Installation
Now that the parts have been introduced properly it is time to install the kit. Firstly I attacked towards the pipes and cut them the right size. The operation was easily done with scissors and force. The hardest part was to measure the length so that the WaterChill would be easy enough to install and that any new pipes wouldn't be needed in the future.
Next it was time for the pump and tank. I decided to attach the tank to the pump because it would be easier to move the system this way. It is also possible to place the tank away from the pump. A metal piece and short piece of pipe are needed when attaching the pump and tank. You can see the metal piece in the picture in the middle of the tank and pump. The manual recommends a minimum of 5,5cm piece of pipe but I decided that the piece is way to long for attaching. The metal piece which holds up the tank is attached with a two-sided piece of tape on the bottom of the tank. The other side is shoved between the pump's body and bottom. The attachment works, but isn't stiff at all, and I think that this is an incomplete feature which could be designed again in the next third revision ;)
After finishing with the pump & tank I moved to the blocks. First to find it's place was the CPU-block, which had an exciting situation waiting. There had been compatibility problems with Epox's main boards with the original cover and the block had an 8RDA3+ waiting for testing. Some practise first and the result was expected. The wide cover hit the condensers left so it was time to do some screwing and attach the other cover.
After attaching the other cover the result was right. The narrower cover enables installation for only Socket A -main boards so P4 and AMD64 platforms do not need the smaller cover. Now that the problems had been won, it was time to move to the attaching step.
The block is attached to the main board with screwable metal plugs. Between the plug and main board there is a place for plastic plugs for avoiding short circuits. It is shame that many main boards have tiny surface mounted components placed near the holes. That was the situation with the test platform Epox as well, you can see the problem in the upper picture. The components handle quite a lot of pressure but if you doubt this you can design them better with a knife or scissors. After every plug was attached should the result be something like the picture above.
Now that the plugs are placed it's time to install the block on the processor so that the plugs go through the holes on the cover. After this you add some springs which are for the tightening with big finger screws. The screws are tightened evenly as long as the screws base touches the metal pieces other end. The attachment is innovated, working and very stable. Shortly, the attaching has been well designed.
After the CPU, I moved a few centimetres down to the northbridge. You push long screws through the main board which do the attachment. The block settles on the NB and the screws get a prick-spring-prick add and a plastic finger screw in this order. After this the screw is tightened until the needed tightness has been achieved. It is recommended to use more pi paste on the NB than on the CPU because it is generally uneven and maybe convex or concave. The block's installation kit includes pi paste but this time I used my own heat paste between the block and NB.
Last but not least, the GPU-block which is attached the same way as the previous NB-block. The only demand is that the graphics card should have holes for screws. This time the GPU-block was installed on an Inno3D's Geforce 4 Ti4200 8x and with 64MB memory.
The block's tightening happens with a plastic screw, a couple of plastic pricks and a spring, just as with the NB-block. After installation the pipe outs point to the outer edge so that the stiff pipes are a lot easier to install and use than normally.
Before assembling everything into one circuit it was time to attach the Sunon fan on the radiator. I thought the operation would be fast but it was furious ;) The screws meant for the operation were like wooden screws with "deep rifles", not like normal machinery screws. Besides this, there weren't any rifles in the screw spots on the radiator so that the screw had to be shoved in with force. After all this, the screw was too long, and after screwing it to the end it broke some lamellas from the radiator. I mentioned this to Asetek and they told me that the newer products have better screws.
Now that the fan was attached I noticed another odd thing. Even though the radiator and fan had been installed on a wooden desk, the radiator had lost some paint. So the paint surface is rather bad on the radiator. But anyway, the problem is only cosmetic. I mentioned also this to Asetek and the replied that they hade noticed the same thing and that they had sent some feedback towards the radiator's manufacturer. So there will be improvements on the paint surface's quality.
Now everything was ready and the pipes were installed. I moved the whole "chaos" on the table. The power- and fan cables found their way to the Control Unit. The PSU was modified to work without the rest of the hardware by attaching the ATX-connectors green and black cable with a piece of cable which was included with the package.
The quick connector's action is seen in the lower picture.
Finally it was time to fill the WaterChill with water. Not like in the picture, the circuit should be checked first without any hardware so that possible water damage would be avoided. Airing took about 5 minutes after the pump stopped its noise. To make the air's exiting faster shaking the blocks and radiator will help.
Performance Experiences
The WaterChill is installed and ready for its first real start. In the beginning I switched the fan to blow with 12V and the noise was what I imagined. My personal opinion is that that this kind of noise makes water chilling worthless so 7V is the only reasonable alternative for the usage voltage. The noise with 7V is acceptable even though it isn't completely silent. The noise is however so little that especially when you assemble it in your case it gets covered with the PSU's and hard drives noise.
After a few hour use I happened to admire the classy WaterChill when I noticed that the graphic card had something wrong. It had bent strongly. After a quick investigation I came up with a conclusion. There were two things which caused the bending. The other one was that there weren't any back panel (platform on a table) which would give the card support. The other guilty one was the pipe between the NB & GPU blocks. It is too short and this and its stiffness caused the bending. Later after I detached the block the graphic card recovered to its normal "shape", so any permanent change didn't happen.
A lot of stiff pipes in a small space. A hot-air blower is a handy devise so that the pipes wouldn't bend in bad angles. After all I didn't end up using any device.
They could be family! :) The clear cover makes the WaterChill's blocks classy.
A black graphics card with a WaterChill GPU-block. The view makes me groan. (The translator agrees;)
I have previously worked with only Sicce's pumps so Hydor's Seltz is a new toy for me. Along with this every Nova I have used have been sunken pumps when the WaterChill's Hydor is used in a dry playground. As writing this article in a quiet apartment I noticed that the pump grumbles a bit which can be heard clearly while sitting next to pump. The grumbling is familiar to hard drives noise, which made me blame my hard drive at first :) It is understandable that sunken pumps are nearly noiseless because they are wrapped up with a thick and down toning liquid layer:) Anyway the noise surprised me a bit, but I am looking forward to that the noise will disappear after installing it to the case.
Asetek WaterChill Complete Kit
