My task for two semesters was the mapping of rack systems, and the simulation of some special cases. During my learning of the features of the rack systems I created an own model of a subrack and made different simulations on it.
I begun my work with literature research, where I familiarized myself with the construction and coolsystem of serverrooms in telecommunication and the rack enclosures. Then I collected some information about one of the most important compontent in the circuits, about oscillators, especially the Oven Controlled Crystal Oscillators – OCXOs, mostly concentrated on the termal problems. Because the timing constraints are the most essentials in syncronized circuits, therefore I paid special attention for the simulations of the oscillator. I also got to know the different methods and programs of thermal simulation.
The Department of Electronic Devices has a number of licences for the FloTHERM CFD program which was selected as tool for the simulation. During my work I’ve learnt to use this program, and I made the simulation in this environment. As the tenth version of the program became avaialble after the first semester of my work the model and the settings of the simulation changed a bit in the newest version.
At the beginning of my work I needed a model for the simulations. So I created a subrack from simple parts, which would be the base of the furter simulations. The parts of the board were only cuboids with dissipation and material. However I’ve managed to made accurate steady-state simulations on it. I also changed the different environmental conditions (such as temperature, air flow rate, etc.), and processed the differences.
The main objective of research was to set up the transient simulations. Here the parameters could be changed in time, so the reality is better represented. First, I tried simple simulations with the base model, then I considered more design points and made more calculations with the program. In the end I made conclusions with the comparison of the cases, where I stated that which tested arrangements should be considered during the development of such systems and when the errors originated from thermal behaviour can be decreased.