Tweet

Posted on 29 June 2019

Remote Sensing Utilising Alternative Energy

Free Bodo's Power Magazines!

 

 

 

Harvest energy from the surroundings

Applications that require remote sensing often utilise alternative energy sources for power. This can be found in systems that monitor vibrations in bridges, the growth of trees in forests and seismic buoys. These energy sources usually are limited in the amount of power available so the design of these systems must be extremely frugal with power. In this article we will examine several methods for designing lower power sensor systems that take advantage of every microwatt yet provide outstanding analogue performance.

By Richard F. Zarr, MTS Technologist, National Semiconductor

 

Our civilised world is only possible through a vast and amazingly complex infrastructure that grows larger everyday. As it ages, weaknesses can appear that could lead to catastrophic events such as a bridge collapse or a wide scale power failure. Engineers are often tasked with devising ways to keep tabs on this network of roads and other crucial systems. One solution that is being deployed utilises active sensors with various telemetry relays via satellite uplinks, terrestrial radio relays or cellular systems to report the status of key elements in the network. By monitoring these elements, it is possible to know the condition of the system and provide required maintenance to prevent failure.

Active Systems Require Power

Active sensing systems provide a great advantage over passive monitoring. For example, passive sensors utilising RFID tags and strain gauges can be mixed with concrete and poured directly into structures - however someone with a RFID reader must excite these tags with near proximity RF fields in order to acquire the data. That means someone must drive to a remote location to monitor the condition of the structure. Active systems however can be retrofitted to older structures and report vibration and stresses via remote control.

The major drawback to active systems is that they require power to run the sensors, microcontrollers and uplink. Batteries can be employed (and often are as backup) however they will deplete with time and require replacement. An alternative is to harvest energy from the surroundings. For instance, a sensor deployed on a bridge that is well travelled can harvest energy from the vibration of passing traffic. This works in all weather and allows longer periods between monitor system maintenance (e.g. battery replacement).

Two things are very important for designers to keep in mind when designing these systems. First and foremost, power is everything. Without it, the system will fail, so efficiency of power conversion and storage is highly important. The second item on the list is the amount of power required to run the system. If the electronics draw more power than is available, the system also fails. So harvesting the most available energy, converting it in the most efficient way and utilising it very frugally is the challenge.

Lowering the Energy Footprint

Power supplies now routinely run at very high efficiencies utilising hysteretic conversion and sleep modes. These designs can provide 90% - 95% efficiency across a very wide range of load. This is important since to operate over long periods, remote sensors sleep a great deal of the time. The power supply however is still busy charging batteries or providing stand-by power to real-time clocks to periodically wake up the system and make a reading.

This method has been utilised for decades to extend the running time of equipment on batteries and still applies today. Figure 1 shows how a typical remote sensor power system might operate. Since reaching a satellite in low earth orbit may take several watts of RF power, the transmission time must be very short. If the power storage is depleted before the communications are complete, the information may be lost. The crucial calculation here is to make sure the power consumed during the sleep phase is low enough to make sure that enough energy is stored (worse case) for the active phase or the system will crash.

PWM Cycle of Power Consumption

The calculations are fairly simple. The total power is:

Total power equation

Where PT is the total power, PA is the active power, PS is the sleep power and DC is the system duty cycle (between 0 and 1). The harvested energy required during the storage phase must be larger than the product of PT and the period (seconds). Even though energy is being used during the depletion phase, the system can still be acquiring energy unless this mode is disabled during active operation. So:

Equation 2

Where EC is the energy harvested in joules during the charging phase, e is the storage efficiency (between 0 and 1), PT is the total power in watts, and T is the cycle period in seconds.

The Problem of Continuous Monitoring

The basic idea shown above is to store more energy than the system needs between the times it consumes it the most. In remote sensing the radio’s power amplifier will most likely be the largest consumer of power during the uplink or connection phase. However, there is a problem with putting the system to sleep in between these updates.

If the system merely needs to read the air temperature and humidity every 15 minutes and report this information, the entire device can sleep with only a real time clock (RTC) running. An RTC uses extremely low power and often has an interrupt pin which is programmable with an interval. This function can bring a system out of sleep (low power mode) to make the measurements, turn on the radio, uplink the data and return to sleep mode.

However, if a system needs to continuously monitor something such as detecting poison gas or transient overload conditions or even accumulating averages then this method does not work. Instead, the system needs to be continuously operating with the lowest power possible. For most microcontrollers this is not an issue, however for the analogue front ends that need to connect to sensors, this can be extremely challenging.

Typically, noise is the enemy. Systems that require a great deal of gain require low noise amplifiers. The challenge here is that low power op-amps typically have higher noise – this is due to the lower tail currents in the first stage of the amplifier. It is a delicate balance between power consumption and lower noise, so the selection of amplifiers that are used continuously must be done with care.

If the system contains an analogue to digital converter (ADC), depending on the frequency of interest, a filter is required to prevent aliasing (or signal mirroring) caused by the sampling. This phenomenon is caused when signals are present at the input of the data converter that are higher in frequency than the half the sampling rate. In sampling theory, this is called the Nyquist Rate which states that to accurately reproduce an analogue signal the rate of sampling must be twice the highest frequency component (or system bandwidth). Anything higher will appear or “alias” as a lower frequency component causing errors in the signal processing. The filter will also require power and if it is a continuous time filter (op-amp based), it will draw power even between the sampling cycles of ADC.

Yet another issue is the offset adjust and other sensor calibration requirements. Some calibration can be done in the microcontroller, but often requires an offset that supplies a stimulus to the sensor or bridge circuit. Gas sensors used to detect toxins can be particularly complex depending on the type of gas being detected. Often sensor manufacturers will provide example circuits depending on the user’s design requirements such as cost, accuracy and end application.

Integration and Tools

Since a large part of the complexity of these monitoring systems resides in the analogue front end, semiconductor manufacturers have begun to integrate much of the functionality into a single device. This has several advantages – one being the control of the signal path from the sensor all the way through the analogue to digital converter. Another advantage is the ability to control power more closely which is a key factor to remote sensing applications.

An example is the LMP91000 AFE Potentiostat which is designed specifically for low-power chemical sensing applications. This device uses a clever architecture that helps designers overcome many of the issues stated above. It contains all the required circuitry to connect directly to either 2 or 3 lead electrochemical cells and provide a calibrated output voltage. Due to the complexity of the many sensors available the LMP91000 is supported by National Semiconductor’s Webench tool called “Sensor AFE Designer”. This tool provides a designer the ability to select an end application such as ammonia sensing, the sensor manufacturer and type and then calculate all the required external components (where there are only a few) and the programming required.

Combine LMP91000 with an ultra-low power microcontroller such as Texas Instrument’s CC430 which includes the analogue to digital converter and most of the radio electronics and a complete remote gas sensing platform  can be created (see Figure 2).

Remote Gas Sensing Application

Conclusions

Remote sensing where the power to run the system is harvested locally can be an amazingly complex design challenge. The energy source can be intermittent and limited, power conversion and storage must be extremely efficient and the circuitry must be ultra-low power. Much of this heavy lifting has been accomplished by semiconductor manufacturers to aid designers in both quickly getting to market as well as meeting all the design challenges found in these sensor systems.

 

 

VN:F [1.9.17_1161]
Rating: 0.0/6 (0 votes cast)

This post was written by:

- who has written 791 posts on PowerGuru - Power Electronics Information Portal.


Contact the author

Leave a Response

You must be logged in to post a comment.