GNGTS 2013 - Atti del 32° Convegno Nazionale

the acquisition device. In order to avoid these errors the input impedance of the acquisition device should be much higher than the source impedance (Keithley, 1998). A four-channel JFET-input operational amplifier (Texas Instruments TL064) is used for this purpose, configured as a buffer amplifier (Jung, 2004). The JFET-input stage has an extremely high input impedance (1TΩ), much higher than the maximum expected contact resistance (in the MΩ range) typically measured in the field (Corry et al. , 1983). Moreover the low output impedance of the operational amplifier significantly reduces the error relative to the input offset current of the acquisition device. The TL064 features low input offset and input bias current thus further reducing the measurement errors. Furthermore, the TL064 is a low- power consumption device with a typical supply current of 200 µA per channel which makes it particularly suitable for a battery-powered system. Printed circuit boards, amplifier boxes, cables and data logger chassis were self-made and all parts were assembled in laboratory. The connections between devices were made by means of shielded multicore cables with multipole connectors. Two 6LR61 9 V batteries (550 mAh capacity) were used to power each TL064 with a current drawing of 0.8 mA and a runtime of about 28 days. One 12 V, 90 Ah battery was used to power the data logger. The current drawing of the data logger was about 200 mA and the battery life of about 18 days. The devices were tested in laboratory feeding a signal through each channel with a HP 33120A signal generator and verifying their performance (gain, distortion, frequency response and noise) with test equipment. The key benefits of the proposed system design are: i) the low bias currents at input stage which prevent the polarization of the electrodes, ii) the low-impedance output of the operational amplifier which decreases the time required for settling of the multiplexed acquisition device, reducing ghosting and crosstalk between channels, iii) the high-impedance input of the amplifier which prevents the source impedance act as a voltage divider across the input of the acquisition device (loading error). Electrodes. The measurements were performed with non-commercial electrodes (Fig. 1d): they were designed and realized in order to be cheap, resistant, high conductivity and easily portable. The basic requirements for the electrodes were durability, long-term stability, and low noise level. Usually non-polarizing porous-pot electrodes (typically Ag-AgCl, Cu- CuSO 4 , or Pb-PbCl 2 ) are used to perform self-potential measurements. Numerous authors have investigated electrode designs that reduce measurement errors and are stable over long periods of time (Perrier et al. , 1997; Clerc et al. , 1998; Petiau, 2000). Measurement errors on the order of several (~5) millivolts should be expected with modern surveying equipment, and this can be significantly reduced by installing the electrodes (semi) permanently (Perrier et al. , 1997; Perrier and Pant, 2005). Since graphite is an excellent electrical conductor and it is not affected by corrosion or electro-chemical effects, graphite electrodes are commonly employed for the application of electric fields such in electrokinetic remediation (Pazzi et al. , 2012). Electrodes made of graphite are also found commercially and used for geophysical monitoring purposes, to eliminate the damaging effects of corrosion and electrochemical degradation with time (Patent US 6,674,286). The electrodes were made of carbon rod 100 mm long with a diameter of 6 mm; at the top a 5 cm carbon rope (®Sigrafil D2-3K Cord, Ø = 19 mm) was fixed with a self-fusing silicon rubber tape and a stainless steel male-connection was protected with heat shrinks. The connection with the amplifier boxes were carried out by means of wires with a stainless steel female-connection on one end. Furthermore, to avoid chemical oxidation all the connections were isolated by means of heat shrinks. Graphite electrodes have been tested in laboratory in a 7-days experiment. The electrodes were placed in an experimental cell (Masi et al. , 2013) filled with quartz sand. The sand was saturated with tap water and the saturation was ensured by recirculating water inside 140 GNGTS 2013 S essione 3.2