Dirk Spreemann

A CMOS integrated voltage and power efficient AC/DC converter for energy harvesting applications

In this paper, a fully CMOS integrated active AC/DC converter for energy harvesting applications is presented. The rectifier is realized in a standard 0.35 µm CMOS process without special process options. It works as a full wave rectifier and can be separated into two stages—one passive and one active. The active part is powered from the storage capacitor and consumes about 600 nA at 2 V supply. The input voltage amplitude range is between 1.25 and 3.75 V, and the operating frequency range is from 1 Hz to as much as several 100 kHz. The series voltage drop over the rectifier is less than 20 mV. Measurements in combination with an electromagnetic harvester show a significant increase in the achievable output voltage and power compared to a common, discrete Schottky diode rectifier. The measured efficiency of the rectifier is over 95%. Measurements show a negligible temperature influence on the output voltage between −40 °C and +125 °C.

Efficient Energy Harvesting With Electromagnetic Energy Transducers Using Active Low-Voltage Rectification and Maximum Power Point Tracking

This paper reports on efficient interfacing of typical vibration-driven electromagnetic transducers for micro energy harvesting. For this reason, an adaptive charge pump for dynamic maximum power point tracking is compared with a novel active full-wave rectifier design. For efficient ultra-low voltage rectification, the introduced active diode design uses a common-gate stage in conjunction with supply-independent biasing. While this active rectifier offers low voltage drops, low complexity and ultra-low power consumption, the adaptive charge pump allows dynamic maximum power point tracking with implicit voltage up-conversion. Hence, efficient energy harvesting with high-resistive transducers, e.g., electromagnetic generators, becomes possible even at buffer voltage levels far above actual transducer output voltages. Both interfaces are fully-integrated in a standard 0.35 μm twin-well CMOS process. The designs are optimized for sub-mW transducer power levels and wide supply voltage ranges. Thus, these presented transducer interfaces are particularly suitable for compact micro energy harvesting systems, such as wireless sensor nodes or medical implants. The active diode rectifier achieves efficiencies over 90% at a wide range of input voltage amplitudes of 0.48 V up to 3.3 V. The adaptive charge pump can harvest with a total efficiency of close to 50%, but very independent of the actual buffer voltage. This charge pump starts operating at a supply voltage of 0.8 V, and has an input voltage range of 0.5 V-2.5 V . Finally, results of harvesting from an actual electromagnetic generator prototype are presented.

Non-resonant vibration conversion

The development of distributed wireless sensor systems for automotive, medical or industrial monitoring applications is one of the aims for MEMS technology. For applications where environmental vibrations are present, the harvesting of this kinetic energy is an opportunity to power remote sensor nodes. For the conversion, typically resonant spring–mass–damper systems are considered. In this paper, a novel non-resonant conversion mechanism is presented. Depending on the geometry of the harvester and the vibration, this conversion mechanism shows a few advantages: low frequencies can be converted, higher or lower modes of vibration will be converted instantaneously, the transducer has 2 DOF for energy conversion and the generation of energy is not limited to a small frequency band. Based on a vibration amplitude of 100 µm, the behavior of a fine-mechanical generator and a MEMS generator has been simulated. The results of the fine-mechanical generator were verified by measurements of a prototype with 1.5 cm3 volume. So far the transducer is capable of producing 0.4–3 mW for vibration frequencies ranging from 30 to 80 Hz.

Numerical optimization approach for resonant electromagnetic vibration transducer designed for random vibration

This paper presents a design and optimization strategy for resonant electromagnetic vibration energy harvesting devices. An analytic expression for the magnetic field of cylindrical permanent magnets is used to build up an electromagnetic subsystem model. This subsystem is used to find the optimal resting position of the oscillating mass and to optimize the geometrical parameters (shape and size) of the magnet and coil. The objective function to be investigated is thereby the maximum voltage output of the transducer. An additional mechanical subsystem model based on well-known equations describing the dynamics of spring–mass–damper systems is established to simulate both nonlinear spring characteristics and the effect of internal limit stops. The mechanical subsystem enables the identification of optimal spring characteristics for realistic operation conditions such as stochastic vibrations. With the overall transducer model, a combination of both subsystems connected to a simple electrical circuit, a virtual operation of the optimized vibration transducer excited by a measured random acceleration profile can be performed. It is shown that the optimization approach results in an appreciable increase of the converter performance.

Energy Harvesting and Chip Autonomy

Energy harvesting micro-generators provide alternative sources of energy for many technical and personal applications. Since the power delivered by such miniaturized devices is limited they need to be optimized and adapted to the application. The associated electronics not only has to operate at very low voltages and use little power it also needs to be adaptive to the fluctuating harvesting conditions. A joint development and optimization of transducer and electronics is essential for improved efficiency.

Coil Topology Optimization for Transducers Based on Cylindrical Magnets

The previous chapters have been concerned with the optimization and comparison of eight different coupling architectures in electromagnetic vibration transducers. In summary, geometrical dimensions were found which yield to a maximum output power and output voltage, respectively. The comparison of the maximum performance limits yield the most efficient architectures which should consequently be favoured in the application whenever possible. However a basic characteristic of all the architectures (independent of the architecture class) is that the topology of the coil has always been predefined to be cylindrical. Hence the underlying optimization approach is strictly speaking a sizing optimization. Obviously this makes sense because cylindrical coils, especially made of enamelled copper wire, are state of the art and easy to fabricate. Moreover the optimized dimensions (especially for the “Magnet across coil” architecture class) show that the resulting coils are rather thin. Consequently there is not much space left for an optimization of the coil topology. But for all that an interesting question arises from this:

Basic Analytical Tools for the Design of Resonant Vibration Transducers

The presented review of existing work on electromagnetic inertial vibration transducers in Chap. 1 shows that there has been much interest in the design of vibration energy harvesting devices. Consequently excellent work has been done by numerous research facilities and a multiplicity of micro– and centimeter scale prototype vibration transducers has been developed. The basic analytical theory behind most of the presented devices is commonly known in the energy harvesting society. It is based on a well understood linear second–order spring–mass–damper system with base excitation. Specific analysis of vibration transducers was first proposed by Williams and Yates [15]. Since then the theory has been modified and described in various ways even though the basic findings are more or less the same. In this respect, an analytical expression for the maximum output power that can be extracted from a certain vibration was derived (also for constrains such as the limitation of the inner displacement of the seismic mass [64]) and the optimization of parameters such as the optimal load resistance or the electromagnetic damping factor was discussed. However, as will be shown, in most of these cases it is rather difficult even impossible to use the results of the analytical modelling directly for the design process of application oriented developments. This is because the theory does not consider geometrical parameters and is based on simplifying assumptions which often do not correlate well with the “real world” (e.g. random vibration instead of harmonic excitation, complex load circuit instead of simple resistance or appreciable magnetic flux leakage instead of homogeneous magnetic field distribution). However the analytical modelling is useful for understanding the influence of the most important system parameters. Furthermore it offers a deeper insight into the overall system behavior.

Power and Voltage Optimization Approach

The previous chapter presented the theoretical basis for the analysis of electromagnetic vibration transducers. Therein closed form solutions for first order power estimations have been obtained based on results from literature. These expressions consider harmonic excitation at one single frequency. Because many realistic vibration sources have a rich spectral content, it has been shown how the results can be used to identify most energetic frequencies and how first order power estimation can be performed also in case of random vibration sources. A popular parameter optimization approach in the analytical analysis considers the ratio of the electromagnetic to parasitic damping factor. A slightly advanced model with a constrained construction volume condition has shown that this commonly discussed optimization approach is important for estimating the maximum possible output power but has only marginal relevance for the design of electromagnetic vibration transducers. As demonstrated in Chap. 2 it is more reasonable to find the geometrical parameters of magnet and coil which yield the maximum output power instead of optimizing the damping ratio. The analytical model for comprehending this takes the magnetic field on the centre axis of the magnet into account which is still a simplifying assumption since leakage field effects are neglected. Beside this, the results are restricted to specific architectures and cannot generally be applied to all the architectures presented in the literature overview in Chap. 1. For this reason a very fundamental question to be answered is how the optimized geometrical parameters can be obtained for the different coupling architectures and do architectures exist that have inherently a higher output power and output voltage generation capability?

Optimization Results and Comparison

The previous chapter introduced and classified the considered electromagnetic coupling architectures. Overall boundary conditions for centimeter scale vibration transducers were defined. In order to compare the performance limits of the architectures these boundary conditions are applied to all architectures. Then an optimization approach was formulated to assess the optimal dimensions with respect to the output power and the output voltage. For architectures with a 6–dimensional search space the evolution strategy optimization technique is used. Furthermore the architecture specific calculation of the magnetic flux gradient was discussed. The calculation of architectures without back iron is based on Maxwell’s equations whereas the calculation of architectures with back iron is based on static magnetic 2–dimensional FEA.

Application Oriented Design of a Prototype Vibration Transducer

Chapter 4 confirmed important results from the optimization approach in Chap. 3: Regardless of the considered coupling architecture there are dimensions for magnet, coil and (if existent) back iron components which result in a maximum output performance. Because there are separate optima for the output power and output voltage every electromagnetic vibration transducer can be designed either as a voltage– or a power source. Moreover there are architectures which inherently have a better output performance. These architectures should be preferred whenever possible. However, because different design constraints will apply even these architectures must be optimized for each application. Therefore, the optimization procedure developed provides a tool for the development of application oriented electromagnetic vibration transducers. To demonstrate the benefit of the optimization approach in the design process this chapter outlines the development of a prototype vibration transducer based on architecture A II. This architecture was chosen because of the simple assembly and the good voltage generation capability. However a basic drawback of this architecture is that the magnetic field is not channeled with back iron part. Hence the oscillator is sensitive against nonlinear forces generated by ferromagnetic components in the environment. Moreover, eddy currents will take place in metal components that are close to the generator reducing its power output.