Deionized water cooled PC

[Conor21]

Probably a stupid idea but is it possible to submerge a PC in deionized water to cool it. I know the water slowly ionizes but could you work against that if you added a deionizing filter system to the whole thing. In combination with a sub zero cooling system it would solve all the condensation issues, add to the cooling and would be extremely beneficial for sustained loads due to the amount of water having massive thermal capacity. In my head this seems like the ultimate cooling solution. 

Please explain why this wouldn't work...

[8tg]

It would have no passive redundancy, mineral oil PCs stay mostly non conductive. If that deionizer fails and you don’t notice your water just slowly becomes conductive.

Pairing it with a sub zero cooling system is also probably not a great idea as it’ll still form ice if you aren’t warming the water and ice can cause other issues even if it’s non conductive, see ice forming in cracks in the sidewalk as an example.

 

Submerged PCs are mostly done for show, they have their own limitations regardless of how you go about them, mineral oil, 3M novec, whatever you’re using. A difficulty with most of them is still getting the heat out of the liquid, but most of the conventional ones follow the same trend of being passively non conductive, not requiring any upkeep to remain non conductive. Deionized water required active maintenance to keep it non conductive and that’s fairly risky.

 

As we’ll even submerged cooling has its limits for hot spots. Note that 3M novec based PCs still use basically water block contact plates and mineral oil PCs use air coolers to disperse heat into moving liquid. A submerged water PC would have to work the same or at least have some consistent flow going on if you were to attach a phase change cooler to it as everything else would just be sitting in slowly moving water, only moving because of temperature changes.

 

Tldr this would work, but weigh your costs and risks vs reward. Submerged PCs are done as a flex, not for any practical purpose.

[Conor21]

3 minutes ago, 8tg said:

It would have no passive redundancy, mineral oil PCs stay mostly non conductive. If that deionizer fails and you don’t notice your water just slowly becomes conductive.

Pairing it with a sub zero cooling system is also probably not a great idea as it’ll still form ice if you aren’t warming the water and ice can cause other issues even if it’s non conductive, see ice forming in cracks in the sidewalk as an example.

 

Submerged PCs are mostly done for show, they have their own limitations regardless of how you go about them, mineral oil, 3M novec, whatever you’re using. A difficulty with most of them is still getting the heat out of the liquid, but most of the conventional ones follow the same trend of being passively non conductive, not requiring any upkeep to remain non conductive. Deionized water required active maintenance to keep it non conductive and that’s fairly risky.

 

As we’ll even submerged cooling has its limits for hot spots. Note that 3M novec based PCs still use basically water block contact plates and mineral oil PCs use air coolers to disperse heat into moving liquid. A submerged water PC would have to work the same or at least have some consistent flow going on if you were to attach a phase change cooler to it as everything else would just be sitting in slowly moving water, only moving because of temperature changes.

 

Tldr this would work, but weigh your costs and risks vs reward. Submerged PCs are done as a flex, not for any practical purpose.

Thanks for explaining :))