Many of today’s modern building automation systems depend on wireless connectivity to ease installation and allow fast modifications or expansion. But in the process of removing wires, designers must rely on batteries as a power source, which affects the total system cost by adding maintenance to periodically replace cells. The challenge in these systems is to build highly efficient power management schemes that, in addition to saving battery life, ensure proper operation in life-safety applications.
Related Articles
- Energy Harvesting Powers Wireless Sensor Networks In Industrial Apps
- Energy Harvesting And Wireless Sensor Networks Drive Industrial Applications
- Power Your Wireless Sensors For 40 Years
Introduction
According to a 2013 “Markets and Markets” forecast1 as well as other sources, the global building automation market will grow to almost $50 billion by 2018. This growth is fueled by the ever-increasing need to make our work and living spaces safer, more comfortable, and efficient.
The expansion of this market is partly due to the introduction of wireless sensors, which allow easy installation, expansion, or modification. A large portion of the installation cost is labor associated with cabling along with rising copper prices. Wireless automation infrastructures can reduce these costs dramatically, though they introduce a long-term cost of ownership—without wires, systems must run on batteries.
This file type includes high resolution graphics and schematics when applicable.
Battery replacement costs vary, but it can be labor-intensive, depending on the location of the sensor. For instance, the sensor could be mounted on a 20-foot ceiling or another hard-to-reach location. In an ideal world, the battery would never be replaced in the useful life of the sensor. However, that is often unrealistic due to the aging of cell chemistry or the sensor’s power drain. For most wireless building automation systems, a five-year cell life is a minimum, and 10 to 15 years is advantageous.
Mitigating Energy Los
Designers can focus on several areas to help extend the period between battery replacements. The first is to understand the RF environment where a sensor will be required to operate. In most cases, it will be in a network with other sensors. The nodes can be arranged as a “mesh” network where some sensors are an end-point or function as repeaters.
This configuration is very energy-efficient due to reduced RF transmit power, which can be the largest consumer of energy in the sensor. By lowering the transmit power and allowing routers to repeat messages, an extremely low-power network can be formed. This topology often is used in wireless networks based on IEEE 802.15.4, such as ZigBee or 6LoWPAN.
An issue with mesh networks is synchronization, as seen with ZigBee. Router nodes must be available at all times. So if a node originates a message or a message is sent to a node, then the message is forwarded or stored and forwarded when the node is available.
For ultra-low-power devices, power is greatly reduced during a “sleep” cycle where most of the device is powered down, including the RF receiver. In this scenario, the router nodes must be powered (for example, ac power) to monitor for messages from the sleeping nodes when they are awake. Here, power consumption is highly asymmetrical. Sensor nodes are extremely low-power and sleep for long periods, and router nodes (which also could be sensors) are powered continuously and have their receivers active.
A complete ultra-low-power network without routers that are line-powered requires very precise timing, which can lead to increased cost and complexity in each node. The entire network wakes within a tight window, communicates, and then sleeps again. The longer the duration between cycles, the tighter the timing requirements. If a node loses synchronization with the network, then it needs to stay awake until the next waking cycle to resynchronize—something to be avoided in a battery-powered network.
In most practical wireless networks, the router is powered continuously from some constant source such as line power or more likely power over Ethernet (PoE). In this case, the nodes have the burden of being as efficient as possible with their battery power. Typically a microcontroller powers down the radio and anything non-essential. In the case of a sensor, all of the analog front-end (AFE) electronics are powered down as well. Then the microcontroller enters a low-power mode using a timer to periodically wake up, power up the system, send any appropriate messages, and then reenter sleep mode.
The duty cycle of these sleep and wake cycles determines the power consumed. The equation shows the power calculation for estimating battery life based on the sleep and wake power consumption:
Run time (T) in hours is based on duty-cycled power states. It is calculated as the ratio of available energy (EA) in watt-hours to the power consumed (P) in watts. This can be further expanded.
Capacity (C) is in amp-hours (Ah), VS and VE are the starting and ending voltages respectively during discharge, PW and PS are power consumed in watts during wake and sleep cycles respectively, and D is the duty cycle (0 to 1) of the waking period. A de-rating factor (α) is used to adjust the battery for a loss in capacity in applications with an extended service life (greater than five years). The efficiency (eff) of the power converter stage is also taken into account since it degrades performance. Typical efficiencies range between 80% (eff = 0.8) to 95% (eff = 0.95).
Part of this design is selecting a power source that will last 10 to 20 years without significant degradation and a power converter (for example, switching regulator) that is extremely efficient at low power. The first problem can be addressed by a battery chemistry of lithium thionyl chloride (Li/SOCl2). This chemistry has been around since the 1970s and has an extremely long service life (10 to over 25 years). Li/SOCl2 has been used successfully in remote meters and other battery-powered wireless systems. Cells have a typical voltage of 3.6 V and exhibit an excellent operating temperature range (–55°C to 125°C).
When using a single lithium-based cell, a sensor-node design may either boost the 3.0- to 3.6-V output to 5.0 V or use a buck-boost converter such as the TPS63001, which has a fixed output of 3.3 V and can deliver up to 800 mA in all conditions (buck or boost). This is important during the active state since RF transmitters may require significant instantaneous current. More importantly, the converter is mostly unloaded during the sleep cycle and must have a feature to automatically enter a pulse frequency mode (PFM) or some other pulse-skipping technique to conserve power.
Another source of energy loss is keeping the microcontroller active during the sleep cycle, even in low-power mode. At a minimum, a timer must be running with the main core shut down to conserve energy, but even this configuration may draw several micro-amps. Even state-of-the-art, low-power microcontrollers such as the MSP430 hover around 0.3 µA of current while in standby mode.
One novel solution is to use timer devices specifically designed for long sleep periods. The TPL5000 from Texas Instruments has programmable dividers to provide wakeup pulses with periods up to 64 seconds. It also has an extremely low power consumption of only 30 nA while running. With extremely long sleep cycles, this can add as much as two years to the service life of a battery-powered wireless sensor (see the figure).
Conclusion
With the proliferation of battery-powered wireless networks, installers and owners are looking for longer battery life—some with requirements for never replacing the battery during the entire network life (25 years or longer). Battery chemistries such as Li/SOCl2, high-efficiency PFM converters, and novel nano-power long-period timers can maintain extremely low power consumption during device sleep cycles, resulting in wireless building automation systems that rival their wired counterparts.
References
1. Building Automation & Controls Market (2013 – 2018): By Product (Lighting, Security & Access, HVAC, Entertainment, Outdoor, Elevator Controls, Building Management Systems (BMS)), Application & Geography (Americas, Europe, APAC, And ROW), marketsandmarkets.com, February 2013, Report Code: SE 1625.
2. For more information about:
• Programmable timers, visit www.ti.com/tpl5000-ca
• High-efficiency switching regulators, use http://www.ti.com/battery-ca
• Wireless connectivity, visit http://www.ti.com/wireless-ca
Richard Zarr is a technologist at Texas Instruments focused on high-speed signal and data path technology. He has more than 30 years of practical engineering experience and has published numerous papers and articles worldwide. He is a member of the IEEE and holds a BSEE from the University of South Florida as well as several patents in LED lighting and cryptography.