Smart meters are rapidly evolving with different architectures (and different regulatory requirements) in the global market. Since they are in the process of being rolled out to utility customers by billions, there is great interest – and great rewards – in successful smart meter design in its most basic form, the meter provides energy and power measurement, data transmission, real-time clock maintenance and data display on the front panel of the meter.
Key design requirements for smart meters include the following: (1) they should operate at low energy to enable them to operate on battery power for extended periods of time, and (2) they must include security features that protect the content of communications and secure stored data. It also provides one-way communication, which enables power suppliers to automatically and remotely read meters using different communication solutions, including RF wireless, power line carriers, and data communication for the General Packet Radio System (GPRS).
Smart meters with Advanced Metering Infrastructure (AMI) architectures provide two-way communication and provide improved reliability and accuracy, as well as monitor outages, and provide the benefit of the ability to disconnect remotely and the option to increase variable tariff full load to incentivize consumers to shift peak loads. Smart meters can also communicate directly with other meters, with internal display units to allow both utilities and their customers to better manage energy consumption.
As implementations and architectures become more complex, meters require more processing power and more flash software stacks, communication protocols, and firmware updates. The meter also has a communication interface. In the U.S., many companies have chosen ZigBee radio stations as a link utility, while in Europe, some utility groups have agreed to use powerline communication nodes. Low power consumption is a basic requirement for smart energy meters, which in turn enable power usage sensing/measurement of MCUs. Low power consumption is also advantageous because even if the meters are powered by a mains, they must be able to use battery power to keep running if the power supply loses the real-time clock (RTC). Smart meter applications for microcontrollers require current and voltage measurements with high-resolution A/D converters; Usually 16-bit or 24-bit A/D conversion speed is not an issue, so a converter can be used. Dual A/D, which usually requires simultaneous measurements, and may require temperature measurement and intrusion detection – must prevent tampering with the meter. The data transmission will most likely need to be encrypted using AES, DES, RSA, ECC or SHA-256. ICs with high EMC rejection reduce the need for external components. and EEPROM may require data logging and storage of calibration data. Metering may be one, two, or three-phase energy meter metering. Single-phase electricity meters are common in most residential applications. This typically has a voltage and one current being measured and it supports low to medium loads. Two-phase meters, which are not common worldwide, and adopted mainly concentrated in Japan, have two voltages and two currents for measurement. Each stage is closed at 180 degrees, and it is usually a large to medium load.
Finally, three-phase measurements are often used in large office spaces and industrial applications. But there are also three different stages with 120 phases of mutual out-of-phase. Three voltages and three currents need to be measured, so instantaneous snapshots of energy consumption and power factor obtained by at least six ADCs are required. The inclusiveness of each A/D programmable gain stage in candidate MCUs is a great aid to sensor interfaces. In energy metering services, an MCU may have to handle a lot of things.
Figure 1 is a block diagram of the functions shown in the center of the processor, which is also the various peripheral processors that may need to be handled in a good smart meter design.
Smart meter block diagram
Figure 1: Block diagram of a typical smart meter.
So, now that we've defined the MCU for smart meter services, where do we find what are the requirements for something like this?
There are several possibilities here. The NXP® EM773FHN33 is an ARM-based Cortex-M0, low-cost, 32-bit, energy metering IC. It runs at 48 MHz and comes with a nested vector interrupt controller, serial line debugging, 32 KB of flash memory and 8 bytes of SRAM. In addition, at its peripheral supplement, the MCU includes an I²C-bus interface, an RS-485/EIA-485 UART, with SSP capability, an SPI interface, three general-purpose counters/timers, up to 25 general-purpose I/O pins and a "metering engine" designed to collect voltage and current inputs of a load to calculate active power, reactive power, apparent power, and power factor. There are two current inputs and one voltage input, and some have a measurement accuracy representing 1 percent. It is available in a 0.85mm HVQFN plastic thermally enhanced, low profile quad flat package with 33 terminals. The energy metering IC is accurate with scalable input sources up to 230 V/50 Hz/16 A and 110 V/60 Hz/20 1%. The 16-bit MCU with a high-resolution ADC is the Texas Instruments MSP430AFE253IPW low-power 16-bit MCU targeting utility metering applications with a single-phase metering analog front end that supports more than 2,400 0.1% accuracy:1 dynamic range. The MSP430AFE253IPW has three 24-bit A/D converters and flash memory for up to 16 bytes, 512 bytes of RAM, and temperature measurement. MCUs also have one, faster, 10-bit A/D. For FS given by 24-bit A/D for accuracy metrics, the maximum is an offset error of 0.2% – which makes for a 19-bit converter. The active mode supply current is only 220A for 1 MHz, 2.2 V and 0.5 A for standby. It operates from -40°C to 85°C. Among them, A/D can be used for tamper-proof functions.
There are nine versions of the MSP430AFE2xx device family (Figure 2), and all have SPI and UART interfaces, LCD controllers, 16-bit timers/PWM, watchdogs, and hardware multipliers. These chips do not have real-time clocks or data encryption.
TI's MSP430AFE2xx family
Figure 2: TI's MSP430AFE2xx family provides SPI and UART interfaces and an LCD controller.
8-bit or 32-bit choice
The 8-bit Freescale MC9S08GW MCU (Figure 3) features a dedicated differential amplifier and up to 16 channels of two 16-bit A/D converters. The device features 64 KB of flash memory, an RTC with tamper protection, an LCD controller with up to 288 segments, and CRC data verification. It runs at 2.15 V at up to 3.6 MHz and 1.8 V at up to 10 MHz. The chip is available in 1010 mm or 1414 mm LQFP packages.
Another possibility for Freescale is their K30's Cortex-M4 type 32-bit MCU with a low-power segment LCD controller for driving up to 320 segments (Figure 3). The PK30X256VLQ100 has a single 6-bit A/D converter, 256 KB of flash memory, an RTC, interrupt controller, and CRC data checksum.
Freescale K30 block diagram
Figure 3: Freescale K30 block diagram.
Microcontroller with LCD driver and low-power mode
Microchip's PIC18F87K90 is a good choice for measurement, although its 24-channel A/D conversion is limited to 12-bit resolution. It has a real-time clock, flash 128 bytes and EEPROM 1 byte, plus LCD driver for 192 pixels and 4 external interrupts. In power-down mode, the IC's supply current is 60°C up to 600 nA. This RTC requires up to 4.6A at 3.3 V and 60°C. The A/D integral linearity error is typically &1 LSB, but 6.0 LSB (maximum) - fairly sprawl. The differential linearity error is specified as 1 typical and +3.0/-1.0 maximum. This is in the industrial temperature range. No encryption or tampering proofing is provided. A different approach taken is that of Analog Devices, whose ADE7880 is not a true MCU but more SOC with a "computational function block" tuned for electronic meter applications. It is a harmonic engine designed for three-phase energy metering and functional adaptive real-time monitoring. The ADE7880 device uses a second-order sigma-delta analog-to-digital converter (ADC), digital integrator, reference circuitry, and all of the signal processing required for total (fundamental and harmonic) active and apparent energy measurements, rms calculations, and fundamental active-only and reactive energy measurements to monitor three user-selectable harmonics, in addition to the fundamental. It automatically tracks the fundamental frequency and provides real-time harmonic measurement updates. Harmonic analysis includes RMS current, RMS voltage, and active, reactive, and apparent power, power factor, as well as harmonic distortion, total harmonic distortion plus noise (THD+N) calculations. The ADE7880 uses a 7-second A/D converter, a digital integrator, reference circuitry, and all the signal processing power required. It supports IEC 62053-21, IEC 62053-22, IEC 62053-23, EN 50470-1, EN 50470-3, ANSI C12.20 and IEC 61000-4-7 standards and requires approximately 25 mA to operate.
summary
The smart meter applications discussed here such as MCUs are very capable of forming the focus of a smart meter system. While some MCUs are with integrated AFEs, in other cases, signal capture and conversion requirements can lead to the use of separate analog front-end chips. In an electricity meter, the AFE senses current and voltage, converts the sensed value into digital form, and then transmits the digital value to the microcontroller. In all cases, other components that will be required for fully intelligent meter operation will be required. Smart meters are essential for peripheral devices such as EEPROM chips and provide line-isolated optocouplers. And, of course, software is required for various data processing functions, including the calculation of the amount of electricity used, and the processing of the customer's energy costs. That said, all the mentioned MCUs available now, and complete smart meter functionality with one or two external integrated circuits, can be achieved – at very low power consumption.
返回顶部