Sigma-Delta ADC Loop Delay Compensation Using Randomized LSB Injection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Continuous-time sigma-delta analog-to-digital converters (ADCs) are sensitive to loop delays, leading to instability and performance loss due to excess loop delay issues, which conventional methods fail to effectively address.
Innovation Solution
The system introduces a random sequence at the least significant bit level to estimate the impulse response of the noise transfer function, allowing for the modification of loop coefficients to compensate for deviations and maintain robust run-time operation by correlating the ADC output with the delayed sequence, and using a fully-adaptive configuration to track variations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional methods are used to handle loop delay in sigma-delta ADCs, then the system structure remains simple, but the ADC becomes unstable and loses performance due to excess loop delay
Solution Approach 1:
The patent implements a feedback mechanism where the impulse response of the noise transfer function is estimated by correlating the ADC output with a delayed random sequence. This estimated impulse response is then fed back to modify the loop coefficients, creating a closed-loop system that automatically compensates for excess loop delay and maintains stability without requiring complex external calibration equipment.
Solution Approach 2:
The system performs self-diagnosis and self-correction by using its own output signal to estimate the loop delay characteristics. The random sequence is injected at the LSB level and the system correlates this with its own output to extract impulse response information, which is then used to adjust its own loop coefficients, making the system self-regulating without external intervention.
2Reliability
If loop coefficients are modified to compensate for excess loop delay, then ADC stability improves, but the system becomes more complex requiring impulse response estimation
Solution Approach 1:
The patent introduces a random sequence as an intermediary signal injected at the LSB level of the quantizer output. This intermediary signal serves as a probe that, when correlated with the ADC output, reveals the impulse response characteristics of the noise transfer function. The random sequence acts as a mediator that enables indirect measurement of the loop delay without requiring direct access to internal signals.
Solution Approach 2:
The system injects a random sequence at the LSB level, which represents a partial action that does not significantly interfere with the main signal path. By using only the LSB portion of the quantizer output for correlation purposes, the system obtains sufficient information about the impulse response while minimizing the impact on overall ADC performance and noise floor.
3Measurement precision
If a random sequence is added at the LSB level to estimate impulse response, then loop delay can be compensated, but the noise floor may be degraded
Solution Approach 1:
The random sequence is injected only at the LSB (least significant bit) level of the quantizer output, representing a partial action that affects only the least significant portion of the signal. This minimizes the overall noise injection while still providing sufficient information for impulse response estimation, as the correlation process extracts useful information from this partial signal modification without significantly degrading the overall noise floor.
Solution Approach 2:
The system separates the impulse response estimation function from the main signal path by extracting only the LSB portion of the quantizer output for correlation purposes. The random sequence is injected at this extracted LSB level, allowing the estimation process to be performed on a separated, minimal signal path that does not directly interfere with the main signal conversion process, thereby isolating the noise impact to a minimal level.
Data Source
AI summary
An analog-to-digital convertor circuit converts the output of a loop filter circuit to a digital signal. A random sequence generation circuit generates a random sequence. Adder circuitry adds the random sequence to the digital signal to generate a randomized digital signal. Noise transfer function impulse response detection circuitry processes the randomized digital signal and the random sequence to determine a noise transfer function impulse response. Loop filter configuration circuitry configures the loop filter circuit based on the noise transfer function impulse response. The random sequence generation circuit may comprises a high-pass sigma delta modulator. The noise transfer function impulse response detection circuitry may determine the noise transfer function impulse response, and the loop filter configuration circuitry may configure the loop filter based on the noise transfer function impulse response. The loop filter configuration circuitry may generate loop filter coefficients based on the noise transfer function impulse response.