Long time ago, when liquid cooling was considered only an expensive hobby for overclocking enthusiasts, a reemerging USA company which has a bitten apple as a logo engineered and manufactured the first (to my knowledge) liquid cooled workstation. The name of the pretty beast was PowerMAC G5, and it’s CPUs where actively cooled by what Apple itself defined a cooling system that “provides a continuous flow of thermally conductive fluid that transfers heat from the processors as they work harder.”

Those times (2004/2005) where pionieristic in the liquid cooling scene, since yet no big manufacturer had started even to think about entering the liquid cooling business. So pioneristic they were, that Apple, among those very few who tried this hard and long way, did make a lot of errors, one of which is still incredibly repeated today, notwithstanding the presence of clear evidence against it.

The error is quite basic: mixing of two chemically incompatible materials inside a sealed loop, and in connection one to each other with an electroconductive means (liquid). What this inevitably brings to is galvanic corrosion, and the final effect is the inesorable, unavoidable corrosion of the weaker and least noble material in favour of the nobler one. In this particolar case, we are talking about, respectively, aluminum anc copper.

Apple did just that: copper blocks to remove the heat from the CPUs, and aluminum radiators. in the same loop, obviously. Apple never admitted this was a problem, they never made a product recall of their Power Macintosh G5/2.5 DP (PCI-X) (June 2004), Power Macintosh G5/2.7 DP (PCI-X) (Early 2005), and Power Macintosh G5 “Quad Core” (2.5) (Late 2005) PCs, but customer forums, dedicated websites and non all reported leaking problems or leaking related problems. Even worse, if not detected in time, this issue could have caused major HW failure, with damaged mainboards, CPUs, and power supplies.

Well, NEC seems too have not learned APPLE’s lesson, and they announced recently their last home theather PC, liquid cooled, naed Valuestar W. While this is another major step in the widspread of liquid cooled computers, looking at the few images and specs of the NEC marvel just sets a “deja vu” mode on inside my human brain. Same intention (putting a liquid cooled computer into mainstream market), same “technology” (small components, crancked up together in a very hard-to-believe small case), same objectives: noiseless computing, cooling performance, higher reliability.

Valuestar W

And same errors: aluminum radiator…

liquid cooling system

aluminum radiator?

Why on earth would a major company on the IT branch like NEC just overlook on this essential component is just a matter of counting one-two-three-money! Well yes, I believe the major point that was brought to the decision table while planning the production of this PC was, indeed, money. Or better, the money they could save using a cheaper quality, aluminum radiator instead of a full copper, or even a brass-copper, a bit more expensive model.

But we know how big companies sometimes (…) only care of profit, and margins. And really dont care about how things should really be made. Or even designed… But I was not expecting that no-one inside NEC had at least once searched in google the terms: “water+cooling+aluminum+copper”

This is epic failure for NEC. Aluminum will corrode in time, no matter what you use as corrosion inhibitor (which does exactly that: it inhibits – i.e. slows down / stop – corrosion, does not prevent it). Its a matter of months, maybe years, ok, but it will happen. And when it does, well, NEC will have to explain why they preferred to spare a few bucks in cheap components rather than preventing a problem everybody in the liquid cooling business knows pretty well, right from the start.

They copy, but never learn…