Home
Ask a Question
Secure Orders
Latest News
Reviews
Projects
Customer Gallery
Guides
FAQ
Live Testing

Radiical Water Chiller Project

Part Three

Probably the biggest single challenge in making a water chiller is finding a solution which will do the job of cooling without freezing. Sure you can get anti freeze solutions that will manage down to -30ish but what coolant is there for temperatures lower than that. The answer is the exotics such as alcohol and other more toxic solutions with varying degrees of explosive/combustible characteristics like Acetone.

We have spent a lot of time evaluating solutions for use in this chiller. First we tried FluidXP. It has all the necessary properties. It is non conductive, has an inbuilt corrosion inhibitor and is non toxic. The unfortunate part is it will only go down to about -12 degrees any further and its mush time as evidenced when we tested its freeze point at -15 degrees. FluidXP remains a viable alternative for low temperature cooling because of its non conductive properties.

FluidXP at -15 degrees Celsius

Anti-freeze will take the solution down to say -30 degrees before the freezing point is reached. The major issue with anti-freeze is that at solution strengths necessary to reach say -30 degrees the heat carrying capacity of the mixture is seriously compromised. What is needed for lower than -30 degrees is an alcohol based solution. We tried a variety of mixture strengths of a commonly used solution using Methylated spirits and demineralised water. At a 50% mixture of Methylated spirits the temperature bottomed out at -40 degrees with marked crystallization of the water as it was driven out.

Methylated spirits and demineralised water at 50% solution

From the previous testing we carried out the maximum temperature we were able to achieve was -45 using a 75% solution of Methylated spirits to water. Clearly the need to go lower with a water chiller is really unnecessary. If temperatures below say -40 degrees were needed it is far simpler to go to direct die phase change.

We have decided to stay with the idea of a small amount of anti freeze, water and methylated spirits solution with our project.

Pull Down Speed Graph

There is little point having a chiller if it will not pull down the temperature from ambient to sub ambient and below quickly. We tested pull down speed on our water chiller measuring the time necessary to pull 1 liter of our metho/water solution to below zero and beyond. For those of you interested in chiller response times here is a pull down to maximum sub zero graph using our 50% Methylated spirits and water solution. The interesting point from our perspective is how fast the solution reaches the subzero zone. (The test was carried out with no insulation present so times will be a lot higher)

To move to the next stage it is necessary to digress and look at the equipment we will be using.

Equipment

Before we can go much further into testing it is important to set up the equipment we will be using in the die simulation testing. We have carried out some performance testing of the various pumps at our disposal with a variety of coolants at various solutions. We have tested Eheim 1048 and 1250 plus Eheim's little compact 600 series. Various offerings from Laing including the MCP600 and DDC-12v (MCP350) pumps. In the next stage we will discuss our findings on pumps and tubing.

Pumps and Tubing

Most pumps used in water cooling are derived from aquarium or pond water use. Recently pumps like the Laing DDC-12v have arrived on the scene. Coming from an industrial background this pump is specifically designed for water transport. The question is would it be able to withstand both the very low temperatures and high viscosities of alcohol based solutions. On the first question the answer is no. The magnetically driven pumps like the Laing DDC-12v and MCP600 simply do not cut the mustard when it comes to highly viscous fluids like those found at such very low temperatures. As the temperature drops the viscosity rises the closer you go towards the freezing point. As a result of viscosity changes the pump simply stops pumping.

What is needed is a pump which will withstand the low temps and cope with increasing viscosity. The Eheim range of pumps will do this very well. Built on a long tradition the Eheim pumps are really second to non in durability and rugged construction. We have had our little Eheim 1048 pumps happily chugging away at anything up to -45 degrees Celsius. The same applies for the big brother in the Eheim range the 1250. Both of these pumps are happy to withstand freezing temps and high viscosities for extended periods of time without apparent ill effect.

The other little gem of an Eheim pump is the Compact 600. In the early testing we installed the pump in an insulated stainless steel beaker and pumped sub zero fluid with the pump fully submerged.The pump was able to handle up to minus 20 degrees without failure. It was only when the coolant began to freeze that the pump stopped.

Pump Choice.

For the purpose of our dry run testing of the chiller on the heat die we will be using the Eheim 1048 pump.

Tubing Choice

The other consideration in the type of solution being used is its long term affect on the tubing. Its is highly unlikely that such solutions would bother chemical resist tubing like Tygon, however, the same may not apply to pump seals which is something to keep in mind. Given that Tygon is able to withstand -30 degrees it is, of course, the tubing of choice.

Testing

We will be pitting the chiller against our die simulator for the first range of tests which will involve estimations of pull down speed at load. The first test was to establish what raw power we had available. We ran our first test with no insulation on tubing or block. The die simulator was set to estimate a load of 130watts. Leakage from the un insulated areas of the block and simulator would have been high but we were encouraged by the performance observed. The results were a very healthy -17.1 degrees at the water reservoir and -7.6 at the simulator core.

The test was then run again but on this occasion we swapped out most of the antifreeze and added additional methylated spirits. After some trials we then insulated up the die simulator and tubing and ran the testing afresh with a nearly complete insulation coverage. At this stage we had reached our design goals and then surpassed them.

Water reservoir temperatures were -21 degrees.

and the core temperature was a very stable -18 degrees.

The final test we were keen to perform was the pull down against heat load test. The die simulator is allowed to run up to a starting temperature of 50 degrees Celsius. The chiller is then turned on and the time taken to reach subzero is measured at timed intervals. The resultant graph showing water temperature is below.

In the next phase we will need to make the final design decision about fitting a reservoir water storage area in the main framework of the chiller or move onto direct block to die chilling.

Next...........The real thing....Live Testing.