A new double FFT-based filter to reduce the effect of 1/f noise spectrum in a tunable diode laser spectrometer (TDLS)

Abstract

A Tunable diode laser spectrometer (TDLS) system has been designed to scan the near-surface atmosphere for ammonia gas over a wide range of distances (10 m to 1 Km). Since the system is designed for space applications, it needs to be small, lightweight, and low power, which dictates the use of relatively low frequency measurement scans. The spectrometer uses a diode laser, which is subject to a large 1/f noise component at these low frequencies. In this work, digital signal processing techniques are used to maximize the measurement sensitivity of a low frequency TDLS system depending on Double Fast Fourier Transform (DFFT-BF) based- filter. Simulations of the 1/f noise spectrum and ammonia gas absorption peak were performed using a sinusoidal waveform to drive the diode laser. A DFFT-BF-BF method is proposed that reduces the average of the error in the gas readings to nearly 50 percent. Because, this method decreases the effect of 1/f noise while keeping the measurement signal relatively constant.