![]() |
|
Design and Analysis of Optical Interconnects Student: Yew Lian Tan Supervisor: Aleksandar Rakic Category: Engineering Thesis Project - Communications In modern digital machines, requirements for high data channel density and bandwidth result in growing interconnection problems. As much as the designers would like to scale the electronics that would be used to implement a high interconnect density and data rate system, the electrical interconnects are facing physical limitations. The evolvement of free-space optical interconnects (FSOI’s) has known to help relieve these limitations. Since then, FSOI’s have been proposed to be implemented on board-to-board and chip-to-chip levels. Despite the many advantages of incorporating FSOI’s into communication systems, issues like diffraction and misalignment are still some of the major problems in current FSOI technology where they can induce crosstalk and insertion loss; thus the need to limit diffraction and control the amount of misalignments in an optical interconnect architecture. A way to do this is to determine the tolerances of every element used in the architecture by computing the insertion loss induced with each type of misalignment. In this thesis, three FSOI architectures have been considered and they are the macrolens design using plano convex lenses, macrolens design using biconvex lenses and a hybrid design, which includes both biconvex microlenses and plano convex macrolenses. By using Code V, an optical modeling tool, we are able to investigate the insertion loss accrued to the architecture as a result of mechanical misalignments and fabrication errors of the different elements in the architecture. A comparison between the two macrolens designs has also been made and it was found out that the tolerance of the lenses varies for different detector sizes; Generally, the insertion loss is smaller for a larger detector size. Furthermore, the sensitivity to misalignments differs between plano convex lenses and biconvex lenses. The hybrid interconnect architecture is also modeled to show the high sensitivity of microlenses to mechanical misalignments and fabrication errors as compared to macrolenses. The findings thus prove that microlenses are much less tolerant to misalignment than macrolenses.
| ||||||||
| feedback | |
| ©2001 The University of Queensland, Australia | |
| ABN: 63 942 912 684 | |
| Authorised by: Secretary & Registrar | |
| Maintained by: webmasters@itee.uq.edu.au | |
| Last Updated: 2 July 2001 | |