Detection and suppression of common-mode noise in capacitive fingerprint sensors

Abstract

Common-mode noise is a major concern within the area of capacitive fingerprint sensing. The measurements necessary for sensing the structure of a fingerprint are extremely small in magnitude and can thus easily be overshadowed by noise. Additionally, the sources of common-mode noise are many, ranging from power supplies to touchscreens. A common way to combat noise is to apply correlated double sampling, a specific application of it was utilized in this thesis. By applying eight sampling capacitors connected in two groups of four to separate leads of a differential ADC, correlated double sampling can be performed. If these capacitors are used properly for sampling the desired signal, it should be theoretically possible to filter out the common-mode noise. By utilizing signal generators, a multitude of different common-mode noise injections were emulated onto a fingerprint sensor. By applying different combinations of the sampling capacitors attempts to detect and suppress the emulated noise sources were made. As the produced noise sources were known, the detection methodologies could be evaluated easily. By comparing the standard deviation of the sensor output data to a noise-free case, the efficiency of the suppression methodologies could be evaluated as well. The methodologies were also applied to overlaid frequencies and actual charger noise. The detection methodologies proved able to detect high-frequency square-wave noise sources with good accuracy, it struggled with other kinds of noise however. The suppression methodologies were able to reduce the standard deviation of the sensor data significantly in all cases. However, while application of the methodologies for detection and suppression designed during the thesis work improved upon the current sensor setup, the solution is not complete. To assure the completeness of the solution it would have to be applied to a real-world system. Until such an application has been performed, all the results have only truly been theoretically verified.