Circling back around to the intro, the coolants included in the roundup cover quite a range of features or selling points. We have the classic distilled water and a simple biocide, vibrant colors, earth friendly, anti-corrosive, dye ready and Nano-particle coolants represented. Each possess something unique that captures the interest of liquid coolers looking to make their build stand out, simplify maintenance or bragging rights on using the latest tech in fluids. Coolants also come with their own set of drawbacks though, dyes and colors stain as well as eventually break down, and glycol tends to show signs of breakdown as well.
While coolant testing appears to be a simple and straightforward endeavor, that is just not reality. Due to the large number of marketing claims there are many different tests both short-term and long-term that would be required to bust or confirm all of the marketing claims. For those reasons, we decided to look purely at thermal performance as stated in the intro of this roundup.
Removing all the noise and getting back to performance, the primary ingredient in every coolant is water, which is why Minnesota Tap water was added into the roundup. The interesting outcome of testing all of these coolants was just how close thermal performance really is, even I expected to see more of a resultant temperature separation amongst the fluids. However, thinking more about each fluid and the fact that water is the primary ingredient in them all the results are not surprising. With our average CPU Core temperatures, we see a spread of 1C and average GPU temperatures spread of 0.7C when we exclude the FluidXP. I exclude FluidXP Nano-fluid from the equation due to the fact that performance was quite awful.
So where does this leave us… In terms of thermal performance, there is a such a small difference between any of the fluids that coolant choice will not make any difference in thermal performance for your setup. Additionally, we can take the previous statement one-step further and apply it to overclocking and with certainty say that coolant choice does not influence your clocks either. Knowing full well what the potential issues are with any fluid that contains additives, yes, I am referring to tube staining and chemical breakdown among others, your coolant choice really comes down to whatever you want to put in your loop and price you are willing to pay.
For me, I go through so much fluid in the lab that cheap distilled water and PT_nuke is my best choice, even Minnesota tap is used on the pump test bench. For those of you who do not have distilled water readily available, Feser Ultra Pure comes to mind… just know you are more than likely buying water from China that was shipped to Germany, then to your favorite reseller only to be shipped to your door. If color is your thing, you have a wide selection and can make your choice based on which coolant best matches the colors in your build. If you are looking for a coolant that has a bit more science and technology rolled in, Ice Dragon Nanofluid is available and has had colors tests with Mayhem dyes to show it can be made into any color you wish if the white is not the look you are after. Same goes for distilled water as well, Mayhem dyes can be added to achieve almost any color if you cannot find a pre-mix that suits your color palette. Just remember, all dyes/color will breakdown and fade over time. Personally, I would avoid the FluidXP Nano-Fluid though as the performance is abysmal. Just to reiterate, thermal performance differences amongst the coolants is very small, hopefully this puts the age-old question to rest.
Before we close the chapter on Coolant Performance, we need to extend our gratitude to Gigabyte for providing a motherboard for system level testing and FrozenCPU, Thermochill and Ice Dragon Cooling for sending coolants over to the lab for testing. Without the help and support of these companies, this test would not have been possible, thanks again! With that said, put the cork back in the bottle, thermal performance testing of coolants is complete.