me a new heatsink, since it works just fine otherwise.
Or i can just buy an extra one, dont think they are too expensive.

 

just buy a spare one from HIDevolution like I did brah

 

I would anyway bought a heatsink to have in reserve. You never know when you'll need one. When you get this, try it. I would show the picture of your applying of the paste to your seller. He should understand that this heasink is something that should be replaced with a better one.

 

How much did it cost?

 

Yeah ill email them today, lets see if they need me to pay for it lol.

 

I had to replace the CPU heat sink on my Panther (P570WM) for the same reason. The original was deformed. I wonder if you shimmed the back plate on the opposite side of the motherboard so the screw stand-offs were not as "tall" on the CPU side of the motherboard if that would help? I wonder the heat sink is bottoming out on the screw stand-offs and holding the heat sink away from the IHS.

 

Yeap, its as you say, the Heat sink is not bottoming out on the IHS towards the bottom. I can try shimming it down a bit with really fine sandpaper. Will the slightly reduced thickness (.5-.75mm) have any effects on the VRM/MOSFET cooling?

 

so the prema sbios and vbios gave me an awesome boost...i benched 4.5ghz on the cpu and 1353mhz / 3605mhz on the gpu using 3dmark vantage scoring 47001. my personal highest score ever achieved with a notebook and putting me on 2nd place behind fox in the 3dmark rankings (i know, vantage is old, but i just like it ^^)
http://www.3dmark.com/3dmv/5443779

@Prema thanks for your absolutely amazing work as always man!

 

YOU LIKE TO LIVE ON THE EDGE? http://forum.notebookreview.com/thr...eforce-drivers-v364-72-findings-fixes.789930/

 

He wants to turn his GPU into a barbecue sandwich no problem

 

Remember he has warranty on his machine. A good chance that this warranty can be used, LOL

 

@Papusan, when I saw this image, the only thing I could think of is you smiling at me while applying my IC Diamond

 

we all have warranty, doesn't mean we will destroy our hardware with a driver that we know has damaged hundreds of GPUs so we can send our machine for a warranty claim and wait for days or weeks without a system.

this is like playing with fire and asking for trouble

 

Would rather scream . I prefer Liquid Ultra over everything else of thermal paste.

 

@bloodhawk @Papusan

guys, when applying IC Diamond, I decided to go with the cross pattern as the provides the least air bubbled and a better spread. But do I need to put anything on the heatsink itself as was the case with CLU or no need?

 

Never use ordinary paste on heatsink

 

@Papusan

Please check this post, I added a new quote to it which might interest you

 

Nope. Unlike CLU , ICD works out because of the quantity.
BTW. Was your heatsink silver when you wiped CLU off ?

 

initially yes, but after using my magical liquids that I showed you before , it became copper again

 

Wow nice. Gotta get me some of that .

 

indeed. I got it from Coollaboratory themselves although their order handling is super slow and they used the crappiest company to ship them to me causing me a lot of hassle. They didn't arrive at my door but instead, reached our local post office, so I had to go there and wait for hours to get the package despite my paying like 80 euros for shipping alone

never ordering anything from them again

 

Ouch. I'll just grab it off of eBay/amazon/frozencpu.
Yep Amazon it.is -https:// www.amazon.com/gp/aw/d/B0058688KA/ref=mp_s_a_1_8?qid=1460765876&sr=8-8&pi=AC_SX236_SY340_FMwebp_QL65&keywords=coollaboratory&dpPl=1&dpID=41yNfwYhYwL&ref=plSrch
Ezpz

 

oh bloody hell I didn't know they had it there....... ezpz japneezi

 

New better Wprime 32M score 4.242 sec http://hwbot.org/submission/3189584_papusan_wprime___32m_core_i7_6700k_4sec_242ms?recalculate=true
@Phoenix please replace my old wPRIME 32M score with the new one on first page?

 

User Phoenix is currently switched off, please post another reminder after 14 hours

 

Dang. You should keep up the leader board. I bench and you have to put in the score!!! What the hell are you going to keep on with the whole 14 hours? Has your wife demanded something of you? OK... I will not put me up in that theme...

 

@Phoenix You got away extra job . New and better wPRIME 32M score 4.234 sec. I moved up in the wprime 32M ranking at HWBOOT. Finally under 4.240 sec with a good margin My higher Wprime 32M score also increased the global points from Hwboot.
http://hwbot.org/submission/3189608_papusan_wprime___32m_core_i7_6700k_4sec_234ms?recalculate=true

4.234 vs 4.235 sec... Very little margin for a better positioning on the ranking, LOL

 

Working on bringing the 'Clevo-gang' also the GTX860M world-record. This essentially already is No.1 for GTX860M & GTX960M:



It's with the latest 364.96 driver...do they ever validate those non WHQL release results or do I need to use another one?
Still missing 19 points for 5k so will have to redo it anyway...
@johnksss @Mr. Fox
That was 1500Mhz on core 1350Mhz on vRAM (yeah I know it sucks) 3.8Ghz on CPU with Cooler Master 120W AC Adpater!
The stock Clevo 120W tripped with the same settings...

 

They do slowly validate each WHQL set, but no they don't do the non-WHQL as that would take a long time and cost a fair bit of money.