Cantilever characterization using Photo sensitive detector

GitHub Link to source code

This project was completed as a part of my semester 1 module and was completed in a team of 7 members. Work was divided into Hardware & signal processing team.

Setup

The whole mechanical setup consisted of two mirrors, a light source, and micro cantilever approximately 5mm with a hinge support and a free end for vibration, and a PSD mounted to receive the reflected beam back.

The signal was processed using a signal conditioning circuit consisting of a trans-impedance stage (using a MCP6002 dual Op-amp) and a differential stage (LM324). The principle on how the PSD works is that it has a resistive layer that acts as a current divider. Hence when a light beam falls on the surface of the PSD, there are two currents generated, each having a value proportional to the distance from the absolute center of the PSD. The role of the first stage is to convert this current signals into voltage signals.

The next task is to process this signals and extract different cantilever parameters like natural frequency and quality factor. For this we used a simple Picoscope 7000 series, and Jupyter notebook for processing the signal. I personally worked on the Fourier transform and the Parabolic interpolation method for finding natural frequency. Both the techniques eventually finds the natural frequency but Parabolic interpolation assumes the fact that the peak detected in the FFT is a gaussian curve, and therefore if you take the log on both sides of a gaussian curve equation and solve for x=x_max you can eventually arrive at the formula for finding the refined peak. The algorithm first finds the data point that corresponds to the peak say k. And then it would fit a parabola between points k-1 and k+1.