Overclock.net banner

1 - 11 of 11 Posts

·
Registered
Joined
·
1,044 Posts
Discussion Starter #1
I'm planning on upgrading my current water cooling setup. Here are my current specs

MCP-355 Pump with Petra's Tech shop Delrin top
7/16" Masterkleer tubing
Koolance GTX285 FC water block
Aquaextreme MP-05LE CPU water block
Swiftech Mini Res
Swiftech MCR-360 Radiator
Yate Loon DSL-Medium 120X25mm fans

I'm cooling

Q6600 G0 in an EVGA 780i
EVGA GTX285 Vanilla

Issues:
  • One I'm having is my Radiator (and presumably my water as well) get's pretty hot after long gaming sessions. I'm dealing with a fairly high ambient temp in my computer room due to our lack of central A/C, if I can get the ambient down the rad will cool down as well.
  • My q6600 loads about 12-15c hotter on cores 0 and 1 with IBT and about 5-7c hotter on P95 than cores 2 and 3. I seem to remember reading that the water block I have has a cooling pin area optimized for Dual core CPUs.... It definitely was purchased before the q6600 became so common. is there a better CPU block I should consider? This block did not come with a back plate. I've been wondering if a back plate would help me get a more solid mount. If I'm upgrading to a new block I want something I'll be able to get new hold down plates for. The Aquaextreme is discontinued as far as I know so I'll need a new block if I change sockets.
  • I have 3 San Ace H1101 120X38mm fans to mount to my radiator and I thought about using some 120X25MM fans gutted out as a shroud. Will a shroud still help my temps if I have the fans mounted in a pull configuration?
  • One last possibly stupid question. Anyone think running a push pull config with the San Ace fans in Pull config and the Yate Loons in a Push config would be a problem? I'm considering hooking them both up to the rad on separate fan controllers so I can run the San Aces in the summer and the Yate Loons in the winter and shut the other set off when not in use. anyone think one set of fans would restrict the others airflow enough to cause problems? anyone tried this before? I'm also wondering if it would make noise when the air passes through the turned off fan.
Thanks I know this is a bunch of questions. I appreciate any help.
 

·
Registered
Joined
·
417 Posts
First of all, that block sucks.

Second, if you have that much of a a temp difference between the cores, it's a TIM issue. A lot of people like to make 2 lines of TIM down each die, some like to make an X, some just like to but a dot over each die. I like the dot method. This shows you where the dies are.



Make sure the block doesn't shift around as you are tightening it- and tighten it evenly in a cross pattern like you would lug nuts. You definitely want a backplate, it keeps the motherboard from bending and makes the mount solid.

Shrouds always help. Martin has done some testing. http://martin.skinneelabs.com/Radiat...ng-Review.html
 

·
Registered
Joined
·
1,376 Posts
Yes shrouds will help. I also recommend an updated block. I Always recommend the HeatKiller 3.0. Any one but my choice is the CU or LT. Identical minus the tops.

You push/pull is fine. Your YL are feeding air to your pulling San Aces. That should be fine. The Sans work great in a pull config. Hope this helps. Ask more if you need more ideas.
 

·
Registered
Joined
·
1,044 Posts
Discussion Starter #4
Quote:

Originally Posted by TopFuel1471 View Post
First of all, that block sucks.
I kinda thought that was part of the issue. it worked GREAT with my old E6600. never saw load temps above 50c. Once I upgraded to the quad my temps have been around 65C loaded with Intel Burn Test (linpack)
more around 60C with p95.

On my TIM application I've been doing a line across the cores, I was thinking of trying using this hard plastic tool I used on my GPU to spread out TIM. Basically cover the whole top with an even thin layer and see if it helps.

The block lacks a backplate but it also uses springs on the mounting screws. Should I maybe get some plastic spacers to replace the springs for a more solid mount? I'm wondering if my crap temps are all related to not getting enough pressure to squeeze the TIM out thin enough.

I have pulled the block several times and it always looks like it's a pretty even spread across the cores but it's tough to say if it's thin enough.

What do you guys think? Should I find a backplate and some spacers to try and get a better mount or just scrap the block and find something better with a back plate?

You mention the Heat Killer 3.0, is that something I can easily upgrade to support new sockets with a new hold down plate? Any idea if that company will be around to support their block with new plates? I like to buy blocks and then use them for a few generations if at all possible.

Also... Backplates... are they something you can buy separate from the block or are they usually a package deal?

EDIT: ok so I answered my own question about the heatkiller 3.0... and WOW .. that is one nice looking block.

I also found their back plates so what I think I might do in this order:

Try spreading the TIM Very thin on my CPU with a plastic tool
Get a heatkiller back plate
Check to see if my mounting kit for my MP-05 is M4 thread, Get M4 screws and thumb screws if not
Add some spacers to replace the springs in the mounting mechanism

If that doesn't work I will already have a backplate and I can order a heatkiller 3.0. 80$ is a little steep but it looks like a really good block for the money.
 

·
Registered
Joined
·
558 Posts
Quote:


Originally Posted by citat3962
View Post

[*] My q6600 loads about 12-15c hotter on cores 0 and 1 with IBT and about 5-7c hotter on P95 than cores 2 and 3.

I have the same issue (albeit not as severe, about 10c). I am not sure about the cause of yours, but mine is not a TIM related problem because it remains constant after reapplying time and time again. Sometimes that stuff just happnes, nothing can be fabricated within exact tolerances, which can cause differences in performance/efficiency.
 

·
Registered
Joined
·
1,044 Posts
Discussion Starter #6
I'm starting to think my sensors may be out of whack because I have reapplied TIM a few times and it's always the same differential of temp between the cores at load.. at idle it's more like 3-5c....
 

·
Registered
Joined
·
1,376 Posts
Buying a block wont drastically change your temps. I dont want you buying it thinking your going to see a huge drop. All it will do is help. Its a good flowing block, and does get great temps. Lowest in the water block round up done by martinm210. It is expensive, but if your not wanting to spend the $80, get the LT HeatKiller, it does just as good and is about $20 cheaper.

If your getting such major differences in temps from one side of the block to the other, try taking a razor blade to the bottom of your block and make sure its flat. You might need a small lapping. Depending on how old it is, and all.
 

·
Registered
Joined
·
1,044 Posts
Discussion Starter #8
The block is about 3 years old. Last night I tried several things.

Rotating the block 180°

Cleaning the surfaces VERY well with Arctic Silver 2 step cleaner and surface purifier (1st time I've used it)

Spread the TIM across the CPU surface in a very thin layer with a thin plastic tool. (first time I've tried this too)

I must have had equal pressure on the original mount because the TIM was thinly spread over the whole CPU when I separated them.

I got it all back together and it took longer for my temps to climb but my temps while gaming are still the same

60-61 Core 0
56-57 Core 1
48-50 Core 2
48-50 Core 3

Pretty close to the exact same temps I was getting before. I really think this is something related to the CPU temp sensors not my waterblock. No doubt I could probably use a newer water block anyway, so I may replace it just to get a quad core oriented block.

I also was looking at the EK Supreme. It looks like it's similar in design to the Heatkiller except it's 40$ instead of 60-80. not sure which I'll get yet. For now the current block will have to do.
 

·
Registered
Joined
·
1,044 Posts
Discussion Starter #10
So I picked up a Heat killer LC and just installed the San Ace fans in push with 25mm shrouds. It's running about 10C cooler across the board but the core 0 is still about 10c hotter than 2 and 3 and about 6 hotter than core 1.

So I had an opportunity to upgrade my CPU to a q9550 so I went for it. I still haven't installed it yet but I'm enjoying the cooling upgrade!

I'm amazed how much air these fans push for so little noise!
 
1 - 11 of 11 Posts
Top