Flash ADC Frequency-Domain Calibration for In-Band Noise
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Delta-sigma analog-to-digital converters (ADCs) face challenges in accurately calibrating flash offsets, particularly in distinguishing noise power from signal power, which affects their performance and requires offline calibration, limiting their ability for continuous and in-field corrections.
Innovation Solution
A calibration system that employs a frequency-domain power measurement technique using a fast Fourier transform (FFT) to adjust comparator offsets, allowing for both foreground and background calibration without disconnecting the ADC from the signal chain, thereby directly minimizing in-band noise and flash offsets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional time-domain calibration methods are used to calibrate ADC comparator offsets, then calibration can be performed, but the ability to distinguish noise power from signal power is limited and offline calibration is required
Solution Approach 1:
The patent transforms the calibration approach from the time domain to the frequency domain by applying Fast Fourier Transform (FFT) to the ADC output data. This dimensional change enables direct measurement of noise power in the frequency spectrum, allowing for continuous in-field calibration while maintaining the ability to distinguish noise from signal components.
2Measurement precision
If offline calibration is performed to calibrate ADC offsets, then calibration accuracy can be achieved, but the ADC must be disconnected from the signal chain
Solution Approach 1:
The patent enables continuous calibration by performing FFT-based noise power measurement directly on the running ADC output while the converter remains connected to the signal chain. This eliminates the need to disconnect the ADC for calibration, allowing the useful conversion action to continue uninterrupted while calibration is performed in the background.
3Adaptability or versatility
If frequency-domain FFT calibration is implemented to enable continuous calibration, then in-field corrections are possible, but additional processing requirements are introduced
Solution Approach 1:
The patent introduces an FFT processing unit as an intermediary component that transforms the raw ADC output data into the frequency domain. This intermediary enables the extraction of noise power measurements without requiring complex modifications to the core ADC architecture, facilitating in-field calibration through a dedicated processing stage.
Data Source
Figure 1~4
Figure 2~3
Figure 5
AI summary
A flash analog-to-digital converter (ADC) includes comparators that convert an analog input signal to a digital output signal. Offsets of these comparators introduce noise and can hurt the performance of the ADC. Thus, these comparators are calibrated using calibration codes. Conventional calibration methods determine these calibration codes by removing the ADC from an input signal. Otherwise, it is difficult to distinguish the noise from the signal in the calibration measurement. In contrast, an embodiment can determine the calibration codes while the ADC converts the input signal to a digital signal. Such an embodiment can be achieved by a frequency-domain technique. In an embodiment employing a frequency-domain power meter, an input signal can be removed from the power measurement. This removal enables accurate measurement of in-band noise without having the measurement be corrupted by input signal power.