Article Preview
TopIntroduction
Large bandwidth and wide power dynamic range are the two experimentally identified characteristics of free oscillations of the Earth (Woodhouse, 2013). The minimum probed displacement using seismeter is about accuracy of 0.1nm. The dynamic range of major earthquake is 220dB, whose range of bandwidth is from 10-5 Hz to 103Hz. Recording propagating seismic waves can be employed to analyze the Earth’s interior (Curtis, 2009). One of requirements of geophysical instrument is detecting and recording shakes of 1nm-10mm in bandwidth of 0.01-10Hz. To collect abovementioned seismic signals with various distance and power, the dynamic range of seismic data acquisition system should exceed 120dB (Shapiro, 2015). High resolution data acquisition system with slow sample rate is hard to achieve due to its volume of code (Doerfler, 2013) To record earthquake with various distances and magnitudes, the dynamic range of seismic data acquisition system should be further improved. A seismic instrument is constituted by seismometers, which is used to collect analog signals, and data acquisition system, which is employed to convert analog signals to digital signals for analysis. The noise level of seismometers is about 1uV when output voltage is ±20V, e.g., the dynamic range of ultra-wideband JCZ-1 seismometer and its successor - very broadband seismometer CTS-1 has exceeded 140dB (Cai, 2007, Cai, 2004). The dynamic ranges of high-resolution seismometers are more than 150dB, e.g., KS-2000 and CMG-40T, (Zeng, 2014). However, common data acquisition system can only achieve signal and noise rate (SNR) of 135dB at 50SPS (sample per second), there is the gap between the dynamic range of data acquisition system and the requirements of seismometer, which will cause low SNR and missing key seismic information stored in seismic wave. The limit of the resolution of data acquisition system can cause inaccuracy results of local measurements for geological and earthquake information.
To improve dynamic range of seismic data acquisition system, high resolution analog - to - digital conversion (ADC) should be employed in it. The resolutions of ADC have developed from 8-bits to 24-bit. Theoretical value of ADC dynamic range of with single 24-bit chips can achieve 144dB. But the realized data acquisition system with single 24-bit ADC only has low dynamic range, i.e., no more than 100-120dB (Bulgakov, 2001, Nash, 2012). 32-bit ADC electronic components have been developed, such as ADS1263 of Texas Instruments (TI) and LTC2500. However, the noise level can’t satisfy the requirements of seismic data acquisition system. Since the geological environments are commonly unpredictable, an optimal solution that uses high resolution seismic data acquisition systems to minimize noise and improve resolution (Vigh, 2014). According to Nyquist sampling theorem, Nyquist-rate data converters are achievable, e.g., successive approximation (SAR) ADC, time intersects ADC, and folding ADC (Huang, 2013, Namgoong, 2015). The convertible rate of abovementioned ADC is equal to Nyquist sample rate. However, the non-ideal conditions of circuit and component matching cause low resolution of Nyquist-rate ADC, which cannot exceed 120dB (Tual, 2016). To realize high resolution ADC for data acquisition system, Σ-△ oversampling and pipeline architecture ADC are needed. The detail principles of those two ADC architectures are elaborated in next section.