ASR&D Physical Sensor Development
ASR&D's primary physical sensor work focuses on wireless measurement of temperature and strain. Strain is best addressed using SAW sensor-tags as wireless interface devices to strain gauges, hence this work is described under the "sensor tag" section of this web site. The goal of ASR&D's temperature sensor work is development of a wireless sensor system consisting of 100 or more coded SAW temperature sensors and an electronic reader for measuring the sensors wirelessly. To date, ASR&D has successfully demonstrated a group of 32 coded SAW devices that are individually identifiable when all the devices are simultaneously operable within the field of view (FOV) of the reader.
In addition, the company has developed a new high sensitivity temperature sensor device structure, which allows the sensor designer to specify the desired temperature sensitivity of the device, rather than having it dictated by the piezoelectric substrate used. ASR&D is also developing a patented wireless electronic reader designed to automatically locate and measure multiple sensors in the FOV. Each of these three areas is discussed in greater detail below. A majority of the work described below has been (or is) supported by SBIR and/or STTR grants from NASA. Summaries of the work under these programs can be found in the NASA Reports section of the web site (under "Resources").
SAW coded sensor sets
Under a recent NASA Phase I SBIR (NNX10RA68P), ASR&D developed a set of coded high-sensitivity temperature sensors that incorporate both time diversity and frequency coding to produce 32 uniquely identifiable sensors. These sensors use a group of functionally orthogonal codes, and incorporate time diversity to allow a small group of "good" codes to be re-used in different time intervals, producing a larger group of individually identifiable sensors. During this six-month Phase I effort, ASR&D demonstrated a 32 device coded set consisting of 4 functionally orthogonal codes, each used 8 times in different time "slots" to produce 32 individually identifiable sensors. The center frequency of these devices was 250 MHz (chosen for good propagation through curing concrete), and the devices had a fractional bandwidth of approximately 25%. Wireless testing of the coded devices during the Phase I program was performed using an Agilent E5070B network analyzer as a signal source. The ENA signal was amplified and transmitted wirelessly using a small fractal antenna from fractus®. Figure 1 shows the antenna used, one of which was also used for each sensor device and one for the receive port of the ENA. The electrically small size of this antenna relative to a 250 MHz signal, combined with the fact that the antenna was designed to operate with a center frequency of 190 MHz (for UHF TV applications) led to significant antenna losses through the test path, and limited the wireless test range to under 16 inches.
Figure 1. Fractal antenna used for wireless tests.
The received signal reflected back from the group of sensors was amplified, passed through a matched filter designed to select a specific sensor code, and measured on the ENA. The red line in Figure 2 shows the wireless response measured for a group of four sensors with codes A, B, C, and D, measured using a matched filter for code B. The blue line corresponds to the response of this group of sensors when sensor B has been removed from the field of view of the ENA measurement system. Clearly, when sensor B is removed, the level of cross-correlation with the three remaining codes is quite low. Similar measurements using different matched filter devices confirmed the ability to utilize matched filters to read each code individually even in the presence of the combined RF response of a group of sensors with multiple codes. Once a set of four "good" codes was identified, sensors that utilize time diversity to "re-use" each code in 8 time slots were built. Figure 3 shows a typical response of a set of eight such sensors with code A. The matched filter response of these sensors (measured one at a time using the same matched filter) produces a set of 8 correlation peaks in the time domain, each in a distinct time slot. Interactions between an individual code at one delay and the same code in the other time slots are fairly low. The time slots are designed to be far enough apart that the change in delay of the correlation peak due to variations in temperature will not produce ambiguity in sensor identification.
Figure 2. Wireless measurement of sensors with four codes (A, B, C, and D) received with a matched filter for code B. Note that when the sensor with code B is removed from the measurement FOV, a low level of cross correlation is observed for the combined signals from the remaining coded sensors.
Figure 3. Measured autocorrelation of sensors with code A and eight different delays. Note the extremely low signal levels outside the narrow time range around the correlation peaks.
These preliminary results demonstrate development a set of 32 individually identifiable sensors that use both time and code diversity. Wireless measurements confirm the ability to detect any of the 32 sensors out of a group of sensors simultaneously operating within the FOV of the measurement system. This approach can be extended to produce sets of 100 individually identifiable sensors, and ongoing efforts at ASR&D encompass both expansion of the number of sensors and wireless system development.
SAW temperature sensors with design-controlled sensitivity
ASR&D has developed a high sensitivity SAW temperature sensor device structure that allows the design engineer to produce sensors with desired target temperature sensitivity. While the possible range of temperature sensitivities and operating temperature ranges for these devices are limited somewhat by material properties of the SAW substrate, there is a wide range of possible sensitivities. As part of the Phase 1 effort, temperature sensors were prototyped on Y-cut Z-propagating lithium niobate (LiNbO3), a piezoelectric substrate with an inherent temperature sensitivity of magnitude 94ppm/°C [Morgan]. Figure 4 shows the measured temperature sensitivity of one such sensor, which demonstrates a temperature coefficient of frequency (TCF) of +384ppm/°C over the range 20°C to 120°C, approximately four times the TCF characteristic of the substrate. Figure 5 illustrates the ability to achieve even higher temperature sensitivity, showing the measured response for a sensor with a TCF of about +680 pm/°C, more than seven times higher than the substrate TCF. Figure 6 illustrates the ability to produce sensors with both positive and negative TCFs, demonstrating a TCF of -165ppm/°C.
Figure 4. Measured temperature response of one SAW sensor on YZ-LiNbO3 shows a TCF of +384ppm/°C over the range 20°C to 120°C, approximately four times the TCF characteristic of the substrate.
Figure 5. Measured temperature response of another SAW sensor on YZ-LiNbO3 shows a TCF of +680ppm/°C, more than seven times the TCF characteristic of the substrate.
Figure 6. Measured temperature response of another SAW sensor on YZ-LiNbO3 during heating and cooling tests, showing a TCF of -165.2 ppm/°C.
The ability to produce negative or positive temperature coefficients over a particular temperature range is dependent on both the wave propagation characteristics of the substrate material and the device design, while the ability to achieve larger or smaller device sensitivity is dependent solely on device design. This provides the designer with the ability to develop sensors ideally suited for specific applications - with targeted device sensitivity for operation over the relevant environmental temperature range.
Wireless reader development
These sensors described above were designed to function with the spread-spectrum wireless sensor interrogation system currently under development at ASR&D. This system utilizes the correlative nature of the codes in each sensor to identify the sensor among a group of functionally orthogonal codes, and time diversity allows a small group of "good" codes to be re-used in different time intervals. Since the system uses mixing with a reference signal and integration that is performed at a single point in time, rather than traditional correlation, this system is inherently relatively immune to the code-collision problems that have plagued orthogonal frequency coded (OFC) and standard SAW coded tag devices. The system is also self-synchronizing, to allow identification of individual sensors with arbitrary RF delay due to position.
The bench-top prototype electronic reader system currently being developed at ASR&D under NASA STTR Phase 2 contract NNX09CB77C is a single-channel reader that can be manually adjusted to select any one sensor out of a group of up to 32 sensors within the field of view (FOV) of the reader. This breadboard system, which consists of a transmitter board and a receiver board, utilizes bench-top power supplies and is read with a digital oscilloscope. ASR&D has recently been awarded another NASA Phase 2 program that will modify the current reader to develop a miniaturized integrated wireless electronic reader capable of autonomously reading up to 32 sensors operating simultaneously within its FOV. A power supply, A/D, and DAQ modules will also be developed.