DEVELOPMENT OF A METHODOLOGY AND SAMPLE PREPARATION TO EVALUATE RELIABILITY OF OIL COOLED DATA CENTER

Abstract

Liquid immersion cooling is not a new concept which is slowly emerging as an alternative solution to conventional chilled air cooling systems for data centers. The main idea is to build a body of knowledge to help industry make more informed regarding mechanical reliability of IT equipment in mineral oil immersed systems. The goal is to develop a testing procedure for evaluating the reliability of electronic packages and components when immersed in mineral oil. PCB’s, chip packages, passive components like capacitors, inductors, resistors, diodes, cables and wiring form the most important parts of a server used in data centers. The main focus is on the PCBs, chip packages, PVC jackets and passive components used in a server. The strategy is to evaluate the change in material properties (young’s modulus, coefficient of thermal expansion (CTE)) and some electrical properties (capacitance, resistance, voltage drop, etc) which determine the mechanical and electrical performance of PCB’S and passive devices respectively after cycling in oil in comparison to control samples in air. Dog-bone shaped samples of PCB’s obtained from open compute servers and also the samples of passive devices and cables generally used in data centers are prepared. Accelerated Thermal Cycle (ATC) test based on ‘ATC JEDEC’ standards which is used for air cooling has some limitations, and as there is a huge difference in a ramp rate of air and oil, this cannot be used for oil cooling. The degradation mechanism of the electronic components of server, PCB’s, and passive devices in particular due to immersion in oil will be investigated. The test samples will be placed in air as well as aged in mineral oil. This methodology explains the analysis of the change in the electrical and mechanical performance in passive devices and PCB’S respectively , compared with that of their performance before cycling.