Extending Seismic Data Acquisition System’s dynamic range by using multiple AD Converters

For a seismic acquisition system, dynamic range is a very import indicator. This indicator reveals the maximum and minimum input signals can be distortionless converted from analog signal to digital signal. The seismic signal dynamic range is very wide, the total dynamic range is more than 160 dB. Nowadays, the broadband seismometers’ noise level is under 1μV, maximum output is exceed ±20V, dynamic range is more than 140dB. But the dynamic range of widely used seismic acquisition system based on ΣΔAD conversion is around 130dB@200SPS. This results in the seismic acquisition system’s 2 shortcomings in real usage: 1) for tiny signal recording, the resolution is not high enough, signal to noise ratio is low; 2) for huge earthquake recording, limiting phenomenon shows up. To achieve the wide dynamic range required by seismic observation, we raised up the multiple analog to digital converters method to extend the acquisition unit’s dynamic range.

Although dynamic range of input seismic signal from seismometer is wide, but in practical application, for tiny signals (like signals below 1 mV), much more higher resolution is needed to improve signal to noise ratio, usually 1μV or even 0.1μV; but for large signals, the resolution around 10 μV to 0.1mV should be enough, the most concerned should be no limiting phenomenon when large signals arrived. Hence, to achieve both high resolution for tiny signal and full input range of large signal, we proposed a using multiple AD converters method to extend seismic acquisition unit dynamic range. That means using different resolutions for different amplitude level in analog to digital conversion: higher resolution for tiny signals and lower resolution for large signals. We divided the input signals to two levels: 0V～100mV and 100mV～40V. The experimental acquisition unit uses ADS1255 as AD converter. The embedded software reads the real time acquiring data, conducts linear compensation, zero offset voltage compensation, sensitivity consistent compensation, and then output 32 bits final results to record. The experimental results shows that when the sample rate is 50 SPS, the self noise RMS is around 0.2μV, the maximum output voltage is ±45V, and the dynamic range reaches 160 dB. That means this method is practical and effective in application.