Research And Design Of Neural Signal Acquisition System Based On SAR ADC

As an electrical signal,neural signal is presented as potential variation when it traval through nerve fiber or muscle fibers.It can transmit at speeds as high as hundreds of miles per hour.Although these electrical signals really exist and can be measured through voltage detection,the electrophysilogical signals of human cells and tissues is actually very faint.There is a large amount of noise in the signal.And it also can be disturbed by the surrounding environment easily.If we directly collect the neural signal without any processing,then the noise-to-signal ratio of electrophysiological signal collected is extremely low and we can not extract desired properties from these signal.So it is important to process the signal when collecting them.Although there are also commercial detection equipment,their price is generally high and have a low flexibility.In view of the above problems,this paper designed a detection and amplification system for weak signals such as electrophysiological signals.The system consists of hardware and software.And the hardware part is mainly divided into the following three modeles: the first is pre amplification mudule of nerve signal.It is used to detect the faint electrophysiological signal and amplifier the signal to an appropriate extent.The second is bandpass filter and second-stage amplification module,which is used to filter the DC and high frequency interference and amplify the signal to the required quantitative range of the back-end ADC.The third is data acquisition module.In this module,a SAR ADC chip that is independently carried out by us own is used to quantize the collected electrophysiological signals.Then a FPGA is used to receive the digital data.Software part mainly uses Quartusii to control the FPGA and uses serial port to store digital data to PC.Finally,digital data is imported to Matlab to filter the power line interference.After completing circuit design and making PCB,the main parameters of the detection system was tested.The test content mainly includes the equivalent noise,gain and bandwidth.Through constant adjustment of the device,final measured input equivalent noise is about 6?V,pass frequency is from 10 Hz to 3kHz.The input signal can be correctly collected by the PC terminal and achieve the accuracy of 10 bit.The work is mainly based on the independently developed low power SAR ADC for detecting weak signals,and has successfully made a low-cost neural signal acquisition system.At the same time,it has also made a forward-looking work for the next step of SOC of neural network acquisition.