Would like to start Water Cooling - Spec me

[GazD]

As the title.

Well my current rig in sig does everything I ask of it, but I am one to tinker. So I have thought about delving into water cooling.

Would like it to be custom as it would be more fun to fit and the overall satisfaction would be bigger once completed.

I would also like to cool my 640GTS and I may throw a Q6600 into the mix in the future.

I would also like to keep my current case if possible Lian Li PC7 +

Dont really have a budget as I dont know much about this subject but obviously the lower the better without skimping on performance too much.

Thanks in advance!

 

[GazD]

Anyone got any help at all?

 

[Matt2889]

http://forums.overclockers.co.uk/showthread.php?t=17793700

 

[CyZMyAssBaby]

You may find weescotts' article helpful and with links http://forums.overclockers.co.uk/sho...php?t=17598033

The thread will give you an insight.

I would highly recommend a D-Tek Fuzion cpu bloc, if you decide to upgrade to a quad core in the future this block will suffice .

D-Tek Fuzion flows significantly better than any other CPU block giving it a distinct advantage in multi-block loop performance

http://www.xtremesystems.org/forums/showthread.php?t=131297

http://www.xtremesystems.org/forums/showpost.php?p=1949570&postcount=84


The D-Tek Fuzsion nozzle accelerator kit has just become available as well for this block, included is the quad core core nozzle accelerator

Clear Acrylic blocks i avoid as they are prone to cracking around the barb area.


The radiator most probably recommended by forum members would be probably a ThermoChill PA120.3 http://www.overclockers.co.uk/showpr...=57&subcat=193.

Rads with less dense cores such as the Thermochill PA series get away with lower speed / quieter fans For example yate loon.


The shroud to accompany this radiator i would also recommend as it eliminates fan dead spots over the rad.


As for hose size/ barbs would be a personal preference either 3/8, or 1/2. There is no difference regarding performance between the two.

As for the pump you could consider a Laing DDC 12v pro water tank (volume 110ml). The res is attached to the pump (110ml). Space saving and removes the need to place a res in a bay or o/s the case.

Regarding tubing i would use Tygon as it is laboratory grade tubing. It is flexible, clear and also kink resistant. Due to its laboratory grade film deposits do not build up and line the inside of the tubing.


Other members will, im sure have further input.

You will find the links in weescotts' guide helpfull with the graphic card blocks helpful as well.

As to fluid i use Feser1 all though there are others on the the market.

 

[GazD]

SNIP

Thanks for your reply, I now know what hardware I am to be looking at, not sure if I could fit a 120.3 rad in my case, I was looking at the 120.2.

 

[Trems]

I'm currently looking at water cooling my PC7 mate, the 120.3 will fit bit you have to subtly mod the case...will post some pics of mine when its done

My only concern is wether or not to fit a resevoir...some say its not needed but i'm thinking i do have a spare 5.25 bay and the more mass of water in the system the better the cooling capacity...and regardless of the system water cooling isnt cheap so i might aswell do it right!

 

[simonnance]

snip

I'd love to see that build.

I'm looking to fit a 120.3 to my PC-A71 eventualy, they are huge rads (430m long?) so i'd really be supprised if you can fit one inside a PC7 without loosing the PSU. A 120.2 might fit the top of the PC7 easier , and you could hang a 120.1 or 120.2 off the back maybe, or a 120.1 behind the front 120mm fan?

A 120.2 might handle the Q6600 and 640GTS ok, and if the temps get highish you could add a 120.1 or 120.2 to the loop later.

As for the res argument, most pump/res combos compromise on flow IMO, and bay reses look good but due to their thin nature arent tha great for bleeding (too easy for air to get sucked back into the loop). Either go for a Swiftech MicroRes, or just go for a Fillport and ignore the res. I'm going for a Fillport only aproach, getting a nice little hole drilled into the top so it looks nice and sleek.

 

[Marci]

the more mass of water in the system the better the cooling capacity...

Incorrect. More mass only has any effect when you get to the point that you have a 4 litre reservoir. Below that, total amount of coolant has no effect whatsoever.

Cooling Capacity is determined via radiator, airflow thru the radiator, and liquid flow thru the radiator. Not amount of water in the system.

 

[WJA96]

Incorrect. More mass only has any effect when you get to the point that you have a 4 litre reservoir. Below that, total amount of coolant has no effect whatsoever.

Is there a quick reason why 4l is the magic cut-off? eg. Something to do with it being a multiple of the number of volumes of a water-cooled system or something?

 

[Trems]

Thats tripe mate, seriously.

Its all about thermodynamics the more of a boundry you have (i.e surface of a water) the more heat exchange you get. Granted the heat exchanger itself "radiator" is the most efficient point to reject heat from the coolant due to the massive increace in boundry area due to the fins but the more water in the system the more natural convection/conduction you will get...

I'll put it in simple terms...go grab one of your pipes or your res...does it feel warm? of course it does, thats heat energy escaping, thus improving your cooling capacity. Granted its not a huge amount but its helping and a res gives a large surface area for this to occur (make the res from copper/silver or even ally and you will seriously increace this effect....better still ad fins and a fan to the res!

God knows where you get 4L from???


I'll give you another example...

Go put 1 litre of water in your kettle and turn it on, time how long it takes to boil.

Now put 2 litres in, time it again...i promise it will take more than twice as long...

This is simply because you are losing heat constantly through convection and conduction...

 

[Marci]

There was testing done a while ago regarding what capacity reservoir was required to seriously delay equilibrium temps to the point where the effect was noticeable... s'on overclockers.com somewhere from about 4 years ago... below that, there was no visual impact on temps or performance.

 

[tsinc80697]

Dont go buying the Fuzion block yet!

EK has just released a new one that is supposedly better, just waiting for stockists and reviews

 

[WJA96]

Dont go buying the Fuzion block yet!

EK has just released a new one that is supposedly better, just waiting for stockists and reviews

I'm hopeful of getting a Supreme in the next week or so, depends when they arrive at the retailer.

 

[tsinc80697]

Been speaking to one and they are expecting to get them sometime next week and hopefully start posting them out Friday at the earliest.
There were problems with the materials last week so they weren't able to take delivery of any

 

[mnemonix]

The cooling effect of water in the pipes and anywhere in the system but the radiator is so small as to be not worth considering, and your example of a kettle supports Marci's argument not your own - it reaches the exact same temperature eventually (as will your cpu), it just takes longer! You're wasting your time if that's the same Marci who's responsible for Thermochill rads

Thats tripe mate, seriously.

Its all about thermodynamics the more of a boundry you have (i.e surface of a water) the more heat exchange you get. Granted the heat exchanger itself "radiator" is the most efficient point to reject heat from the coolant due to the massive increace in boundry area due to the fins but the more water in the system the more natural convection/conduction you will get...

I'll put it in simple terms...go grab one of your pipes or your res...does it feel warm? of course it does, thats heat energy escaping, thus improving your cooling capacity. Granted its not a huge amount but its helping and a res gives a large surface area for this to occur (make the res from copper/silver or even ally and you will seriously increace this effect....better still ad fins and a fan to the res!

God knows where you get 4L from???


I'll give you another example...

Go put 1 litre of water in your kettle and turn it on, time how long it takes to boil.

Now put 2 litres in, time it again...i promise it will take more than twice as long...

This is simply because you are losing heat constantly through convection and conduction...

 

[simonnance]

Now put 2 litres in, time it again...i promise it will take more than twice as long...

This is simply because you are losing heat constantly through convection and conduction...

That is total tripe....... thermodynamic bull faeces.

The reason 2L of water takes longer to warm up is that specific heat capacity (the amount of energy required to heat something up is measured in J g−1 K−1, thus the more mass (and therefore volume) of water you have, the more energy (in joules) is required to heat it up.

The reason it will take more than 2x as long to heat up is that in order for a set surface area of heating element to heat up 2x the volume of water large amount of water must circulate via convection in order for it all to reach the same temperature. Sme heat is lost by conduction to the container, and convection to the air, but rather low amounts compared to that stored in the water. In a Watercooling system this convection if performed by the pump, rotating the water through the system so that fresh, cool, water is always in contact with the blocks..... the rate at which modern pumps push the water will give very little time for the water to transfer its heat to other water, and in fact the closed system means that after a very short period an equilibrium will be reached when the temperature of water in any resevoir is constant.

Water is useful as a coolant as it has a high ability to absorb energy without large increases in temperature..... and as a result it is equaly hard to REMOVE the energy (heat) from the water..... thus why we use rather big radiators with nice powerful fans, with pumps pushing a LOT of volume through. Unless you have a massive passive radiator with a LOT of surface area, then you arent going to get a lot of cooling benefit.

The cooling effects of the pipes and resevoir will be negligible, and to be honest

 

[marscay]

Dont go buying the Fuzion block yet!

EK has just released a new one that is supposedly better, just waiting for stockists and reviews

it's all hype until they're in the hands of some independant ppl to test.

aesthetically it's a nice looking block, i really think the fuzion is ugly to look at & GTX not much better

 

[CyZMyAssBaby]

Snip

it's all hype until they're in the hands of some independant ppl to test.

Don't believe the hype...I agree must see the evidence first.

 

[WJA96]

it's all hype until they're in the hands of some independant ppl to test.

aesthetically it's a nice looking block, i really think the fuzion is ugly to look at & GTX not much better

Don't believe the hype...I agree must see the evidence first.

I agree totally, and I don't think either tsinc80697 or myself were suggesting anything other than a wait and see policy.

I've ordered one because I need a block and I've tried the FuZion (with and without nozzles), the Apogee GT and GTX and the Cuplex XT Di, so why not try the Supreme too?

The concept is definitely different, and different can be catastrophically disastrous or spectacularly successful. I hope for Eddy's sake that it's a good 'un as anything else will seal him into a box that says "Great GPU Blocks" rather than "Great Water Cooling Blocks".

And, like all of EK's stuff, it looks the dogs danglies...

 

[tsinc80697]

Yeah, definitely worth trying if you've dabbled in other blocks so its something I'm happy to wait for aslong as its sorted by a week or so's time
EK stuff is normally good so unless something has been radically changed without testing I can't see it being any worse than a Fuzion, and its cheaper!