My First Build Underwater

Associate
Joined
11 Nov 2008
Posts
6
Hi guys and gals

My first post here on the forum though I have been stalking it for a few weeks trying to pick up as much info as I can for you lot.

So far I have managed to build myself a gaming rig with the following spec:

Antec P182
Intel Q6600 Energy Efficient @ 3.0GHz (9x 333FSB)
Artic Cooling Freezer 7 Pro
Abit IX38 Quad GT Mainboard
2gb Corsair Dominator DDR2 - 8500
Gainward GeForce 9800GT 1GB
Corsair 650W Modular PSU
2 x 250gb Samsung Spinpoint HDD's

Currently temps are not to good (I think) with my processor between 40-45c idle with a 1.38 Vcore.

I want to try and get the Q6600 around the 3.6GHz mark and thats got me looking into water cooling.

Components wise, for the loop I think i've got it all sussed. a rough spec for this will be -

XSPC Delta V3 775 CPU block
Thermochill PA120.3 radiator
Laing DDC Ultra with XSPC Top
XSPC Bay mounted reservoir

and then all the other bits and pieces.

Finally here is where I am seeking your advice!!

1)If I were to add a VGA and NB blocks to the loop will the thermochill rad be up to the job??

2)I can see no 9800GT VGA blocks on the web. Will any 9800 GTX full cover block be compatible with my card??

3)Suggestions on a NB block for my mobo and what passive cooler could I use on the SB seeing as I will be removing the current one??

4)Everyone seems to have differing opinions on loop order. Will we finally, once and for all, be able to establish what is the best??

5)How do you go about bleeding the water cooling loop to remove any air bubbles etc?? Any experience from other people using the XSPC Bay reservoir??


Thanks in advance for all your help guys and I hope I haven't ranted on to much in what seems a helluva big post for my introduction lol.

Joey B
 

RJC

RJC

Don
Joined
29 May 2005
Posts
29,009
Location
Kent
Regards your temps, what software are you using to monitor and what are your full load temps.

At idle your temps look fine.
 
Associate
OP
Joined
11 Nov 2008
Posts
6
That would be CPU-Z and Speedfan

I've got Prime 95 installed. how long should I look at running it to take full load temps????????


Ohh and "claimed"??? what do you mean by that lol
 

RJC

RJC

Don
Joined
29 May 2005
Posts
29,009
Location
Kent
Download Core temp and Real temp to double check Speed fan, this software will read the actual core diodes of each core.

Max temp idealy around the 65c - 70c~ at full load.
 
Associate
OP
Joined
11 Nov 2008
Posts
6
Yup Core Temp is giving roughly the same as Speedfan, low 40's.

And for rodo the VID is 1.3V. Is that right I'm unsure on what the VID affects?

Going to Prime test it on Blend setting for 15min ( I assume that long enough for load temps) and i'll let you know the results to see where I will need to go from there to achieve my 3.6GHz target.

EDIT: Prime 95 was run for 41 min to give these temps

30GHzFullLoad.jpg
 
Last edited:
Associate
Joined
19 Oct 2002
Posts
1,308
Location
Northumberland
Get real temp and follow the calibration instructions to the t. This will give your true core temps to work from then go from there. Also 1.36V looks a bit too much to get 3Ghz to me? I'd look to keep reducing them volts. With my 9650 i worked from the low end and worked my way up one notch at a time until it finally ran stable in prime. I have it at 3.6 with only around 1.2v. I always find what the minimum stable volts are for a given OC. I'm not one for pushing a CPU with overvolting. I like to get the most out of a CPU i can on stock volts if i can. This helps with temps and lifespan of the CPU. So.. Get Real temp, calibrate it, then find the minimum volts for desired OC as this will help with the temps the most.
 
Last edited:
Associate
Joined
21 Jan 2005
Posts
762
Location
Manchester
Hi guys and gals

My first post here on the forum though I have been stalking it for a few weeks trying to pick up as much info as I can for you lot.

So far I have managed to build myself a gaming rig with the following spec:

Antec P182
Intel Q6600 Energy Efficient @ 3.0GHz (9x 333FSB)
Artic Cooling Freezer 7 Pro
Abit IX38 Quad GT Mainboard
2gb Corsair Dominator DDR2 - 8500
Gainward GeForce 9800GT 1GB
Corsair 650W Modular PSU
2 x 250gb Samsung Spinpoint HDD's

Currently temps are not to good (I think) with my processor between 40-45c idle with a 1.38 Vcore.

I want to try and get the Q6600 around the 3.6GHz mark and thats got me looking into water cooling.

Components wise, for the loop I think i've got it all sussed. a rough spec for this will be -

XSPC Delta V3 775 CPU block
Thermochill PA120.3 radiator
Laing DDC Ultra with XSPC Top
XSPC Bay mounted reservoir

and then all the other bits and pieces.

Finally here is where I am seeking your advice!!

1)If I were to add a VGA and NB blocks to the loop will the thermochill rad be up to the job??

Yes, the 120.3 can take 220w heat

2)I can see no 9800GT VGA blocks on the web. Will any 9800 GTX full cover block be compatible with my card??

As far as I am aware they will fit

3)Suggestions on a NB block for my mobo and what passive cooler could I use on the SB seeing as I will be removing the current one??

EK or Swiftech are both good, ZAlman passive cooler

4)Everyone seems to have differing opinions on loop order. Will we finally, once and for all, be able to establish what is the best??

RAD, RES, PUMP, CPU, NB, GPU, RAD

5)How do you go about bleeding the water cooling loop to remove any air bubbles etc?? Any experience from other people using the XSPC Bay reservoir??

Fill it up with liquid, switch it on, top up as needed, gently rock the computer to get bubbles out of rad.


Thanks in advance for all your help guys and I hope I haven't ranted on to much in what seems a helluva big post for my introduction lol.

Joey B
 
Associate
Joined
21 Jan 2005
Posts
762
Location
Manchester
Agree with all of the above except for loop order. I think the general consenus is that it doesn't really matter. Just do what's easiest.

Not true, max flow rate is after pump, most cpu blocks are impingment blocks, therefore, work best after pump.

Res or T bar creates disturbance in the fluid and works best when not behind the rad, as this reduces rad efficiency.

CPU is generally the most overclocked component in the loop (but not always) and should be first in the block queue for maximum efficiency.
 
Soldato
Joined
7 Feb 2004
Posts
8,114
Location
North East
Not true, max flow rate is after pump, most cpu blocks are impingment blocks, therefore, work best after pump.

Res or T bar creates disturbance in the fluid and works best when not behind the rad, as this reduces rad efficiency.

CPU is generally the most overclocked component in the loop (but not always) and should be first in the block queue for maximum efficiency.

Given the loop is closed, the flow rate is constant throughout the loop :)

Very few blocks are impingement any more due to the multiple dies.
 
Associate
Joined
21 Jan 2005
Posts
762
Location
Manchester
Care to provide a source to backup your theory. I'll happily be proven wrong.

I wont go into the thermodynamics but needless to say I am a sceptic and have come to my conclusions by trial and error. I have been watercooling computers for quite a while, in fact years. There are one or two people that I respect in this field and so I do not tread on any toes at Overclockers I will name them and you can do the rest. Petras and Marci. You can google as you see fit.

I have had many WC loops, have tried many, many combinations, have tried countless, blocks, rads and tubes as a hobby and I can tell you, these guys know their onions.

My conclusion is this, if you look at your own loop, try changing it around and see what happens because I guarantee you all things are not equal. It is not possible to have a driving force (pump) then add restrictions, then have a driving force (pump) = constant. Race car hits the straight, max speed, hits first chicane, next bend, next bend, next bend, hits the straight again. Yes, constant laptime, no constant speed.
 
Soldato
Joined
25 Oct 2007
Posts
6,911
Location
Los Angeles
Fair enough mate - i agree with Brad with regards to res and pump order but i am of the opinion that if i was to put my CPU before my NB in my primary loop (see sig) the difference would be relatively negligible. I.e.: 1-2 degrees. I guess it's down to personal preference as to whether you want to be that anal about temps.

I'm also pretty sure (correct me if i'm wrong) that flow rate / heat dump is not linear in the sense that the higher the flow rate the more heat you dump. I'm pretty sure that once you get to a certain rate within your loop the difference in rate with regards to your temps is again negligible. I'm not sure what this value is but I'm positive i read it somewhere on martins liquid lab. There are also other factors to take into account such as the efficiency of your block and radiator.
 
Last edited:
Back
Top Bottom