Radiical Water Chiller Project

Radiical Water Chiller Project

Part Four

The testing completed on the die simulator its is on to real life testing on the computer.

The Test Unit

The test unit is a Abit IG-80 Motherboard with an Intel Prescott 3Ghz CPU.

The Test

The first test is the pull down from ambient. This test involves the computer being switched on with the compressor. In this test no pre cooling of the coolant before switch on is performed. In a real world situation this testing would most likely be the commonest representation of the normal use of the chiller.

As can be seen from the graphic the normal system temperatures remain as additional heat sources inside the case. It is nice to know that temperature inaccuracies occur at both the above and sub zero temperature levels with the onboard Abit motherboard sensors. Below is a picture illustrating the temperature on the underside of the water block to compare with the reported temperature in the web temp graphic.

And the reservoir temperature.

Condensation

It is important to mention to anyone wishing to phase change or as in this case water chill that the major issues you will have stem from condensation. Providing an air tight seal around most socket areas particularly the LGA775 socket is a major headache. It can be done but it is a bit of a pain. Socket areas once sealed should provide high levels of condensation freedom but they are only part of the issue. Tubing even when heavily insulated can and does find a way of transmitting sufficient cooling capacity to drop the air around the tubing to the dew point. Once that occurs it is condensation time. Probably the easiest way of defeating this condensation is with the good old fan. Small 80mm low noise fans work wonders on driving warm air across condensation points . These rapidly drive off any formed condensation. When the computer is switched off there remains a lot of cooler than air coolant left in the chiller lines and this can and does present a different problem.

The cooler water in the tubing tends to continue cooling after the pump and compressor have been switched off. Without the heat from the CPU and with no fans going condensate forms around the CPU area insulation and can drip onto the video card. To stop this occurring it is necessary to devise a different condensation strategy. Enter the condensation tray.

Condensation Tray

The tray simply hooks over the mounting pins for the water block and sits between the water block mounting plate and the CPU socket effectively covering the top surface of the video card.

Condensation Tray in Place

Although the picture is not very clear you can see the plastic tray sitting over the video card stopping any unwanted drips. We insulated inside the socket with dielectric grease and around the socket we used an adhesive backed form of the insulation. A piece was cut to be placed around the LGA775 socket area.

Adhesive backed foam

The outline was traced onto some more sheeting.

Traced copy

Then cut out and another piece was stuck to that building up to three layers of insulation. The water block we used was a Delrin topped version of our SRX water block. It was prepared with the thicker 10mm sheeting and then one adhesive layer to provide a little more protection.

Water block insulation

Finally the mounting plate to attach the water block to the motherboard was fitted to the water block with nylon screws to reduce conducted cooling and onto the motherboard it goes. We conducted two runs of the SRX using 10mm and 12mm tubing. The picture below illustrates the 10mm setup for the SRX block.

Pays to do a couple of trial fittings of the water block and check for errant bits of foam getting in the way of good contact. When all that is done hopefully the condensation will be kept to a minimum.

Results

So how did we do in the final stages of sub zero testing using the 12 mm (1/2") tubing. Highest sub zero was a little over -17 degrees at the block.

and at the reservoir a non to shabby -24 degrees

Conclusions

This first trial of water chilling has raised a few interesting points. One of which is the need for a reservoir at all in the process of water chilling. Reservoirs are bulky and add additional unwanted weight. The advantages are that they supply a storage area for increased pre cooled coolant which places less strain on the compressor switching on and off. The question that needs to be answered is can the water chilling occur in real time using a direct die to water block connection.

Next we will look at water chilling using direct die to water block contact. For that of course we need a direct die block. Enter Radiical's own evaporator.

Next.........Direct Die cooling of a water block