Page 294 - System on Package_ Miniaturization of the Entire System

268 Cha pte r F i v e

render this kind of design approach unpractical. Also, the low-frequency RF packaging
design process often requires scalable equivalent circuits for the package itself. As the
problems associated with the integration involve more and more factors to be considered,
the design and optimization of such systems requires more comprehensive and
sophisticated tools. The current design and optimization methods, using the commercially
available electromagnetic simulators, do not take into account the specific effect of each
of the factors involved in the design process, the degree to which these factors interact
with each other, and their ranges of values. Only this type of thorough understanding
of the entire system can enable the optimization and synthesis of any module under
different given conditions. For example, a combination of design of experiments (DOE)
and response surface methods (RSM) can be implemented [17]. In such cases, first the
factors that affect the performance of the system and the output figures of merit have to
be identified. The next step involves the design of a factorial experiment with center
points based on a design space for these factors to determine the effect of each of the
parameters, identify their interaction, and determine which ones are significant for each
of the outputs. The experiment is run using electromagnetic simulations and/or
microwave measurements, and the outputs are recorded and input into statistical
analysis software. After the statistical analysis of the data, significant factors are
identified for all figures of merit, and then RSM statistical methodology is applied for
optimization. The result is an explicit set of equations that show how the outputs
depend on the input variables, which are used to simultaneously optimize the figures
of merit. Within the design space of the experiment, the optimized figures of merit and
the required design parameters are identified.
The nonlinearity of the system, combined with the lack of analytical input-output
description, suggests the use of soft computing algorithms also. Genetic algorithms can
be utilized as an optimization method of this kind. These algorithms search the
parameter space stochastically generating solutions that are close to the optimal. They
are efficient for problems where small perturbations in the optimal solution lead to an
abrupt increase of the error.
These techniques can be applied to any type of design, especially in complex RF
microsystems and packages where the number of factors increases and it is extremely
difficult to optimize using only electromagnetic simulators. It gives a thorough understanding
of the system behavior and integrates geometrical, material, and functional parameters
altogether. The approach is generic and independent of the choice of the electromagnetic
simulator and statistical analysis software.

5.4.2 RF Substrate Materials Technologies
As mentioned before, central to the theme of the SOP approach is the development of
highly miniaturized systems, novel integrating technologies, and the suitable material
and component technologies with which to integrate. The substrate material platforms
should provide excellent high-frequency electrical properties, mechanical and chemical
resistance, and thin-film multilayer capabilities, and be cost competitive. The prominent
packaging technologies, which can satisfy all these requirements, fall into two
categories—ceramic substrates [18] and organic substrates [19].
The substrate technologies that include ceramics and organics are discussed in Chapter 7
in detail. Ceramics were the primary focus until 2000, but organic technologies began to
provide a combination of low cost and high performance to generate both homogenous and
heterogeneous multilayer SOP architectures [20]. Ceramic substrates include low-temperature