Advantages of Quasi-Optical Grids (Content was taken out of the Chapter 2 in J.C. Chiao's Ph.D. Thesis.Reference numbers were deleted.)

A grid of many planar devices quasi-optically coupled in free space does not require the construction of single-mode waveguides and can potentially overcome the power limits of conventional single/dual-diode multipliers. Rutledge and Schwarz first demonstrated the idea of integrating solid-state devices into a periodic grid as a multimode detector array. The approach of using quasi-optical grids to combine power provides many advantages and these advantages become more attractive when the frequencies increase:

(1) Higher power. By integrating a large number of devices together into the grid, very large powers can be achieved. Because the input power is distributed onto the entire grid, each device only handles certain amount of input power which will increase the total saturation power and the dynamic range. To increase the power, we simply increase the number of devices and the size of the grid. (2) Higher frequency. Quasi-optical system does not require the construction of single-mode waveguides or single-mode resonators, therefore, the limitation of operating wavelengths due to the machining difficulties is eliminated. Also there is no transmission line like microstrip or other planar waveguides needed, the limitation of the roll-off frequencies or higher-mode cut-on frequencies then disappears. (3) Less loss. Because the power is combined in free space, losses associated with waveguide walls and feed networks are eliminated. Conduction loss, dielectric loss and radiation loss which limit the operating frequencies using traditional planar-waveguide components will be reduced in a quasi-optical system. (4)Easier tuning. Quasi-optical systems use mirrors, dielectric slabs, metal-patterned grids or Fabry-Perot interferometers for tuning the impedances, it is easier than for most of the power combiners based on microstrip lines or waveguides. The impedance tuning is independent of the active grid itself, there is no need to change the design or re-fabricate the active devices or lumped elements on the circuits when tuning the performance of the entire system. The tuning elements are easier to be replaced or redesigned without changing the active grids.

(5) Simpler analysis. Although quasi-optical circuits look more like optical devices, they can be modelled with reasonable accuracy using simple transmission-line and lumped-element components by assuming unit cells in a large array. This provides two major advantages. First, unit cells decide the performance of the whole grid. Design changes such as power requirements or operating frequencies can be done by simply changing the number of the unit cells or scaling of the unit cells, unlike microstrip circuits or waveguides which require a new iteration of analysis and layout. Second, simple transmission-line models allow the well-developed commercial computer-aided design tools for conventional microwave circuits such as Hewlett-Packard Microwave and RF Design Systems (MDS) and (Spice) to be applied to the design and modelling of quasi-optical circuits. Also the unit-cell feature makes it possible to use the EMF technique, which was developed by Eisenhart and Khan to derive the impedance of a waveguide mounting structure and later extended by Dr. Weikle , to determine the characteristics of the grids. Conventional numerical techniques such as the finite-element methods or conjugate gradient methods which use an iterative approach require large amounts of computer time and memory. This method reduces the computing time and effort greatly.

(6) Monolithic. Either active or passive quasi-optical circuits are suitable for monolithic manufacture using planar photolithographic techniques and existing semiconductor fabrication technologies. Also no transmission lines or waveguides are used and that makes fabrication simpler and cost less.

(7) Failure tolerance. Intuitively, we expect a large number of devices in an array provides a better failure tolerance. Kim shows that a 10% failure of devices in a grid-amplifier only causes a 1-dB drop in output power. This means that grids could be more reliable than a single-device system.

Created by J.C. Chiao