So, who's up for another installment of jam jar watercooling with a twist? In the twist of that it doesn't actually use a jam jar! This whole article is all theory, and I infact did not put this to practise because of obvious reasons explained later.

As you can see from the diagram, this is basically the same as Jam Jar Mark 1. The difference being this time, it does not require evaporation. The main problem with those being, the CPU would only ever get to a temperature that was similar to that was necessary to make the water evaporate. This time, it works on convection currents.

To prevent loss of transfer as there was last time by using a metal plate, heatsink and jam jar lid, this time the method is different. This system works on a heatsink inset into acrylic sheet walls.

If you can imagine a small 8x8 (cm) sheet of 5mm acyrlic sheet. Cut a square out the middle. Using a heatsink with an inset (i.e the fins don't go to the edge), epoxy this into place. Shown left.

Then of course, epoxy the rest of the sides on. On the left hand side wall, a copper tube is to be inserted - this is an inlet pipe of cold water. This copper pipe goes to the top piece. It coils around many a time before going into the sheet. On the Copper tubing is a fan. The whole thing is filled with water.

The idea is this. The heat from the CPU warms up the heatsink. The now hot water travels upwards and into the copper pipe. As it goes round the coil, the fan makes the copper tube act as a radiator. The now cold water goes down the pipe and into the heatsink to once again be heated up.

There are a few problems to be taken into account when doing this. Firstly - weight. The previous Jam Jar experiments where done on Celerons - these cores are exceedingly tough! My AMD Thunderbird is fragile. It could quite easily chip. Imagine the weight of all this water on it, pulling down. Because remember the cpu is 90degrees in the case. Thats a lot of leverage. If it didn't chip/crack the core - it would make certain bits of it not touch correctly. Secondly, just like the jam jar coolers, it just ain't gonna work well! Well, it might.. But I think air cooling would beat it considerably.

Conclusion

Well, just some pretty pictures to look at basically!

Spode