Two types of three-component (3C) sensors based on Micro-Electro Mechanical Systems (MEMS) technology are commercially available. We conducted a test comparing data from these two MEMS and data from conventional geophones. Although these sensors have different designs, we found that the differences between the processed datasets were small. Normalized root-mean-square analysis indicates that the sensor type may be less significant than other factors such as weather and shot repeatability.
The availability of MEMS-based 3C receivers has led to an increase in multicomponent seismic acquisition on land. 3C single sensors that tolerate tilt and measure tilt angles have greatly simplified field operations compared to systems with coiled geophones (Gibson et al., 2003). These characteristics coupled with increased bandwidth provide improved data quality. There are currently two MEMS sensors available, VectorSeis® (Maxwell et al., 2001) and DSU3 (Farine et al., 2003). These sensors employ different designs. To determine if there are data quality and operational efficiency differences between the two MEMS types, we conducted a field test.
The test data were acquired on a 2D line in the heavy oil province of northern Alberta. 278 stations of each sensor type were planted at an interval of 7.5 m in shallow drilled holes. The source was one-eighth kg charges at 15 m depth and 22.5 m interval. Manufacturer representatives attended the test to ensure proper operation of their systems. A preference to acquire the data concurrently had to be compromised in lieu of each manufacturer’s requirement to keep their equipment proprietary from the other. To accommodate this, two shot holes were drilled at every source location and the lines recorded on consecutive days; VectorSeis on day 1 and DSU on day 2. Due to the relatively small charge size it was judged that 4.5 m was a sufficient separation to avoid the first shot effecting the second. Acquiring the data on separate days with different shots introduced a margin of error in the subsequent evaluation. However, the conditions were similar enough for one to be confident that any significant difference in data quality would be above the margin of error introduced in the field experiment. In addition to the MEMS, data from geophones were also recorded. Out of the geophone data, 9 conventional groups of 6 10Hz geophones were available for comparison.
Figures 1 and 2 show comparisons of stacked PP and PS data. Sensor specification differences such as instrument noise level, dynamic range and overdrive behavior led to anticipation of larger variations in the data than were actually observed. It is not clear which sensor provides better stacked data. However, we think that the difference in data quality is below the margin of error, considering the two data sets were acquired on different days.
The data from the 9 geophone groups that were left in place overnight on the two days of the test provided an opportunity to assess the repeatability of land systems and subsequently compare the cumulative effects of several factor such as different sensor types, different shots repeated in almost the same locations, and random noise.