Differential Pair Mismatch Detection Through Spectral Error Separation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Analog circuits with differential element pairs often suffer from mismatch issues, leading to sub-optimal performance due to resistance or capacitance mismatches, and conventional solutions like oversizing or chopping introduce area and bandwidth trade-offs or signal fold-back problems.
Innovation Solution
Spectral separation of mismatch-induced errors in analog circuits using an analog-to-digital converter (ADC) and a controller to measure and compensate for errors in differential element pairs, allowing for precise identification and correction of mismatches without the drawbacks of conventional methods.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If differential pair elements are oversized to reduce mismatch percentage, then manufacturing precision improves, but area consumption increases and bandwidth decreases due to increased parasitic capacitance
Solution Approach 1:
The patent segments the mismatch error from the main signal by assigning them to different frequency domains. The desired signal remains at baseband while mismatch-induced errors are modulated to higher frequencies through chopping, enabling separate processing and filtering of the error components.
Solution Approach 2:
The patent changes the frequency parameter of the mismatch error signal by applying chopping modulation. This transforms the static DC offset error into a time-varying signal at the chopping frequency, allowing it to be distinguished from and separated from the baseband desired signal through frequency-domain differentiation.
2Manufacturing precision
If conventional chopping is applied to move mismatch-induced offset to chopping frequency, then manufacturing precision improves, but signal fold-back occurs where input signal energy around chopping frequency folds back to DC
Solution Approach 1:
The patent extracts the mismatch error component from the combined signal by utilizing frequency-domain separation. Through chopping modulation, the error signal is shifted to a distinct frequency band where it can be independently identified and removed using band-pass filtering, preventing it from contaminating the baseband signal.
Solution Approach 2:
The patent implements a feedback mechanism where the output signal is fed back through a band-pass filter tuned to the chopping frequency to extract the mismatch error component. This extracted error is then used to generate a compensation signal that is subtracted from the main output, thereby eliminating the harmful fold-back effect.
3Manufacturing precision
If spectral separation with ADC monitoring is used to measure mismatch-induced error, then manufacturing precision improves, but device complexity increases due to additional ADC and controller components
Solution Approach 1:
The patent replaces complex analog mismatch compensation circuitry with a digital processing approach. By using an ADC to convert the analog output to digital form and then applying digital filtering and processing to extract and compensate for mismatch errors, the system achieves high precision while utilizing well-established digital signal processing techniques.
Data Source
AI summary
A method for use in an analog circuit having a plurality of differential pairs of elements, wherein for each pair of the plurality of differential pairs of elements, the elements of the pair are designed to match but may have mismatch that induces error. The method includes, for each pair of at least two pairs of the plurality of differential pairs of elements: spectrally separating the mismatch-induced error of the pair from mismatch-induced error of a remainder of the plurality of differential pairs of elements, monitoring, by an analog-to-digital converter (ADC), an output of the analog circuit, and analyzing the monitored output to measure the mismatch-induced error of the pair.