Digitizer Noise Reduction for Pseudo-Periodic Waveforms
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing oscilloscopes and test instruments introduce random noise and spurious distortions during signal digitization, particularly affecting pseudo-periodic time-domain waveforms, which are not effectively addressed by current noise reduction techniques such as averaging, oversampling, or signal modeling due to limitations in handling uncorrelated noise and inconsistent noise statistics.
Innovation Solution
A method involving the use of multiple digitizers to sample a signal under test (SUT) independently, pairing the waveforms based on their repeating pattern, calculating covariance to estimate signal variance, and scaling the samples to adjust variances, thereby reducing noise and reconstructing a noise-reduced waveform.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple digitizers are used to sample the SUT independently, then noise reduction is improved, but device complexity increases
Solution Approach 1:
The patent divides the measurement system into multiple independent digitizer channels, each sampling the SUT separately. This segmentation allows uncorrelated noise from each digitizer to be treated independently, enabling noise reduction through statistical processing while maintaining the ability to capture the complete signal waveform.
Solution Approach 2:
The patent creates multiple copies of the SUT waveform by routing the signal to multiple digitizers simultaneously. Each digitizer produces an independent copy of the waveform, and these copies are then processed together to eliminate uncorrelated noise, effectively replicating the signal multiple times for statistical analysis.
2Measurement precision
If the signal is split into multiple copies and routed to multiple digitizers, then noise reduction is improved, but signal amplitude is reduced
Solution Approach 1:
The patent merges multiple digitized waveform copies back together through statistical processing. After each digitizer independently samples the signal, the system combines the results by calculating covariance and reconstructing the waveform, thereby recovering the full signal amplitude while eliminating uncorrelated noise from individual digitizers.
Solution Approach 2:
The patent employs feedback mechanisms where the system continuously monitors the waveforms from multiple digitizers and adjusts the reconstruction process based on the observed signal characteristics. This feedback allows the system to maintain accurate signal representation while compensating for amplitude reductions that may occur during the splitting and digitization process.
3Measurement precision
If averaging is used to reduce noise, then noise reduction is improved, but it cannot handle pseudo-periodic signals with uncorrelated components
Solution Approach 1:
The patent applies local quality analysis by examining the statistical properties of each waveform segment independently. Instead of applying a uniform averaging process, the system analyzes the covariance structure of each local region of the waveform, allowing it to handle pseudo-periodic signals with uncorrelated components by treating each segment's noise characteristics individually.
Solution Approach 2:
The patent changes the statistical parameters used for noise reduction from simple mean averaging to covariance-based estimation. By calculating the covariance matrix of the waveform segments and using this information to reconstruct the signal, the system can adapt to different signal types including pseudo-periodic waveforms, maintaining noise reduction effectiveness across diverse signal characteristics.
4Measurement precision
If oversampling is used to reduce noise, then noise reduction is improved, but large sample rates are required achieving large noise reduction
Solution Approach 1:
The patent transitions from time-domain oversampling to a multi-dimensional approach by using multiple simultaneous digitizer channels. Instead of increasing the sample rate in the time domain, the system adds another dimension by measuring the signal across multiple independent digitizers simultaneously, achieving noise reduction through statistical processing of these parallel measurements rather than through high-rate sequential sampling.
Data Source
AI summary
A method and system provide for measuring a repeating waveform of a SUT. The method includes repeatedly sampling first and second copies of the SUT to provide first SUT waveforms including first noise introduced by a first digitizer and second SUT waveforms including second noise introduced by a second digitizer; pairing the first and second SUT waveforms to provide corresponding pairs of first and second digital samples; organizing the pairs of first and second digital samples into groups of sample pairs corresponding to sampling times; for each group, calculating a covariance of the first and second digital samples to estimate a signal variance of the SUT, and scaling the first and second digital samples to preserve a mean of the group while adjusting a variance of the group to match the estimated signal variance of the SUT at the corresponding sampling time; and reassembling the groups into the SUT waveform.