Power Modeling in mm-Wave and Terahertz CMOS Transmitters for Wireless Network-on-Chip

Mohammad Shahmoradi,Korkut Kaan Tokgoz,Eduard Alarcon,Sergi Abadal

2024-11-02

Abstract:Wireless Network-on-Chip (WNoC) systems, which interconnect chips using wireless links, face significant challenges in area and power consumption. To tackle these constraints, behavioral models (BMs) are crucial for assessing system performance under various conditions and optimizing parameters like data throughput and power consumption. Building transceivers (TRXs) physically is costly and time-consuming, making modeling a more practical approach. This paper develops a power consumption model for the sub-blocks of a WNoC transmitter (TX) at the chip level. By integrating these BMs with MATLAB, we aim to create a power model for TXs in WNoC architectures, optimized for CMOS technology operating at millimeter-wave and terahertz frequencies.

Signal Processing

What problem does this paper attempt to address?

This paper attempts to solve the power consumption problem in the Wireless Network - on - Chip (WNoC) system, especially in CMOS transmitters in the millimeter - wave (mm - Wave) and terahertz (THz) frequency bands. Specifically, the paper aims to develop a behavioral model (BM) to predict the power consumption of these high - frequency - band transmitters at different operating frequencies.

### Main problems:
1. **Power consumption and area constraints**: The WNoC system faces significant power consumption and area limitations. Physically constructing transmitters (TRXs) is costly and time - consuming. Therefore, it is necessary to evaluate system performance and optimize parameters such as data throughput and power consumption through modeling.
2. **Limitations of existing models**: Existing behavioral models usually focus on specific frequency bands or single components, ignoring system - level interactions and unable to fully understand the power consumption trends in the WNoC system.

### Solutions:
- **Develop a comprehensive model**: The paper proposes a new behavioral modeling method to predict the power consumption of integrated transmitters in the WNoC system, especially for the millimeter - wave and terahertz frequency bands. This model covers the power consumption dynamics of key transmitter components (such as power amplifiers, mixers, and oscillators) and analyzes the impact of operating frequencies on power consumption.
- **MATLAB integration**: By integrating these behavioral models with MATLAB, researchers can create an optimized transmitter power consumption model suitable for millimeter - wave and terahertz frequencies operating in CMOS technology.

### Key innovation points:
- **Cross - band applicability**: The model is not only applicable to the current millimeter - wave frequency band but also extends to the terahertz frequency band, providing support for future 6G and higher - version wireless communications.
- **System - level optimization**: The model takes into account each sub - module in the entire transmission link, ensuring system - level power consumption optimization rather than just a single component.

### Formula summary:
- Formula for power - added efficiency (PAE):
\[
\text{PAE}=\frac{P_{\text{out}}-P_{\text{in}}}{P_{\text{DC}}}
\]
where \(P_{\text{out}}\) is the output power, \(P_{\text{in}}\) is the input power, and \(P_{\text{DC}}\) is the DC power consumption.

- Formula for the figure - of - merit (FOM) of an oscillator:
\[
\text{FOM}_{\text{Osc}}=P_N + 10\log_{10}\left(\frac{P_{\text{DC}}}{1\text{mW}}\right)-P_{\text{out}}- 20\log_{10}\left(\frac{f_0}{\Delta f}\right)
\]
where \(P_N\) is the phase noise (dBc/Hz), \(f_0\) is the oscillation frequency (GHz), and \(\Delta f\) is the frequency offset (MHz).

- Formula for the conversion gain (CG) of a mixer:
\[
\text{CG (dB)}=10\log_{10}\left(\frac{P_{\text{RF, out}}}{P_{\text{IF, in}}}\right)
\]

Through these models and formulas, the paper provides designers with an effective tool to optimize the power consumption efficiency of the WNoC system without relying on expensive physical prototypes.