Coherent Averaging for RF EVM Measurement Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for measuring error vector magnitude (EVM) and adjacent channel leakage ratio (ACLR) in RF devices are prone to inconsistencies and errors due to noise spikes, phase noise, and instrument-related issues, leading to unreliable measurements.
Innovation Solution
A method and system that utilize a signal analyzer with a processor to access repeated copies of OFDM signals, time align and de-rotate them, coherently sum and average them, discard outliers based on variance thresholds, and determine the EVM and noise power to isolate the contribution of the device under test, while nulling out unused tones and separating components due to the analyzer and source.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional EVM and ACLR measurement methods are used, then the measurement process is simple and direct, but the measurements are prone to inconsistencies and errors due to noise spikes, phase noise, and instrument-related issues
Solution Approach 1:
The patent segments the measurement process by dividing the received signal into multiple individual symbol measurements, then processing each symbol separately through time alignment, phase de-rotation, and coherent averaging. This segmentation allows for better control and reduction of noise and spike-related errors while maintaining measurement accuracy.
Solution Approach 2:
The patent applies preliminary actions by performing time alignment and phase de-rotation on each symbol before coherent averaging. These preparatory steps ensure that symbols are properly synchronized and phase-corrected prior to aggregation, which significantly reduces measurement errors and improves EVM and ACLR accuracy.
2Reliability
If repeated copies of OFDM signals are coherently summed and averaged, then noise and spike-related errors are reduced, but the measurement process becomes more complex and computationally intensive
Solution Approach 1:
The patent employs periodic action by utilizing repeated copies of OFDM signals and performing coherent averaging across multiple periods. This periodic repetition allows noise and random spikes to be averaged out, significantly improving measurement consistency and reliability while the structured approach maintains computational efficiency.
Solution Approach 2:
The patent uses copying by acquiring and processing multiple copies of the same OFDM signal. By creating and analyzing multiple replicas of the signal through repeated transmissions or symbol repetitions, the system can coherently average the copies to suppress random noise and improve measurement reliability without requiring additional hardware complexity.
3Measurement precision
If time alignment and phase de-rotation are performed on each symbol, then measurement accuracy is improved, but the processing time and computational load increase
Solution Approach 1:
The patent replaces mechanical/time-consuming manual alignment processes with digital signal processing techniques. Time alignment is achieved through digital correlation and offset estimation, while phase de-rotation is performed through digital multiplication with conjugate phase references. This substitution of digital processing for mechanical adjustment significantly reduces processing time while maintaining high measurement precision.
Data Source
AI summary
An apparatus and method: access repeated copies of an OFDM output signal produced by a device in response to corresponding repeated copies of an OFDM input signal; for each copy of the OFDM output signal, time align the OFDM output symbols to the OFDM input symbols, and de-rotate a phase of the OFDM output signal with respect to the OFDM input signal; coherently sum and average the copies of the OFDM output signal; determine a variance of each copy of the OFDM output signal, and an ensemble variance of all of the copies of the OFDM output signal; discard copies of the OFDM output signal whose variance differs from the ensemble variance by more than a threshold amount to produce a qualified set of copies; determine a mean value of the qualified set of copies; and determine a total noise power of the qualified set of copies from the mean value.