Thermal sensors based on the seebeck effect

Over the few next decades, the personal vehicle may evolve into a device distinct from what exists today. This thesis considers energy and communication systems among vehicles that utilize two developing technologies: battery-powered drive trains and accurate sensors. We use linear optimization to construct separate algorithms for networks of Plug-in Electric Vehicles (PEVs) and networks of Sensor-equipped Vehicles (SVs). Plug-in electric vehicles will have flexible charging options, and may be capable of transmitting electricity back to the grid (i.e., discharging). We construct an automated mechanism for a fleet of PEVs that efficiently organizes distributed energy trading to benefit both the consumers and the electric utilities. A linear programming model of the fleet provides a composite valuation, which can be used in an online environment managed by a fleet aggregator to allocate feasible energy exchange schedules that decrease the peak electricity demand and reduce the cost to consumers. The resulting charging and discharging schedules are assigned to tens of thousands of vehicles instantly as they plug into the grid and are robust to unexpected events in driving patterns. We give empirical results based on electricity and gasoline pricing, electricity demand, vehicle characteristics, and driving behaviors. Sensor-equipped vehicles use accurate sensors to detect their surroundings; such technology has been used to help improve traffic flow and safety. We consider a network of SVs that measure the distances between one another. In this framework, a semidefinite programming (SDP) relaxation is an efficient computational method to solve the sensor network localization problem of determining the locations of vehicles on the road given some of the pair-wise distances between them. We present two results about the SDP relaxation that can be applied to localizing sensor-equipped vehicles. First, we provide the first non-asymptotic bound on the required sensor communication range to ensure a unique localization of the sensors. Second, we show that if the graph induced from a network of n sensors in dimension d is a (d + 1)-lateration graph with points in general position, then the graph admits a positive semidefinite stress matrix with rank n, and hence the SDP relaxation will produce the correct localization of the points.

To what degree.

A high-frequency root-mean-square (r.m.s.) converter composed of two matched thermal elements and on-chip circuitry has been fabricated in a foundry CMOS process with post-process micromachining techniques. An anisotropic undercut etch in conjunction with an electrochemical etch stop is used to form the thermal elements of the r.m.s. converter, which are suspended single-crystal silicon regions containing temperature-sensing diodes heated by polysilicon resistors. One of the elements is heated by the a.c. signal, while an integrated continuous-time feedback network maintains a second element at the same temperature. The converter has a packaging-limited — 3 dB bandwidth of 415 MHz, an r.m.s. dynamic range of 53 dB, nonlinearity of 1%, a quiescent power dissipation under 1 mW, and occupies an area of approximately 400 μm × 400 μm. The thermoelements have a thermal resistance of up to 37 000 K W −1 in air. In addition to the r.m.s. converter itself, an improved wet etchant chemistry is presented. Ammonium persulfate, when added to tetramethyl ammonium hydroxide (TMAH) silicon etchant, improves the silicon etch rate and the (100) surface quality. This etch chemistry should be useful for a broad range of CMOS-compatible micromachining applications.

Diode-based thermal r.m.s. converter with on-chip circuitry fabricated using CMOS technology

A novel approach is presented for measuring the parameters of a nonstationary high-frequency signal. This type of signal was experienced in actual recordings of voltage waveforms of a large industrial load in power systems. The approach is based on the extended Kalman filtering technique. The parameters to be measured are the magnitude, frequency, and rate of decay. The results indicate that the proposed algorithm optimally tracks the frequency, its time variation, and all other parameters. >

Measurement of the parameters of slowly time varying high frequency transients

An integrated electrostatic true rms converter has been developed. It uses a compatible surface micromachining post-process, which combines the mechanical structure with standard electronic readout circuitry, thereby creating an integrated measurement system. The converter uses two identical bridge structures, which are capacitively driven. The system generates an rms voltage V/sub rms/, which forces the second bridge structure to give the same deflection as the one driven by the ac-input voltage. When a feedback loop is used, the converter can achieve a high accuracy. The converter features a high bandwidth compared to existing fully electronic converters and a very high input impedance compared to existing thermal converters. >

/pdf/integrated-micromachined-electrostatic-true-rms-to-dc-n46ox00zwp.pdf

Integrated micromachined electrostatic true RMS-to-DC converter

Development of a new monolithic thermal sensor, a unique high-impedance attenuator, a precision sample-and-hold circuit, and a novel ac-error-correction technique have brought significant improvements in the accuracy of ac measurements. A precision thermal truerms measurement system is described that offers a transfer stability of 25 parts/106 in the frequency band 40 Hz-kHz. A microprocessor-based system uses a nonvolatile memory for software calibration; this enhances dc, ac, and resistance measurement accuracy, including production and repair.

Benjamin T. Brodie

Papers

Recent Advances in Precision AC Measurements