Table of Contents
Noise shaping is a key technique used in delta-sigma converters to improve their resolution by pushing quantization noise out of the band of interest. Understanding the design calculations involved helps in optimizing converter performance for various applications.
Fundamentals of Noise Shaping
In delta-sigma converters, noise shaping involves the use of feedback loops and filters to control the spectral distribution of quantization noise. The goal is to minimize in-band noise while increasing out-of-band noise, which can be filtered out later.
Design Calculations for Noise Shaping
The effectiveness of noise shaping depends on the order of the noise transfer function (NTF). Higher-order NTFs provide more aggressive noise shaping but can introduce stability issues. The basic calculation involves selecting the NTF order and calculating the loop filter parameters.
For an NTF of order n, the noise transfer function typically follows a polynomial form. The noise power in the in-band region can be estimated using the formula:
S_in = (Δ² / 12) * (π / fs)^{2n}
Practical Insights
Implementing noise shaping requires balancing between noise suppression and stability. Designers often simulate the NTF response to ensure the desired noise profile without risking instability. Component tolerances and non-idealities also affect the actual noise performance.
Practical techniques include using digital filters to further reduce in-band noise and optimizing the loop filter design for specific application requirements. Regular testing and iterative adjustments are essential for achieving optimal results.
- Choose the appropriate NTF order based on stability considerations.
- Calculate loop filter parameters carefully to match the desired noise shaping profile.
- Simulate the noise transfer function before hardware implementation.
- Account for non-idealities in real-world components.
- Use digital filtering to refine noise performance post-conversion.