Search

This article presents a multiple-die-embedded glass package that supports a thermal management solution for millimeter-wave (mmWave) applications. The package includes dies with different thicknesses embedded into isolated cavities created on a single glass substrate. In the package, partial cavities are used for embedding thin and low-power chips as well as passive components, while through cavities are used for thick and high-power chips. Connection to the embedded dies is facilitated through via transitions and transmission lines. Thermal management is integrated on the backside of the dies embedded in through cavities. This article discusses the design of interconnects along with the fabrication process for the package. The measured package broadband interconnect loss is less than 1 dB/mm from 40 MHz to 80 GHz, with a panel warpage of $145~\mu $ m at $30~^{\circ }$ C. A cross section of the package is also presented. The multidie-embedded glass package exhibits low-loss broadband performance and the ability to integrate thermal solutions, suggesting significant potential for module-level mmWave applications.

This article demonstrates ultra-compact, high-performance filters with footprint smaller than ${0.602}\lambda _{0} \times 0.125\lambda _{0}$ , operating in 39-GHz frequency band, on ultra-thin alumina ribbon ceramic (ARC) substrate, a newly developed material technology at Corning Inc. Bandpass filters (BPFs) with three topologies, i.e.,edge-coupled, hairpin, and interdigital, together with quasi-lumped elliptical low-pass filters (LPFs) are designed on 80- $\mu \text{m}$ -thick ARC substrates and fabricated using a semi-additive patterning (SAP) process. Characterized from 14 to 50 GHz using a vector network analyzer (VNA), filters on ARC exhibit excellent model-to-hardware correlation and outstanding performance with respect to passband insertion loss (IL), return loss, selectivity, and out-of-band rejection. Moreover, they depict a very flat group delay within the passband with a maximum variation up to 0.058 ns. Showing an IL less than 1.43- and 0.95-dB at the cut-off frequency and mid-band frequency, respectively, LPFs and BPFs demonstrate a size reduction of up to 93% and 86%, respectively, compared to their counterparts on other substrates, while outperforming them as well. These highly miniaturized, integrated filters with high $Q$ -factor are best-suited for RF front-end modules (FEMs) and mobile handsets in 5G and millimeter-wave (mm-wave) frequency ranges, indicating the potential of ARC for such applications.

Neural networks (NNs) are quite attractive in creating surrogate models for many signal integrity (SI) applications. NN-based surrogate models offer the benefits of reducing the design cycle time and providing the designer with a quick prototype that can efficiently analyze the performance of the SI task. This article, therefore, proposes a new end-to-end learning approach for surrogate modeling using complex-valued NNs, incorporating higher functionality and better representation. This approach introduces a deep complex dense network ( $\mathbb {C}$ DNet), which is built with complex dense blocks to support complex operations using complex-valued weights, and a physically consistent layer to enforce passivity and causality constraints. We also present a robust inverse multiobjective optimization method to minimize the modeling error and optimize the design space parameters. The results show that our model outperforms state-of-the-art deep surrogate models when tasked with forward and inverse learning for a relatively small amount of data. The effectiveness of the proposed approach is demonstrated through two SI design applications, where the model is used to predict broadband $S$ -parameters and obtain optimal design space parameters given the desired target specifications.