Abstract
This paper presents the design and implementation of a fully differential analog front-end which pre-processes the acquired Electro Encephalogram (EEG) signals before digitization. It consists of an analog multiplexer and a filter which can efficiently condition real-time EEG signals. The circuit has been designed so as to be compatible with a 128 multi-channel data acquisition system and hence, a 128 multi-channel analog multiplexer has been designed which aids in selecting the signal to be processed. Further, EEG signals are known to exist in the frequency range of 0-5 KHz and amplitude range of 100nv to 10mv. Hence, an eighth order low-pass filter with a bandwidth of 5 KHz, a stop band attenuation of -105dB, and a pass-band gain of 35.56dB has been implemented using cascaded bi-quads. The entire filter possesses fully-differential architecture because of its ability to effectively suppress power-line interference and on-chip noise. Chebyshev implementation of Ackerberg-Mosseberg bi-quad helped in conserving the chip area and aided in effective noise suppression. The schematics of the above mentioned circuits have been simulated in the 90nm technology.