analyze_error_autocorr

Overview

analyze_error_autocorr computes the autocorrelation of ADC errors to reveal temporal patterns. White noise has zero autocorrelation at all lags; correlated errors indicate memory effects, settling issues, or periodic disturbances.

Syntax

from adctoolbox import analyze_error_autocorr

# Basic usage
result = analyze_error_autocorr(signal, max_lag=100, show_plot=True)

# With custom parameters
result = analyze_error_autocorr(signal, max_lag=200, resolution=12)

Parameters

• signal (array_like) — Input ADC signal

• max_lag (int, default=100) — Maximum lag for autocorrelation

• resolution (int, optional) — ADC resolution in bits

• show_plot (bool, default=False) — Display autocorrelation plot

• ax (matplotlib axis, optional) — Axis for plotting

Returns

Dictionary containing:

• autocorr — Autocorrelation values

• lags — Lag values

• error — Error signal used

Interpretation

Autocorrelation Pattern	Likely Cause
Peak at lag=0 only	White noise (ideal)
Decay over few samples	Low-pass filtering, bandwidth limit
Periodic peaks	Switching artifacts, clock coupling
Slow decay	1/f noise, drift

Use Cases

• Distinguish white noise from colored noise

• Detect memory effects in pipelined ADCs

• Identify periodic disturbances

See Also

• analyze_error_pdf — Error distribution

• analyze_error_spectrum — Frequency domain view

• analyze_error_envelope_spectrum — AM modulation

References

1. B. Razavi, “Principles of Data Conversion System Design,” IEEE Press, 1995