I have recently started to watch some videos posted by LinusTechTips, JayzTwoCents and some others - all YouTubers that make videos mostly about computer hardware reviews. As much as I am sometimes entertained, the serious lack of scientific rigour makes me cringe sometimes.

These are not purely entertainment channels. They are supposed to be informative and accurate with some humour sprinkled atop. What is the point of hardware reviews if they contain factual errors or incorrect conclusions due to poor scientific methodology? No matter how much you want to be entertaining, if your main focus is to provide an unbiased opinion on something that is based on (true) facts, then you need to get the science nailed down.

I am going to tackle a specific video today. To best understand my rant, please view the full video (17m13s) first.

Done? Good. Read on.

Lets start with the title: "Are front mounted radiators bad for your PC?". There is NO mention in that title, of any specifics other than the mounting location of the radiator (implied meaning: front mounted radiator taking cold air in and expelling hot air inside the case vs. top mounted radiator that takes warm air in from the case and expel it via the top vent to the outside). Already that should have been made clear as it is only implied, and not stated clearly.

Using a radiator implies using liquid cooling for the components. He explicitly only liquid cooled the CPU - not the GPU (which was air cooled). Although in his conclusion he mentions that including the GPU in the loop might affect it somewhat, he implies that even then it should not be a significant issue as it will only raise the motherboard components' temperature.

My system has both CPU and 2 x GPU liquid cooling. When playing StarCraft - which is not very demanding on CPU or GPU, with a front mounted 240mm radiator I measured an intake temperature of 24.1°C and an outlet temperature of 35.9°C. I measured the outlet temperature not like he did, but rather directly on the inside of the case where the air is expelled from the radiator into the case. This is more accurate. So the temperature differential is 11.8°C.

He measured a temperature difference for intake temperature vs. expelled temperature of 0.39°C when he stressed his CPU to the maximum. Two flaws in his test that I spotted immediately:

1. By measuring the case outlet temperature and not the temperature of the radiator outflow side, he muddied the results by side effects such as differences in air pressure due to thermal differences causing more or less air to be drawn in / expelled from the case from other holes and vents.
2. By not considering the effect when also cooling the GPU, he did not perform a generic test that could apply as widely as the title implies. This test is only relevant for CPU liquid cooling. He did state this in his conclusion, but then his title should have been amended.

So - is it an issue to raise the air temperature by 11.8°C due to a front radiator mount? I'd say absolutely. It is the same for the motherboard AND CPU AND GPU as if the ambient temperature was suddenly 11.8°C warmer. And we all know how servers and PCs react to summertime vs wintertime temperatures. Remember, if the air temperature inside the case is 11.8°C warmer, it is NOT just the motherboard components that will be affected like he stated. The CPU and GPU has each a total temperature of ambient + internally generated. Internally generated stays constant (for this test), but we just raised the ambient. Now there is even more heat to expel from the radiator as it needs to deal with the 11.8°C raise as well - eventually this reaches a steady state but it will definitely cause some oscillation and amplification of temperatures.

Whether this will break your system or not depends on your configuration. But I suspect even with dual 1080 Ti GPUs and an overclocked i7-6850K it will not break anything, however you will definitely not reap the most benefit of the liquid cooling setup. I would strongly recommend against a front mounted radiator (taking cold air from the outside) unless you also have additional radiators expelling air to the outside. In an ideal situation you want to bring cold air in the case with fans, keep the internal case temperature at ambient and expel the liquid's temperature via radiators pushing air out of the case. This will never be achieved but one can approach it. Do take note that the obvious solution - to mount the radiator as an exhaust on top will not work if the case temperature is much higher than ambient - radiators need cool air in otherwise they do not function efficiently.

PS: I know I used an IR temperature meter - but if you remember basic thermodynamics, the air temperature will be in equilibrium with the temperature of its surroundings provided those surroundings do not generate internal heat - such as the blades of the fans.