Why doesn't anyone consider cooling servers and CPUs with a cryogenic cooler that uses the heat generated by the...
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
servers/CPUs and produces cooling on the spot -- i.e., within the racks?
Cryocoolers, or cryogenic coolers, are heat engines based on the principles of Stirling engines; they've been used for decades in many industries. A Stirling-engine-based cryogenic cooler is a theoretical option for the data center, but they need to be large enough to be effective. And with increasing densities of equipment such large CPU-mounted systems are not acceptable.
Cryocoolers, however, can be useful for in-cabinet or in-row cooling. But the problem is being able to route enough hot air to the cryocooler to run it effectively, and then use the cold junction generated to create a closed-loop cooling system. This would require more powered fans and ducting to ensure air-flows are working adequately.
A power-free type of Stirling engine system needs a temperature differential to work effectively; ambient temperatures of 77 degrees Fahrenheit (25 degrees Celsius), for instance, will require a warmer air temperature to drive the cycle that will stress the components of a data center environment.
Applying power to drive the compressor provides a standard refrigerant approach, as with the Cold Head StirlingCool-10, a two-part assembly that needs 400 W to operate at full capability and is rated for cooling a 10-W thermal load to 80 degrees Kelvin. Server chips have higher thermal output than this (around 80 to 150 W), but don't need as much cooling.
Stirling Cryogenics -- the spin-off from Philips N.V. that began manufacturing cryogenic coolers in 1960 -- focuses on liquefaction and other low-temperature systems; it doesn't seem to have any capability for data center cooling. At the moment, I would say that the costs of implementing a Stirling-engine-based cryocooler system would be far too high and with little overall benefit for a data center.
About the author:
Clive Longbottom is the co-founder and service director of IT research and analysis firm Quocirca, based in the U.K. Longbottom has more than 15 years of experience in the field. With a background in chemical engineering, he's worked on automation, control of hazardous substances, document management and knowledge management projects.
Dig Deeper on Data center design and facilities
Related Q&A from Clive Longbottom
Our small colocation provider's other customers have been under DDoS attacks that affected our infrastructure. The colo null routes the bad IP ...continue reading
Should facilities and IT teams agree on one cohesive DCIM tool for everything in the data center? Are DCIM vendors offering that?continue reading
What tools can I use to manage capacity to accommodate further VM growth in a complex IT environment?continue reading
Have a question for an expert?
Please add a title for your question
Get answers from a TechTarget expert on whatever's puzzling you.