Kevin Explores Water Chilling – Part I

So let me guess. You saw the title of this article and you got all excited about how you are going to be able to drop the temperature of the water circulating throughout your system. Unfortunately, you can click the "back" button of your browser right now. This was a failure. However, this is only Part I. I have some other ideas and I will be exploring them in the near future. Look at this article as a "how to not chill your water" article.

I had been talking with Dennis from Dangerden for a while via e-mail, testing out some of the Dangerden products and commenting on them. When the topic of water chilling arose, he was kind enough to construct me two copper waterblocks to be used as a water chiller. As you can see in the picture below, the blocks were long with two nylon bolts on either side. The idea was to sandwich peltiers in between them. I would have two completely separate water cooling loops. One would serve the purpose of cooling all the system components. The other would chill the water of the first loop. $90 later I had 3 peltiers in hand. These peltiers are expensive because at 5v, they work at 50w. That is a pretty high wattage for such low voltage peltiers. I sandwiched these babies together with some RTV and got ready for business.

I'm using an Enermax 450w power supply to power these TECs. Xjinn from Pro/Chat suggested that I use a standalone power supply that would give the peltiers a little bit more juice. I figured that I would first see if the chiller worked. If so, I'd move to higher power. I tested it many different ways. However, the way that I wanted it to work in my system would be that water would go into the chiller and come out a lot cooler. I thought it might work considering that when I would run the chiller with no water passing through the cold side, it would look like this.

And within a minute or so, the cold side would get REALLY cold.

Below is a pic of some testing I did a while back. There are no radiators. I just put ice in the water that was circulating through the hot side. I also had water circulating throughout the cold side in a reservoir. Now after a while, the reservoir circulating through the cold side of the blocks got REALLY cold, colder than the ice water. However, I had no temperature measuring equipment at that time.

This test was interesting and gave me a false sense of hope, but this was not how the chiller was to work in my system. I wanted water to go in the chiller and come out colder.

My bathroom has two sinks in it and I figured that I would fill a sink up with water and pump water from one sink to the other. I measured the temperature of the sink water and I measured the temperature of the water coming out of the chiller. Unfortunately, as you can see from the pictures, there was only a .5 degree difference in temperature. 150w of chilling power got me a mere .5 degree change. Whoop-dee-fuckin-doo… The water would go in at 25.7 degrees and come out at 25.2. I tried filling the sink with warmer water, but I still got the same results.

Well, it looks like I learned a lot from this experiment. Water cannot cool itself as well as it can heat itself. When a waterblock is cooling a CPU, the water absorbs the heat so easily. However, when the water passes through a cold source, it basically shrugs it off. I think that the water needs to be exposed to the cold source for a longer period of time in order for it to significantly chill. In other words, I don't think that putting higher powered peltiers on these blocks is the answer. I need to find a new design. I am actually working on one right now (hence the title… Part I). I also learned that this type of chilling may be effective if the water is constantly circulated through the chiller. When I had a reservoir of water circulating through the chiller, the water got quite cold. This did take a long time though.

I'd like to give a big shot out to all of the sponsors that helped to fund this experiment. Dennis from Dangerden was really interested in the idea of a water chiller so he went out of his way to custom make these two blocks for me. Chip from overclock-watercool was nice enough to send me a Big Momma radiator to cool the water on the hotside of the loop if it were to work. Cole from Caseetc sent over a bunch of Panaflos to work with the Big Momma. And of course a huge shot out to the peeps at Coolerguys for hooking me up with the Comp-u-nurses.

As I alluded to this before, my endeavors in water chilling are nowhere near finished. I'm ready to start my next design any day now. But before that, my Kyro II looks like it would be happier with one of these extra 50w peltiers sitting on it ;).