Time-to-Digital Converter Oversampling for Lower DPLL Phase Noise

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Time-to-digital converters (TDCs) in digital phase-locked loops (DPLLs) contribute significantly to phase noise, limiting the spectral purity of radio frequency clocks and requiring increased power consumption to reduce noise, which can introduce adverse effects.

Innovation Solution

Implementing a multi-edge phase comparison technique that performs TDC phase comparisons at a higher rate, processing TDC codes in the digital domain, and down-sampling to a lower reference clock rate to oversample analog and quantization noise, thereby reducing in-band noise contributions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If TDC phase comparisons are performed at a higher rate to reduce noise, then phase noise performance is improved, but power consumption increases

Engineering Contradiction:
Improvephase noise performanceVSAvoidpower consumption
Core Design Contradiction:
Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The patent segments the TDC operation into two distinct domains: a high-frequency domain for performing phase comparisons at a higher rate to reduce noise, and a lower-frequency domain for processing the results. This segmentation allows the system to achieve better phase noise performance through oversampling while keeping the power consumption of the processing circuitry at manageable levels by not running all circuits at the high frequency continuously.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent implements dynamic operation by allowing the TDC to operate at different frequencies in different domains. The comparison circuit operates at a higher frequency to capture more phase information, while the processing circuit dynamically processes this information at appropriate rates. This dynamic approach enables the system to adapt its operating frequency to the specific requirements of noise reduction versus power consumption.

Inventive Principle:
Principle #15Dynamics

2Measurement precision

If TDC architecture is redesigned to reduce noise, then phase noise performance is improved, but device complexity increases

Engineering Contradiction:
Improvephase noise performanceVSAvoidTDC architecture complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent extracts the noise reduction function from the TDC architecture itself and relocates it to the signal processing domain. Instead of modifying the TDC hardware to reduce noise, the system performs phase comparisons at a higher rate and then applies digital signal processing techniques to filter and process the results. This extraction approach reduces phase noise without requiring complex architectural changes to the TDC itself.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent substitutes mechanical/architectural modifications of the TDC with digital signal processing methods. Rather than redesigning the TDC circuitry to reduce noise, the system uses digital filtering and processing of the high-rate comparison results to achieve noise reduction. This substitution of digital processing for hardware modification simplifies the overall architecture while achieving the desired performance.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS9851696B2Circuit, a time-to-digital converter, an integrated circuit, a transmitter, a receiver and a transceiver
Publication Date: 2017.12.26 INTEL CORP
  • US9851696B2 patent drawing
  • US9851696B2 patent drawing
  • US9851696B2 patent drawing

AI summary

A circuit according to an example includes a controllable oscillator configured to generate an output signal based on a control signal, an input signal processing circuit configured to receive a reference signal and configured to generate a sequence of digital values indicative of a phase relation between the reference signal and the output signal or a signal derived from the output signal, and a digital data processing circuit configured to generate a sequence of processed values at a lower frequency than a frequency of the sequence of the digital values, each processed value being based on a plurality of the digital values of the sequence of digital values, wherein the control signal is based on the sequence of processed values.