Auto-adjustable low-signal processing technique based on programmable mixed-signal SoCs

Abstract

Today several electronic applications are demanding linear ICs that operate on a single supply rail, with the same performance of dual supply parts. Power consumption is now the key parameter for those battery operated systems and in some instances, more important than cost. In fact, through the effort of the manufacturers, doing more with less, is possible to find notable performance OpAmp's; parameters such as the offset voltage, leakage current and the precision of the gain are increasingly improved. In addition, in applications where the low-signal processing is required, as in many of the applications above described; in some cases the use of reconfigurable circuits is necessary to improve and highlight the performance of these OpAmp's. Then, the performance of a stage can be improve on the fly. This paper presents a design approach for amplification and active filter stages, with auto adjustable features, for signal processing greater than 100 mV and up to 1 KHz. The proposed approach uses a successive approximation algorithm for setting gain stages, and algorithms based on Wavelets-Mallat discrete thresholding and power spectral density of the signal for tuning the active filters. Its implementation based on a reconfigurable mixed-signal circuit, PSoC1, is also presented and validated. The results show that the proposed helps to improve the dynamic range and SNR of a signal, without the need to redesign and implementation processes of these stages, very commons in low-signal processing. These implementation can be used in applications where signals with the above conditions, need to be conditioned. For example: biomedical and semiconductor sensor signals.