Skip to Main Content
This paper presents a low power, low voltage and low frequency bandpass filter implementation of a continuous wavelet transform (CWT) for use with physiological signals in the electroencephalogram (EEG) range (1-150 μV, 1-70 Hz bandwidth). Experimental results are presented for a 1 V, 7th order gmC filter based CWT with filter center frequencies ranging from 1 to 64 Hz. Low power and low frequency operation is achieved by biasing the transconductor transistors at low current levels in the deep weak inversion region. The resulting increased mismatch and reduced bandwidth are compensated for at the topology level. The filter has a 43 dB dynamic range and a 60 pW power consumption. This power consumption is three orders of magnitude lower than existing CWT implementations and assessed via a suitable figure of merit the performance is better than all considered bandpass filters. The improvement in the state-of-the-art originates from the close integration of the application requirements, CWT theory, bandpass filter design theory, and low transconductance transconductor design. These topics are described in detail.