Загрузка...

Adaptive Federated Learning for Industrial Wireless Energy Optimization

🚀 Paper Title: Adaptive Federated Learning for Industrial Wireless Energy Optimization | MWCNT-TSV Crosstalk & Power Reduction

In this video, we explain our research on next-generation 3D IC interconnects using MWCNT-based Through-Silicon Vias (TSVs) with polymer dielectric liners to reduce crosstalk, power consumption, PDP, and EDP. We introduce a novel TSV structure where MWCNT bundles act as conductive elements and polymers like Polyimide, PPC, and BCB serve as insulating liners to improve signal integrity.

Publisher: Jack Sparrow Publishers
Journal : International Journal of Research and Development in Engineering Sciences (IJRDES) , , www.ijrdes.com
e-ISSN: 2582-4201

Paper Link : https://ijrdes.com/paper-view/comparative-analysis-of-advanced-shielding-techniques-for-crosstalk-mitigation-in-multi-walled-carbon-nanotube-tsvs-100651937
DOI : https://doi.org/10.63328/IJRDES-V7RI6P6

Abstract: This paper proposes a novel technique for reducing crossover disturbances, consumption of electricity, Power Delay Product (PDP), and Energy-Delay Product (EDP) in MWCNT-based through-silicon vias (TSVs) with polymer dielectric liners. A new TSV structure is introduced, which uses MWCNT bundles as conductive elements and polymer liners such as polyimide, Polypropylene Carbonate (PPC), and Benzo Cyclobutene (BCB) as insulators. An analogous electrical circuit model is developed to assess crosstalk effects in connected TSVs powered by a ternary inverter. Simulations with the HSPICE tool show that BCB-based TSVs significantly improve performance, especially at lower TSV heights, by up to 30.21% over single-walled carbon nanotube (SWCNT) TSVs. Additionally, BCB-lined TSVs exhibit superior crosstalk suppression, especially at larger TSV pitch values, with a 40.03% enhancement at 0.5 to 3 μm pitch compared to SWCNT TSVs. The study further explores the impact of TSV height and dielectric constant variations, revealing that BCB provides a 43.8% reduction in crosstalk delay over conventional SiO₂ liners. These findings highlight the potential of polymer liners, particularly BCB, in improving signal integrity and overall performance in next-generation 3D ICs, making them a promising alternative to traditional materials.

✅ Key Highlights (Results):

* Developed an analogous electrical circuit model to analyze crosstalk in connected TSVs powered by a ternary inverter
* HSPICE simulations validate performance improvements
* BCB-based TSVs show up to 30.21% performance improvement (especially at lower TSV heights) compared to SWCNT TSVs
* Enhanced crosstalk suppression by 40.03% at 0.5–3 μm TSV pitch vs. SWCNT TSVs
* 43.8% reduction in crosstalk delay compared to conventional SiO₂ liners
* Study includes the impact of TSV height and dielectric constant variations

📌 Why this matters:
These results demonstrate that polymer liners especially BCB can significantly improve signal integrity, energy efficiency, and reliability in advanced 3D ICs, making them promising alternatives to traditional dielectric materials.

🎯 Applications:

* 3D ICs and advanced packaging
* High-speed VLSI interconnects
* Low-power, high-performance computing hardware
* Signal integrity improvement in dense TSV arrays
* Next-gen semiconductor design and research

🛠️ Tools/Methods Used:
MWCNT/SWCNT TSV Modeling • Equivalent circuit analysis • HSPICE simulation • Crosstalk evaluation • PDP/EDP analysis

👍 Like | 🔔 Subscribe | 📤 Share

#FederatedLearning #IndustrialWireless #EnergyOptimization #TSV #3DIC #MWCNT #SWCNT #Crosstalk #LowPowerVLSI #HSPICE #VLSIDesign #SignalIntegrity #Semiconductor #EngineeringResearch

Видео Adaptive Federated Learning for Industrial Wireless Energy Optimization канала Jack Sparrow Publishers
Яндекс.Метрика
Все заметки Новая заметка Страницу в заметки
Страницу в закладки Мои закладки
На информационно-развлекательном портале SALDA.WS применяются cookie-файлы. Нажимая кнопку Принять, вы подтверждаете свое согласие на их использование.
О CookiesНапомнить позжеПринять