Wireless Connectivity Plays a Leading Role in Smart Meter Networks

As financial and environmental concerns over energy consumption drive deployment of smart grid technologies, engineers face increasing requirements for information by stakeholders across the grid. Energy producers and distributors require accurate data to inform capital investors as well as to optimize generating capacity, distribution infrastructure, and peak-load management. At the outer edges of the grid, homes and businesses require data on pricing and availability of low-cost power.

To gather, analyze, and distribute all of the required data, the grid relies on three key subsystems: data aggregators, smart meters, and residential/building energy gateways. On the utility side, data aggregators collect energy usage data from smart meters through automated meter reading (AMR) or advanced metering infrastructure (AMI) networks. On the consumer side, energy gateways manage local power utilization, providing mechanisms for building operators and homeowners to lower the cost of power consumption. While each serves a critical function for producers and consumers, the energy meter remains the focal point for power usage measurement and data distribution.

Smart energy meter design presents unique requirements for wireless connectivity on the grid side and on the consumer side. These designs require maximum interoperability within consumer energy management wireless networks, while ensuring maximum range and carrier-grade reliability for wireless communications on the grid side – requirements that dictate use of different approaches optimized for each environment. For engineers, available solutions from Analog Devices, Freescale Semiconductor, Texas Instruments, and others offer a ready solution for ensuring full satisfaction of the diverse needs for wireless connectivity of smart meters.

Sophisticated functionality

In its basic form, a smart meter design will comprise subsystems for energy measurement and for communications. A variety of specialized metrology solutions are available (see the TechZone article “AFE ICs Simplify High-Precision Energy-Metering Design"). Because metrology firmware may need to be qualified and securely isolated from ancillary meter functionality, smart meter designs often combine metrology processing chips or subsystems with an additional host processor to handle application requirements (Fig.1). For some applications, this additional host might have sufficient remaining headroom available after application requirements are accounted for to handle complex communications protocol stacks associated with standards such as ZigBee.

Figure 1: Energy meters combine metrology engines with increasingly sophisticated communications subsystems – often aided by an additional applications processor or dedicated communications processor. (Courtesy of Silicon Labs.)

In delivering the required bits of data to producers and consumers, energy meter designers can face communications requirements that differ significantly by country, region, neighborhood, and even specific location. Despite the promise of a universal standard for such connectivity, engineers must ensure smart meters offer sufficient functionality and flexibility to handle different connectivity mechanisms.

For communications, ZigBee has gained rapid acceptance as a dominant solution, but other mechanisms, including Wi-Fi, M-Bus, PLM, and sub-GHz ISM, are used for connecting on the consumer side. On the producer side, sub-GHz ISM has emerged as a dominant communications mechanism and serves as the basis for standardization in the Smart Utility Networks 802.15.4g standard. However, other communications mechanisms involving cellular communications are used, including GPRS and WiMax as well as wired mechanisms, particularly using OFDM power line modems.

Grid-side connectivity

For AMR or AMI grids, smart meter communications requires maximum range to ensure complete coverage over all meters within a particular geographical locale – a particular challenge in utility meter networks because meters are often located out of sight and in positions typically not particularly well suited for wireless communications. In particular, range is a critical requirement in these networks, where doubling range increases the coverage area by a factor of four, potentially reducing cost by the same factor because fewer grid access points or aggregators are needed. Consequently, proprietary sub-GHz ISM-based approaches remain particularly viable solutions for energy meter connectivity.

Sub-GHz ISM offers greater range with the same link power budget while reducing co-existence issues with other radio sources. The increased range available with sub-GHz methods becomes evident in examining the Friis transmission equation:



Where
P_t = transmitted power
P_r = received power
G_t = transmitter antenna gain
G_r = receiver antenna gain
λ = wavelength
d = distance between transmitter and receiver

Thus, for a fixed transmit power, P_t, the received power will decrease with the square of the distance, d, and increase with the square of the wavelength, λ. For engineers, this translates into reduced power requirements for wireless applications using lower frequencies.

Furthermore, proprietary communications protocols can strip away more generalized layers and even unneeded elements in packet envelopes. As a result, engineers can utilize more modest MCU-based configurations because proprietary radio protocols will typically require less memory and fewer processor cycles thanks to the reduced overhead. For example, compared to a fully-compliant ZigBee design, typically requiring a 128 Kbyte stack, a sub-GHz solution could get by with stack less than 8 Kbytes in size. Combined with the lower processing requirements, the final design would cost substantially less than the full ZigBee-compliant alternatives.

Engineers can take advantage of a broad array of sub-GHz transceivers for building effective communications subsystems. Furthermore, among available low-power sub-GHz solutions, MCU-based SoCs offer plenty of memory to handle implementations of sophisticated proprietary protocols. For example, the Silicon Labs Si102x/3x Low Power Wireless Microcontroller series combines an ISM EXRadioPRO transceiver with an 8051-compatible MCU and up to 128 Kbytes of on-chip flash memory with 4 or 8 Kbytes of on-chip RAM.

Consumer connectivity

For connecting to home networks in general, and energy gateway devices in particular, ZigBee Smart Energy Profile (SEP) 2.0 offers the required set of features, as well as vendor independence needed to facilitate use across multiple home appliances and the like. For designing compliant energy meters, engineers can choose from a growing array of system-on-chip (SoC) ZigBee solutions (see the TechZone article "Wireless MCUs Simplify Design of Smart Meters and Energy Management Systems").

As with other devices in this class, the Freescale Semiconductor MC1322x Platform-in-a-Package (PiP) offers specific features needed to meet ZigBee processing demands. For example, the MC1322x family includes hardware blocks to accelerate IEEE 802.15.4 mechanisms including AES encryption/decryption and MAC acceleration. The MC1322x advanced security module (ASM) is a dedicated on-chip hardware subsystem designed to accelerate encryption/decryption using the Advanced Encryption Standard (AES). The ASM provides Cipher Block Chaining (CBC) and Counter encryption, completing each in separate encryption sequences of 13 clock cycles.

Designed to offload the CPU from baseband protocol processing tasks, the MC1322x's dedicated MAC accelerator (MACA) includes a sequencer/controller, TX and RX packet buffers, DMA block, frame check sequence (FCS) generator/checker, control registers, and associated timers (Figure 2). The MACA handles packet assembly for packet transmission and preamble recognition and FCS for packet reception – as well as packet acknowledgement, poll sequences, and clear channel assessment (CCA) independent of the ARM core. In fact, the included DMA circuit moves data between the MACA buffers and RAM without any CPU intervention.

Figure 2: The Freescale MC1322x platform MAC accelerator (MACA) provides dedicated hardware that offloads low-level packet processing tasks from the host processor. (Courtesy of Freescale Semiconductor.)

For communications, the software development effort can easily overtake the hardware effort thanks to the relative ease of designing with available SoCs. ZigBee IC manufacturers address this concern through a wide variety of software development capabilities, ranging from ZigBee stacks delivered as object code and software libraries delivered as source code to comprehensive software development toolkits. For example, Freescale’s packages in support of ZigBee development include a graphical user interface (GUI) and various code bases comprising wireless networking libraries, application templates, and sample applications (Figure 3).

Figure 3: ZigBee IC manufacturers typically offer full-feature software development environments that provide debug tools, protocol software stacks, and code libraries. (Courtesy of Freescale Semiconductor.)

Freescale offers a MAC stack for both proprietary networks and ZigBee stacks. The Freescale Simple Media Access Controller (SMAC) is a basic C-based code stack delivered in source code. With a stack code size fewer than 4 Kbytes, the SMAC enables engineers to develop efficient, MC1322x-based proprietary RF transceiver applications. In addition, Freescale has an 802.15.4 Standard MAC as object code, plus a full ZigBee stack with support for ZigBee Smart Energy Profile and ZigBee Home Automation Profile, among others. Still another stack, the SynkroRF stack supports proprietary 802.15.4 applications requiring only about 32 Kbytes of memory. A SynkroRF demo kit is also available to designers.

Smart Utility Networks

Using sub-GHz transceivers as standalone devices or integrated in SoCs, engineers can build proprietary networks using available basic MAC stacks or using simple stacks based on IEEE 802.15.4 MAC/PHY standards. The 802.15.4g Smart Utility Networks standard extends basic 802.15.4 communications with capabilities designed specifically to meet the specialized requirements for smart meter grid connectivity to ensure interoperability of millions of devices in mesh, star, point-to-point, or any other topology.

Existing standards such as 802.11 Wi-Fi and 802.16 WiMAX fall short in features such as operating range, considered critical for smart meter networking. For example, Wi-Fi and WiMAX are optimized to achieve maximum data rate at a particular bandwidth channel at the expense of range.

Existing wireless protocols are also less effective in supporting burst asynchronous communications from nodes. The density of wireless smart meters can be high, with thousands of devices per square kilometer in urban areas. At the same time, the nature of the communications from these devices is often event driven, potentially occurring simultaneously and at high priority.

An example of this behavior occurs during service outage detection, when an electric meter loses mains power and emits a 300 ms "last gasp" message. Poll-based or connection-oriented protocols will likely miss traffic particularly when confronted with asynchronous bursts from many nodes within a short period of time. Similarly, utility meter networks can require frame sizes of 1,500 in length, which exceeds the limits of classical 802.15.4 protocols. Also, 802.15.4 error detection is based on a two-byte CRC, which could result in undetected errors in utility meter networks expected to face large numbers of packet in normal operation.

802.15.4g was designed to address these concerns and ensure reliable communications for the largest possible numbers of devices consistent with local regulations and the need to ensure data rates of at least 40 kbits/s. Engineers can take advantage of available devices such as the Analog Devices ADF7023 ISM transceiver IC or the Texas Instruments CC1120 ISM transceiver IC targeting 802.15.4g-based applications.

The Analog Devices ADF7023 ISM transceiver IC is designed for low-power operation in the worldwide license-free ISM bands at 433, 868, and 915 MHz. At 50 kbits/s, the ADF7023 radio offers receiver sensitivity of –106.5 dBm and transmitter output power that is programmable in the range –20 to 13.5 dBm. The transmit RF synthesizer comprises a VCO and fractional-N PLL, while the receiver uses an automatic frequency control (AFC) loop that enables the PLL to find and correct any RF frequency errors in the recovered packet. Analog Devices offers a patent-pending image rejection calibration scheme designed to eliminate the need for an external RF source or for any further user intervention once initiated. Engineers can store calibration results in nonvolatile memory for use on subsequent transceiver power-up events.

The device supports generic packet formats that are fully programmable. In transmit mode, the device can add administrative bytes to the payload data stored in packet RAM. In receive mode, the communications processor can detect packets, store them, and interrupt the host processor for further action. To handle requirements beyond those supported within device firmware, the device offers a serial port (sport) mode that allows engineers to put the host processor in full control of packet structure, bypassing the device's packet management features entirely. This flexible operation allows engineers to support a variety of specialized packet methods such as longer CRC and packet lengths specified in 802.15.4g.

The Texas Instruments CC1120 ISM transceiver IC requires only a few external components to provide a complete wireless solution (Figure 4). The IC features receiver sensitivity of –110 dBm at 50 kbits/s and programmable output power up to 16 dBm in 0.4-dB steps. The device features a fractional-N-based frequency synthesizer with direct synthesis of the RF frequency on the Tx stage and automatic gain control on the Rx stage. Designed for flexible support of proprietary and standard protocols, the device allows engineers to program the device to support arbitrarily long packet lengths even while still using the device's packet-handling hardware support functionality.

Figure 4: Along with decoupling capacitors (not shown), the Texas Instruments CC1120 ISM transceiver IC requires only a few external components to deliver a complete wireless solution in ISM frequency bands. (Courtesy of Texas Instruments.)

Summary

For engineers, smart meters present significantly different requirements for wireless connectivity on the grid side versus the consumer side. While interoperability concerns dominate on the consumer side, range and reliability remain of most concern on the grid side. By exploiting the advantages of ZigBee SEP for consumer requirements and sub-GHz wireless connectivity for grid communications, engineers can meet demands for interoperable and robust smart meter connectivity.