Advanced Linear Devices Inc., www.aldinc.com
Design engineers once viewed physical phenomena such as vibrations as nuisances to be eliminated. Today, though, they increasingly look at the occurrence of vibrations as an opportunity to harvest useful energy. The field of energy harvesting (EH) is maturing rapidly. Ultimately, the goal is to capture energy from any number of low-state sources, besides vibrations, including heat and light, wind, biomechanical movement, and even ambient radio frequency (RF) and electromagnetic (EM) radiation.
In fact, certain segments of energy harvesting (EH) technology aren’t that novel anymore. On a larger scale, wind and solar energy have been sources of harvestable energy for quite some time. Today’s real technical challenge is capturing extremely low levels of energy (on the order of microwatts) that are a by-product of normal operation. The process of capturing these tiny amounts of energy has been a tricky proposition. Designers who attempt this task often find that the energy required to operate the capturing unit exceeds the amount of energy available to be captured. Thus the challenge has been to reduce the power consumed by energy-harvesting electronics and make it practical to capture and utilize sources of extremely low power.
One factor that complicates the task of harvesting low levels of energy is that many widely used EH energy generators produce only low voltages, on the order of fractions of a volt. To be of practical use, the generator output must undergo a low-voltage boost, or a step up. The necessary low-voltage (LV) modules have only recently come on the scene and are largely responsible for the uptick in the EH industry.
At present, there are two approaches to low-voltage step-up that seem to be bubbling to the top. Both have merit, and the basis for selecting either one is based on a number of factors that specific applications may demand.
One approach is based on a “chip package” component that must be specifically designed-in to the end application. The basic idea is to use commercially available off-the-shelf (OTS) voltage regulator circuits and adapt them to specific energy harvesting sources. One problem is it may be tough to find OTS regulators that can be configured to work with the super-low voltage and current levels that characterize some kinds of EH sources.
So the first challenge to this approach is that the chip must be designed-in by someone with analog electronic circuit design expertise. Second, because it is a microchip, it must become part of the design package. If the designer wants to make the chip a stand-alone module, it must contain its own power source and power conditioning electronics. Third, it requires a relatively high input voltage, in EH terms, of at least 300 mV. This minimum threshold voltage makes it tough to harvest energy from ultra-low voltage energy sources because a significant amount of energy is lost in the voltage boost process. This is true especially if low levels of voltage or energy are present most of the time.
The second approach is to use a LVB (Low Voltage Booster) module between the EH module and the application requiring power. To the application, the LVB looks, electrically, like a battery and presents similar voltage characteristics. The designer simply must focus on the typical battery design parameters, such as output voltage qualities, Vh and Vl, and impedance at the application’s input. The difference between the two approaches is analogous to whether one wants to design a power supply rather than buy a power supply to add to a finished product.
Currently, there are several devices available that can work at 300 mV and above. The real challenge is in capturing low energy levels from sources of energy that generate 100 mV and lower. Systems that work at higher voltages can’t begin harnessing energy directly from nano and microwatt-level sources without first allowing the harvested energy to build up to a minimum threshold. It is critical that this minimum threshold be kept as low as possible, in terms of both voltage and energy levels, because harvesting systems can’t capture energy until the minimum threshold level is reached. Conversely, any energy generated below this level can’t be captured.
Sources that fall in the very-low-output range include photovoltaic (PV), piezoelectric, thermoelectric, biomechanical and even ambient radiation. Each has different qualities that bear on how harvesting systems can handle the energy each generates.
PV is a well understood and mature method of generating electrical power. Though PV arrays find common use in large-scale energy storage applications, single solar cells are well-suited for low-power EH applications. A single solar cell can optimally generate 400 to 500 mV. Once it starts conducting, the optimal power-generating voltage for the single solar cell drops to almost half that level. However, PV has well-known limitations. Obviously, light must be present to generate energy, and dust, dirt, clouds, and other sources of interference can reduce PV output.
Low-voltage piezoelectric sources that generate energy from vibration continue to gain traction. Strain or pressure on PZ crystals generates electricity in the low-milliwatt range. Such strain can come from any number of sources – motion, low-frequency seismic vibrations, acoustic noise, and vibration from engines or impact, such as the heel of a shoe hitting the ground.
One edge-of-the-envelope application for PZ energy harvesting is in micro-scale devices, as for harvesting energy from hydraulic systems. Here, pressurized hydraulic fluid drives a reciprocating piston attached to three piezoelectric elements. The pressure fluctuations are converted into alternating current.
Thermoelectric generators (TEGs) consist of two dissimilar materials and a junction between them that creates a thermal gradient. Most applications use several of these devices ganged together because generated voltage is 100 to 200 μV/K per junction. TEGs can typically capture milliwatts’-worth of energy from industrial equipment, structures, and even the human body. One nice thing about TEGs is that they can capture low heat differentials. Their key limitation is the low voltages they generate, especially when limited by heat sink capabilities.
The human body is capable of generating a wide range of energy that can be harvested. Joint movement, body heat and impact (walking, for example) are all potential energy sources. Human breath can be captured to provide air pressure, and sound or voice-box vibrations are also promising potential sources. So the capture of biomechanical energy can employ a mix of EH technologies.
Most ambient radiation in the form of radio-frequency waves (RF) contains little salvageable energy. However, RF is still an abundant source of energy. The challenge is in capturing workable amounts of it with a low-profile collector system. At present, the size of the collector system is an issue. It generally takes a sizeable antenna, perhaps specially made, to collect meaningful amounts of RF energy. RF collectors for EH are still a research topic.
It might seem absurd to say that a 300 mV threshold for an EH system isn’t “low. ” A more precise statement is that it isn’t low enough to take advantage of numerous free energy sources. Thus the need for a low-voltage “booster” module that can bring these near-zero energy sources up to the voltage levels required by available EH modules.
As of the writing of this article, there is only one low-voltage booster module commercially available. It is manufactured by Advanced Linear Devices and is designed to work with their EH300 modules. With the booster module in front of standard EH modules, the input source voltage can be below 100 mV and input power as low as 1 µW. This is thanks to the use of zero threshold MOSFETs that support the design of analog circuits operating on less than one microwatt. An onboard self-starting oscillator operates at a resonance frequency that depends on the source output impedance, source voltage level, and resonating components onboard the booster module. The oscillator waveform is coupled to a transformer, also resident on the module, which outputs an ac signal with power proportional to the input power level.
There are two significant parameters that define EH module voltage – minimum operating voltage (MOV) and minimum startup voltage (MSV). The MSV should be as close to zero as possible to handle the widest variety of energy sources. Of course, the output power is a function of the input power. Harvesting circuits work at their maximum efficiency and power transfer when the output impedance of the EH module matches the input impedance of the energy generating souorce to which it will be integrated.
Conventional complimentary metal-oxide (CMOS) technologies have been hard pressed to work in nanowatt or low-level microwatt environments. Practical circuits have been designed to work down to around 0.5 V, but with several constraints. The ALD LV module works in this range because of proprietary technology pertaining to the use of precision gate trimming during manufacture. Technology employed in the booster module is capable of stepping input voltages from as low as 60 mV up to an average of 6 V; i.e, a factor of 100.
One reason it is difficult to field EH systems is that many energy harvesting transducers weren’t originally designed as energy sources. Most thermal and vibration harvesters, for example, are sensors that have been repurposed to harvest energy. EH transducers will likely improve as harvesting becomes a mainstream technology. There is also a movement to bring disparate technologies together on a single substrate through use of such advanced techniques as three-dimensional transistors and wafer structures. This integration will usher in electronics that are more power efficient.
Such trends promise to help make energy harvesting a more widely used source of power for a plethora of electronic platforms. Perhaps, one day, harvested energy may even be the predominant source of power for mobile and remote sensor devices.
Inside an EH voltage booster
Chips that step up the output voltage from energy harvesting transducers have at their core EPAD MOSFET arrays. EPAD MOSFET stands for “Electrically Programmable Analog Device matched Pair MOSFET.” EPAD is an analog technology innovation by ALD which produces a CMOS MOSFET whose threshold voltage and on-resistance qualities can be electrically programmed to a precise level. Once programmed, the set parameters are indefinitely stored within the device even after power is removed.
EPAD technology employs a floating gate structure which can be precision trimmed to produce tightly controlled electrical qualities. A model of an EPAD is simply that of a regular MOSFET device, with all the electrical qualities as indicated on the datasheet. The floating gate can just be regarded as an embedded part of the EPAD MOSFET. Thus an EPAD can be modeled as a MOSFET with a floating gate between the gate control terminal and the MOSFET device. This floating gate can receive tiny, controlled bursts of electrons and trap them indefinitely.
The voltage booster contains a self-starting oscillator circuit that includes an on-board transformer that couples to a dedicated EPAD MOSFET Array. An input decoupling capacitor integrates and filters the input signal to drive the transformer’s primary winding core. An input ground voltage also turns on an EPAD MOSFET Array through the connection of a resistor to its Gate Input. A current flows through the primary winding of the transformer, coupling and developing a corresponding current in the secondary winding.
Upon being energized, a voltage develops across the secondary winding of the transformer. A small coupling resistor-capacitor network then provides negative feedback from the secondary winding to drive the EPAD MOSFET to an ‘off’ state. This RC network then charges the gate voltage of the EPAD MOSFET until it is again in an ‘on’ state. Once the EPAD MOSFET is turned on again, the cycle repeats itself and the circuit oscillates at a frequency that is determined by the source generator’s impedance, the output loading, the parameters of the RC network, and qualities of the EPAD MOSFET array and of the transformer. This ‘natural’ frequency also varies with varying input source impedance and the input voltages at the source, as well as the changing output loading.
The EH4205 module begins operating as soon as enough energy is available for the oscillator to start oscillating. This minimum self-starting energy starts boosting voltages at such a low energy level the module generally can capture very low levels of energy spurts before many other industry low-voltage booster modules would begin to function. For select members of the EH4200 MLVB Series of Modules, the oscillator can initiate oscillation at less than 1uW average input power.
The EH4205 primary output is an ac output that can go directly to the inputs of an EH300 series Energy Harvesting Module through a two-wire connection. Designers can feed the chip ac output to a bridge rectifier to produce a fullwave rectified dc output. The output of the full-wave rectifier can be used to drive an output dc load such as a trickle charger or to power an electronic circuit directly.