The Design And Optim Of LTCC Filter Based On Implicit Space Mapping

Recent years, with people’s increasing demand of miniaturized mobile devices, Low Temperature Co-fired Ceramic(LTCC), as an excellent packaging technique, becomes more and more popular among engineers. It buries various passive components into its multilayer ceramic substrates, and then co-fired to3 D structure, which is smaller in volume. This technique has been applied to make balun, filters and antennas. This paper mainly describes the design of band-pass filter based on LTCC technique, and also provide the improved implicit space mapping algorithm to solve the issues of time-consuming and optimization direction during the simulation. The main work of this paper is as follows:1 Based on the Chebyshev low-pass filter prototype, the band-pass filter lumped circuit is obtained through frequency conversion, component transformation and J/K inverted converter. Then the four-layer LTCC band-pass filter is accomplished with the theory of transmission zeros and transmission line.2 Has done some research on space mapping algorithm, especially the implicit space mapping. Provide the method of frequency parameter extraction to solve the nonuniqueness during the parameter extraction process. A preliminary verification of this algorithm is made through an example of parallel coupled microstrip filter3 Use the improved algorithm exploiting frequency parameter extraction to iteratively optimize the model, of which the fine model is constructed in HFSS and the coarse model is constructed in ADS. And finally obtain the band-pass filter that satisfies the design specifications. We also make a comparison to initial implicit space mapping, the frequency implicit space mapping converges after 5 iterations and the initial implicit space mapping after 7 iterations. The overall size of the filter is 2.5mm×2.0mm×0.9mm. Its centralfrequency is 2.45 GHz, with the bandwidth of 100 MHz. It can be applied to Bluetooth and Wireless Local Area(WLAN) network.The simulation results demonstrate that the frequency implicit space mapping converges faster, and both algorithms can solve the problem of time-consuming and optimization direction better than the traditional approach of parameter sweeping.