Application Note Poly-Phase Energy Metering IC M90E36A APPLICATION OUTLINE This document describes system application issues when using the M90E36A (polyphase energy metering ICs) to design poly-phase energy meters. The M90E36A is applicable in class 0.5S or class 1 poly-phase meter design and also supports three-phase four-wire (3P4W, Y0) or three-phase three-wire (3P3W, Y or Δ) connection modes. The M90E36A can also be used in harmonic meter design. The M90E36A uses 3.3V single power supply. In a typical 3P4W design, there are three transformers®ulators to provide power supply. The AC power supply outputs 3.3V to chip digital power supply DVDD after rectifier and voltage regulation. The analog power supply AVDD should be connected directly to digital power supply DVDD. The M90E36A has on-chip power-on-reset circuit. The RESET pin should be connected to DVDD through a 10kΩ resistor and a 0.1μF filter capacitor to ground. The M90E36A has highly stable on-chip reference power supply. The Vref pin should be decoupled with a 10μF capacitor and a 0.1μF ceramic capacitor. The M90E36A employs 16.384MHz as the system frequency. The M90E36A has built-in crystal oscillator circuit and 10pF matching capacitance. Users only need to connect a 16.384MHz crystal between OSCI and OSCO pins in application. The M90E36A provides a 4-wire SPI interface (CS, SCLK, SDI and SDO) for external MCU connection. MCU can perform chip configuration and register reading/writing through SPI. The M90E36A also supports Master mode SPI, which is named Direct Memory Access (DMA) mode. In DMA mode, The M90E36A streams out ADC sampling raw data to external MCU at an up to 1800kbps rate. The M90E36A provides four energy pulse output pins: active energy pulse CF1, reactive energy pulse CF2 (can also be configured as apparent energy pulse), fundamental energy pulse CF3 and harmonic energy pulse CF4. They can be used for energy metering calibration and can also be connected to MCU for energy accumulation. The M90E36A provides three zero-crossing pins ZX0, ZX1 and ZX2 which can select different phase’s voltage or current as inputs. The M90E36A provides three output pins IRQ0, IRQ1 and WarnOut to generate interrupt and warn out signals at different levels. The default application in this document is 3P4W, otherwise it will be specially indicated. Atmel-46104A-SE-M90E36A-ApplicationNote_050514 Ta bl e o f C o n t en ts 1 HARDWARE REFERENCE DESIGN ........................................................................................... 4 1.1 3P4W Application ................................................................................................................... 4 1.1.1 1.1.2 1.1.3 1.1.4 1.1.5 Schematics (Current Transformer (CT)) ................................................................................... 4 BOM (CT) ................................................................................................................................. 5 Schematics (Rogowski) ............................................................................................................ 6 BOM (Rogowski) ....................................................................................................................... 7 Circuit Description ..................................................................................................................... 7 1.2 3P3W Application ................................................................................................................... 9 1.2.1 1.2.2 1.2.3 Schematics ............................................................................................................................... 9 BOM ........................................................................................................................................ 10 Circuit Description ................................................................................................................... 10 2 INTERFACE ................................................................................................................................ 11 2.1 SPI ........................................................................................................................................ 11 2.2 DMA...................................................................................................................................... 12 3 POWER MODES ........................................................................................................................ 13 3.1 Normal Mode ........................................................................................................................ 13 3.2 Partial Measurement Mode................................................................................................... 13 3.3 Detection Mode..................................................................................................................... 13 3.4 Idle Mode .............................................................................................................................. 14 3.5 Transition and Application of Power Modes ......................................................................... 14 4 CALIBRATION ........................................................................................................................... 17 4.1 Calibration Method................................................................................................................ 17 4.2 Calibration in Normal Mode .................................................................................................. 17 4.2.1 4.2.2 4.2.3 4.2.4 4.2.5 4.2.6 4.2.7 4.2.8 Measurement/Metering Startup Command (Configstart/Calstart/HarmStart/AdjStart) ........... 18 PL Constant Configuration (PL_Constant) ............................................................................. 20 Metering Method Configuration (MMode0) ............................................................................. 21 PGA Gain Configuration (MMode1) ........................................................................................ 22 Offset Calibration of Voltage/ Current/ Power ......................................................................... 23 Voltage/ Current Measurement Calibration ............................................................................ 24 Energy Metering Calibration ................................................................................................... 25 Fundamental Energy Metering Calibration ............................................................................. 26 4.3 Calibration in Partial Measurement Mode............................................................................. 28 4.3.1 4.3.2 4.3.3 4.3.4 4.3.5 4.3.6 Partial Measurement Configuration (PMConfig) ..................................................................... 28 Sampling Cycle Configuration (PMAvgSamples) ................................................................... 28 PGAgain Configuration (PMPGA) ........................................................................................... 28 Current Offset Calibration ....................................................................................................... 29 Current Measurement Calibration ........................................................................................... 29 Special Application of Partial Measurement Function ............................................................ 30 4.4 Calibration in Detection Mode............................................................................................... 31 4.4.1 4.4.2 Current Detection Module Configuration ................................................................................ 31 Current Detection Threshold Calibration ................................................................................ 32 M90E36A [APPLICATION NOTE] Atmel-46104A-SE-M90E36A-ApplicationNote_043014 2 5 FUNCTION REGISTERS CONFIGURATION ............................................................................ 33 5.1 Startup Current Configuration............................................................................................... 33 5.2 SAG Function ....................................................................................................................... 34 5.3 Reserved Register/ Address and Reserved bits................................................................... 35 5.3.1 5.3.2 5.3.3 Reserved Register/ Address ................................................................................................... 35 Reserved Register Bits ........................................................................................................... 35 Reserved Bits in the FUNC_EN1 Register ............................................................................. 35 6 TEMPERATURE COMPENSATION .......................................................................................... 36 6.1 On-chip Temperature Sensor Configuration......................................................................... 36 6.2 Temperature Compensation Based on ADC Sampling Channel.......................................... 36 6.3 Temperature Compensation Based on reference Voltage ................................................... 38 7 HARMONIC ANALYSIS ............................................................................................................. 39 7.1 DFT Engine........................................................................................................................... 39 7.2 Obtain Harmonic Analysis of Above 32nd ............................................................................ 42 M90E36A [APPLICATION NOTE] Atmel-46104A-SE-M90E36A-ApplicationNote_043014 3 D C B A IB- IB+ IA- IA+ UN UC UB UA 1 1 R53 2.4 R51 2.4 R44 2.4 R38 2.4 1K R55 GND 1K R49 1K R46 GND 1K R35 240K GND R26 240K 240K 240K R25 R17 240K 240K R16 R4 R3 240K R29 240K R20 240K R7 240K R30 240K R21 240K R8 C20 18nF C18 18nF C16 18nF C11 18nF IBN IBP IAN IAP 240K R31 240K R22 240K R9 2 R54 2.4 R52 2.4 R45 2.4 R39 2.4 R32 1K R23 1K R10 1K C8 18nF VCP 1K R56 GND 1K R50 1K R47 GND 1K VBP C6 18nF R36 GND GND GND C3 18nF VAP Option for N-line current sampling IN- IN+ IC- IC+ Voltage Sampling 240K R28 240K R19 240K R6 Current Sampling (with CT) 240K R27 240K R18 240K R5 2 3 3 C21 18nF C19 18nF C17 18nF C12 18nF INN INP ICN ICP C9 10uF GND R40 1K GND 4 C13 18nF C10 0.1uF IAP IAN IBP IBN ICP ICN INP INN C5 0.1uF DVDD33 GND AVDD33 AGND I1P I1N I2P I2N I3P I3N I4P I4N Vref AGND 0.1uF 0.1uF C4 10uF C2 C1 R2 10K 5 C14 18nF R43 1K GND C15 18nF 16.384MHz X1 DMA NC PM1 PM0 TEST IRQ1 IRQ0 WarnOut CF4 CF3 CF2 CF1 R11 10K DVDD33 36 35 34 33 32 31 30 29 28 27 26 25 R12 10K R13 10K CF4 CF3 CF2 CF1 5 ZX2 ZX1 ZX0 IRQ1 IRQ0 WarnOut CF4 CF3 CF2 CF1 PM1 PM0 DMA SDI SDO SCLK CS 6 6 Connect to MCU R14 10K Poly Phase Metering AFE Chip (ATM90E36A) R41 1K U5 ATM90E36A 1 2 3 4 5 6 7 8 9 10 11 12 AVDD33 C7 0.1uF 4 48 47 46 45 44 43 42 41 40 39 38 37 DVDD33 DGND NC NC DGND DVDD18 VPP RST SDI SDO SCLK CS Atmel-46104A-SE-M90E36A-ApplicationNote_050514 V1P V1N V2P V2N V3P V3N NC OSCI OSCO ZX0 ZX1 ZX2 M90E36A [Application Note] VAP GND 510 R33 U4 510 R24 U3 510 R15 U2 510 R1 U1 GND GND GND GND PS2501 PS2501 PS2501 PS2501 1 2 3 4 5 CON-5 1 2 3 4 5 JP1 8 D4 CF1 D3 CF2 D2 CF3 D1 CF4 GND 7 Poly Phase Metering AFE ATM90E36A (3P4W with CT) 1 1.0 of 1 A3 8 ATM90E36A_3P4W_CT.SchDoc Title: File: * Project: Revision: Page: 10:59:16 * Document Number: Date: Size: * ApprovedBy: 3/19/2014 Felix Yao DrawnBy: * SMART ENERGY Atmel China CheckedBy: 510 R48 510 R42 510 R37 510 R34 Energy Pulse Output Indicate CF4 CF3 CF2 CF1 Energy Pulse Output (Isolated with Optocoupler) CF4 CF3 CF2 CF1 7 D C B A 3P4W APPLICATION VBP 1.1 VCP HARDWARE REFERENCE DESIGN 13 14 15 16 17 18 19 20 21 22 23 24 4 1 1.1.1 Schematics (Current Transformer (CT)) 1.1.2 BOM (CT) Table-1 3P4W BOM (CT) Component Type Designator Quantity Parameter Tolerance SMT Capacitor C3 C6 C8 C11 C12 C13 C14 C15 C16 C17 C18 C19 C20 C21 C1 C4 C5 C7 C10 14 18nF ±10% X7R (anti-aliasing filter capacitor) 5 0.1μF ±10% X7R 2 8 10μF 2.4Ω ±10% X7R ±1% 1/8W 25ppm 8 510Ω ±5% 1/8W 100ppm 14 1kΩ ±1% 1/8W 25ppm (anti-aliasing filter resistor) 5 21 10kΩ 240kΩ ±5% 1/8W 100ppm ±1% 1/8W 25ppm LED SMT Optocoupler Crystal C2 C9 R38 R39 R44 R45 R51 R52 R53 R54 R1 R15 R24 R33 R34 R37 R42 R48 R10 R23 R32 R35 R36 R40 R41 R43 R46 R47 R49 R50 R55 R56 R2 R11 R12 R13 R14 R3~R9, R16~R22, R25~R31 D1 D2 D3 D4 U1 U2 U3 U4 X1 4 4 1 PS2501 16.384MHz ±20ppm IC Connector U5 JP1 1 1 M90E36A CON-5 - SMT Resistor M90E36A [APPLICATION NOTE] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 5 D C B A IB- IB+ IA- IA+ UN UC UB 1 C22 18nF 240K 240K R28 240K R19 240K R6 240K R29 240K R20 240K R7 240K R30 240K R21 240K R8 1K R53 1k IBN IBP IAN 1K R54 1K R52 1K R45 1K R39 2 C8 18nF 1K C29 GND 18nF R56 C27 18nF 1K R50 1K C25 GND 18nF R47 1K VCP C6 18nF VBP C3 18nF R36 GND GND GND C23 18nF R32 1K R23 1K R10 1K VAP Option for N-line current sampling IN- IN+ IC- IC+ 240K R31 240K R22 240K R9 Current Sampling (with Rogowski Coil) C20 18nF C28 GND 18nF R55 C16 18nF C11 18nF IAP Voltage Sampling C18 18nF 1K 240K R27 240K R18 240K R5 C26 18nF R49 1K 1K R51 1K R44 1K 1K C24 GND 18nF R46 R35 R38 GND R26 240K 240K 240K R25 R17 240K R16 R4 240K 3 C21 18nF C19 18nF C17 18nF C12 18nF INN INP ICN ICP C9 10uF GND R40 1K GND 4 C13 18nF C10 0.1uF IAP IAN IBP IBN ICP ICN INP INN C7 0.1uF C5 0.1uF AVDD33 AGND I1P I1N I2P I2N I3P I3N I4P I4N Vref AGND 0.1uF 0.1uF C4 10uF C2 C1 R2 10K 5 C14 18nF R43 1K GND C15 18nF 16.384MHz X1 DMA NC PM1 PM0 TEST IRQ1 IRQ0 WarnOut CF4 CF3 CF2 CF1 R11 10K DVDD33 36 35 34 33 32 31 30 29 28 27 26 25 R12 10K R13 10K CF4 CF3 CF2 CF1 5 ZX2 ZX1 ZX0 IRQ1 IRQ0 WarnOut CF4 CF3 CF2 CF1 PM1 PM0 DMA SDI SDO SCLK CS 6 6 Connect to MCU R14 10K Poly Phase Metering AFE Chip (ATM90E36A) R41 1K U5 ATM90E36A 1 2 3 4 5 6 7 8 9 10 11 12 AVDD33 DVDD33 GND VAP 4 VBP R3 3 VCP 2 48 47 46 45 44 43 42 41 40 39 38 37 Atmel-46104A-SE-M90E36A-ApplicationNote_050514 DVDD33 DGND NC NC DGND DVDD18 VPP RST SDI SDO SCLK CS M90E36A [Application Note] V1P V1N V2P V2N V3P V3N NC OSCI OSCO ZX0 ZX1 ZX2 6 13 14 15 16 17 18 19 20 21 22 23 24 UA 1 GND 510 R33 U4 510 R24 U3 510 R15 U2 510 R1 U1 GND GND GND GND PS2501 PS2501 PS2501 PS2501 1 2 3 4 5 CON-5 1 2 3 4 5 JP1 8 D4 CF1 D3 CF2 D2 CF3 D1 CF4 GND 7 8 Poly Phase Metering AFE ATM90E36A (3P4W with Rogiwski Coil) 1.0 of 1 ATM90E36A_3P4W_Coil.SchDoc Title: File: * 1 A3 Project: Revision: Page: 10:59:07 * Document Number: Date: Size: * ApprovedBy: 3/19/2014 Felix Yao DrawnBy: * SMART ENERGY Atmel China CheckedBy: 510 R48 510 R42 510 R37 510 R34 Energy Pulse Output Indicate CF4 CF3 CF2 CF1 Energy Pulse Output (Isolated with Optocoupler) CF4 CF3 CF2 CF1 7 D C B A 1.1.3 Schematics (Rogowski) 1.1.4 BOM (Rogowski) Table-2 3P4W BOM (Rogowski) Component Type Designator Quantity Parameter Tolerance SMT Capacitor C3 C6 C8 C11 C12 C13 C14 C15 C16 C17 C18 C19 C20 C21 C22 C23 C24 C25 C26 C27 C28 C29 C1 C4 C5 C7 C10 22 18nF ±10% X7R (anti-aliasing filter capacitor) 5 0.1μF ±10% X7R 2 8 10μF 510Ω ±10% X7R ±5% 1/8W 100ppm 14 1kΩ ±1% 1/8W 25ppm (anti-aliasing filter resistor) 5 21 10kΩ 240kΩ ±5% 1/8W 100ppm ±1% 1/8W 25ppm LED SMT Optocoupler Crystal C2 C9 R1 R15 R24 R33 R34 R37 R42 R48 R10 R23 R32 R35 R36 R38 R39 R40 R41 R43 R44 R45 R46 R47 R49 R50 R51 R52 R53 R54 R55 R56 R2 R11 R12 R13 R14 R3~R9, R16~R22, R25~R31 D1 D2 D3 D4 U1 U2 U3 U4 X1 4 4 1 PS2501 16.384MHz ±20ppm IC Connector U5 JP1 1 1 M90E36A CON-5 - SMT Resistor 1.1.5 Circuit Description The recommended circuit for the M90E36A three-phase four-wire (3P4W) application is as shown in 1.1.1 Schematics (Current Transformer (CT)). The M90E36A can use CT and Rogowski coil in current sampling. The recommended circuit for 3P4W application with Rogowski coil is as shown in 1.1.3 Schematics (Rogowski). It is recommended to use two-order filtering when sampling with Rogowski coil. The other parts are the same as the CT application circuit. The recommended type of Rogowski coil is: PA3202NL (Pulse Electronics). Poly-phase voltage is sampled over resistor divider network with recommended ratio of 240KΩ x 7:1KΩ. The anti-aliasing filter capacitor is recommended to be 18nF. Poly-phase current and N line current are sampled over CT. The CT ratio and load resistance should be selected based on the actual metering range. The anti-aliasing filter resistance/capacitor is suggested to be 1KΩ/18nF for the current sampling circuit. The CF1~CF4 pins are provided with driving capacity of 8mA which can drive LED and optocoupler parallelly. The other digital pins are provided with driving capacity of 3mA which can drive optocoupler directly. M90E36A [APPLICATION NOTE] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 7 Application note: how to select CT and CT load resistance Condition: M90E36A ADC input voltage range is 120μVrms ~ 720mVrms M90E36A ADC input gain PGA_GAIN = 1, 2, 4 Assume: Metering range of the energy meter is Imin ~ Imax CT current output ratio is N:1 CT load resistance is RCT So the parameters meet the formula as below: 120 μVrms < PGA_GAIN × RCT × I min N PGA_GAIN × RCT × I max < 720mVrms N 8 M90E36A [Application Note] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 D C B A IA- IA+ UC UB 1 240K R44 2.4 1K R46 GND 1K R35 240K 240K R38 2.4 R26 R25 GND R4 240K 240K R29 240K R7 240K R30 240K R8 C16 18nF C11 18nF IAN IAP IC- IC+ Voltage Sampling 240K R28 240K R6 240K R31 240K R9 2 Current Sampling (with CT) 240K R27 240K R5 R45 2.4 R39 2.4 R32 1K R10 1K 1K R47 GND 1K VCP C8 18nF R36 GND GND C3 18nF VAP 3 C17 18nF C12 18nF ICN ICP GND C9 10uF GND R40 1K GND 4 C13 18nF C10 0.1uF IAP IAN IBP IBN ICP ICN INP INN C5 0.1uF DVDD33 GND AVDD33 AGND I1P I1N I2P I2N I3P I3N I4P I4N Vref AGND 0.1uF 0.1uF C4 10uF C2 C1 R2 10K 5 R43 1K GND C15 18nF 16.384MHz X1 DMA NC PM1 PM0 TEST IRQ1 IRQ0 WarnOut CF4 CF3 CF2 CF1 R11 10K DVDD33 36 35 34 33 32 31 30 29 28 27 26 25 R12 10K R13 10K CF4 CF3 CF2 CF1 5 ZX2 ZX1 ZX0 IRQ1 IRQ0 WarnOut CF4 CF3 CF2 CF1 PM1 PM0 DMA SDI SDO SCLK CS 6 6 Connect to MCU R14 10K Poly Phase Metering AFE Chip (ATM90E36A) U5 ATM90E36A 1 2 3 4 5 6 7 8 9 10 11 12 AVDD33 C7 0.1uF 4 VAP R3 3 VBP 2 VCP UA 1 48 47 46 45 44 43 42 41 40 39 38 37 DVDD33 DGND NC NC DGND DVDD18 VPP RST SDI SDO SCLK CS V1P V1N V2P V2N V3P V3N NC OSCI OSCO ZX0 ZX1 ZX2 13 14 15 16 17 18 19 20 21 22 23 24 GND 510 R33 U4 510 R24 U3 510 R15 U2 510 R1 U1 GND GND GND GND PS2501 PS2501 PS2501 PS2501 1 2 3 4 5 CON-5 1 2 3 4 5 JP1 8 D4 CF1 D3 CF2 D2 CF3 D1 CF4 GND 7 Poly Phase Metering AFE ATM90E36A (3P3W with CT) 1 1.0 of 1 A3 8 ATM90E36A_3P3W_CT.SchDoc Title: File: * Project: Revision: Page: 10:58:57 * Document Number: Date: Size: * ApprovedBy: 3/19/2014 Felix Yao DrawnBy: * SMART ENERGY Atmel China CheckedBy: 510 R48 510 R42 510 R37 510 R34 Energy Pulse Output Indicate CF4 CF3 CF2 CF1 Energy Pulse Output (Isolated with Optocoupler) CF4 CF3 CF2 CF1 7 D C B A 1.2 3P3W APPLICATION 1.2.1 Schematics M90E36A [APPLICATION NOTE] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 9 1.2.2 BOM Table-3 3P3W BOM Component Type Designator Quantity Parameter Tolerance SMT Capacitor C3 C8 C11 C12 C13 C15 C16 C17 C1 C4 C5 C7 C10 8 18nF 5 0.1μF ±10% X7R (anti-aliasing filter capacitor) ±10% X7R 2 4 8 10μF 2.4Ω 510Ω ±10% X7R ±1% 1/8W 25ppm ±5% 1/8W 100ppm C2 C9 R38 R39 R44 R45 R1 R15 R24 R33 R34 R37 R42 R48 R10 R32 R35 R36 R40 R43 R46 R47 R2 R11 R12 R13 R14 R3~R9, R25~R31 8 1kΩ 5 14 10kΩ 240kΩ ±1% 1/8W 25ppm (anti-aliasing filter resistor) ±5% 1/8W 100ppm ±1% 1/8W 25ppm LED SMT Optocoupler Crystal D1 D2 D3 D4 U1 U2 U3 U4 X1 4 4 1 NEC2501 16.384MHz ±20ppm IC Connector U5 JP1 1 1 M90E36A CON-5 - SMT Resistor 1.2.3 Circuit Description This circuit is the recommended circuit for the M90E36A three-phase three-wire (3P3W) application. Phase B is the reference ground in 3P3W application. In 3P3W system, Uab stands for Ua, Ucb stands for Uc and there is no Ub. Phase B voltage, phase B current and N line sampling current are not needed in 3P3W application. Pin 5, 6, 9, 10, 15 and 16 should be connected to GND. If DMA function is not used, pin 36 should also be connected to GND. All NC pins should be left open. The other parts of 3P3W application circuit are similar to 3P4W and can be treated in the same way. 10 M90E36A [Application Note] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 2 INTERFACE The M90E36A provides a four-wire SPI interface (CS, SCLK, SDI and SDO). The interface can be configured to two modes by the DMA_CTRL pin: Slave mode and Master mode. 2.1 SPI The SPI interface in Slave mode is mainly used for register read/write operation. A complete SPI read/write operation is of 32 bits, which contains 16-bit address and 16-bit data. In the 16-bit address, bit0 ~ bit9 correspond to valid register address A0 ~ A9, and bit10 ~ bit14 are reserved (these bits are don’t-care). Bit15 indicates the SPI operation is read or write. SPI Operation Description Highest Bit (Bit15) Read Write Read register data Write data to register 1 0 The transmission of address and data bits is from high to low, which means MSB first and LSB last. Note that the M90E36A read/write only supports single address operation, rather than continuous read or write. The M90E36A has a special register LastSPIData [0FH] for recording the last SPI read/write data. This register can be used for data check for SPI read/write operation. When the system is in strong interference situation, the disturbance signal may cause SPI communication disorder and result in SPI read/write error. In this case, LastSPIData can be used to check the correctness of SPI read/write and strengthen system robustness. For read-clear registers, if the read data is different from the LastSPIData data, the actual data can be obtained by reading the LastSPIData register repeatedly. LastSPIData application is as shown in Figure-1 and Figure-2: SPI Read Data Buffer Read LastSPIData LastSPIData == Buffer ? Y N Buffer=LastSPIData Read LastSPIData LastSPIData == Buffer ? N Y End Figure-1 LastSPIData Application (Read) M90E36A [APPLICATION NOTE] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 11 SPI Write Data Buffer Read LastSPIData LastSPIData == Buffer ? N Y End Figure-2 LastSPIData Application (Write) 2.2 DMA For details please refer to the “SPI/DMA Interface” chapter in the M90E36A datasheet. 12 M90E36A [Application Note] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 POWER MODES 3 Four power modes are supported which correspond to four kinds of power consumption. The power mode is configured by PM1/PM0 pins. 3.1 PM1 PM0 Power Modes Power Consumption 1 1 0 0 1 0 1 0 Normal mode Partial Measurement mode Detection mode Idle mode High ↓ Low NORMAL MODE In Normal mode, all function blocks are active except for the current detector block. All registers can be accessed, including the registers related to Partial Measurement mode and Detection mode. 3.2 PARTIAL MEASUREMENT MODE In Partial Measurement mode, only three-phase current sampling and the related blocks are active. SPI communication is normal in this mode, but only partial measurement related registers and some special registers can be accessed by external MCU.The accessible registers in Partial Measurement mode are listed as below: Address Name Address Name Address Name 00H 01H 03H 07H 0EH 0FH SoftReset SysStatus0 FuncEn0 ZXConfig DMACtrl LastSPIData 14H 15H 16H 17H 18H 19H PMOffsetA PMOffsetB PMOffsetC PMPGA PMIrmsA PMIrmsB 1AH 1BH 1CH 1DH PMIrmsC PMConfig PMAvgSamples PMIrmsLSB There is a special enable control bit ReMeasure (bit14 of PMConfig) for Partial Measurement mode. When the control bit is enabled, sampling and measurement are proceeded at the sampling period determined by the PMAvgSamples[1CH] register. Measure function is automatically shut off upon measurement completion. It needs to be enabled again if to measure again. Upon measurement completion, the IRQ0 pin outputs high level. MCU can judge whether measurement is completed through IRQ0. IRQ0 is cleared when the control bit (ReMeasure) is enabled again or partial measurement mode is exited. There is also a special “Busy” indication bit PMBusy (bit0 of PMConfig) for Partial Measurement mode. MCU can also judge whether measurement is completed through the PMBusy bit. Accuracy of current measurement in Partial Measurement mode is the same as Normal mode, because reference power supply module is active. 3.3 DETECTION MODE In Detection mode, only the current detector is active and all the registers can not be accessed by external MCU. In this mode, each I/O is in specific state (for details refer to datasheet) and SPI is disabled. So the control and threshold registers for Detection mode need to be programmed in Normal mode before entering Detection mode. Once these related registers are written, there is no need to re-configure them when switching between different power modes. Detection mode related registers are listed as below: M90E36A [APPLICATION NOTE] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 13 Address Name 10H 11H 12H 13H DetectCtrl DetectThA DetectThB DetectThC Current detection is achieved with low power comparators. Two comparators are supplied for each phase on detecting positive and negative current. When any single-phase current or multiple-phase current exceeds the configured threshold, the IRQ0 pin is asserted high. When all three phase currents exceed the configured threshold, the IRQ1 pin is asserted high. The IRQ0/IRQ1 state is cleared when entering or exiting Detection mode. The all three phase currents are considered as the currents of three current channels I1~I3. As there is no phase B current in 3P3W application, IRQ1 will not be asserted high even if both phase A and phase C current exceed the configured threshold. 3.4 IDLE MODE In Idle mode, all the modules are disabled and all the registers can not be accessed. In this mode, each I/O is in a specific state (for details refer to datasheet) and SPI is disabled. All register values are lost except for current detection related registers. 3.5 TRANSITION AND APPLICATION OF POWER MODES The four power modes are controlled by the PM0 and PM1 pins. In application, any power mode transition goes through Idle mode to avoid register value confusion or system status uncertainty in mode transition. All function modules are disabled in Idle mode while the related modules will be enabled after switching from Idle mode to other mode, which is equivalent to reset to the function modules, thus ensuring normal operation of the function modules. It needs to reload registers to ensure normal operation when switching from Idle mode to Normal mode or Partial Measurement mode, while no need to reload registers when switching from Idle mode to Detection mode. Power mode transition is shown as Figure-3: Normal Mode PM1:PM0 = 11 Need to reload all register values Idle Mode PM1:PM0 = 00 All the register values will be lost except for the Detection mode related registers Detection Mode PM1:PM0 = 01 Detection Mode related register value will be kept Partial Measurement Mode PM1:PM0 = 10 Need to reload Partial Measurement related registers Figure-3 Power Mode Transition 14 M90E36A [Application Note] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 Note: For description convenience, the intermediary Idle mode will be omitted when refering to power mode transition. The most typical application of power mode transition is no-voltage detection for power meter. The so-called no-voltage state is when all phase voltages are less than the voltage threshold but the load current is greater than the configured current value (such as 5% of rated current). In no-voltage state, the power meter usually uses backup battery for power supply. The system needs to enter low power mode and perform measurement and recording for novoltage state periodically. The recommended flow for power meter with the M90E36A is as below: 1 Set the current detection threshold to be the minimum load current (such as 5% of rated current) required in no-voltage state. 2 When no-voltage happens, the system enters Idle mode; 3 The system enters Detection mode every once in a while (such as 5s); 4 Once the load current is greater than the configured value, the system enters Partial Measurement mode to measure and record the load current; 5 The system returns to Idle mode after measurement and recording are completed; 6 The system enters Partial Measurement mode every once in a while (such as 60s) to measure and record the load current. System voltage sag Enter Partial Measurement mode Enter Idle mode N Measure current>minimum load current? Y Delay 5s Record current value (no-voltage event) Enter Detection mode Enter Idle mode N IRQ0/IRQ1 output high level? Delay 60s Y Figure-4 Application of Detection Mode and Partial Measurement Mode M90E36A [APPLICATION NOTE] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 15 Application note: Design principle for current detection threshold It is recommended to do system design based on current detection threshold of 3mVrms. Example: Assume: The requirement is that the minimum load current detected is 5% of rated current. Current specification is 5(60)A; The minimum load current is Id, which corresponds to a 3mVrms ADC input signal. The parameters meet the following relations: Minimum Detection Load Current Id Rated Current In Maximum Current Imax 5% In 3mVrms 250mA In 60mVrms 5A 12In 720mVrms 60A ratio to rated current corresponding ADC input signal actual current 16 M90E36A [Application Note] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 4 CALIBRATION 4.1 CALIBRATION METHOD Normally voltage, current, mean power, phase angle, frequency and so on are regarded as measurement values, while active energy, reactive energy and so on are regarded as metering value. Measurement and metering function both need calibration before normal use as shown in below table: Power Mode Normal mode Parameter voltage/current Need Calibration √ Calibration Method offset/gain calibration power/frequency/phase angle/ power factor X -- THD X -- full-wave energy metering √ fundamental energy metering √ offset/gain/phase angle calibration offset/gain calibration harmonic energy metering X -- Partial Measurement mode current measurement √ offset/gain calibration Detection mode current detection √ threshold calibration In typical application of three-phase power meter, voltage, current and full-wave energy must be calibrated. The others can be calibrated according to actual application, no need to calibrate if no use. The calibration flow follows the sequence of measurement first then metering. Metering calibration is realized by first calibrating gain and then calibrating phase angle compensation, only single-point calibration is needed over the entire dynamic range. Reactive does not need to be calibrated since it is guaranteed by chip design. Frequency, phase angle and power factor do not need calibration, since their accuracy is guaranteed by chip design. 4.2 CALIBRATION IN NORMAL MODE The basic functions, such as measurement, metering, harmonic analysis and so on are only active in Normal mode. So calibration in Normal mode is basic and a must. The related registers need to be configured before calibration. Calibration flow is as shown in Figure-5. M90E36A [APPLICATION NOTE] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 17 Calibration Initialize Voltage/Current Offset Calibration Voltage/Current Gain Calibration Power Offset Calibration Energy Gain Calibration Phase Angle Calibration End Figure-5 Active Energy Metering Calibration Flow in Normal Mode 4.2.1 Measurement/Metering Startup Command (Configstart/Calstart/HarmStart/AdjStart) Startup command registers have multiple valid settings for different operation modes. Startup Register Value Usage 6886H Rower up state Operation 5678H Calibration Similar like 6886H, This state blocks checksum checking error generation. Writing with this value trigger a reset to the associated registers. 8765H Operation Checksum checking is enabled and if error detected, IRQ/Warn is asserted and Metering stopped. Other Error It is the value after reset. This state blocks checksum checking error generation Force checksum error generation and system stop. The default value for these registers is '6886H' after power-on reset. At this time, measurement functions can be started but metering functions can not. The measurement/ metering functions will be started when related startup registers are set to '5678H' or '8765H'. If other values are written to these registers, the corresponding measurement/ metering functions will be disabled, the corresponding checksum and CSxErr bits will be set and the WarnOut pin will output high level. 18 M90E36A [Application Note] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 Startup Register/ Address Register Address Range (CSx Calculation Range) ConfigStart / 30H CalStart / 40H HarmStart / 50H 31H ~ 3AH 41H ~ 4CH 51H ~ 56H AdjStart / 60H 61H ~ 6EH Register Function CSx / Address Function Startup on Reset function configuration energy metering calibration fundamental/ harmonic energy metering calibration measurement value calibration CS0 / 3BH CS1 / 4DH CS2 / 57H -not startup not startup CS3 / 6FH startup When '5678H' is written, the registers resume to their power-on values and metering/measurement functions are started without checksum check. When '8765H' is written, the registers do not resume to their power-on values, but checksum will be checked. If the written checksum is the same as the system self generated checksum, normal metering/measurement functions will be started. If they are different, metering/measurement functions will not be started, the corresponding CSxErr bits are set and the WarnOut pin outputs high level. Note that if CS2 is not correct, when the startup register (xxxStart) is 8765H, only harmonic measurement and metering functions will be disabled. But if CS0, CS1 or CS3 are not correct, all measurement and metering functions will be disabled. The written checksum means the value MCU (or other external processor) writes to the addresses 3BH/4DH/57H/6FH through SPI. The value acquired by MCU reading through SPI is the checksum generated internally. When the startup register (xxxStart) is 5678H or 8765H, the M90E36A will calculate checksum automatically. As long as there is any register change, the corresponding CSx value will be updated immediately. So in application the MCU process can be simplified by reading the CSx registers first to get the correct checksum, then writing the checksum directly back to the CSx registers. Address Name 01H SysStatus0 1 2 3 4 Bit15 ~ Bit0 Bit15 Bit7 URevWn Bit14 CS0Err Bit6 IRevWn Bit13 Bit5 - Bit12 CS1Err Bit4 - Bit11 Bit3 SagWarn Bit10 CS2Err Bit2 PhaseLos eWn Bit9 Bit1 - Bit8 CS3Err Bit0 - CS0Err: indicates CS0 checksum status 0: CS0 checksum correct (default) 1: CS0 checksum error. The WarnOut pin is asserted at the same time. CS1Err: indicates CS1 checksum status 0: CS1 checksum correct (default) 1: CS1 checksum error. The WarnOut pin is asserted at the same time. CS2Err: indicates CS2 checksum status 0: CS2 checksum correct (default) 1: CS2 checksum error. The WarnOut pin is asserted at the same time. CS3Err: indicates CS3 checksum status 0: CS3 checksum correct (default) 1: CS3 checksum error. The WarnOut pin is asserted at the same time. In application, it is recommended to set all the startup registers (xxxStart) to 8765H, and timely check the startup registers (xxxStart) and checksum status indicate bits (CSxErr) in order to judge whether system is in normal operation. M90E36A [APPLICATION NOTE] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 19 xxxStart == 8765H ? N Re-initialize registers Y CSxErr == 0 ? N Y Figure-6 Check the Effectiveness of Register Value 4.2.2 PL Constant Configuration (PL_Constant) Energy accumulation and metering are usually referenced by energy unit, such as kWh. However, within the M90E36A, energy calculation or accumulation are based on energy pulse (CF). kWh and CF are connected by Meter Constant (MC, such as 3200 imp/kWh, which means each kWh corresponds to 3200 energy pulses). The chip’s PL_Constant is a parameter related to MC. One PL_Constant corresponds to 0.01CF. PL_Constant should be configured according to different MC in application. The M90E36A provides four energy pulse outputs: active energy pulse CF1, reactive energy pulse or apparent energy pulse CF2, fundamental energy pulse CF3 and harmonic energy pulse CF4. Their Meter Constants are all set by PL_Constant in union rather than separately. The PL_Constant registers consist of the PLconstH[31H] and PLconstL[32H] registers, corresponding to high word and low word of PL_Constant respectively. PL_Constant is calculated as below: PL_Constant = 450,000,000,000 / MC 450,000,000,000: Constant MC: Meter Constant, unit is imp/KWh, imp/Kvarh or imp/KVA Example: Calculation of PL_constant Assume: Meter Constant MC = 3200 Thus: PL_constant = 450,000,000,000 / 3200 = 140,625,000 (Hex is 8614C68H) so the registers are set as below: PLconstH[31H] = 0861H PLconstL[32H] = 4C68H 20 M90E36A [Application Note] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 4.2.3 Metering Method Configuration (MMode0) The M90E36A can be used in difference systems and metering modes, which can be configured by the MMode0[33H] register. Address Name 33H MMode0 1 2 3 4 5 6 7 8 Bit15 ~ Bit0 Bit15 Bit7 CF2varh Bit14 Bit6 CF2ESV Bit13 I1I3Swap Bit5 - Bit12 Freq60Hz Bit4 ABSEnQ Bit11 HPFOff Bit3 ABSEnP Bit10 didtEn Bit2 EnPA Bit9 001LSB Bit1 EnPB Bit8 3P3W Bit0 EnPC I1I3Swap: this bit defines phase mapping for I1 and I3 0: I1 maps to phase A, I3 maps to phase C (default) 1: I1 maps to phase C, I3 maps to phase A Note: I2 always maps to phase B. In PCB layout, the M90E36A may be placed on the top layer or bottom layer. The two placements create input current cross between phase A and C, affecting assembly of the whole meter. The influence can be eliminated by adjusting this control bit. Note that the swapping of I1and I3 only changes ADC channels and does not affect the chip’s internal data processing. Freq60Hz: grid operating line reference frequency 0: 50Hz (default) 1: 60Hz The M90E36A is applicable in 50 Hz or 60 Hz power grid. The M90E36A uses different calculation parameters in data processing according to different grid frequency. To improve the accuracy of measurement and metering, please set this control bit according to the real power grid frequency. HPFOff: HPF enable control bit 0: enable HPF (default) 1: disable HPF Besides measuring the voltage/current RMS in 50Hz or 60Hz (AC) power grid, the M90E36A can also measure the mean current value of DC condition. HPF should be disabled when using DC measurement functions. didtEn: enable integrator for didt current sensor 0: disable integrator for didt current sensor; use CT sampling for current channel (default) 1: enable integrator for didt current sensor; use Rogowski coil sampling for current channel The M90E36A supports sampling over CT or Rogowski coil. Please set this control bit according to the real current sampling means. Note that different sampling circuit should be adopted when using Rogowski coil. 001LSB: energy register LSB configuration for all energy registers 0: 0.1CF (default) 1: 0.01CF 3P3W: connection type for three-phase energy meter 0: 3P4W connection (default) 1: 3P3W connection The M90E36A uses different phase sequence judgment for different connection. Please set this control bit according to the real connection type. CF2varh: CF2 pin source configuration 0: apparent energy 1: reactive energy (default) CF2ESV: this bit is to configure the apparent energy computation type when the CF2 pin is set as apparent energy output. This control bit is also used to configure the apparent energy computation type when calculating power factor (PF). 0: All-phase apparent energy arithmetic sum (default) M90E36A [APPLICATION NOTE] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 21 9 10 11 12 13 1: All-phase apparent energy vector sum ABSEnQ: configure the calculation method of total (all-phase-sum) reactive energy and power 0: total reactive energy equals to all-phase reactive energy arithmetic sum (default) 1: total reactive energy equals to all-phase reactive energy absolute sum ABSEnP: configure the calculation method of total (all-phase-sum) active energy and power 0: total active energy equals to all-phase active energy arithmetic sum (default) 1: total active energy equals to all-phase active energy absolute sum EnPA: this bit configures whether Phase A is counted into the all-phase sum energy/power (P/Q/S) 0: Corresponding Phase A not counted into the all-phase sum energy/power (P/Q/S) 1: Corresponding Phase A to be counted into the all-phase sum energy/power (P/Q/S) (default) EnPB: this bit configures whether Phase B is counted into the all-phase sum energy/power (P/Q/S) 0: Corresponding Phase B not counted into the all-phase sum energy/power (P/Q/S) 1: Corresponding Phase B to be counted into the all-phase sum energy/power (P/Q/S) (default) EnPC: this bit configures whether Phase C is counted into the all-phase sum energy/power (P/Q/S) 0: Corresponding Phase C not counted into the all-phase sum energy/power (P/Q/S) 1: Corresponding Phase C to be counted into the all-phase sum energy/power (P/Q/S) (default) Application note: Common configuration of MMode0 (a) 3P4W, grid frequency 50Hz, MMode0 = 0087H (b) 3P3W, grid frequency 50Hz, MMode0 = 0185H 4.2.4 PGA Gain Configuration (MMode1) The MMode1 register is used to configure PGA gain of ADC sampling channel, making chips applicable to meter designs of different current specifications. Address Name 34H MMode1 1 2 Bit15 ~ Bit0 Bit15 Bit14 DPGA_GAIN Bit7 Bit6 PGA_GAIN (I4) Bit13 Bit12 PGA_GAIN (V3) Bit5 Bit4 PGA_GAIN (I3) Bit11 Bit10 PGA_GAIN (V2) Bit3 Bit2 PGA_GAIN (I2) DPGA_GAIN: digital PGA gain for the 4 current channels 00: Gain = 1 (default) 01: Gain = 2 10: Gain = 4 11: Gain = 8 PGA_GAIN (V1~V3, I1~I4): analog PGA gain for seven ADC channels 00: 1X (default) 01: 2X 10: 4X 11: N/A Application note: Configuration principle of PGA gain (a) Ensure that the ADC channel analog input signal should be within the dynamic range of 0~720mVrms (b) Configure PGA gain to be the maximum value within the whole dynamic range 22 M90E36A [Application Note] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 Bit9 Bit8 PGA_GAIN (V1) Bit1 Bit0 PGA_GAIN (I1) 4.2.5 Offset Calibration of Voltage/ Current/ Power In application, the input signal is often influenced by the interference signal. This interference will enter data processing module through ADC and high-pass filter, not only producing errors to the voltage/current RMS and power calculation, but also affecting accuracy of the energy metering. The M90E36A provides offset calibration function to voltage, current and power, reducing the influence of the interference signal to measurement/metering accuracy. Every phase’s voltage/current offset calibration should be proceeded individually. Take phase A for example, the signal source is: Ub=Uc=Un, Ua=0, Ia=0. The calibration flow of voltage/current offset is as below: a. Read measurement registers (32 bits). It is suggested to read several times to get the average value; b. Right shift the 32-bit data by 7 bits (ignore the lowest 7 bits); c. Invert all bits and add 1 (2’s complement); d. Write the lower 16-bit result to the offset register Every phase’s power offset calibration should be proceeded individually. Take phase A for example, the signal source is: Ua=Ub=Uc=Un, Ia=0. Set the input source to be 0, the calibration flow of power offset is as below: a. Read measurement registers (32 bits). It is suggested to read several times to get the average value; b. Calculate: register value x 100,000 / 65,536 c. Right shift the caculated result data by 8 bits (ignore the lowest 8 bits); d. Invert all bits and add 1 (2’s complement); e. Write the lower 16-bits result to the offset register The relationship between offset register and measurement register is as below: Item Data Width voltage rms current rms mean power 32 bits 32 bits 32 bits Data Align 9 bits 9 bits 8 bits 16 bits 16 bits 16 bits Offset register Minimum Unit 7 bits 7 bits 8 bits 0.02mV 0.2μA 0.00256W The corresponding offset register and measurement value registers are shown as below: Offset Registers Voltage Current All-wave Power fundamental power Measurement Value Registers Address Register Name Address Register Name Address Register Name 63H 67H 6BH 64H 68H 6CH 6EH 41H 42H 43H 44H 45H 46H 51H 52H 53H UoffsetA UoffsetB UoffsetC IoffsetA IoffsetB IoffsetC IoffsetN PoffsetA QoffsetA PoffsetB QoffsetB PoffsetC QoffsetC PoffsetAF PoffsetBF PoffsetCF 0D9H 0DAH 0DBH 0DDH 0DEH 0DFH 0D8H 0B1H 0B5H 0B2H 0B6H 0B3H 0B7H 0D1H 0D2H 0D3H UrmsA UrmsB UrmsC IrmsA IrmsB IrmsC IrmsN1 PmeanA QmeanA PmeanB QmeanB PmeanC QmeanC PmeanAF PmeanBF PmeanCF 0E9H 0EAH 0EBH 0EDH 0EEH 0EFH 0C1H 0C5H 0C2H 0C6H 0C3H 0C7H 0E1H 0E2H 0E3H UrmsALSB UrmsBLSB UrmsCLSB IrmsALSB IrmsBLSB IrmsCLSB PmeanALSB QmeanALSB PmeanBLSB QmeanBLSB PmeanCLSB QmeanCLSB PmeanAFLSB PmeanBFLSB PmeanCFLSB M90E36A [APPLICATION NOTE] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 23 4.2.6 Voltage/ Current Measurement Calibration Measurement calibration means the calibration of voltage rms (Urms) gain and current rms (Irms) gain. Measurement calibration is the premise of energy metering calibration. 1 Voltage/current offset (Uoffset/Ioffset) calibration: For calibration method, please refer to 4.2.5 Offset Calibration of Voltage/ Current/ Power. No need of calibration if the voltage/current offset is very small in general application. 2 Voltage/current gain calibration: The three phases’ calibration can be proceeded simultaneously. The signal source is: Ua=Ub=Uc=Un, Ia=Ib=Ic=In(Ib). The calibration method is as below: a. Read voltage/current value of the external reference meter, and also read the voltage/current measurement value from chip registers; b. Calculate the voltage/current gain: Voltage Gain = Current Gain = reference voltage value voltage measurement value x 52800 reference current value x 30000 current measurement value c. Write the result to the corresponding voltage/current gain registers Note: voltage/current gain calibration is not necessarily proceeded when gain register is the default value. That is, when the first calibration result is not ideal, there is no need to reset the gain register to the default value. Calibration can be performed again based on the current value. The formula is as below: New Voltage Gain = New Current Gain = reference voltage value voltage measurement value reference current value current measurement value x existing voltage gain x existing current gain The corresponding voltage/current gain register and measurement value registers are shown as below: Gain Register Voltage Current Measurement Value Registers Address Register Name Address Register Name Address Register Name 61H 65H 69H 62H 66H 6AH 6DH UgainA UgainB UgainC IgainA IgainB IgainC IgainN 0D9H 0DAH 0DBH 0DDH 0DEH 0DFH 0D8H UrmsA UrmsB UrmsC IrmsA IrmsB IrmsC IrmsN1 0E9H 0EAH 0EBH 0EDH 0EEH 0EFH - UrmsALSB UrmsBLSB UrmsCLSB IrmsALSB IrmsBLSB IrmsCLSB - Application Note: (a) Voltage rms is unsigned and the minimum unit 1LSB of the UrmsA/UrmsB/UrmsC registers is 0.01V. Only the higher 8 bits of the UrmsALSB/UrmsBLSB/UrmsCLSB registers are valid, the lower 8 bits are always 0, and 1LSB is 0.01/256 V. (b) Current rms is unsigned and the minimum unit 1LSB of the IrmsA/IrmsB/IrmsC registers is 0.001A; Only the higher 8 bits of the IrmsALSB/IrmsBLSB/IrmsCLSB registers are valid, the lower 8 bits are always 0, and 1LSB is 0.001/256 A. 24 M90E36A [Application Note] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 Example: Voltage gain calibration Assume: The initial value of phase A voltage gain register UgainA is 0CE40H (52800) Reference meter output voltage is 220.00V Voltage rms register readout UrmsA = 5BA0H (23456) The higher 8 bits of voltage LSB register readout UrmsALSB = 4EH (78) Thus: voltage measured value = (UrmsA x 0.01) + (UrmsALSB x 0.01 / 256) = (23456 x 0.01) + (78 x 0.01 / 256) =234.563 V voltage gain = 220.00 / 234.563 x 52800 = 49521.88 = 0C172H So the register can be set to: UgainA = 0C172H 4.2.7 Energy Metering Calibration Only active energy is required for energy calibration. There is no need to calibrate reactive energy, the accuracy of which is guaranteed by chip design. Metering calibration flow is gain first then phase angle. Active energy pulse output (CF1) should be connected to the pulse input port of the calibration bench during calibration. Energy metering should be calibrated at In (Ib). 1 Power offset (Poffset/Qoffset) calibration For calibration method please refer to 4.2.5 Offset Calibration of Voltage/ Current/ Power. No need of calibration if the power offset is very small in general application. 2 Gain calibration Every phase’s gain calibration should be proceeded individually. Take phase A for example, the signal source is: Ua=Ub=Uc=Un, Ia=In(Ib), Ib=Ic=0, PF=1.0. The calibration method is as below: a. Read the energy error value e from calibration bench; b. Calculate the gain; Gain = Complementary ( 3 -ε x 215 ) 1+ ε c. Write the result to the corresponding gain registers. Phase angle calibration. Take phase A for example, the signal source is: Ua=Ub=Uc=Un, Ia=In(Ib), Ib=Ic=0, PF=0.5L.The calibration method is as below: a. Read the energy error value ep from calibration bench; b. Calculate the phase angle error; AngleError = ε p ∗ Gphase Gphase is a constant. When grid frequency is 50Hz, Gphase=3763.739. When grid frequency is 60Hz, Gphase=3136.449 c. Write the result to the corresponding phase angle error registers. The phase angle registers are signed and MSB of 1 indicates a negative value. The corresponding gain register and phase angle registers are shown as below: Phase A Phase B Phase C Address Register Name 47H 48H 49H 4AH 4BH 4CH GainA PhiA GainB PhiB GainC PhiC M90E36A [APPLICATION NOTE] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 25 Example: Energy gain and phase angle calculation The condition is that power factor PF=1.0, current is Ib, energy error ε is -13.78%, so: -ε/(1+ε)=0.159823707, gain = int(0.159823707* 2^15)=5237.10=1475H Write 1475H to the gain register. After gain calibration, energy error εP is 0.95% in the condition of PF=0.5L, current is Ib and frequency is 50Hz, so: phase angle = εP*3763.739 =0.0095*3763.739=35.75553=24H; Write 24H to the phase angle register. 4.2.8 Fundamental Energy Metering Calibration For fundamental energy metering calibration, only gain and offset calibration is needed. There is no need to calibrate phase angle. Fundamental energy pulse output (CF3) should be connected to the pulse input port of the calibration bench during calibration. The startup register for fundamental energy metering calibration is HarmStart [50H]. Calibration related registers are 51H~56H. During calibration, please only start and configure these registers rather than other registers. Fundamental energy metering calibration is similar to energy metering calibration. 1 Fundamental power offset (PoffsetxF) calibration For fundamental power offset calibration, only active power error needs to be calibrated per phase individually. Take phase A for example, the signal source is: Ua=Ub-Uc=Un, Ia=0. For calibration method please refer to 4.2.5 Offset Calibration of Voltage/ Current/ Power. No need of calibration if the power offset is very small in general application. 2 Fundamental energy gain calibration Every phase’s fundamental energy calibration should be proceeded individually. Take phase A for example, the signal source is: Ua=Ub=Uc=Un, Ia=In(Ib), Ib=Ic=0, PF=1.0. The calibration method is as below: a. Read the error value e from the external reference meter; b. Calculate the gain; Gain = Complementary ( -ε x 215 ) 1+ ε c. Write the result to the corresponding gain registers. The corresponding fundamental energy gain registers are shown as below: 26 M90E36A [Application Note] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 Address Register Name 54H 55H 56H PGainAF PGainBF PGainCF Fundamental Energy Calibration Startup Fundamental Power Offset Calibration Fundamental Energy Gain Calibration End Figure-7 Fundamental Energy Metering Calibration Flow M90E36A [APPLICATION NOTE] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 27 4.3 CALIBRATION IN PARTIAL MEASUREMENT MODE Partial measurement related registers are 14H~1DH. There are no specific startup and checksum registers. Therefore, the related register should be cleared and configured with initial values before calibration. 4.3.1 Partial Measurement Configuration (PMConfig) Address Register Name 1BH PMConfig 1 2 3 4 Bit15 ~ Bit0 Bit15 Bit7 - Bit14 ReMeasure Bit6 - Bit13 MeasureStartZX Bit5 - Bit12 MeasureType Bit4 - Bit11 Bit3 - Bit10 Bit2 - Bit9 Bit1 - Bit8 Bit0 PMBusy ReMeasure: enable another measurement cycle 0: not enable (default) 1: trigger another measurement cycle The partial measurement module is one-time triggered, that is, once the ReMeasure bit is set, current measurement is performed one time then turned off. If measurement is required again, the ReMeasure bit should be set again. MeasureStartZX: configure start of measurement 0: Measurement start immediately after the ReMeasure bit is set (default) 1: Measurement start from zero-crossing point after the ReMeasure bit is set MeasureType: indicate the measurement type 0: RMS measurement (default) 1: Mean Value (DC Average) measurement PMBusy: indicate the measure ‘Busy’ status 0: Measurement done (default) 1: Measurement in progress, ‘Busy’ 4.3.2 Sampling Cycle Configuration (PMAvgSamples) The PMAvgSamples[1CH] register is used to configure the partial measurement sampling cycle. The unit is the number of ADC sampling within a partial measurement cycle. The default grid frequency is 50Hz, so at the ADC sampling rate of 8K, the default value of PMAvgSamples is 160 (0A0H). 4.3.3 PGAgain Configuration (PMPGA) The ADC PGAgain Configuration in Partial Measurement mode is similar to the MMode1 configuration in Normal mode. Address Register Name 17H PMPGA 1 2 28 Bit15 ~ Bit0 Bit15 Bit14 DPGA_GAIN Bit7 Bit6 - Bit13 Bit12 Bit5 Bit4 PGA_GAIN (I3) DPGA_GAIN: DPGA gain for four current sampling channels 00: Gain = 1 (default) 01: Gain = 2 10: Gain = 4 11: Gain = 8 PGA_GAIN(I1~I3): PGA gain for three ADC sampling channels 00: Gain = 1 (default) 01: Gain = 2 10: Gain = 4 11: N/A M90E36A [Application Note] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 Bit11 Bit10 Bit3 Bit2 PGA_GAIN (I2) Bit9 Bit8 Bit1 Bit0 PGA_GAIN (I1) 4.3.4 Current Offset Calibration Considering the influence of the interference signal, Partial Measurement mode also supports current offset calibration. The calibration method is slightly different from the current offset calibration in Normal mode. Take the rms measurement for example, the calibration method is as below: a. Set the input source to be 0; b. Set MeasureType=1 and ReMeasure=1 to start one current measurement; c. Read the current registers (16 bits) after measurement completion d. Repeat step b and c. It is suggested to read many times to get the average value; e. Invert all bits and add 1 (2’s complement); f. Write the result to the corresponding offset register Offset Registers Measurement Value Registers Address Register Name Address Register Name 14H 15H 16H PMoffsetA PMoffsetB PMoffsetC 18H 19H 1AH PMIrmsA PMIrmsB PMIrmsC Partial Measurement module also provides the measurement value LSB register PMIrmsLSB. This register value has different definition in rms measurement (AC) and Mean Value measurement (DC) as shown below: Address Register Name 1DH PMIrmsLSB Bit15 ~ Bit0 Bit15 Bit7 Bit14 Bit13 Bit6 Bit5 IrmsBLSB Bit12 Bit4 Bit11 Bit3 Bit10 Bit9 IrmsCLSB Bit2 Bit1 IrmsALSB Bit8 Bit0 In rms measurement, the PMIrmsLSB register value is the LSB of the measurement value. In mean value measurement, this register value is the MSB of the measurement value. 4.3.5 Current Measurement Calibration The M90E36A only has current measurement function in Partial Measurement mode, so current gain needs to be calibrated. Current gain calibration for three phases can be proceeded simultaneously. Take rms measurement for example, the calibration method is as below: a. Set MeasureType=0 and ReMeasure=1 to start one current measurement; b. Read the current value from the signal source (reference meter) and read the value of current registers (16 bits); c. Calculate the current gain Current gain = reference current value / current measurement value d. Partial Measurement mode has no special current gain register, so the calculated result should be saved in MCU or external memory. Note that the partial measurement module is enabled in both Normal and Partial Measurement mode. That means the partial measurement module can measure current value in both modes. M90E36A [APPLICATION NOTE] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 29 Partial Measurement Statup Current offset calibration Partial Measurement Statup Current gain calibration End Figure-8 4.3.6 Partial Measurement Calibration Flow Special Application of Partial Measurement Function In Normal mode, the current rms measurement uses 16 cycles as the measurement period, that means current measurement period is 320ms when grid frequency is 50Hz. If there is a need to measure current at special period, start partial measurement function in Normal mode, and proceed specific period current measurement by configuring the partial measurement sampling period register PMAvgSamples. It is noted that employing specific period current measurement function will not influence the energy metering and parameters measurement functions in Normal mode, since the partial measurement module is also active in Normal mode. The partial measurement module can proceed RMS measurement (AC) and Mean Value measurement (DC). The mean value measurement can also be proceeded in Normal mode by starting partial measurement function. Application note: How to measure DC current signal DC measurement function is proceeded by partial measurement module. The flow of DC measurement is as below: a. Disable HPF (HPFOff=1) b. Set MeasureType to 1 c. Set ReMeasure to start current measurement once d. Detect the IRQ0 pin or the PMBusy bit to judge whether measurement is completed e. Read the PMIrmsA/PMIrmsB/PMIrmsC/PMIrmsLSB registers to get measurement result f. Repeat step c~e if to measure again 30 M90E36A [Application Note] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 4.4 CALIBRATION IN DETECTION MODE Current detection is realized by low power consumption comparators. The comparator outputs low level when the external current is lower than the configured threshold; The comparator outputs high level when the external current is higher than the configured threshold, as shown in Figure-9. Current Input Current Threshold IRQ Output DAC Figure-9 Current Detection Principle 4.4.1 Current Detection Module Configuration Six current threshold comparators can be configured for the current detection module to detect positive and negative current of three phases. These six threshold comparators can be enabled and disabled by the control bits, as shown below: Address Register Name 10H DetectCtrl 1 2 Bit15 ~ Bit0 Bit15 Bit7 - Bit14 Bit6 - Bit13 Bit5 PDN3 Bit12 Bit4 PDN2 Bit11 Bit3 PDN1 Bit10 Bit2 PDP3 Bit9 Bit1 PDP2 Bit8 Bit0 PDP1 Bit10 Bit9 Bit8 Bit2 Bit1 Bit0 PDN3/2/1: Control bits for negative detector of channel 3/2/1; 0: Detector enable (default) 1: Detector disable PDP3/2/1: Control bits for positive detector of channel 3/2/1; 0: Detector enable (default) 1: Detector disable Each of the six current threshold comparators has its own register configuration as shown below: Address Register Name 11H 12H 13H DetectThA DetectThB DetectThC 1 2 Bit15 ~ Bit0 Bit15 Bit7 - Bit14 Bit13 Bit12 Bit6 Bit5 Bit4 Bit11 CalCodeN Bit3 CalCodeP CalCodeN: negative detector threshold, 7-bit width. 7’b000-0000 corresponds to minimum threshold Vc=-4.28mV=-3.03mVrms 7’b111-1111 corresponds to maximum threshold Vc=12.91mV=9.14mVrms CalcodeP: positive detector threshold, definition is the same as CalCodeN. M90E36A [APPLICATION NOTE] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 31 4.4.2 Current Detection Threshold Calibration Because of the low power consumption consideration and the manufacturing process, the current detection threshold is different from different chips. Therefore, calibration is needed due to the offset of chip’s DAC output (less than ±5mVrms). The threshold current range is 2mVrms ~ 4mVrms within which the current detection module (low power consumption comparator) can detect accurately. It is recommended to proceed system design according to current detection threshold of 3mVrms. In Detection mode, all registers are not accessible, so the current threshold registers need to be configured in Normal mode first before entering Detection mode. Dichotomy is suggested in current detection threshold calibration. The recommended calibration flow is as shown in Figure10. Reference source outputs current signal that needs detection (such as 5%Ib) Th_max = 80H Th_min = 00H Th_temp = 40H Th_max, Th_min, Th_temp: variable DetectThx: Threshold register Switch to Normal mode DetectThx = Th_temp Th_min = Th_temp Th_temp = (Th_max – Th_min) / 2 Th_max = Th_temp Th_temp = (Th_max – Th_min) / 2 Switch to Detection mode N N Th_max – Th_temp == 1 ? Y IRQ0 output high level? N Th_temp – Th_min == 1 ? Y Y current detection threshold = Th_temp current detection threshold = Th_min End Figure-10 Current Detection Threshold Calibration Flow 32 M90E36A [Application Note] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 5 FUNCTION REGISTERS CONFIGURATION 5.1 STARTUP CURRENT CONFIGURATION The registers which related to startup current configuration is shown as below: 1 2 3 4 Address Register Name Description 35H PStartTh All-phase Active Startup Power Threshold. 36H QStartTh All-phase Reactive Startup Power Threshold. 37H SStartTh All-phase Apparent Startup Power Threshold. 38H PPhaseTh Each-phase Active Startup Power Threshold. 39H QPhaseTh Each-phase Reactive Startup Power Threshold. 3AH SPhaseTh Each-phase Apparent Startup Power Threshold. Due to system interference when current is 0, small signal may be gernerated in current sampling channel, producing a certain amount of energy and affecting the measurement and metering accuracy. To avoid this, the M90E36A provides the each-phase startup power configuration/judgment function. PPhaseTh, QPhaseTh and SPhaseTh are used to judge the startup power of each phase (A/B/C). Take active power for example, when a single phase input power is smaller than the configured PPhaseTh value, the input active power of that phase will be set to 0 by force, that means input to the next process is 0. Otherwise the signal will be streamed to the next process. Note that the threshold are configured separately to active(P), reactive(Q) and apparent (S). The compared value is (|P|+|Q|). PStartTh, QStartTh and SStartTh are used to judge all-phase startup power. Take active power for example, when allphase-sum power is less than the configured PStartTh value, energy accumulation will not start. Otherwise energy accumulation will start. Calculation methods of the two register groups are the same. The formula is as below: Register value = N / 0.00032, (N is the configured power threshold). Example: Startup Current Configuration Assume: meter voltage is 220V, current specification 5(100)A, active startup current is 0.1% Considering the accuracy of current measurement in small-current state, it is recommended to configure the allphase startup current threshold to be 50% of startup current (also can configure based the actual conditions). Assume the startup threshold of each-phase power is configured to be 10% of startup current. so: All-phase Active Startup Power Threshold = 3 x 5 x 0.1% x 50% x 220 = 1.65W Each-phase Active Startup Power Threshold = 5 x 0.1% x 10% x 220 = 0.11W Register values are: PStartTh[35H] = 1.65 / 0.00032 = 5156 = 1424H PPhaseTh[38H] = 0.11 / 0.00032 = 344 = 158H M90E36A [APPLICATION NOTE] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 33 Power Threshold |P|+|Q|> PPhaseTh? A/B/C Phase Active Power from DSP 0 Total Active Power 3 phases ABS > PStartTh? + 1 Phase Active Energy Metering 0 0 0 Power Threshold |P|+|Q|> QPhaseTh? Phase Reactive Power from DSP 0 Total Reactive Power 3 phases 0 1 Total Reactive Energy Metering ABS > QStartTh? + 1 1 Phase Reactive Energy Metering 0 0 0 Power Threshold |P|+|Q|> SPhaseTh? Phase Apparent Power from DSP 0 0 Total Apparent Power 3 phases A/B/C Total Active Energy Metering 1 0 A/B/C 0 1 ABS > SStartTh? + 0 1 Total (arithmetic sum) Apparent Energy Metering 1 1 0 0 0 0 Phase Apparent Energy Metering Figure-11 Metering Startup Handling 5.2 SAG FUNCTION Sag detect function is provided in M90E36A. The threshold of sag detection is configured through the SagTh register (08H). All three voltage phases use the same threshold. The threshold equation is as below: SagTh = Vth × 100 × 2 2 × Ugain / 32768 Vth: the voltage threshold to be configured; Ugain: the gain after calibration The default value of the SagTh register (08H) is 0000H. It is suggested to set the SagTh register (08H) appropriately even if the Sag function is not used in application. The reasons are as follows: In normal mode, all measurement values are calculated based on the average cycle out of 16 voltage cycles. And the voltage cycle makes use of the internal zero-crossing signal, which is different from the output on the ZX2 / ZX1 / ZX0 pins. In a 3P4W system, this internal zero-crossing signal is based on phase A voltage Ua firstly. If phase A is in sag, phase C voltage Uc is switched. If phase A and phase C are both in sag, phase B voltage Ub is switched. In a 3P3W system, the internal zero-crossing signal is based on phase A voltage Uab firstly. If phase A is in sag, phase C voltage Ucb is switched. In either 3P4W and 3P3W application, if all phases are in sag, the average cycle of 16 voltage cycles is calculated in accordance with the configured reference frequency (50 Hz or 60 Hz), i.e. 320ms for 50Hz system and 266.7ms for 60Hz system. However, for energy meter or power instrument with auxiliary power supply, if the SagTh register is not configured (the default power-on value is 0), the M90E36A will not enter sag even if there is no signal on voltage circuits due to the interference noise. Thus the wrong average cycle might be used to calculate measurement values such as current rms. 34 M90E36A [Application Note] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 5.3 RESERVED REGISTER/ ADDRESS AND RESERVED BITS 5.3.1 Reserved Register/ Address The M90E36A has many reserved registers and non-listed address areas besides the registers listed in datasheet. These reserved registers are not open to users. Please don’t operate on the address outside of the datasheet. Access to reserved register/address needs special operation because normal operation can not make change or impact to these register values. 5.3.2 Reserved Register Bits Some fields of defined registers in the datasheet are labeled as ‘reserved’. In application, all the reserved bit fields shall be written with ‘0’ when those bits have to be written, and shall be masked out (hence ignored) upon read. 5.3.3 Reserved Bits in the FUNC_EN1 Register The most registers’ reserved bits will return ‘fixed 0’ upon read. There is only one exception: the FUNC_EN1 register (04H). This register has three reserved bits. Those bits may return non-fixed value upon read. Users need to mask out or ignore those bits when using the read-back value of this register. Please note that the non-fixed return value does not indicate any abnormal working condition of the internal hardware logic. Users just need to ignore those bit fields. Address Name 04H FuncEn1 Bit15 ~ Bit0 Bit15 INOv1En Bit7 RevQchgTEn Bit14 INOv0En Bit6 RevQchgAEn Bit13 Reserved Bit5 RevQchgBEn Bit12 Reserved Bit4 RevQchgCEn Bit11 Bit10 THDUOvEn THDIOvEn Bit3 Bit2 RevPchgRevPchTEn gAEn Bit9 DFTDone Bit1 RevPchgBEn Bit8 Reserved Bit0 RevPchgCEn M90E36A [APPLICATION NOTE] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 35 6 TEMPERATURE COMPENSATION The M90E36A itself embodies good temperature characteristic. Considering that the external components might be affected by temperature in application, the M90E36A also provides compensation function for external temperature drift. A series of special registers should be configured for temperature compensation. These registers are located in special addresses, and access to these registers should be strictly carried out as the following. 6.1 ON-CHIP TEMPERATURE SENSOR CONFIGURATION The M90E36A provides a built-in temperature sensor. Due to the manufacturing process, the temperature sensor might be somewhat different for different chips. Therefore the on-chip temperature sensor should be configured before temperature compensation. The configuration method is as below: a. Write AA55H to address 2FDH b. Write 5122H to address 216H c. Write 012BH to address 219H d. Write 0000H to address 2FDH Read the Temp[0FCH] register directly to get the current temperature after configuration completed. Please note that, the temperature sensor will sense the temperature of the chip and it may have a few degrees of difference between the chip junction temperature and ambient temperature. 6.2 TEMPERATURE COMPENSATION BASED ON ADC SAMPLING CHANNEL The temperature compensation method is as below: a. Write AA55H to address 2FCH b. Write the temperature coefficient to be compensated to address 270H c. Write the fiducial point temperature of the temperature coefficient to address 27BH d. Write 0000H to address 2FCH The reference point temperature of the temperature coefficient is generally configured to be the temperature in calibration. That is, in calibration, read the Temp[0FCH] register first to get the current temperature and then save it as the reference temperature. Address 270H 27BH 36 Register Name TempCompGain Bit 15:8 Read/Write Read/Write 7:0 - - Temperature compensation coefficient, bit 7 is the sign bit, unit is ppm/ ℃ 15 Read/Write 0 1: enable temperature compensation 0: disable temperature compensation Reserved bit. Readout value is 0 TempCompRef Default Value Description 0 Reserved bit. Readout value is 0. 14:9 - - 8:0 Read/Write 19H M90E36A [Application Note] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 The reference temperature of the temperature coefficient, bit 8 is the sign bit, unit is ℃ Example: Error Test data before temperature compensation is as below: 0.08 0.06 0.04 0.02 0 -0.02 -50 0 50 Temp 100 -0.04 -0.06 -0.08 Error Linear Compesation After linearization: The reference temperature (temperature in calibrating) is 25 ℃ , metering error is 0.0000%; The error at 85 ℃ point is 0.06% So the temperature coefficient is calculated as below: (0.06% - 0.0000%) / (85 ℃ - 25 ℃ ) = 10ppm/ ℃ The temperature coefficient to be compensated is -10ppm/ ℃ . Register configuration is as below: a. Write AA55H to address 2FCH b. Write FFF6H to address 270H c. Write 8019H to address 27BH d. Write 0000H to address 2FCH M90E36A [APPLICATION NOTE] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 37 6.3 TEMPERATURE COMPENSATION BASED ON REFERENCE VOLTAGE On-chip high-precision reference voltage is provided with excellent low temperature coefficient. But in application, what should be considered is the temperature drift of the whole system. Therefore, the M90E36A specially provides temperature compensation based on reference voltage to minimize temperature drift caused by the on-board components. Note that, as voltage and current ADC sampling adopt the same reference voltage, compensation on the reference voltage will bring double effect on power and energy. Temperature compensation on reference voltage is proceeded with every 8 ℃ as a segment. In application, it is suggested to test on a small batch of components from the same lot to get the best temperature compensation coefficient. And then use this compensation coefficient as a fixed value to be written directly to register in production. The temperature compensation method is as below: a. Write AA55H to address 2FDH b. Write the reference voltage coefficient of segmented compensation to addresses 202H~209H c. Write the curvature of segmented compensation curve to address 201H d. Write 0000H to address 2FDH In normal condition, reference voltage is 1200mV. The unit of reference voltage compensation is 0.020mV in this compensation method. The default value of the corresponding compensation registers is the ideal value in chip design. In application, only incremental adjustment is needed. Address 201H 38 Register Name BGCurveK Bit 15:4 Read/Write - 3:0 Read/Write Read/Write Default Value Description Reserved bit, readout value is 0 0 Reference voltage temperature compensation curve. The bit3 is assumed as sign bit, range is -8 to +7. Scaling factor = 1+ register value*1/8. So the scaling factor will be from 0, with step of 1/8, all the way to 1+7/8. 0 Compensation coefficient on the 25 ℃ temperature point 202H BG_TEMP_P12 15:8 7:0 Read/Write 1 Compensation coefficient on the 17 ℃ temperature point 203H BG_TEMP_P34 15:8 Read/Write 6 Compensation coefficient on the 41 ℃ temperature point 7:0 Read/Write 2 Compensation coefficient on the 33 ℃ temperature point 204H BG_TEMP_P56 15:8 Read/Write 25 Compensation coefficient on the 57 ℃ temperature point 7:0 Read/Write 13 Compensation coefficient on the 49 ℃ temperature point 205H BG_TEMP_P78 15:8 Read/Write 53 Compensation coefficient on the 73 ℃ temperature point 7:0 Read/Write 39 Compensation coefficient on the 65 ℃ temperature point 206H BG_TEMP_N12 15:8 Read/Write 16 Compensation coefficient on the 1 ℃ temperature point 7:0 Read/Write 6 Compensation coefficient on the 9 ℃ temperature point 207H BG_TEMP_N34 15:8 Read/Write 54 Compensation coefficient on the -15 ℃ temperature point 7:0 Read/Write 34 Compensation coefficient on the -7 ℃ temperature point 208H BG_TEMP_N56 15:8 Read/Write 117 Compensation coefficient on the -31 ℃ temperature point 7:0 Read/Write 79 Compensation coefficient on the -23 ℃ temperature point 209H BG_TEMP_N78 15:8 Read/Write 205 Compensation coefficient on the -47 ℃ temperature point 7:0 Read/Write 159 Compensation coefficient on the -39 ℃ temperature point M90E36A [Application Note] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 7 HARMONIC ANALYSIS 7.1 DFT ENGINE The built-in DFT computation engine supports 2nd-32nd order harmonic analysis function for 6 channels. The calculation error will be enlarged when input signal is small. To address this issue, a prescaler is designed and placed before the DFT engine to amplify the signal to be calculated. The designed ADC sampling rate is 8kHz. Harmonic analysis adopts 4096 sampling points for DFT computation which takes around 0.5s once. Considering there are many DFT computation outputs, the DFT computation engine is closed by default to ensure external MCU can read DFT data from the same calculation. The engine needs to be enabled to startup, and it will automatically shut off after completing calculation for one time. Calculation results are stored in registers. External MCU can read these registers to get calculation results through SPI interface. Harmonic measurement accuracy is guaranteed by chip design after voltage/current calibration. The control registers of DFT computation engine is as below: Address 1D0H 1D1H 1 2 3 4 Register Name DFT_SCALE DFT_CTRL Bit 15 Read/Write Read/Write Default Value 0 Description 14:13 Read/Write 0 Voltage scale for phase C. 12:11 Read/Write 0 Voltage scale for phase B. 10:9 Read/Write 0 Voltage scale for phase A. 8:6 Read/Write 0 Current scale for phase C. 5:3 Read/Write 0 Current scale for phase B. 2:0 Read/Write 0 Current scale for phase A. 15:1 - 0 Reserved bit. 0 Read/Write 0 0: Disable DFT engine 1: Enable DFT engine 0: Enable Hanning window 1: Disable Hanning window The function of Hanning window is to bring periodicity to ADC sampling signal in DFT computation to achieve the exact calculation result. Please enable Hanning window in general application. Voltage scale, InputGain= 2^Scale 00: Gain = 1 01: Gain = 2 10: Gain = 4 11: Gain = 8 Current scale, InputGain= 2^Scale 000: Gain = 1 001: Gain = 2 010: Gain = 4 011: Gain = 8 100: Gain = 16 101: Gain = 32 110: Gain = 64 111: Gain = 128 DFT engine switch: DFT computation engine is enabled after setting the DFT_CTRL bit. This bit will be cleared automatically after completing calibration for one time. In application, this bit can be used to judge whether DFT computation is completed. M90E36A [APPLICATION NOTE] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 39 DFT application is as below: a. Set DFT computation engine and write 2A49H to the DFT_SCALE [1D0H] register (Assume gain of voltage and current is two) b. Start DFT computation engine and write 001H to the DFT_CTRL [1D1H] register c. Check DFT_CTRL. If DFT_CTRL=0, DFT computation is completed (about 0.5s) d. Read register value and get harmonic component and fundamental voltage/current value after transition Harmonic Component (%) = Register Value / 163.84 Register address 100H~1BFH Fundamental Current = Register Value x 3.2656 2Scale x 1000 Register address 1C0H, 1C2H, 1C4H Fundamental Voltage = Register Value x 3.2656 2Scale x 100 Register address 1C1H, 1C3H, 1C5H For description of the related registers, please refer to datasheet. Example: Assume: Meter ‘s nominal voltage is 220V, nominal current is 5A The signal source outputs 10% of 5th order harmonic component for phase A voltage and 40% of 5th order harmonic for phase A current. Register values are as follows after DFT computation engine: [103H] = 0671H (1649) [123H] = 19C9H (6601) so the measured harmonic component is : Voltage 5th order harmonic component = 1649 / 163.84 = 10.0647 (means 10.0647%) Current 5th order harmonic component = 6601 / 163.84 = 40.2893 (means 40.2893%) In application, the THD+N threshold for three-phase voltage and three-phase current can be configured. The M90E36A can judge whether the THD+N value is greater than the configured threshold by checking the corresponding status bits or the IRQ output signal. If the THD+N value is greater than the configured threshold, DFT computation engine can be started, and analysis and recording can be processed to the harmonic signal. 40 M90E36A [Application Note] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 Set THD+N monitor threshold N Check THD+N detector Y Start DFT function N Detect DFT completion identification Y Read the DFT computation result directly Figure-12 Harmonic Analysis Application Flow M90E36A [APPLICATION NOTE] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 41 7.2 OBTAIN HARMONIC ANALYSIS OF ABOVE 32ND Harmonic analysis of 32nd to 42nd can be obtained as below: 1. Read the [1D7H] register (divided by 1024, current frequency can be achieved); 2. Write the value of the [1D7H] register to the 1D4H register; 3. Calculate the value of the [1D7H] register *15.872, then convert to hex. Write the high word to the [1D2H] register and the low word to the [1D3H] register; 4. Write 0x101 to the [1D1H] register to start DFT calculation; 5. Read the 33rd - 42nd order component of each phase’s voltage and current in the 2nd - 11th registers when DFT calculation is complete after 0.5s (typical). For example, if the current frequency is 50Hz, the value of the [1D7H] register is 51200 (or 0xC800). Write 0xC800 to the [1D4H] register. Meanwhile 51200*15.872=812646.4=0xC 6666, so write 0x0C to the [1D2H] register and 0x6666 to the [1D3H] register. Of course, if not consider the influence of frequency changes to harmonic analysis, step 3 can be simplified as: write 0xC to the [1D2H] register and 0x6666 to the [1D3H] register. It should be pointed out that because of the multiplex of some registers, the 90E36/36A can not provide harmonic analysis of the 2nd - 32nd and the 33rd - 42nd simultaneously. The common practice is to read the 2nd - 32nd harmonic analysis first, and then read the 33rd - 42nd harmonic analysis. Hence their corresponding intervals are different. The THD data of each phase’s voltage/current is analyzed based on current calculated harmonic. 2nd - 32nd harmonic analysis: 31 THDI = | X (k ) | 2 k =1 | X ( 0) | 2 × 100% 33rd - 42nd harmonic analysis: 41 THD II = | X (k ) | k = 32 | X (0) |2 2 × 100 % So the THD for 2nd - 42nd order harmonic: THDT = THDI 2 + THDII 2 42 M90E36A [Application Note] Atmel-46104A-SE-M90E36A-ApplicationNote_050514 Revision History Doc. Rev. Date 46104A 5/5/2014 Comments Initial release. X X X X Atmel Corporation 1600 Technology Drive, San Jose, CA 95110 USA T: (+1)(408) 441.0311 F: (+1)(408) 436.4200 | www.atmel.com © 2014 Atmel Corporation. All rights reserved. / Rev.: Atmel-46104A-SE-M90E36A-ApplicationNote_050514. Atmel®, Atmel logo and combinations thereof, Enabling Unlimited Possibilities®, and others are registered trademarks or trademarks of Atmel Corporation or its subsidiaries. Other terms and product names may be trademarks of others. DISCLAIMER: The information in this document is provided in connection with Atmel products. No license, express or implied, by estoppel or otherwise, to any intellectual property right is granted by this document or in connection with the sale of Atmel products. EXCEPT AS SET FORTH IN THE ATMEL TERMS AND CONDITIONS OF SALES LOCATED ON THE ATMEL WEBSITE, ATMEL ASSUMES NO LIABILITY WHATSOEVER AND DISCLAIMS ANY EXPRESS, IMPLIED OR STATUTORY WARRANTY RELATING TO ITS PRODUCTS INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTY OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, OR NON-INFRINGEMENT. IN NO EVENT SHALL ATMEL BE LIABLE FOR ANY DIRECT, INDIRECT, CONSEQUENTIAL, PUNITIVE, SPECIAL OR INCIDENTAL DAMAGES (INCLUDING, WITHOUT LIMITATION, DAMAGES FOR LOSS AND PROFITS, BUSINESS INTERRUPTION, OR LOSS OF INFORMATION) ARISING OUT OF THE USE OR INABILITY TO USE THIS DOCUMENT, EVEN IF ATMEL HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES. Atmel makes no representations or warranties with respect to the accuracy or completeness of the contents of this document and reserves the right to make changes to specifications and products descriptions at any time without notice. Atmel does not make any commitment to update the information contained herein. Unless specifically provided otherwise, Atmel products are not suitable for, and shall not be used in, automotive applications. Atmel products are not intended, authorized, or warranted for use as components in applications intended to support or sustain life. SAFETY-CRITICAL, MILITARY, AND AUTOMOTIVE APPLICATIONS DISCLAIMER: Atmel products are not designed for and will not be used in connection with any applications where the failure of such products would reasonably be expected to result in significant personal injury or death (“Safety-Critical Applications”) without an Atmel officer's specific written consent. Safety-Critical Applications include, without limitation, life support devices and systems, equipment or systems for the operation of nuclear facilities and weapons systems. Atmel products are not designed nor intended for use in military or aerospace applications or environments unless specifically designated by Atmel as military-grade. Atmel products are not designed nor intended for use in automotive applications unless specifically designated by Atmel as automotive-grade.