Watercooling experiment (based on my system)

Woody...!!!

Woody...!!!

Woody...!!!

Permabanned
Joined
18 Oct 2002
Posts
2,231
Location
Stoke on Trent
Please do not comment on my experiment until sunday 12th of july.

then i would be gratful of your thoughts, i an going to start from the fans i have now, from both fans running, check the temps, then with just 1 fan on and check the temps, then with no fans on and check the temps...

then i will use the same fans with shrouds, then i will use the same fans on there own blow cold air from outside my case through the rad, then try it again with shrouds, and then by some Scythe S-FLEX 1600RPM Ultra Quiet 120mm Fans and try these, all the time reporting my findings with each experiment. also the fans will be running in 7v formation for quietness, plus a fan at 12v as a rule only knocks about 2 to 3c off the heat

the programme's i will use are coretemp and prime, i will be running prime just for 30 mins just to get my load temps to high, not to make the computer stable.

my ambient test room is 7' x 9' and the door is closed, and i have an air condition unit set to 20c i have this running for an hour before testing...

Its not going to be an exact scientific test, just a test to help me choose the best and coolist temp for my set up

my system

gigabyte ex58 extreme
intel I7 920
6gb 16000 ocz blade (3x 2gb)
ek supreme cpu block
Swiftech Laing D5 Vario MCP655 12V DC Pump
Swiftech MCR220 radiator

The tim i am using is Zalman super thermal grease zm-stg1 (i like this stuff because it is so easy to put a thin coat of tim on the cpu with the brush provided)
 
Last edited:
The fans i have now are

Ruilian Science RDM1225s (which came with the swiftech water cooling kit)

Description/Primary features: 120mm class fan, with RPM sensor.
Construction: Low noise sleeve bearing.
Current: .23A
Speed: 2,000 RPM
Static Pressure: 2.7 mmH20
Volume: 81.3 CFM
Audible Noise 36 dB/A @ 12 VDC, 26 dB/A @ 7 VDC


This set up is with both fans on, blowing from inside the case, through the rad, to the outside top of the case...

temps are

Idol 37,37,37,34
load 58,58,58,54
 
Last edited:
ecat said:
:D

10 mins is not long enough to establish the load temp. of a water loop. The results will not be comparable, this is especially true if the pc off time / loop cool down time / base loop temperature varies between runs.

The only temperature you're monitoring is the cpu temperature ? In this case one of the most important considerations is ambient temperature control.
Click to expand...

stop being padantic, lol

so what is a good time an hour
 
Now the same set up but with only 1 fan on the rad, again blowing from the inside of the case through the rad to the outside top of the case.

temps are

idol 40,39,40,37
load 64,64,64,61
 
Last edited:
ecat said:
;)

All thing been equal the test with the most impressive results is always the first of the day, this is when the test/PC starts with the lowest loop temp. I'd say give it at least 30 mins, ymmv. I don't know about CoreTemp but Prime places different loads on the cpu during different tests, always worth keeping an eye on what it's up to.

Ambient is the killer factor in all temperature testing, it's strange how so many tests ignore it. I usually chart cpu, water, ambient and mb temperature for cross checking results, but keeping a note of ambient against each test should suffice.

Happy testing, I love experiments :D
Click to expand...

i have altered my test to 30 mins, and i will be checking all the time, i dont want to break my pc, i just want what is the best for cooling, my ambient test room is 7' x 9' and the door is closed, and i have an air condition unit set to 20c i have this running for an hour before testing...
 
Last edited:
shadowscotland said:
The Air con should help even out the ambient temps and keep ecat happy ;)
And have fun with the testing.
Click to expand...

Its not going to be an exact scientific test, just a test to help me choose the best and coolist temp for my set up, and when i am happy to which is the coolist i want to overclock then
 
This time i am running my water cooling set up without any fans running at all (just my case fans, that will run in all the tests) so the water cooling is running with just the water pump and rad, and i must say i am quite surprised with the idol temps.

idol 48,48,49,45
load 86,86,86,84
 
people were telling me the swiftech rad's were not very good, i must say its holding its own at the moment.

right now for the next stage, time to take the rad out and set the fans to sucking (pulling) in from outside the case through the rad into the closed case.
 
bifday2k said:
nice, as i'll be rebuilding my loop soonish this might come in handy, test if fans on both sides in push/pull is really worth it too if u can with ur set up, if not no worries...
Click to expand...

Well i will try, i am a bit skint at the monent thats why i have to wait till saturday to finish the experiment, i am ashamed to say that is when i get my dole money, so that is when i can buy my fans
 
JonJ678 said:
Half an hour probably not enough time either. I stressed my system on a horizontal 240 feser with no fans. Temperature rose linearly with no signs of cuving over until I stopped it at 75 degrees component temp. It then took a very long time to come back down.

I ran furmark, and watched the line. It'll rise linearly, curve, and approach the horizontal asymptotically. Once I was certain it wasn't going to go up another degree I took that reading. That wasn't scientific either mind :)
Click to expand...

from what i have seen the temps rise for the first 15 mins, and then even off but i keep an eye on things and still run the test for 30 mins, i could be wrong but i dont think the temps are going to alter much if anything after a 30 min period, but again i am not doing an experiment to stress the parts of my pc, just to check the temps of my rad and watercooling through everyday use, and lets face it most of us only use our pc's in everyday use, i even checked the temps playing games last night and they were cooler than when i was testing with prime, so i would say in fact in real life we only use about a quarter of our pc's full potential in everyday use
 
Right then, all changed around, so now i have the air being pulled in from the outside through the rad and into the case, these are the results with just 1 fan running @ 7v

idol 37,37,37,33
load 60,60,60,57

as you can see it is a 4c drop in temp blowing from out side the case
 
This time i have both fans on 7v sucking from the outside and through the rad into my case...

idol 34,34,35,31
load 55,55,55,51

Well it looks like the cpu and rad run cooler sucking from outside the case through the rad into the case, but then the inside of the case gets warmer, but that is easily rectified by either getting a better exhaust fan or in fact adding another exhaust fan so you have 2 fans sucking the heat out of your case in which case this will also help to cool the graphics card and mobo.

now tomorrow i will try sucking from outside the case but this time i will be adding shrouds to the fans (TFC Xtender - Radiator Shroud 120 - GREEN) i must admit these shrouds from feser are really well made and look the part.
 
JonJ678 said:
Too true. Students (like me) can't claim jobseekers either, rough times.

If you run two exhaust at 7V, and two intake at 7V, it wont balance as the exhaust aren't pushing through a radiator. Not sure that affects anything though. Are you sticking with the swiftec after all?
Click to expand...

i am going to put the shrouds on today and i will see what happens, i must say the swiftech radiator is doing its job, and doing it well, just to clarify i have the cpu and northbridge being cooled by my system (gigabyte ex58 extreme has a built in northbridge cooling block)

having said that a double thickness rad can only make things better, and i do like feser products, they just scream quality
 
Bradley8988 said:
Tbh you should be doin this with your CPU at 4Ghz or more so we can really see what it can do.
Click to expand...

Well i have never done much overclocking, and once i have found the best solution for me (which i know now will be sucking from outside the case) then i am going to go for a 3.8ghz overclock
 
hit a little bit of trouble today, the pins that came with the feser shrouds don't fit my rad, so i wil have to wait till tomorrow so i can get some pins...

If OcUK had the reser rads in stock i would have just gone down there and bought one, guess i am lucky that they are only a mile from me.
 
You must log in or register to reply here.
Back
Top Bottom