ideas‎ > ‎

Parallel thermoelectric water-cooling loop

Here I consider the possibility of a relatively simple and inexpensive addition to a regular PC water cooling loop. Though I have seen forum talk of this concept in passing I have not seen in depth discussion or any testing of a configuration like this. I plan to investigate the concept further and possibly build a prototype platform.

A thermoelectric (peltier) device is sandwiched between 2 additional water-blocks in the cooling loop for the purpose of pumping heat from the intake to the output of the primary water-block. The result is lower water-block intake temperature and higher radiator intake temperature increasing the effectiveness of both components.

This configuration has the potential to bring some components below ambient temperature and below dew point. At the very least this will need to be monitored and prevented or mitigated.
Temperatures below zero seem unlikely to be possible considering the operating temperature range of available peltier devices and driving both sides of the primary heat load.
There have been projects that placed the cold plate of the peltier directly on the CPU. As CPU power dissipation rates have increased over the last decade this has become far less feasible and unpopular.
There have been projects that placed the cold plate of the peltier prior to the primary water-block and then used a secondary loop or heat-sink on the hot plate. This has generally been successful and can be configured for below dew point or below freezing operation. 

Given the extra heat added to the loop by the peltier device itself, will the improvement in radiator efficiency be enough or will additional surface area be required to prevent the entire loop operating at increased temperature?
Is the primary block ΔT already so great that there isn't sufficient headroom for peltier operation?

Jan 3, 2014, 2:25 AM