My small watercooling test and roundup

[specialtech]

This is my slightly unscientific guide to the kits we have and how well they perform. Its worth nothing that this is a very rough test but should give you an idea of what you can expect in various situations

The test

A file copy of 1gb on the same drive to another directory of a video file and then a 2 pass xvid compression test to get things nice and toasty.

Whats being tested

2 x 80mm rads no fans - idle

2 x 80mm rads no fans - load

2 x 80mm with fans on rad - idle

2 x 80mm with fans on rad - load

1 x 120mm rad and fan - idle

1 x 120mm rad and fan - load

1 x 360mm rad no fans - idle

1 x 360mm rad no fans - load

1 x 360mm rad 3 x fans - idle

1 x 360mm rad 3 x fans - load

1 x Waterchiller titan 150 - idle

1 x Waterchiller titan 150 - load

Test Kit

Silverstone LC16 Silver

Amd 4000+ Venice Core

2 x 512 Twinmos Memory

Fanless X700 Ati Card

Terratec Tv Tuner card

Wireless Nic

2 x Maxtor 250gb Hdd's

The Watercooling kit, all alphacool of course . The pump is the Eheim 12v 600 Station square pump, cpu block is the NexXxoS XP light and for fun we also cooled the northbridge as well. I dont have the NB results because i fubared Speedfan but I have the results, just need to work out what one is what.

Test conditions.

After each test the case lid is taken off, everything allowed to stand and then the new kit installed. New kit is then ran for 10 mins to allow the kit to do its cooling and get it to real world levels. Case fans havent been used apart from the first test with the 80mm units. Reason being is that this is to make it as silent as possible. Fans were left in but off to allow a small part of air flow.

No cards were taken out to try and give a more pratical real world result of kit in situation

Variables

The hard drives just wouldnt cool down properly in a short amount of time. The test was done for the whole day with a steady ambient room temp.

Results...

Tests 1 to 4 Results

Wow, fans make a big difference, was expecting that but not so much a difference. The 10 degrees difference from no fans to being fanless on the overall cooling ratio and the 23 degree difference on the hot day that was tested in was no doubt impressive. A bit more tubing and hassle to install but overall, its a working system thats proved it can be used on any case with 2 x 80mm fans at the back

Test 5 and 6

9 Degrees difference on cpu cooling and an overall result of 10 degrees cooling when on a full load isnt bad. Obviously if you have a 120mm rad slot on the back of the case a starter kit with a 120mm rad is more than adquate for a 4000 chip like the one we test with

Test 7 and 8

A lower and more stable idle temp but the load temp was a bit surprising. Amongst the scoring on test the 360 rad is one of the hottest and one of the coolest. This is more than due to the fan and no fan test. Why test like this, to see if a easy, fanless system is viable.

Test 9 and 10

The overall temp was a nice 37 no load and only 41 on load, thus proving that the bigger the rad the better the cooling. There was only a 9 degree difference between the load an no load test on the cpu so plenty of room to upgrade. Downside to this unit was the 3 fans being ran at the same time so it was amongst the nosiest.

Test 11 and 12

Waterchilling. Its cool (pardon the pun) although the most expensive kit here on test and the noisiest (it is a fridge after all) it performed brilliantly.... Almost too brilliantly... see below

Just 23 degrees on load isnt bad. This was at the hottest day of the year and in direct sunlight (we left it testing whilst we went to the pub for a quick one)

So conclusions time.

Fans are the key here but they dont have to be manic mad loud fast fans, something that gets the air circulating is the best thing.

Value for money the 120 kit and just the 80mm kits weigh in quite nicely. What we have show here is the relation between size and performance (sorry couldnt resist). I'll get some pricing up and some more stats on it all to give a better relation of performance v cost

 

[JN]

You just sat there with condensation about to drip off those pipes!!

 

[Ham]

Yeah, nice little test. but that really water? Not like bubble effect piping or anything.

 

[specialtech]

that is 100% water and i was moping the pipes as quick as it was forming. Ive got the high flow tubing in with me now, so as soon as I get the block, im going to big bore a system and cool it

 

[PV5150]

OMG I couldn't just sit there and watch that condensation; and you went to the pub!!!. Electricity and conductive liquids (i.e water don't mix), unfortunately twisted experienced this not so long ago when his wc setup leaked.

 

[Ham]

leaked.

I thought he got cumbersom and knocked off a pipe?

 

[specialtech]

It wasnt intentional lol. I really didnt think that it would get that cool in the system to that sort of level where condensation was rapidly forming. Whats interesting is that the chiller is the basic model thats a pelitier unit. The 250 which uses gas is in later this week. Cant wait to try that out

 

[JN]

Initially i thought the waterchillers were a very good idea, however, if they produce this much condensation and only manage to cool you're cpu to 23oc then it's not really worth it imo.

Unlike phase change, there isn't any real way to stop condensation forming on your hosing (apart from raising the water temperature), which kinda defeats the idea.

Am i missing something?

 

[FragTek]

Wow insane condensation, good thing you didn't fry anything!!

Waterchillers usually pull of MUCH MUCH better temps than this, usually near 0*c. Most chillers are done with phase units and a custom evap coil inside of a large cooler that's tapped for the tubing to come in/out.

 

[specialtech]

Remember that this chiller wasnt the gas cooled one that we have coming in, this is a peltier based cooler. The next chiller up here works on a gas called "R134A" so im expecting a LOT better results from cooling

Also on this test we set the water chiller temp to 10 degrees as, due to the condensation, I really didnt want to push more than that. Its going back in for another bout of testing this week so if you have any ideas then let me know and Ill post them here.

Considering that the max temp that the cpu hit on the day was about 67 degrees approx, is it really that bad? We are hopefully getting some phase change stuff in shortly

If you have tests you want me to run, pop them down here and Ill see what I can do. I would try some clocking but Ive never really got into that side of things yet..

**UPDATE**

Did some digging around the other night and its seems that if your water in the pipes is more than 10 degrees below the room temp then condensation will start to form. Just finding out now the ins and outs of this. Does bring up the question on what you would have to do to the pipes to stop condensation.

Also the price on the chillers is about to drop this week as well. The 250 unit is going to be about 330ukp. Not sure how that compares to a gas unit.

 

[Raven]

Most chillers are done with phase units and a custom evap coil inside of a large cooler that's tapped for the tubing to come in/out.

Fragtek, do you know how good a chiller made for aquariums would work ?

they can be found to atleast 1100w from what i seen and is made to run 24/7

EDIT:

they are not for extrem cooling but down to atleast 15c

 

[PV5150]

We are hopefully getting some phase change stuff in shortly

Awesome, you guys are really looking out for the enthusiasts...props to you :worship: It'll be cool to see what 'chilly' stuff you get in, keep us posted.

Fragtek, do you know how good a chiller made for aquariums would work ?

they can be found to atleast 1100w from what i seen and is made to run 24/7

Crikey 1100W, that's some serious moolah spent on power bills. Especially running it 24/7...

 

[Raven]

Crikey 1100W, that's some serious moolah spent on power bills. Especially running it 24/7...

good cooling costs

their are smaller versions too, down to 150w but those use peltier

 

[specialtech]

Im going to give the Thermaltake silent water a whirl later on today on a slightly different rig to see how it performs.

Ill run vista as a bench test, seems to keep the cpu at 100% most of the time

 

[Counter Terrorist Zombie]

Does bring up the question on what you would have to do to the pipes to stop condensation.

Put the pipe in a pipe, with an exit somewhere for the condensation/water.