We all know about submersion mineral oil cooling. One of my favorites is legoman666's beautiful case. but it has its drawbacks: its messy it can damage certain rubbers used on certain boards it dissolves TIM its not very efficient but it does have a few big advantages: its non conductive its naturally anti-microbial and anti-corrosive so i have been wondering: how bad would it do when used in a standard closed-loop cooling system. this would eliminate most of the disadvantages while retaining all of the advantages. googling any combination of "mineral oil" and "pc" only brings up submersed systems, so i decided to try it myself. my guinea pig is an old pentium 4 524 3.06ghz with HT, 1.5gb DDR2 RAM, ATI X1650 512mb GPU. the cooling kit used is a knock-off larkooler "extreme g1/4" which uses a solid copper block, 240 rad, and a pump rated for 1.98gpm or 7.5lpm. i tested using the usual protocol when testing CPU coolers. power system on and let idle temp stabilize for 15 minutes or so. then run 2 workers of prime95 large FFTs until temps reach equilibrium. i tested first with the stock intel cooler, then mineral oil, then water (after flushing the loop). recorded temps are as follows: -stock cooler (auto fan control)- idle: 53c load: 69c -closed loop with mineral oil coolant- idle: 46c load: 65c closed loop with water coolant: idle: 37c load: 45c as you can see, the mineral oil barely beat the stock intel cooler on this non-overclocked chip. i knew it would perform worse than the water, but i had higher hopes for it than this. this unfortunately busts mineral oil as a possible option for a non-conductive, non-corrosive coolant in any performance PC. all in all, a productive and educational saturday night!