Skip to content

Systematic high-level design of a fifth order Continuous-Time CRFF Delta Sigma ADC

MetadataDetails
Publication Date2021-02-21
AuthorsM. Germain, F. Rarbi, O. Rossetto
InstitutionsUniversitĂŠ Grenoble Alpes
AnalysisFull AI Review Included

This paper details the systematic high-level design and simulation of a Continuous-Time Delta Sigma Analog-to-Digital Converter (CT ΔΣ ADC) optimized for high-speed energy measurement systems.

  • Core Objective: Design a robust, high-level model for a fifth-order CT ΔΣ ADC targeting 10-bit Effective Number of Bits (ENOB) resolution.
  • Target Application: Energy measurement systems used in particle identification, specifically utilizing diamond detectors, requiring wide bandwidth (40 MHz).
  • Architecture Choice: A Cascaded Resonators Feedforward (CRFF) topology was selected for its ability to achieve high Signal-to-Quantization-Noise Ratio (SQNR) at a low OverSampling Ratio (OSR=8).
  • Methodology: A Model-Based Design approach was implemented using MATLAB scripts integrated with SIMULINK graphical behavioral models (incorporating Schreier’s Toolbox).
  • Synthesis Achievement: The ideal model achieved a simulated ENOB of 11.9 bits (73.5 dB SQNR), exceeding the 10-bit target specification.
  • Robustness Analysis: Parametric Monte Carlo simulations (10,000 iterations) were performed to analyze the impact of component dispersion (R/C process variation) on loop coefficients.
  • Key Finding: Under 20% process variation and 1% mismatch error, the mean SQNR dropped significantly to 51.7 dB, highlighting the critical sensitivity of the loop coefficients (a1, a2, b1, c1) to fabrication non-idealities.
ParameterValueUnitContext
Target ENOB Resolution10bitsRequired for energy measurement systems
Simulated ENOB (Ideal)11.9bitsAchieved in behavioral simulation
Signal Bandwidth (BW)40MHzTarget input frequency range
Sampling Frequency (fs)640MHzDetermined by OSR
OverSampling Ratio (OSR)8-Low ratio chosen for wideband design
Loop Filter Order (L)5-Fifth-order modulator architecture
Quantizer Resolution (N)3bitsModulator internal ADC resolution
Theoretical SQNR80dBCalculated maximum performance
Simulated SQNR (Ideal)73.5dBAchieved in behavioral simulation
Mean SQNR (Dispersion)51.7dBResult of 10,000 iterations with Δp = 20%, δm = 1%
Integrator DC Gain40dBSpecified for all five Opamp-RC integrators
Integrator Saturation450mVMaximum output swing limit for integrators
Linearity Target<0.2%Required specification

The design and analysis relied on a systematic, high-level modeling approach combining established tools with custom scripting:

  1. Model-Based Design (MBD): The CT ΔΣ modulator was implemented as a graphical behavioral model in SIMULINK, based on the general definition of the CRFF architecture.
  2. Synthesis Tool Integration: Schreier’s Toolbox [6] was used to synthesize the initial, ideal loop coefficients (ai, bi, ci, gi) required to meet the 40 MHz BW and 10-bit ENOB specifications.
  3. Custom Scripting: MATLAB scripts were developed and integrated to control the SIMULINK model, allowing for automated simulation runs and systematic analysis of non-idealities.
  4. Component Dispersion Modeling: The variation of loop coefficients (Xi) due to fabrication process was modeled using the equation: Xi = Xsi * (1 + Δpi + δmi).
    • Δpi represents process variation error (tested up to 20%).
    • δmi represents mismatch error (tested up to 1%).
  5. Parametric Monte Carlo Simulation: 10,000 iterations were run, varying each loop coefficient independently according to a normal Gaussian distribution within the defined error range (e.g., Δp = 20%, δm = 1%).
  6. Performance Extraction: For each iteration, the script injected the dispersed coefficients, ran the SIMULINK simulation with a 40 MHz sine wave input, and calculated the resulting SQNR and ENOB from the digital output spectrum.
  7. Critical Parameter Identification: Through iterative simulations, the coefficients a1, a2, b1, and c1 were identified as the most critical parameters, requiring the tightest design tolerance (e.g., reducing Δp to 10% and δm to 0.5%) to maintain acceptable SQNR.

The systematic design methodology and the resulting high-performance CT ΔΣ ADC architecture are relevant to several demanding engineering fields:

  • Particle Physics and Detection: Used specifically for energy measurement systems in new generation detectors (e.g., based on diamond), requiring fast, precise signal digitization from front-end electronics.
  • High-Speed Data Acquisition (DAQ): Applicable in systems requiring wideband conversion (40 MHz BW) at high sampling rates (640 MHz) for scientific instrumentation and testing.
  • Front-End Electronics (FEE) Design: Provides a robust, systematic method for designing complex analog control and signal processing blocks, reducing design time and effort during the schematic phase.
  • CMOS Analog Integrated Circuits (IC): The methodology allows designers to estimate the impact of process variation (R/C time constant shifts) early in the design flow, ensuring manufacturability and yield for Opamp-RC integrator-based circuits.
  • Telecommunications: Potential use in high-speed receivers or base stations where wide bandwidth and moderate resolution are required, and where low OSR is advantageous.
View Original Abstract

In this paper we present the development of a Systematic high level design model based on MATLAB scripts. It is integrated into a graphical behavioral model toolbox for the synthesis and simulation of a Continuous-Time Delta Sigma ADC. For this, we decided to use a Model-based design approach which it is adopted to address problems associated with designing complex control and signal processing systems such as the case of Continuous-Time Delta Sigma ADC. The goal of our study is the design of a 10 bit ENOB ADC for energy measurement systems used in particle identification through a new generation of detectors based on diamond. Results of the synthesis of a proposed fifth order Continuous-Time Delta Sigma ADC modulator for 10-bit ENOB ADC based on a Cascaded Resonators Feedforward architecture and simulations of the dispersion of its components (due to fabrication process) using the proposed tool are demonstrated and discussed.