Split PWM for Analog Crossbar Arrays With Shorter Integration Time
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional pulse width modulation in analog crossbar arrays results in increased pulse duration with higher bit precision, leading to temporal inefficiencies and significant quantization errors, which are not adequately addressed by existing methods.
Innovation Solution
The method involves splitting a multi-bit input into chunks with different significance factors, converting each chunk into a pulse width modulated signal, scaling the partial results, and accumulating them to maintain accuracy without excessively prolonging pulse duration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If larger bit precision is used in pulse width modulation, then measurement precision is improved, but duration of action increases
Solution Approach 1:
The patent divides the multi-bit input signal into multiple bit-sliced segments, where each segment is processed independently with a shorter integration time. This segmentation allows the system to achieve high bit precision through multiple quick measurements rather than one long measurement, thereby resolving the contradiction between precision and duration.
Solution Approach 2:
The patent implements continuous processing by overlapping integration periods across different bit slices. While one bit slice is being integrated, preparation for the next slice begins, ensuring continuous useful action without idle waiting time. This maintains high precision requirements while preventing pulse duration from becoming excessively long.
2Measurement precision
If integration time period is extended to reduce quantization error, then measurement precision is improved, but productivity decreases
Solution Approach 1:
By segmenting the integration process into bit-sliced time periods, the patent reduces the integration time required for each individual measurement while maintaining overall precision through multiple segments. This segmentation enables faster processing (higher productivity) without sacrificing the precision needed to minimize quantization errors.
Solution Approach 2:
The patent employs periodic bit-sliced integration where each bit slice is processed in discrete time periods. This periodic action allows the system to accumulate precision across multiple periods rather than requiring one excessively long continuous integration, thereby improving temporal efficiency while reducing quantization errors.
3Measurement precision
If bit-sliced integration is used to scale output, then measurement precision is improved, but loss of time increases
Solution Approach 1:
The patent segments the output scaling operation into bit-sliced components, where each slice is scaled and integrated independently over a shorter time period. This segmentation achieves accurate output scaling without requiring a single long integration time, thereby reducing time loss while maintaining precision.
Solution Approach 2:
The patent performs preliminary scaling of each bit-sliced segment before integration. By pre-scaling the segments according to their significance weights, the system reduces the complexity and time required for the final integration step, thereby maintaining measurement precision while minimizing time loss.
Data Source
AI summary
A computer-implemented method, according to one embodiment, includes: causing a multi-bit input to be split into two or more chunks, where each of the two or more chunks include at least one individual bit. Each of the two or more chunks are also converted into a respective pulse width modulated signal, and a partial result is generated in digital form for each of the respective pulse width modulated signals. Each of the partial results are scaled by a respective significance factor corresponding to each of the two or more chunks, and the scaled partial results are also accumulated.