Convolution-Based Drop Detector Signal Similarity Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing print apparatuses face challenges in accurately determining the operational status of printhead nozzles due to reliance on threshold-based methods, which can lead to false designations of nozzle functionality and are vulnerable to electrical noise, affecting image quality and maintenance efficiency.
Innovation Solution
The implementation of a drop detector system that uses a convolution module to normalize and compare drop detector signals with pre-calibrated print agent ejection signatures, allowing for accurate determination of nozzle performance and operational status through similarity analysis, reducing the impact of noise and system degradation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If threshold-based methods are used to determine nozzle operational status, then the detection process is simple, but the accuracy of nozzle status determination deteriorates due to false designations and vulnerability to electrical noise
Solution Approach 1:
The patent replaces the mechanical/electrical threshold-based detection system with a signal processing system that uses convolution and correlation analysis. Instead of comparing raw detector signals against fixed electrical thresholds, the system convolves the detector signal with a reference ejection signature to generate a correlation output, which is then compared to a threshold. This substitution of the detection mechanism fundamentally improves accuracy by accounting for signal variations and noise while maintaining relative simplicity through software-based processing.
Solution Approach 2:
The patent applies preliminary action by pre-calibrating the system to establish a reference ejection signature representing normal nozzle operation. This reference signature is obtained during a calibration phase when the nozzle is known to be functioning correctly. By having this pre-established reference, the system can later compare actual detector signals against it using convolution, enabling accurate detection of deviations from normal operation without requiring complex real-time analysis of multiple parameters.
2Ease of operation
If threshold-based detection is used, then the system is easier to operate, but reliability deteriorates due to false positives and negatives from electrical noise
Solution Approach 1:
The system replaces simple threshold comparison with convolution-based signal processing that inherently filters electrical noise. The convolution operation with a reference ejection signature acts as a matched filter, enhancing the signal components that match the expected ejection pattern while suppressing random noise. This substitution maintains ease of operation through automated processing while dramatically improving reliability by reducing false positives and negatives caused by electrical interference.
Solution Approach 2:
The patent implements feedback by using the correlation output from convolution as a basis for determining nozzle status. The system continuously monitors the correlation output and uses it to feedback control the determination of nozzle operational state. This feedback mechanism allows the system to adapt to varying conditions while maintaining reliable detection, as the correlation output provides continuous information about how well the current signal matches the expected ejection signature, enabling robust decision-making even in noisy environments.
Data Source
Figure 1~2
Figure 3A~3C
Figure 4~5
AI summary
In an example, a print apparatus includes a printhead carriage to receive a printhead comprising a print agent ejection nozzle, a drop detector to acquire a signal indicative of variations in a parameter detected by the drop detector over a period of drop detection, a memory to hold a print agent ejection signature, and processing circuitry. The processing circuitry includes a convolution module to convolve the drop detector signal with the print agent ejection signature, and the processing circuitry is to determine, from an output of the convolution module, an indication of similarity between the drop detector signal and the print agent ejection signature.