A Frequency Transformation for Co-Designed Multi-Passband Multi-Embedded-Notch RF Filters
Abstract : A class of multi-resonant RF filtering stage that exhibits a multi-band bandpass filtering transfer function with embedded in-band notches is presented. It is derived from the application of a composed lowpass-to-multi-passband/multi-stopband frequency transformation that converts the normalized lowpass prototype into the proposed multi-passband/multi-embedded-notch filtering cell. Thus, when employed in higher-order filtering networks, high-selectivity multi-band bandpass-filter (BPF) counterparts with embedded in-band stopbands are synthesized. Furthermore, two different approaches for out-of-band transmission-zero (TZ) generation in these type of filters to achieve sharp-rejection characteristics into them are detailed. The operational foundations of the engineered multi-passband/multi-embedded-notch filtering stage are described, along with several illustrative first-to-multi-order theoretical design examples with/without stopband TZs. Moreover, for experimental-demonstration purposes, a proof-of-concept microstrip prototype of a third-order dual-band BPF with two and one embedded notches in its lower and upper transmission bands, respectively, is manufactured and characterized.
? Therefore, for supporting the generation of uppaal cora models, we also use an existing uppaal metamodel developed at the University of Paderborn.
? The models conforming to three metamodels explained earlier, are transformed to uppaal cora models automatically via model transformations in the framework.
? We reuse the existing SDF model of the application without making any modifications.
? Using the framework, we derive the energy-optimal schedule on the new hardware platform, for time per graph iteration constraint of 650 ms.
? To alleviate these issues, we propose a notch filter. The blockers should ideally see a short circuit at the input of the filter structures.
? The main drawback of those structure comes from the required input signal impedance. In the previous demonstrations, for the technique to be most effective, the driving source should have a high impedance characteristic, or equivalently the input signal should be in current domain.
? Up to now it seems this idea pushed the issue of inductor design from the RF domain to DC. However, there is a key difference.
? This issue has been addressed in the measurement by adjusting the phase of the positive and negative filter’s clocks relative to each other.
• In this paper, we have presented a model-driven framework for HW-SW codesign of dataflow applications. In our framework, we have proposed a reusable set of three coherent metamodels for HW-SW co-design domain.
• To provide interoperability among domains, we have defined a reusable set of extensible model transformations.
• We have already described our method of using priced timed-automata for the purpose of energy optimization in, and therefore out of scope of this paper. Rather the novelty of this paper is the design prospects of using MDE.
• Rather than Java, which is a general-purpose language, we use etl that is specifically designed as a domainspecific language for model transformations.
? Technology scaling reduces the available voltage headroom for the CMOS blocks, hence degrades their linearity performance.
? A high power blocker causes fluctuations at the RF side of the filter and degrades its performance through the nonlinearity of the switches.
? The architecture is compatible with scaling as the performance improves with switch performance .
? In the present prototype, the blocker signal is used directly to drive the notch filter, while in an actual system an injection locked oscillator can be used to automatically track and lock to the blocker.
? The blocker is used to drive the gates of the switches in the filter as well. By looking at the output spectrum, the desired signal as well as the attenuation of the blocker is evident.
|