Digital Frequency Synthesizer Phase-to-Amplitude Conversion

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Direct Digital Frequency Synthesizers (DDFS) face challenges in achieving a high Spurious Free Dynamic Range (SFDR) with low hardware complexity, often requiring high circuit complexity and computational power.

Innovation Solution

A method for phase-to-amplitude conversion using a combination of linear interpolation and a second-order sinusoidal term to correct errors, reducing the need for large ROM look-up tables and arithmetic circuitry, thereby achieving efficient storage and computational performance.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If a wide phase accumulator is used to satisfy frequency resolution requirements, then frequency resolution is improved, but LUT size increases significantly

Engineering Contradiction:
Improvefrequency resolutionVSAvoidLUT size
Core Design Contradiction:
Measurement precisionVSQuantity of substance

Solution Approach 1:

The phase-to-amplitude conversion is segmented into multiple stages: first a coarse conversion using a smaller LUT, then refinement using interpolation and error correction terms. This divides the single large LUT operation into multiple smaller operations that achieve the same result with reduced storage requirements.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The invention changes the representation parameters by using difference tables and error correction terms instead of storing complete amplitude values. The LUT stores only differential values and correction coefficients, which require fewer bits while maintaining the same conversion accuracy.

Inventive Principle:
Principle #35Parameter changes

2Quantity of substance

If polynomial approximation techniques are used to reduce LUT size, then storage requirements are reduced, but computational complexity increases

Engineering Contradiction:
Improvestorage requirementsVSAvoidcomputational complexity
Core Design Contradiction:
Quantity of substanceVSDevice complexity

Solution Approach 1:

The invention replaces complex polynomial arithmetic with simpler operations: lookup table access, addition, and bit-shifting. The error correction mechanism uses basic arithmetic operations instead of complex multiplication and division, significantly reducing the computational burden while maintaining accuracy.

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

Solution Approach 2:

The invention introduces intermediate difference tables and error correction terms as mediators between the phase accumulator and final amplitude output. These intermediaries break down the complex conversion into manageable steps that can be executed with simple arithmetic operations.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Measurement precision

If large LUT and arithmetic circuitry are used to achieve high SFDR, then SFDR performance is improved, but hardware complexity increases

Engineering Contradiction:
ImproveSFDR performanceVSAvoidhardware complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The high-precision conversion is segmented into multiple passes: a coarse conversion followed by error correction passes. Each pass uses simpler circuitry, but the combination achieves the same SFDR performance as a single complex conversion stage would require.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The invention uses feedback through error correction terms that compensate for inaccuracies in the coarse conversion. The difference tables and correction mechanisms provide feedback loops that iteratively improve the amplitude accuracy without requiring a single large complex LUT.

Inventive Principle:
Principle #23Feedback

Data Source

PatentEP2126659B1A method for use in a digital frequency synthesizer
Publication Date: 2018.07.25 TELEFONAKTIEBOLAGET LM ERICSSON (PUBL)
  • EP2126659B1 patent drawingFigure 1
  • EP2126659B1 patent drawingFigure 2
  • EP2126659B1 patent drawingFigure 3

AI summary

A method for use in a digital frequency synthesizer (100), the method comprising phase to amplitude conversion (120) of an output value of a phase accumulator (110) in said synthesizer (100), said conversion being carried out as an approximation (y) of a phase value (x) which corresponds to said output amplitude value, the method being characterized in that the approximation comprises a combination of a linear interpolation value and a second order sinusoidal value, the second order sinusoidal value being used as an error term to correct for errors in the linear interpolation value.