Sign in with
Sign up | Sign in
Your question

Fractal R2 watercooled [pics inside]

Last response: in Overclocking
Share
April 17, 2012 1:03:11 PM

Some days ago I started on watercooling a Fractal R2 PC, this is my 2nd computer to be watercooled and I'm much more happy with how smooth things went and how neat it is compared to my Haf 932 watercooled.

I was originaly planning on having the radiator internally but having it externa allows for slower fan and noise so having the heat get outside of case.

I did not get cpu block because it is not really necessary to watercool this cpu.
The air cooler is massive but I was able to nicely reroute tubing, with cpu block it would look more spacy ofcourse but also needs more pump power.

The pump and shoggy sandwich is not attached anyway to case, instead it is hold in place by the tubing, hopefuly this works OK.

Ive yet been able to fill it with water because I haven't figured out how to get my pump working, worst case scenarion I need to order a brand new pump and wait 1-2weeks.







a b à CPUs
a c 78 K Overclocking
April 17, 2012 4:43:45 PM

[:lutfij:3]
i think you have one of the cheapest external rad mount advice on Tom's :)  nice work with the case - is it primed and running or are you yet to fill it with liquid?

and have you gone past the temp issue you had when running triple crossfire?

build looks nice, but cablemanagement can do better - sorry modo i said before you :p 
April 17, 2012 5:04:18 PM

Thanks for the kind words Lutfij :) 

Yeah the external rad mount was cheap, I think the pictures tell pretty much how it's done (1x M3 metal male/female spacer combination + 3washers and 2 nuts to make sure rad is not punctured). I used 8spacers but I think 4 might be enough depending how much space is needed from PC case.

I did not look into the tripple crossfire bridge anymore, it took alot of time, energy and pain so I'll put that project for some other time, I either did not install it properly or it's deffective.

Related resources
April 17, 2012 5:06:09 PM

I did not get to test it yet as I have difficulties powering on the OEM Laing DDC-1t but the seller is supporting me and I think should eventually figure out the correct wiring for this molex.
a b à CPUs
a c 78 K Overclocking
April 17, 2012 5:15:41 PM

forget the OEM pump and get a MCP35x or Alpahcool VPP single...or that same family of pumps... :)  would like to see this thing in action and finally breathing some life after all that trouble you've gone.

* youre welcome!
a b à CPUs
a b K Overclocking
April 18, 2012 5:37:42 AM

is that the same system jest without the bridge and 1 card missing?

what is the Y for in this loop?
April 18, 2012 6:13:43 AM

No, I should have stated earlier, It's 2 different PC's, they both have same model motherboard.

The Y is instead of reservoir, so I screw in a long tube that I use to fill the loop, then once filled and bled I would remove the long tube and seal the short Y fill tubing.
a b à CPUs
a b K Overclocking
April 18, 2012 7:23:13 AM

a single tube for filling and bleeding dose that work well for you, I use a 2 tube system for filling and bleeding here is a look.

http://img204.imageshack.us/img204/9879/fillsystem002.j...

I use a make shift res. like a 2 litter container with this to fill and bleed.

by the way you have done a nice job on your builds, mine don't look so hot. but they are functional guess looks aren't what I worry about when I'm building them, too preoccupied with how to make it more efficient first. but when I get it working the way I won't, I never go back and clean it up. the experimental radiator project, I'm currently working on is a first for me, I am trying to make it look nice and function right at the same time. are you still wonting to try that pump, or is the company sending you a pin out for that pump, so that you can wire it correctly.
April 18, 2012 11:21:45 AM

This is actually my 2nd PC watercooled ever and I have yet to test how the single tube works with the Y-fitting, I have looked others build and seems bleeding it might take longer vs reservoir, but some users report it to be easy to fill/bleed, I'm hoping it will go easy.

Thanks for the compliments on my builds, I choose rotary fittings this time around which I found easier to work with, obviously there is bigger risk of leaks.
a b à CPUs
a c 78 K Overclocking
April 18, 2012 11:42:40 AM

leaks will not occur if you're careful and patient with the setup. far too often its manhandling that leads to all the bad words :p 
a b à CPUs
a b K Overclocking
April 18, 2012 11:49:15 AM

well you could build the double tube set up with compression fittings and bits power T's and shut off valves, but that is expensive
April 18, 2012 12:21:00 PM

Do you guys think if I screw in a 1.5meter tubing to fill tubing and then fill it completely full and let the pump fill the loop and radiator full of water is a good idea?

Obviously the tubing is thin so the air might be difficult to have get out of the fill tubing compared to reservoir which collects air.

Also I still did not figure out pump :( 
a b à CPUs
a c 78 K Overclocking
April 18, 2012 1:36:03 PM

so no res?
April 19, 2012 7:24:50 AM

Lutfij said:
so no res?


correct, this build has no reservoire.
a b à CPUs
a c 78 K Overclocking
April 19, 2012 12:57:38 PM

well , then the compromise is that it'll take longer to fill/bleed system as opposed to having a res in the loop. A better placement of the y joint will help with the bleeding.
April 19, 2012 1:12:04 PM

I wonder why I'm not allowed to edit my posts here on toms forums?

I finally got the pump running and have filled the loop with water there are some airbubbles still in gpu's, do these disappear if I run 24-48h the system?

Maybe I should keep moving the case to make the airpockets move aswell.


Too long filltubing, wet floor and difiiculties filling:


This is better, I should have had a bit longer fill tubing, now I must have 2 short filltubings attached to each other for easier fill/bleeding:


This bottle turned out to be a life saver, so glad I bought it:


All connections seem to be sealed well:
a b à CPUs
a b K Overclocking
April 19, 2012 1:55:30 PM

ok i understand how 1 line can fill the system but how do you bleed the air out of the system?
a b à CPUs
a c 78 K Overclocking
April 19, 2012 3:10:25 PM

instead of a T, its a Y and the open end on the right of the Y will collect the air as the system is being primed.
April 19, 2012 3:25:22 PM

Any tips how to secure the pump to the decoupling and case so it does not move?

I'm not able to bleed properly because the pump would fall if I rotate case.
a b à CPUs
a c 78 K Overclocking
April 19, 2012 3:43:42 PM

a pump should always be "mounted" not left on its own while bleeding.
April 19, 2012 3:53:02 PM

Well it does not move because the tubing holds it ok.

But I can't bleed because if I rotated the case the pump would move and fall.
a b à CPUs
a c 78 K Overclocking
April 19, 2012 4:28:27 PM

please re-read my post
a b à CPUs
a c 324 K Overclocking
April 19, 2012 4:48:16 PM

I'm not sure why you went this route, but it's a decent idea to have your pump mounted in some way or another. I mentioned before that your sound/vibration dampening idea was far more than what you needed, but it seems that you wanted to move forward with it. No harm in that- testing out new ideas is good, but I was trying to save you the headaches.
a b à CPUs
a b K Overclocking
April 19, 2012 5:18:08 PM

a little double sided tape could solve the problem.
April 20, 2012 8:50:15 AM

toolmaker_03 said:
a little double sided tape could solve the problem.


I think this is the only good way to decouple this oem Laing DDC-1t 10w because there is not screw holes to install it.

I currently put electric tape around the pump, hopefuly the electric tape does not melt from the heat?
a b à CPUs
a c 324 K Overclocking
April 20, 2012 4:05:04 PM

A pump shouldn't get that hot to melt electrical tape. I know DDC's typically run a little warm compared to D5's, but even then, that's scary. I know a lot of people that use heatsinks for the bottom of their DDC's and then elevate the mount via standoffs and screws.
a b à CPUs
a c 78 K Overclocking
April 20, 2012 4:10:54 PM

swiftech had an option for their pumps too :/ 
a b à CPUs
a c 78 K Overclocking
April 20, 2012 4:52:53 PM

ay, ye be right matey!
a b à CPUs
a c 324 K Overclocking
April 20, 2012 6:52:34 PM

You think I'd lie to you guys? :) 
a b à CPUs
a b K Overclocking
April 20, 2012 6:56:46 PM

+ 1 for NO
a b à CPUs
a c 78 K Overclocking
April 20, 2012 7:21:47 PM

rubix_1011 said:
You think I'd lie to you guys? :) 
:lol:  nay matey!
April 24, 2012 5:11:01 PM

The Gpu core temps are 55c and the other gpu 52c.

Looking at Laing specifications the pump has maximum system temperature 60c.

I guess I'm still on the safe side?
a b à CPUs
a c 324 K Overclocking
April 24, 2012 5:51:56 PM

Seems about 10C high to me.
April 25, 2012 1:40:14 PM

rubix_1011 said:
Seems about 10C high to me.


why is it bad to have 55c?
a b à CPUs
a c 324 K Overclocking
April 25, 2012 2:08:01 PM

My GPUs at load never exceed 45C at load for extended periods of time (4-5 hours) using GPU processing in BOINC @100%. This isn't as rough as Furmark or other benching software, but it's considerably more than most games push cards.
April 25, 2012 2:40:22 PM

I think Crysis 2 with DX11 texture or BF3 can push harder then furmark
a b à CPUs
a c 324 K Overclocking
April 25, 2012 2:44:18 PM

My point is that you still shouldn't see temps that high. If your GPUs are pushing 50% or 100%, they really shouldn't be more than 1-2C difference in temps; you aren't increasing the heat load in watts by a significant amount that would cause a difference of 10C.
April 25, 2012 4:36:51 PM

Ah ok I got it.

I think the only component that can take damage is the pump then?
I can't see how 55c harms the tubing or any other part of watercooling?

I'ts a 10w Laing DDC-1t that I'm running at 63% so it should be about 7.5volts or 6.3watts I think.
a b à CPUs
a b K Overclocking
April 25, 2012 4:49:07 PM

i dont think that is what he is trying to say, the problem is the temps are too high, and the way that the blocks are installed might be the problem. I have ugly computers and no air flow through the case, it is basically a passive system. not to mention the fact that the video cards I have, run hot, the 8800 ultra under air cooling ran at idol 135 F under water cooling 113 F. so there is defiantly something wrong here, but what it is I can't say for shore, without knowing how you installed the block and with what material you used for that perpose.
a b à CPUs
a c 78 K Overclocking
April 25, 2012 5:20:01 PM

Quote:
and with what material you used for that perpose.
i had to giggle at this :) 
April 25, 2012 5:23:50 PM

well on air it would be 80c+, so 55c is atleast 25c less compared to air.

Considering the other GPU is 52c I would think the water temperature is 52c or less, so the pump should be safe.
a b à CPUs
a b K Overclocking
April 25, 2012 5:31:04 PM

well your right, if that is what it is on air, then that is about the best you will get on water about 25 F difference, that is a hot card, what is it again?
a b à CPUs
a b K Overclocking
April 25, 2012 5:47:27 PM

so you're talking about load temps then, and not at idol. and the card stabilizes at that temp, well that is pretty good.
a b à CPUs
a c 78 K Overclocking
April 25, 2012 6:03:08 PM

no, the temps rubix was mentioning is that the temps between both cards should be a degree of two each other and not what grand is seeing is also high.
a b à CPUs
a c 324 K Overclocking
April 25, 2012 8:36:35 PM

Water temp isn't the same as the reported GPU or CPU temps.
a b à CPUs
a c 78 K Overclocking
April 25, 2012 10:18:28 PM

true
a b à CPUs
a c 324 K Overclocking
April 26, 2012 3:00:19 AM

It was mainly to address the concerns here:

Quote:
Considering the other GPU is 52c I would think the water temperature is 52c or less, so the pump should be safe.
April 26, 2012 2:08:59 PM

well it stabilized at 54c and 52c so I guess the cards are only 2c different then?

So what is the water temp, I would imagine it being lower then 52c or same so 52c.
It can't be higher then 52c because then the GPU would have to be more then 52c.
!