Polyphase Multifunction Energy Metering IC
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E
Information furnished by Analog Devices is believed to be accurate and reliable. However, no
responsibility is assumed by Analog Devices for its use, nor for any infringements of patents or other
rights of third parties that may result from its use. Specifications subject to change without notice. No
license is granted by implication or otherwise under any patent or patent rights of Analog Devices.
Trademarks and registered trademarks are the property of their respective owners.
One Technology Way, P.O. Box 9106, Norwood, MA 02062-9106, U.S.A.
Tel: 781.329.4700 www.analog.com
Fax: 781.461.3113 ©2010–2011 Analog Devices, Inc. All rights reserved.
FEATURES
Highly accurate; supports EN 50470-1, EN 50470-3,
IEC 62053-21, IEC 62053-22, and IEC 62053-23 standards
Compatible with 3-phase, 3- or 4-wire (delta or wye), and
other 3-phase services
Supplies total (fundamental and harmonic) active, reactive
(ADE7878, ADE7868, and ADE7858 only), and apparent
energy, and fundamental active/reactive energy (ADE7878
only) on each phase and on the overall system
Less than 0.1% error in active and reactive energy over a
dynamic range of 1000 to 1 at TA = 25°C
Less than 0.2% error in active and reactive energy over a
dynamic range of 3000 to 1 at TA = 25°C
Supports current transformer and di/dt current sensors
Dedicated ADC channel for neutral current input (ADE7868 and
ADE7878 only)
Less than 0.1% error in voltage and current rms over a
dynamic range of 1000 to 1 at TA = 25°C
Supplies sampled waveform data on all three phases and on
neutral current
Selectable no load threshold levels for total and
fundamental active and reactive powers, as well as for
apparent powers
Low power battery mode monitors phase currents for
antitampering detection (ADE7868 and ADE7878 only)
Battery supply input for missing neutral operation
Phase angle measurements in both current and voltage
channels with a typical 0.3° error
Wide-supply voltage operation: 2.4 V to 3.7 V
Reference: 1.2 V (drift 10 ppm/°C typical) with external
overdrive capability
Single 3.3 V supply
40-lead lead frame chip scale package (LFCSP), Pb-free
Operating temperature: −40°C to +85°C
Flexible I2C, SPI, and HSDC serial interfaces
APPLICATIONS
Energy metering systems
GENERAL DESCRIPTION
The ADE7854/ADE7858/ADE7868/ADE7878 are high
accuracy, 3-phase electrical energy measurement ICs with serial
interfaces and three flexible pulse outputs. The ADE78xx devices
incorporate second-order sigma-delta (Σ-∆) analog-to-digital
converters (ADCs), a digital integrator, reference circuitry, and
all of the signal processing required to perform total (fundamental
and harmonic) active, reactive (ADE7878, ADE7868, and
ADE7858), and apparent energy measurement and rms calcu-
lations, as well as fundamental-only active and reactive energy
measurement (ADE7878) and rms calculations. A fixed function
digital signal processor (DSP) executes this signal processing.
The DSP program is stored in the internal ROM memory.
The ADE7854/ADE7858/ADE7868/ADE7878 are suitable for
measuring active, reactive, and apparent energy in various 3-phase
configurations, such as wye or delta services, with both three
and four wires. The ADE78xx devices provide system calibration
features for each phase, that is, rms offset correction, phase
calibration, and gain calibration. The CF1, CF2, and CF3 logic
outputs provide a wide choice of power information: total active,
reactive, and apparent powers, or the sum of the current rms
values, and fundamental active and reactive powers.
The ADE7854/ADE7858/ADE7868/ADE7878 contain wave-
form sample registers that allow access to all ADC outputs. The
devices also incorporate power quality measurements, such as
short duration low or high voltage detections, short duration
high current variations, line voltage period measurement, and
angles between phase voltages and currents. Two serial interfaces,
SPI and I2C, can be used to communicate with the ADE78xx. A
dedicated high speed interface, the high speed data capture
(HSDC) port, can be used in conjunction with I2C to provide
access to the ADC outputs and real-time power information.
The ADE7854/ADE7858/ADE7868/ADE7878 also have two
interrupt request pins, IRQ0 and IRQ1, to indicate that an enabled
interrupt event has occurred. For the ADE7868/ADE7878, three
specially designed low power modes ensure the continuity of
energy accumulation when the ADE7868/ADE7878 is in a tam-
pering situation. See for a quick reference chart listing
each part and its functions. The ADE78xx are available in the
40-lead LFCSP, Pb-free package.
Table 1
Table 1. Part Comparison
Part No. WATT VAR
IRMS,
VRMS,
and
VA di/dt
Fundamental
WATT and
VAR
Tamper
Detect
and Low
Power
Modes
ADE7878 Yes Yes Yes Yes Yes Yes
ADE7868 Yes Yes Yes Yes No Yes
ADE7858 Yes Yes Yes Yes No No
ADE7854 Yes No Yes Yes No No
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 2 of 96
TABLE OF CONTENTS
Features .............................................................................................. 1
Applications....................................................................................... 1
General Description......................................................................... 1
Revision History ............................................................................... 3
Functional Block Diagrams............................................................. 4
Specifications..................................................................................... 8
Timing Characteristics .............................................................. 11
Absolute Maximum Ratings.......................................................... 14
Thermal Resistance .................................................................... 14
ESD Caution................................................................................ 14
Pin Configuration and Function Descriptions........................... 15
Typical Performance Characteristics ........................................... 17
Test Circuit ...................................................................................... 19
Terminology .................................................................................... 20
Power Management........................................................................ 21
PSM0—Normal Power Mode (All Parts)................................ 21
PSM1—Reduced Power Mode (ADE7868, ADE7878 Only)21
PSM2—Low Power Mode (ADE7868, ADE7878 Only)....... 21
PSM3—Sleep Mode (All Parts) ................................................ 22
Power-Up Procedure.................................................................. 24
Hardware Reset........................................................................... 25
Software Reset Functionality .................................................... 25
Theory of Operation ...................................................................... 26
Analog Inputs.............................................................................. 26
Analog-to-Digital Conversion.................................................. 26
Current Channel ADC............................................................... 27
di/dt Current Sensor and Digital Integrator .............................. 29
Voltage Channel ADC ............................................................... 30
Changing Phase Voltage Datapath........................................... 31
Power Quality Measurements................................................... 32
Phase Compensation ................................................................. 37
Reference Circuit........................................................................ 39
Digital Signal Processor............................................................. 39
Root Mean Square Measurement............................................. 40
Active Power Calculation.......................................................... 44
Reactive Power Calculation—ADE7858, ADE7868, ADE7878
Only.............................................................................................. 49
Apparent Power Calculation..................................................... 54
Waveform Sampling Mode ....................................................... 57
Energy-to-Frequency Conversion............................................ 57
No Load Condition.................................................................... 61
Checksum Register..................................................................... 63
Interrupts..................................................................................... 64
Serial Interfaces .......................................................................... 65
ADE7878 Evaluation Board...................................................... 72
Die Version.................................................................................. 72
Silicon Anomaly ............................................................................. 73
ADE7854/ADE7858/ADE7868/ADE7878 Functionality
Issues ............................................................................................ 73
Functionality Issues.................................................................... 73
Section 1. ADE7854/ADE7858/ADE7868/ADE7878
Functionality Issues.................................................................... 74
Registers List ................................................................................... 75
Outline Dimensions ....................................................................... 93
Ordering Guide .......................................................................... 93
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 3 of 96
REVISION HISTORY
4/11—Rev. D to Rev. E
Changes to Input Clock FrequencyParameter, Table 2 ..............10
Changes to Current RMS Offset Compensation Section ..........42
Changes to Voltage RMS Offset Compensation Section ...........44
Changes to Note 2, Table 30...........................................................77
Changes to Address 0xE707, Table 33 ..........................................80
Changes to Table 45 ........................................................................87
Changes to Table 46 ........................................................................88
Changes to Bit Location 7:3, Default Value, Table 54.................92
2/11—Rev. C to Rev. D
Changes to Figure 1...........................................................................4
Changes to Figure 2...........................................................................5
Changes to Figure 3...........................................................................6
Changes to Figure 4...........................................................................7
Changes to Table 2 ............................................................................8
Changed SCLK Edge to HSCLK Edge, Table 5 ...........................13
Change to Current Channel HPF Section ...................................28
Change to di/dt Current Sensor and Digital Integrator
Section ..............................................................................................30
Changes to Digital Signal Processor Section...............................39
Changes to Figure 59 ......................................................................44
Changes to Figure 62 ......................................................................47
Changes to Figure 65 ......................................................................49
Changes to Figure 66 ......................................................................52
Changes to Line Cycle Reactive Energy Accumulation Mode
Section and to Figure 67.................................................................53
No Load Detection Based On Total Active, Reactive Powers
Section ..............................................................................................61
Change to Equation 50 ...................................................................63
Changes to the HSDC Interface Section......................................70
Changes to Figure 87 and Figure 88 .............................................71
Changes to Figure 89 ......................................................................72
Changes to Table 30 ........................................................................77
11/10—Rev. B to Rev. C
Change to Signal-to-Noise-and-Distortion Ratio, SINAD
Parameter, Table 1 .............................................................................9
Changes to Figure 18 ......................................................................18
Changes to Figure 22 ......................................................................19
Changes to Silicon Anomaly Section............................................72
Added Table 28 to Silicon Anomaly Section, Renumbered
Tables Sequentially..........................................................................73
8/10—Rev. A to Rev. B
Changes to Figure 1 ..........................................................................4
Changes to Figure 2 ..........................................................................5
Changes to Figure 3 ..........................................................................6
Changes to Figure 4 ..........................................................................7
Change to Table 8............................................................................16
Changes to Power-Up Procedure Section....................................23
Changes to Equation 6 and Equation 7........................................33
Changes to Equation 17 .................................................................43
Changes to Active Power Offset Calibration Section.................45
Changes to Figure 63 ......................................................................46
Changes to Reactive Power Offset Calibration Section .............49
Changes to Figure 82 ......................................................................65
Added Silicon Anomaly Section, Renumbered Tables
Sequentially......................................................................................71
3/10—Rev. 0 to Rev. A
Added ADE7854, ADE7858, and ADE7878 .................. Universal
Reorganized Layout ........................................................... Universal
Added Table 1, Renumbered Sequentially.....................................1
Added Figure 1, Renumbered Sequentially...................................3
Added Figure 2 ..................................................................................4
Added Figure 3 ..................................................................................5
Changes to Specifications Section ..................................................7
Changes to Figure 9 ........................................................................14
Changes to Table 8 ..........................................................................14
Changes to Typical Performance Characteristics Section .........16
Changes to Figure 22 ......................................................................18
Changes to the Power Management Section ...............................20
Changes to the Theory of Operation Section..............................25
Changes to Figure 31 and Figure 32 .............................................27
Change to Equation 28...................................................................47
Changes to Figure 83 ......................................................................66
Changes to Figure 86 ......................................................................68
Changes to the Registers List Section...........................................72
Changes to Ordering Guide...........................................................91
2/10—Revision 0: Initial Version
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 4 of 96
FUNCTIONAL BLOCK DIAGRAMS
PGA1
PGA1
PGA1
PGA3
PGA3
PGA3
1.2V
REF
DIGITAL SIGNAL
PROCESSOR
AVG AINAPHCAL
HPFDIS
[23:0]
HPF
AIGAIN
HPFDIS
[23:0]
DIGITAL
INTEGRATOR
HPF
POR LDO LDO
X
2
AIRMS
LPF
AIRMSOS
X
2
AVRMS
LPF
LPF
AVRMS OS
AVAGAIN
AWGAIN DFC
CFxDEN
:
DFC
CFxDEN
:
DFC
CFxDEN
:
52426 25174
7
8
9
22
12
13
14
19
18
39
37
38
36
32
29
35
34
33
3
2
27
28
23
6
SPI/I
2
C
I
2
C
RESET REF
IN/OUT
VDD AGND AVDD DVDD DGND
CLKIN
CLKOUT
IAP
IAN
VAP
IBP
IBN
VBP
ICP
ICN
VCP
VN
PM0
PM1
CF1
CF2
CF3/HSCLK
IRQ0
IRQ1
SCLK/SCL
MOSI/SDA
MISO/HSD
SS/HSA
HSDC
ADE7854
ADC
ADC
ADC
ADC
ADC
ADC
08510-204
TOTAL ACTIVE/APPARENT
ENERGIES AND VOLTAGE/
CURRENT RMS CALCULATION FOR
PHASE C
(SEE PHASE A FOR DETAILED
DATA PATH)
TOTAL ACTIVE/APPARENT
ENERGIES AND VOLTAGE/
CURRENT RMS CALCULATION FOR
PHASE B
(SEE PHASE A FOR DETAILED
DATA PATH)
PHASE A,
PHASE B,
AND
PHASE C
DATA
AWATTOS
Figure 1. ADE7854 Functional Block Diagram
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 5 of 96
PGA1
PGA1
PGA1
PGA3
PGA3
PGA3
1.2V
REF
DIGITAL SIGNAL
PROCESSOR
COMPUTATIONAL
BLOCK FOR
TOTAL
REACTIVE POWER
AVGAINAPHCAL
HPFDIS
[23:0]
HPF
AIGAIN
HPFDIS
[23:0]
DIGITAL
INTEGRATOR
HPF
POR LDO LDO
X2AIRMS
LPF
AIRMSOS
X2AVRMS
LPF
LPF
AVRMSOS
AVAGAIN
AVARGAIN AVAROS
DFC
CF1DEN
:
DFC
CF2DEN
:
DFC
CF3DEN
:
52426 25174
7
8
9
22
12
13
14
19
18
39
37
38
36
32
29
35
34
33
3
2
27
28
23
6
SPI/I2C
I2C
RESET REFIN/OUT VDD AGND AVDD DVDD DGND
CLKIN
CLKOUT
IAP
IAN
VAP
IBP
IBN
VBP
ICP
ICN
VCP
VN
PM0
PM1
CF1
CF2
CF3/HSCLK
IRQ0
IRQ1
SCLK/SCL
MOSI/SDA
MISO/HSD
SS/HSA
HSDC
ADE7858
ADC
ADC
ADC
ADC
ADC
ADC
08510-203
TOTAL ACTIVE/REACTIVE/
APPARENT/ENERGIES AND
VOLTAGE/CURRENT
RMS CALCULATION FOR PHASE B
(SEE PHASE A FOR DETAILED
DATA PATH)
TOTAL ACTIVE/REACTIVE/
APPARENT/ENERGIES AND
VOLTAGE/CURRENT
RMS CALCULATION FOR PHASE C
(SEE PHASE A FOR DETAILED
DATA PATH)
PHASE A,
PHASE B,
AND
PHASE C
DATA
AWGAIN AWATTOS
Figure 2. ADE7858 Functional Block Diagram
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 6 of 96
PGA1
PGA1
PGA1
PGA3
PGA3
PGA3
1.2V
REF
TOTAL ACTIVE/REACTIVE/
APPARENT/ENERGIES AND
VOLTAGE/CURRENT
RMS CALCULATION FOR PHASE B
(SEE PHASE A FOR DETAILED
DATA PATH)
TOTAL ACTIVE/REACTIVE/
APPARENT/ENERGIES AND
VOLTAGE/CURRENT
RMS CALCULATION FOR PHASE C
(SEE PHASE A FOR DETAILED
DATA PATH)
DIGITAL SIGNAL
PROCESSOR
COMPUTATIONAL
BLOCK FOR
TOTAL
REACTIVE POWER
AVGAINAPHCAL
HPFDIS
[23:0]
HPF
NIGAIN
HPFDIS
[23:0]
HPF
AIGAIN
HPFDIS
[23:0]
DIGITAL
INTEGRATOR
HPF
DIGITAL
INTEGRATOR
POR LDO LDO
X
2
AIRMS
LPF
AIRMSOS
X
2
X
2
AVRMS
LPF
LPF
AVRMSOS
AVAGAIN
LPF
NIRMS
NIRMSOS
AWGAIN AWATTOS
AVARGAIN AVAROS
DFC
CF1DEN
:
DFC
CF2DEN
:
DFC
CF3DEN
:
52426 25174
7
8
9
22
12
13
14
19
18
39
37
38
36
32
29
35
34
33
3
2
27
28
23
6
SPI/I
2
C
I
2
C
RESET REF
IN/OUT
VDD AGND AVDD DVDD DGND
CLKIN
CLKOUT
IAP
IAN
VAP
IBP
IBN
VBP
ICP
ICN
VCP
VN
PM0
PM1
CF1
CF2
CF3/HSCL
K
IRQ0
IRQ1
SCLK/SCL
MOSI/SDA
MISO/HSD
SS/HSA
HSDC
ADE7868
ADC
ADC
ADC
ADC
ADC
ADC
PGA2
15
16
INP
INN
ADC
08510-202
PHASE A,
PHASE B,
AND
PHASE C
DATA
Figure 3. ADE7868 Functional Block Diagram
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 7 of 96
PGA1
PGA1
PGA1
PGA3
PGA3
PGA3
1.2V
REF
TOTAL/FUNDAMENTAL ACTIVE/
REACTIVE ENERGIES, APPARENT
ENERGY AND VOLTAGE/CURRENT
RMS CALCULATION FOR PHASE B
(SEE PHASE A FOR DETAILED
DATA PATH)
TOTAL/FUNDAMENTAL ACTIVE/
REACTIVE ENERGIES, APPARENT
ENERGY AND VOLTAGE/CURRENT
RMS CALCULATION FOR PHASE C
(SEE PHASE A FOR DETAILED
DATA PATH)
DIGITAL SIGNAL
PROCESSOR
COMPUTATIONAL
BLOCK FOR
TOTAL
REACTIVE POWER
AVGAINAPHCAL
HPFDIS
[23:0]
HPF
NIGAIN
HPFDIS
[23:0]
HPF
AIGAIN
HPFDIS
[23:0]
DIGITAL
INTEGRATOR
HPF
DIGITAL
INTEGRATOR
POR LDO LDO
X
2
AIRMS
LPF
AIRMSOS
X
2
X
2
AVRMS
LPF
LPF
AVRMSOS
AVAGAIN
LPF
COMPUTATIONAL
BLOCK FOR
FUNDAMENTAL
ACTIVE AND
REACTIVE POWER
AWGAIN
NIRMS
NIRMSOS
AWATTOS
AVAROS
AFWATTOS
AVARGAIN
AFWGAIN
AFVARGAIN AFVAROS
PHASE A,
PHASE B,
AND
PHASE C
DATA
DFC
CF1DEN
:
DFC
CF2DEN
:
DFC
CF3DEN
:
52426 25174
7
8
9
22
12
13
14
19
18
39
37
38
36
32
29
35
34
33
3
2
27
28
23
6
SPI/I
2
C
I
2
C
RESET REF
IN/OUT
VDD AGND AVDD DVDD DGND
CLKIN
C
LKOUT
IAP
IAN
VAP
IBP
IBN
VBP
ICP
ICN
VCP
VN
PM0
PM1
CF1
CF2
CF3/HSCLK
IRQ0
IRQ1
SCLK/SCL
MOSI/SDA
MISO/HSD
SS/HSA
HSDC
ADE7878
ADC
ADC
ADC
ADC
ADC
ADC
PGA2
15
16
INP
INN
ADC
08510-201
Figure 4. ADE7878 Functional Block Diagram
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 8 of 96
SPECIFICATIONS
VDD = 3.3 V ± 10%, AGND = DGND = 0 V, on-chip reference, CLKIN = 16.384 MHz, TMIN to TMAX = −40°C to +85°C.
Table 2.
Parameter1, 2 Min Typ Max Unit Test Conditions/Comments
ACCURACY
Active Energy Measurement
Active Energy Measurement Error
(per Phase)
Total Active Power 0.1 % Over a dynamic range of 1000 to 1, PGA = 1, 2, 4;
integrator off
0.2 %
Over a dynamic range of 3000 to 1, PGA = 1, 2, 4;
integrator off
0.1 %
Over a dynamic range of 500 to 1, PGA = 8, 16;
integrator on
Fundamental Active Power (ADE7878
Only)
0.1 % Over a dynamic range of 1000 to 1, PGA = 1, 2, 4;
integrator off
0.2 %
Over a dynamic range of 3000 to 1, PGA = 1, 2, 4;
integrator off
0.1 %
Over a dynamic range of 500 to 1, PGA = 8, 16;
integrator on
Phase Error Between Channels Line frequency = 45 Hz to 65 Hz, HPF on
Power Factor (PF) = 0.8 Capacitive ±0.05 Degrees Phase lead 37°
PF = 0.5 Inductive ±0.05 Degrees Phase lag 60°
AC Power Supply Rejection VDD = 3.3 V + 120 mV rms/120 Hz, IPx = VPx =
±100 mV rms
Output Frequency Variation 0.01 %
DC Power Supply Rejection VDD = 3.3 V ± 330 mV dc
Output Frequency Variation 0.01 %
Total Active Energy Measurement
Bandwidth
2 kHz
REACTIVE ENERGY MEASUREMENT
(ADE7858, ADE7868, AND ADE7878)
Reactive Energy Measurement Error
(per Phase)
Total Active Power 0.1 % Over a dynamic range of 1000 to 1, PGA = 1, 2, 4;
integrator off
0.2 %
Over a dynamic range of 3000 to 1, PGA = 1, 2, 4;
integrator off
0.1 %
Over a dynamic range of 500 to 1, PGA = 8, 16;
integrator on
Fundamental Active Power (ADE7878
Only)
0.1 % Over a dynamic range of 1000 to 1, PGA = 1, 2, 4;
integrator off
0.2 %
Over a dynamic range of 3000 to 1, PGA = 1, 2, 4;
integrator off
0.1 %
Over a dynamic range of 500 to 1, PGA = 8, 16;
integrator on
Phase Error Between Channels Line frequency = 45 Hz to 65 Hz, HPF on
PF = 0.8 Capacitive ±0.05 Degrees Phase lead 37°
PF = 0.5 Inductive ±0.05 Degrees Phase lag 60°
AC Power Supply Rejection VDD = 3.3 V + 120 mV rms/120 Hz, IPx = VPx =
±100 mV rms
Output Frequency Variation 0.01 %
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 9 of 96
Parameter1, 2 Min Typ Max Unit Test Conditions/Comments
DC Power Supply Rejection VDD = 3.3 V ± 330 mV dc
Output Frequency Variation 0.01 %
Total Reactive Energy Measurement
Bandwidth
2 kHz
RMS MEASUREMENTS
I rms and V rms Measurement
Bandwidth
2 kHz
I rms and V rms Measurement Error
(PSM0 Mode)
0.1 % Over a dynamic range of 1000 to 1, PGA = 1
MEAN ABSOLUTE VALUE (MAV)
MEASUREMENT (ADE7868 AND ADE7878)
I mav Measurement Bandwidth (PSM1
Mode)
260 Hz
I mav Measurement Error (PSM1 Mode) 0.5 % Over a dynamic range of 100 to 1, PGA = 1, 2, 4, 8
ANALOG INPUTS
Maximum Signal Levels ±500 mV peak Differential inputs between the following pins:
IAP and IAN, IBP and IBN, ICP and ICN; single-
ended inputs between the following pins: VAP
and VN, VBP and VN, VCP and VN
Input Impedance (DC)
IAP, IAN, IBP, IBN, ICP, ICN, VAP, VBP,
and VCP Pins
400 kΩ
VN Pin 130 kΩ
ADC Offset Error ±2 mV PGA = 1, uncalibrated error, see the Terminology
section
Gain Error ±4 % External 1.2 V reference
WAVEFORM SAMPLING Sampling CLKIN/2048, 16.384 MHz/2048 = 8 kSPS
Current and Voltage Channels See the Waveform Sampling Mode section
Signal-to-Noise Ratio, SNR 70 dB PGA = 1
Signal-to-Noise-and-Distortion Ratio,
SINAD
60 dB PGA = 1
Bandwidth (−3 dB) 2 kHz
TIME INTERVAL BETWEEN PHASES
Measurement Error 0.3 Degrees Line frequency = 45 Hz to 65 Hz, HPF on
CF1, CF2, CF3 PULSE OUTPUTS
Maximum Output Frequency 8 kHz WTHR = VARTHR = VATHR = PMAX = 33,516,139
Duty Cycle 50 % If CF1, CF2, or CF3 frequency > 6.25 Hz and
CFDEN is even and > 1
(1 + 1/CFDEN)
× 50%
If CF1, CF2, or CF3 frequency > 6.25 Hz and
CFDEN is odd and > 1
Active Low Pulse Width 80 ms If CF1, CF2, or CF3 frequency < 6.25 Hz
Jitter 0.04 %
For CF1, CF2, or CF3 frequency = 1 Hz and
nominal phase currents are larger than 10% of
full scale
REFERENCE INPUT
REFIN/OUT Input Voltage Range 1.1 1.3 V Minimum = 1.2 V − 8%; maximum = 1.2 V + 8%
Input Capacitance 10 pF
ON-CHIP REFERENCE Nominal 1.207 V at the REFIN/OUT pin at TA = 25°C
PSM0 and PSM1 Modes
Reference Error ±2 mV
Output Impedance 1.2 kΩ
Temperature Coefficient 10 50 ppm/°C Maximum value across full temperature range
of −40°C to +85°C
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 10 of 96
Parameter1, 2 Min Typ Max Unit Test Conditions/Comments
CLKIN All specifications CLKIN of 16.384 MHz
Input Clock Frequency 16.22 16.384 16.55 MHz
Crystal Equivalent Series Resistance 30 200
CLKIN Input Capacitance 20 pF
CLKOUT Output Capacitance 20 pF
LOGIC INPUTS—MOSI/SDA, SCLK/SCL, SS,
RESET, PM0, AND PM1
Input High Voltage, VINH 2.0 V VDD = 3.3 V ± 10%
Input Low Voltage, VINL 0.8 V VDD = 3.3 V ± 10%
Input Current, IIN −8.7 µA Input = 0 V, VDD = 3.3 V
3 A Input = VDD = 3.3 V
100 nA Input = VDD = 3.3 V
Input Capacitance, CIN 10 pF
LOGIC OUTPUTSIRQ0, IRQ1, MISO/HSD VDD = 3.3 V ± 10%
Output High Voltage, VOH 2.4 V VDD = 3.3 V ± 10%
ISOURCE 800 µA
Output Low Voltage, VOL 0.4 V VDD = 3.3 V ± 10%
ISINK 2 mA
CF1, CF2, CF3/HSCLK
Output High Voltage, VOH 2.4 V VDD = 3.3 V ± 10%
ISOURCE 500 µA
Output Low Voltage, VOL 0.4 V VDD = 3.3 V ± 10%
ISINK 2 mA
POWER SUPPLY For specified performance
PSM0 Mode
VDD Pin 3.0 3.6 V Minimum = 3.3 V − 10%; maximum = 3.3 V + 10%
IDD 24 26.8 mA
PSM1 and PSM2 Modes (ADE7868 and
ADE7878)
VDD Pin 2.4 3.7 V
IDD
PSM1 Mode 6.0 mA
PSM2 Mode 0.2 mA
PSM3 Mode For specified performance
VDD Pin 2.4 3.7 V
IDD in PSM3 Mode 1.7 A
1 See the Typical Performance Characteristics section.
2 See the Terminology section for a definition of the parameters.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 11 of 96
TIMING CHARACTERISTICS
VDD = 3.3 V ± 10%, AGND = DGND = 0 V, on-chip reference, CLKIN = 16.384 MHz, TMIN to TMAX = −40°C to +85°C. Note that dual
function pin names are referenced by the relevant function only within the timing tables and diagrams; see the Pin Configuration and
Function Descriptions section for full pin mnemonics and descriptions.
Table 3. I2C-Compatible Interface Timing Parameter
Standard Mode Fast Mode
Parameter Symbol Min Max Min Max Unit
SCL Clock Frequency fSCL 0 100 0 400 kHz
Hold Time (Repeated) Start Condition tHD;STA 4.0 0.6 s
Low Period of SCL Clock tLOW 4.7 1.3 µs
High Period of SCL Clock tHIGH 4.0 0.6 µs
Set-Up Time for Repeated Start Condition tSU;STA 4.7 0.6 µs
Data Hold Time tHD;DAT 0 3.45 0 0.9 µs
Data Setup Time tSU;DAT 250 100 ns
Rise Time of Both SDA and SCL Signals tR 1000 20 300 ns
Fall Time of Both SDA and SCL Signals tF 300 20 300 ns
Setup Time for Stop Condition tSU;STO 4.0 0.6 µs
Bus Free Time Between a Stop and Start Condition tBUF 4.7 1.3 µs
Pulse Width of Suppressed Spikes tSP N/A1 50 ns
1 N/A means not applicable.
t
F
t
r
t
HD;DAT
t
HD;STA
t
HIGH
t
SU;STA
t
SU;DAT
t
f
t
HD;STA
t
SP
t
SU;STO
t
r
t
BUF
t
LOW
SDA
SCL
K
START
CONDITION
REPEATED START
CONDITION
STOP
CONDITION
START
CONDITION
08510-002
Figure 5. I2C-Compatible Interface Timing
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 12 of 96
Table 4. SPI Interface Timing Parameters
Parameter Symbol Min Max Unit
SS to SCLK Edge tSS 50 ns
SCLK Period 0.4 40001 s
SCLK Low Pulse Width tSL 175 ns
SCLK High Pulse Width tSH 175 ns
Data Output Valid After SCLK Edge tDAV 100 ns
Data Input Setup Time Before SCLK Edge tDSU 100 ns
Data Input Hold Time After SCLK Edge tDHD 5 ns
Data Output Fall Time tDF 20 ns
Data Output Rise Time tDR 20 ns
SCLK Rise Time tSR 20 ns
SCLK Fall Time tSF 20 ns
MISO Disable After SS Rising Edge tDIS 200 ns
SS High After SCLK Edge tSFS 0 ns
1 Guaranteed by design.
MSB LSB
LSB IN
INTERMEDIATE BITS
INTERMEDIATE BITS
t
SFS
t
DIS
t
SS
t
SL
t
DF
t
SH
t
DHD
t
DAV
t
DSU
t
SR
t
SF
t
DR
MSB IN
MOSI
MISO
SCLK
SS
08510-003
Figure 6. SPI Interface Timing
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 13 of 96
Table 5. HSDC Interface Timing Parameter
Parameter Symbol Min Max Unit
HSA to HSCLK Edge tSS 0 ns
HSCLK Period 125 ns
HSCLK Low Pulse Width tSL 50 ns
HSCLK High Pulse Width tSH 50 ns
Data Output Valid After HSCLK Edge tDAV 40 ns
Data Output Fall Time tDF 20 ns
Data Output Rise Time tDR 20 ns
HSCLK Rise Time tSR 10 ns
HSCLK Fall Time tSF 10 ns
HSD Disable After HSA Rising Edge tDIS 5 ns
HSA High After HSCLK Edge tSFS 0 ns
MSB LSBINTERMEDIATE BITS
t
SFS
t
DIS
t
SS
t
SL
t
DF
t
SH
t
DAV
t
SR
t
SF
t
DR
HSD
HSCL
K
HSA
08510-004
Figure 7. HSDC Interface Timing
2mA I
OL
800µA I
OH
1.6V
TO OUTPUT
PIN C
L
50pF
08510-005
Figure 8. Load Circuit for Timing Specifications
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 14 of 96
ABSOLUTE MAXIMUM RATINGS
TA = 25°C, unless otherwise noted. THERMAL RESISTANCE
Table 6.
Parameter Rating
VDD to AGND −0.3 V to +3.7 V
VDD to DGND −0.3 V to +3.7 V
Analog Input Voltage to AGND, IAP,
IAN, IBP, IBN, ICP, ICN, VAP, VBP, VCP,
VN
−2 V to +2 V
Analog Input Voltage to INP and INN −2 V to +2 V
Reference Input Voltage to AGND −0.3 V to VDD + 0.3 V
Digital Input Voltage to DGND −0.3 V to VDD + 0.3 V
Digital Output Voltage to DGND −0.3 V to VDD + 0.3 V
Operating Temperature
Industrial Range −40°C to +85°C
Storage Temperature Range −65°C to +150°C
Junction Temperature 150°C
Lead Temperature (Soldering, 10 sec) 300°C
θJA is specified equal to 29.3°C/W; θJC is specified equal to
1.8°C/W.
Table 7. Thermal Resistance
Package Type θJA θ
JC Unit
40-Lead LFCSP 29.3 1.8 °C/W
ESD CAUTION
Stresses above those listed under Absolute Maximum Ratings
may cause permanent damage to the device. This is a stress
rating only; functional operation of the device at these or any
other conditions above those listed in the operational sections
of this specification is not implied. Exposure to absolute
maximum rating conditions for extended periods may affect
device reliability.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 15 of 96
PIN CONFIGURATION AND FUNCTION DESCRIPTIONS
NOTES
1. NC = NO CONNECT.
2. CREATE A SIMILAR PAD ON THE PCB UNDER THE
EXPOSED PAD. SOLDER THE EXPOSED PAD TO
THE PAD ON THE PCB TO CONFER MECHANICAL
STRENGTH TO THE PACKAGE. DO NOT CONNECT
THE PADS TO AGND.
11
NC
12
IBN
13
ICP
15
INP
17
REF
IN/OUT
16
INN
18
VN
19
VCP
20
NC
14
ICN
NC
PM0
PM1
RESET
DVDD
DGND
IAP
IAN
IBP
NC
VAP
AVDD
AGND
VDD
CLKIN
CLKOUT
IRQ0
NC
VBP
NC
33 CF1
34 CF2
35 CF3/HSCL
K
36 SCLK/SCL
37 MISO/HSD
38 MOSI/SDA
39 SS/HSA
40 NC
32 IRQ1
31 NC
08510-106
1
2
3
4
5
6
7
8
9
10
23
24
25
26
27
28
29
30
22
21
ADE78xx
TOP VIEW
(Not to Scale)
Figure 9. Pin Configuration
Table 8. Pin Function Descriptions
Pin No. Mnemonic Description
1, 10, 11, 20,
21, 30, 31, 40
NC No Connect. These pins are not connected internally.
2 PM0 Power Mode Pin 0. This pin, combined with PM1, defines the power mode of the
ADE7854/ADE7858/ADE7868/ADE7878, as described in Table 9.
3 PM1 Power Mode Pin 1. This pin defines the power mode of the ADE7854/ADE7858/ADE7868/ADE7878
when combined with PM0, as described in Table 9.
4 RESET Reset Input, Active Low. In PSM0 mode, this pin should stay low for at least 10 µs to trigger a
hardware reset.
5 DVDD This pin provides access to the on-chip 2.5 V digital LDO. Do not connect any external active
circuitry to this pin. Decouple this pin with a 4.7 µF capacitor in parallel with a ceramic 220 nF
capacitor.
6 DGND Ground Reference. This pin provides the ground reference for the digital circuitry.
7, 8 IAP, IAN Analog Inputs for Current Channel A. This channel is used with the current transducers and is
referenced in this document as Current Channel A. These inputs are fully differential voltage inputs
with a maximum differential level of ±0.5 V. This channel also has an internal PGA equal to the ones
on Channel B and Channel C.
9, 12 IBP, IBN Analog Inputs for Current Channel B. This channel is used with the current transducers and is
referenced in this document as Current Channel B. These inputs are fully differential voltage inputs
with a maximum differential level of ±0.5 V. This channel also has an internal PGA equal to the ones
on Channel C and Channel A.
13, 14 ICP, ICN Analog Inputs for Current Channel C. This channel is used with the current transducers and is
referenced in this document as Current Channel C. These inputs are fully differential voltage inputs
with a maximum differential level of ±0.5 V. This channel also has an internal PGA equal to the ones
on Channel A and Channel B.
15, 16 INP, INN Analog Inputs for Neutral Current Channel N. This channel is used with the current transducers and
is referenced in this document as Current Channel N. These inputs are fully differential voltage
inputs with a maximum differential level of ±0.5 V. This channel also has an internal PGA, different
from the ones found on the A, B, and C channels. The neutral current channel is available in the
ADE7878 and ADE7868. In the ADE7858 and ADE7854, connect these pins to AGND.
17 REFIN/OUT This pin provides access to the on-chip voltage reference. The on-chip reference has a nominal
value of 1.2 V. An external reference source with 1.2 V ± 8% can also be connected at this pin. In
either case, decouple this pin to AGND with a 4.7 µF capacitor in parallel with a ceramic 100 nF
capacitor. After reset, the on-chip reference is enabled.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 16 of 96
Pin No. Mnemonic Description
18, 19, 22, 23 VN, VCP, VBP, VAP Analog Inputs for the Voltage Channel. This channel is used with the voltage transducer and is
referenced as the voltage channel in this document. These inputs are single-ended voltage inputs
with a maximum signal level of ±0.5 V with respect to VN for specified operation. This channel also
has an internal PGA.
24 AVDD This pin provides access to the on-chip 2.5 V analog low dropout regulator (LDO). Do not connect
external active circuitry to this pin. Decouple this pin with a 4.7 µF capacitor in parallel with a
ceramic 220 nF capacitor.
25 AGND Ground Reference. This pin provides the ground reference for the analog circuitry. Tie this pin to the
analog ground plane or to the quietest ground reference in the system. Use this quiet ground
reference for all analog circuitry, for example, antialiasing filters, current, and voltage transducers.
26 VDD Supply Voltage. This pin provides the supply voltage. In PSM0 (normal power mode), maintain the
supply voltage at 3.3 V ± 10% for specified operation. In PSM1 (reduced power mode), PSM2 (low
power mode), and PSM3 (sleep mode), when the ADE7868/ADE7878 is supplied from a battery,
maintain the supply voltage between 2.4 V and 3.7 V. Decouple this pin to AGND with a 10 µF
capacitor in parallel with a ceramic 100 nF capacitor. The only modes available on the ADE7858 and
ADE7854 are the PSM0 and PSM3 power modes.
27 CLKIN Master Clock. An external clock can be provided at this logic input. Alternatively, a parallel resonant
AT-cut crystal can be connected across CLKIN and CLKOUT to provide a clock source for the
ADE7854/ADE7858/ADE7868/ADE7878. The clock frequency for specified operation is 16.384 MHz.
Use ceramic load capacitors of a few tens of picofarad with the gate oscillator circuit. Refer to the
crystal manufacturer’s data sheet for load capacitance requirements.
28 CLKOUT A crystal can be connected across this pin and CLKIN (as previously described with Pin 27 in this
table) to provide a clock source for the ADE7854/ADE7858/ADE7868/ADE7878. The CLKOUT pin can
drive one CMOS load when either an external clock is supplied at CLKIN or a crystal is being used.
29, 32 IRQ0, IRQ1 Interrupt Request Outputs. These are active low logic outputs. See the Interrupts section for a
detailed presentation of the events that can trigger interrupts.
33, 34, 35 CF1, CF2,
CF3/HSCLK
Calibration Frequency (CF) Logic Outputs. These outputs provide power information based on the
CF1SEL[2:0], CF2SEL[2:0], and CF3SEL[2:0] bits in the CFMODE register. These outputs are used for
operational and calibration purposes. The full-scale output frequency can be scaled by writing to the
CF1DEN, CF2DEN, and CF3DEN registers, respectively (see the Energy-to-Frequency Conversion
section). CF3 is multiplexed with the serial clock output of the HSDC port.
36 SCLK/SCL
Serial Clock Input for SPI Port/Serial Clock Input for I2C Port. All serial data transfers are synchronized
to this clock (see the Serial Interfaces section). This pin has a Schmidt trigger input for use with a
clock source that has a slow edge transition time, for example, opto-isolator outputs.
37 MISO/HSD Data Out for SPI Port/Data Out for HSDC Port.
38 MOSI/SDA Data In for SPI Port/Data Out for I2C Port.
39 SS/HSA Slave Select for SPI Port/HSDC Port Active.
EP Exposed Pad
Create a similar pad on the PCB under the exposed pad. Solder the exposed pad to the pad on the
PCB to confer mechanical strength to the package. Do not connect the pads to AGND.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 17 of 96
TYPICAL PERFORMANCE CHARACTERISTICS
–0.40
–0.35
–0.30
–0.25
–0.20
–0.15
–0.10
ERROR (%)
–0.05
0
0.05
0.10
0.01 0.1 1 10 100
PERCENTAGE OF FULL-SCALE CURRENT (%)
+25°C, POWER FACTOR = 1.0
+85°C, POWER FACTOR = 1.0
–40°C, POWER FACTOR = 1.0
0
8510-301
Figure 10. Total Active Energy Error As Percentage of Reading (Gain = +1,
Power Factor = 1) over Temperature with Internal Reference and Integrator Off
–0.25
–0.20
–0.15
–0.10
–0.05
0
0.05
0.10
0.15
45 47 49 51 53 55 57 59 61 63 65
LINE FREQUENCY (Hz)
ERROR (%)
POWER FACTOR = 1
POWER FACTOR = +0.5
POWER FACTOR = –0.5
08510-305
Figure 11. Total Active Energy Error As Percentage of Reading (Gain = +1)
over Frequency with Internal Reference and Integrator Off
–0.15
–0.10
–0.05
0
0.05
0.10
0.15
ERROR (%)
0.01 0.1 1 10 100
PERCENTAGE OF FULL-SCALE CURRENT (%)
V
DD
= 2.97V
V
DD
= 3.30V
V
DD
= 3.63V
08510-306
Figure 12. Total Active Energy Error As Percentage of Reading (Gain = +1,
Power Factor = 1) over Power Supply with Internal Reference and Integrator Off
–0.50
–0.30
0.30
–0.10
0.10
–0.40
–0.20
0
0.20
0.40
0.50
0.001 0.01 0.1 1
PERCENTAGE OF FULL-SCALE CURRENT (%)
ERROR (%)
+85°C, POWER FACTOR = 1.0
+25°C, POWER FACTOR = 1.0
–40°C, POWER FACTOR = 1.0
08510-308
Figure 13. Total Active Energy Error As Percentage of Reading (Gain = +16,
Power Factor = 1) over Temperature with Internal Reference and Integrator On
–0.50
–0.40
–0.30
–0.20
–0.10
0
0.10
0.20
0.30
0.40
0.01 0.1 1 10 100
PERCENTAGE OF FULL-SCALE CURRENT (%)
ERROR (%)
+85°C, POWER FACTOR = 0
+25°C, POWER FACTOR = 0
–40°C, POWER FACTOR = 0
08510-311
Figure 14. Total Reactive Energy Error As Percentage of Reading (Gain = +1,
Power Factor = 0) over Temperature with Internal Reference and Integrator Off
–0.25
–0.20
–0.15
–0.10
–0.05
0
0.05
0.10
45 47 49 51 53 55 57 59 61 63 65
LINE FREQUENCY (Hz)
ERROR (%)
POWER FACTOR = 0
POWER FACTOR = +0.5
POWER FACTOR = –0.5
08510-315
Figure 15. Total Reactive Energy Error As Percentage of Reading (Gain = +1)
over Frequency with Internal Reference and Integrator Off
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 18 of 96
–0.30
–0.20
–0.10
0
0.10
0.20
0.30
0.01 0.1 1 10 100
PERCENTAGE OF FULL-SCALE CURRENT (%)
ERROR (%)
V
DD
= 2.97V
V
DD
= 3.30V
V
DD
= 3.63V
08510-316
Figure 16. Total Reactive Energy Error As Percentage of Reading (Gain = +1,
Power Factor = 0) over Power Supply with Internal Reference and Integrator Off
–0.40
–0.30
–0.20
–0.10
0
0.10
0.20
0.30
0.40
0.50
0.60
0.1 1 10 100
PERCENTAGE OF FULL-SCALE CURRENT (%)
ERROR (%)
+85°C, POWER FACTOR = 0
+25°C, POWER FACTOR = 0
–40°C, POWER FACTOR = 0
08510-318
Figure 17. Total Reactive Energy Error As Percentage of Reading (Gain = +16,
Power Factor = 1) over Temperature with Internal Reference and Integrator On
–0.25
–0.20
–0.15
–0.10
–0.05
0
0.05
0.10
0.15
0.20
ERROR (%)
45 47 49 51 53 55 57 59 61 63 65
LINE FREQUENCY (Hz)
POWER FACTOR = –0.5
POWER FACTOR = 1.0
POWER FACTOR = +0.5
0
8510-335
Figure 18. Fundamental Active Energy Error As Percentage of Reading
(Gain = +1) over Frequency with Internal Reference and Integrator Off
–0.30
–0.20
–0.10
0
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.1 1 10 100
PERCENTAGE OF FULL-SCALE CURRENT (%)
ERROR (%)
08510-338
+25°C, POWER FACTOR = 1.0
+85°C, POWER FACTOR = 1.0
–40°C, POWER FACTOR = 1.0
Figure 19. Fundamental Active Energy Error As Percentage of Reading
(Gain = +16) over Temperature with Internal Reference and Integrator On
–0.25
–0.20
–0.15
–0.10
–0.05
0
0.05
0.10
0.15
45 47 49 51 53 55 57 59 61 63 65
LINE FREQUENCY (Hz)
POWER FACTOR = –0.5
POWER FACTOR = 1.0
POWER FACTOR = +0.5
ERROR (%)
08510-345
Figure 20. Fundamental Reactive Energy Error As Percentage of Reading
(Gain = +1) over Frequency with Internal Reference and Integrator Off
–0.40
–0.30
–0.20
–0.10
0
0.10
0.20
0.30
0.40
0.50
0.1 1 10 100
PERCENTAGE OF FULL-SCALE CURRENT (%)
+25°C, POWER FACTOR = 1.0
–40°C, POWER FACTOR = 1.0
+85°C, POWER FACTOR = 1.0
ERROR (%)
08510-348
Figure 21. Fundamental Reactive Energy Error As Percentage of Reading
(Gain = +16) over Temperature with Internal Reference and Integrator On
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 19 of 96
TEST CIRCUIT
SAME AS
CF2
PM0
0.22µF
4.7µF
MOSI/SDA
MISO/HSD
SCLK/SCL
CF3/HSCLK
CF2
CF1
REF
IN/OUT
CLKOUT
CLKIN
PM1
RESET
IAP
IAN
IBP
IBN
ICP
ICN
VN
VCP
VBP
VAP
2
23
3
4
7
8
9
12
13
14
18
19
22
39
38
37
36
35
34
33
32
29
17
28
27
ADE78xx
24 26 5
AVDD
VDD
DVDD
625
DGND
AGND
0.22µF
4.7µF
0.1µF
4.7µF
20pF +
+ +
3.3V
20pF
16.384MHz
SAME AS
VCP
SAME AS
VCP
SAME AS
IAP, IAN
SAME AS
IAP, IAN
18nF
1k
1k
1k
10k
10k
1.5k
1k18nF
3.3V
3.3
V
1µF
18nF
18nF
SS/HSA
IRQ1
IRQ0
08510-099
Figure 22. Test Circuit
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 20 of 96
TERMINOLOGY
Measurement Error
The error associated with the energy measurement made by the
ADE7854/ADE7858/ADE7868/ADE7878 is defined by
Measurement Error =
%100
7878 ×
EnergyTrue
EnergyTrueADEbyRegisteredEnergy (1)
Phase Error Between Channels
The high-pass filter (HPF) and digital integrator introduce a
slight phase mismatch between the current and the voltage
channel. The all digital design ensures that the phase matching
between the current channels and voltage channels in all three
phases is within ±0.1° over a range of 45 Hz to 65 Hz and ±0.2°
over a range of 40 Hz to 1 kHz. This internal phase mismatch
can be combined with the external phase error (from current
sensor or component tolerance) and calibrated with the phase
calibration registers.
Power Supply Rejection (PSR)
This quantifies the ADE7878 measurement error as a percen-
tage of reading when the power supplies are varied. For the ac
PSR measurement, a reading at nominal supplies (3.3 V) is
taken. A second reading is obtained with the same input signal
levels when an ac signal (120 mV rms at 100 Hz) is introduced
onto the supplies. Any error introduced by this ac signal is
expressed as a percentage of reading—see the Measurement
Error definition.
For the dc PSR measurement, a reading at nominal supplies
(3.3 V) is taken. A second reading is obtained with the same
input signal levels when the power supplies are varied ±10%.
Any error introduced is expressed as a percentage of the
reading.
ADC Offset Error
This refers to the dc offset associated with the analog inputs to
the ADCs. It means that with the analog inputs connected to
AGND, the ADCs still see a dc analog input signal. The magni-
tude of the offset depends on the gain and input range selection
(see the Typical Performance Characteristics section). However,
the HPF removes the offset from the current and voltage channels
and the power calculation remains unaffected by this offset.
Gain Error
The gain error in the ADCs of the ADE7858/ADE7868/ADE7878/
ADE7854 is defined as the difference between the measured
ADC output code (minus the offset) and the ideal output code
(see the Current Channel ADC section and the Voltage Channel
ADC section). The difference is expressed as a percentage of the
ideal code.
CF Jitter
The period of pulses at one of the CF1, CF2, or CF3 pins is
continuously measured. The maximum, minimum, and average
values of four consecutive pulses are computed as follows:
Maximum = max(Period0, Period1, Period2, Period3)
Minimum = min(Period0, Period1, Period2, Period3)
Average = 4
3210 PeriodPeriodPeriodPeriod ++
+
The CF jitter is then computed as
%100×
=Average
MinimumMaximum
CFJITTER (2)
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 21 of 96
POWER MANAGEMENT
The ADE7868/ADE7878 have four modes of operation, deter-
mined by the state of the PM0 and PM1 pins (see Table 9). The
ADE7854/ADE7858 have two modes of operation. These pins
provide complete control of the ADE7854/ADE7858/ADE7868/
ADE7878 operation and can easily be connected to an external
microprocessor I/O. The PM0 and PM1 pins have internal pull-
up resistors. See Table 11 and Table 12 for a list of actions that are
recommended before and after setting a new power mode.
Table 9. Power Supply Modes
Power Supply Modes PM1 PM0
PSM0, Normal Power Mode 0 1
PSM1, Reduced Power Mode1 0 0
PSM2, Low Power Mode1 1 0
PSM3, Sleep Mode 1 1
1 Available in the ADE7868 and ADE7878.
PSM0—NORMAL POWER MODE (ALL PARTS)
In PSM0 mode, the ADE7854/ADE7858/ADE7868/ADE7878
are fully functional. The PM0 pin is set to high and the PM1 pin
is set to low for the ADE78xx to enter this mode. If the ADE78xx
is in one of PSM1, PSM2, or PSM3 modes and is switched into
PSM0 mode, then all control registers take the default values with
the exception of the threshold register, LPOILVL, which is used
in PSM2 mode, and the CONFIG2 register, both of which
maintain their values.
The ADE7854/ADE7858/ADE7868/ADE7878 signal the end of
the transition period by triggering the IRQ1 interrupt pin low and
setting Bit 15 (RSTDONE) in the STATUS1 register to 1. This bit is
0 during the transition period and becomes 1 when the transition is
finished. The status bit is cleared and the IRQ1 pin is set back to
high by writing to the STATUS1 register with the corresponding
bit set to 1. Bit 15 (RSTDONE) in the interrupt mask register
does not have any functionality attached even if the IRQ1 pin goes
low when Bit 15 (RSTDONE) in the STATUS1 register is set to 1.
This makes the RSTDONE interrupt unmaskable.
PSM1—REDUCED POWER MODE (ADE7868,
ADE7878 ONLY)
The reduced power mode, PSM1, is available on the ADE7868
and ADE7878 only. In this mode, the ADE7868/ADE7878
measure the mean absolute values (mav) of the 3-phase currents
and store the results in the AIMAV, BIMAV, and CIMAV 20-bit
registers. This mode is useful in missing neutral cases in which
the voltage supply of the ADE7868 or ADE7878 is provided by an
external battery. The serial ports, I2C or SPI, are enabled in this
mode; the active port can be used to read the AIMAV, BIMAV,
and CIMAV registers. It is not recommended to read any of the
other registers because their values are not guaranteed in this
mode. Similarly, a write operation is not taken into account by
the ADE7868/ADE7878 in this mode.
In summary, in this mode, it is not recommended to access any
register other than AIMAV, BIMAV, and CIMAV. The circuit
that measures these estimates of rms values is also active during
PSM0; therefore, its calibration can be completed in either PSM0
mode or in PSM1 mode. Note that the ADE7868 and ADE7878
do not provide any register to store or process the corrections
resulting from the calibration process. The external microprocessor
stores the gain values in connection with these measurements
and uses them during PSM1 (see the Current Mean Absolute
Value Calculation—ADE7868 and ADE7878 Only section for
more details on the xIMAV registers).
The 20-bit mean absolute value measurements done in PSM1,
although available also in PSM0, are different from the rms
measurements of phase currents and voltages executed only in
PSM0 and stored in the xIRMS and xVRMS 24-bit registers. See
the Current Mean Absolute Value Calculation—ADE7868 and
ADE7878 Only section for details.
If the ADE7868/ADE7878 is set in PSM1 mode while still in the
PSM0 mode, the ADE7868/ADE7878 immediately begin the
mean absolute value calculations without any delay. The xIMAV
registers are accessible at any time; however, if the ADE7878 or
ADE7868 is set in PSM1 mode while still in PSM2 or PSM3
modes, the ADE7868/ADE7878 signal the start of the mean
absolute value computations by triggering the IRQ1 pin low.
The xIMAV registers can be accessed only after this moment.
PSM2—LOW POWER MODE (ADE7868, ADE7878
ONLY)
The low power mode, PSM2, is available on the ADE7868 and
ADE7878 only. In this mode, the ADE7868/ADE7878 compare
all phase currents against a threshold for a period of 0.02 ×
(LPLINE[4:0] + 1) seconds, independent of the line frequency.
LPLINE[4:0] are Bits[7:3] of the LPOILVL register (see Table 10).
Table 10. LPOILVL Register
Bit Mnemonic Default Description
[2:0] LPOIL[2:0] 111 Threshold is put at a value
corresponding to full scale
multiplied by LPOIL/8.
[7:3] LPLINE[4:0] 00000 The measurement period is
(LPLINE[4:0] + 1)/50 sec.
The threshold is derived from Bits[2:0] (LPOIL[2:0]) of the
LPOILVL register as LPOIL[2:0]/8 of full scale. Every time
one phase current becomes greater than the threshold, a
counter is incremented. If every phase counter remains below
LPLINE[4:0] + 1 at the end of the measurement period, then
the IRQ0 pin is triggered low. If a single phase counter becomes
greater or equal to LPLINE[4:0] + 1 at the end of the measure-
ment period, the IRQ1 pin is triggered low. illustrates
how the ADE7868/ADE7878 behave in PSM2 mode when
Figure 23
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 22 of 96
LPLINE[4:0] = 2 and LPOIL[2:0] = 3. The test period is three
50 Hz cycles (60 ms), and the Phase A current rises above the
LPOIL[2:0] threshold three times. At the end of the test period,
the IRQ1 pin is triggered low.
PHASE
COUNTER = 1
PHASE
COUNTER = 2
PHASE
COUNTER = 3
LPLINE[4:0] = 2
IRQ1
IA CURREN
LPOIL[2:0]
THRESHOLD
08510-008
Figure 23. PSM2 Mode Triggering IRQ1 Pin for LPLINE[4:0] = 2
(50 Hz Systems)
The I2C or SPI port is not functional during this mode. The PSM2
mode reduces the power consumption required to monitor the
currents when there is no voltage input and the voltage supply
of the ADE7868/ADE7878 is provided by an external battery. If
the IRQ0 pin is triggered low at the end of a measurement period,
this signifies all phase currents stayed below threshold and,
therefore, there is no current flowing through the system.
At this point, the external microprocessor sets the ADE7868/
ADE7878 into Sleep Mode PSM3. If the IRQ1 pin is triggered
low at the end of the measurement period, this signifies that at
least one current input is above the defined threshold and
current is flowing through the system, although no voltage is
present at the ADE7868/ADE7878 pins. This situation is often
called missing neutral and is considered a tampering situation,
at which point the external microprocessor sets the ADE7868/
ADE7878 into PSM1 mode, measures the mean absolute values
of phase currents, and integrates the energy based on their values
and the nominal voltage.
It is recommended to use the ADE7868/ADE7878 in PSM2
mode when Bits[2:0] (PGA1[2:0]) of the gain register are equal
to 1 or 2. These bits represent the gain in the current channel
datapath. It is not recommended to use the ADE7868/ADE7878
in PSM2 mode when the PGA1[2:0] bits are equal to 4, 8, or 16.
PSM3—SLEEP MODE (ALL PARTS)
The sleep mode is available on all parts (ADE7854, ADE7858,
ADE7868, and ADE7878). In this mode, the ADE78xx has most
of its internal circuits turned off and the current consumption is
at its lowest level. The I2C, HSDC, and SPI ports are not func-
tional during this mode, and the RESET, SCLK/SCL, MOSI/SDA,
and SS/HSA pins should be set high.
Table 11. Power Modes and Related Characteristics
Power Mode All Registers1
LPOILVL,
CONFIG2 I2C/SPI Functionality
PSM0
State After Hardware Reset Set to default Set to default I2C enabled All circuits are active and
DSP is in idle mode.
State After Software Reset Set to default Unchanged Active serial port is unchanged if lock-
in procedure has been previously
executed
All circuits are active and
DSP is in idle mode.
PSM1—ADE7878, ADE7868 Only Not available Values set
during PSM0
unchanged
Enabled Current mean absolute
values are computed and
the results are stored in
the AIMAV, BIMAV, and
CIMAV registers. The I2C or
SPI serial port is enabled
with limited functionality.
PSM2—ADE7878, ADE7868 Only Not available Values set
during PSM0
unchanged
Disabled Compares phase currents
against the threshold set
in LPOILVL. Triggers IRQ0
or IRQ1 pins accordingly.
The serial ports are not
available.
PSM3 Not available
Values set
during PSM0
unchanged
Disabled Internal circuits shut down
and the serial ports are
not available.
1 Setting for all registers except the LPOILVL and CONFIG2 registers.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 23 of 96
Table 12. Recommended Actions When Changing Power Modes
Next Power Mode
Initial Power
Mode
Recommended Actions
Before Setting Next
Power Mode PSM0 PSM1 PSM2 PSM3
Stop DSP by setting the run
register = 0x0000.
Current mean absolute
values (mav) computed
immediately.
Disable HSDC by clearing
Bit 6 (HSDEN) to 0 in the
CONFIG register.
xIMAV registers can be
accessed immediately.
Mask interrupts by setting
MASK0 = 0x0 and
MASK1 = 0x0.
PSM0
Erase interrupt status flags
in the STATUS0 and STATUS1
registers.
Wait until the IRQ0
or IRQ1 pin is
triggered
accordingly.
No action
necessary.
Wait until the IRQ1 pin
is triggered low.
PSM1—
ADE7878,
ADE7868 Only
No action necessary.
Poll the STATUS1
register until Bit 15
(RSTDONE) is set to 1.
Wait until the IRQ0
or IRQ1 pin is
triggered
accordingly.
No action
necessary.
Wait until the IRQ1 pin
is triggered low.
Wait until the IRQ1 pin
triggered low.
Poll the STATUS1
register until Bit 15
(RSTDONE) is set to 1.
Current mean absolute
values compute at this
moment.
PSM2—
ADE7878,
ADE7868 Only
No action necessary.
xIMAV registers may be
accessed from this
moment.
No action
necessary.
Wait until the IRQ1 pin
is triggered low.
Wait until the IRQ1 pin is
triggered low.
Poll the STATUS1
register until Bit 15
(RSTDONE) is set to 1.
Current mav circuit
begins computations at
this time.
PSM3 No action necessary.
xIMAV registers can be
accessed from this
moment.
Wait until the IRQ0
or IRQ1 pin is
triggered
accordingly.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 24 of 96
POWER-UP PROCEDURE
POR TIMER
TURNED ON
ADE78xx
POWERED UP
ADE78xx
ENTER PSM3
MICROPROCESSOR
SETS ADE78xx
IN PSM0
MICROPROCESSOR
MAKES THE
CHOICE BETWEEN
I
2
C AND SPI
RSTDONE
INTERRUPT
TRIGGERED
40ms26ms
0V
3.3V – 10%
2.0V ± 10% ADE78xx
PSM0 READY
08510-009
Figure 24. Power-Up Procedure
The ADE7854/ADE7858/ADE7868/ADE7878 contain an on-
chip power supply monitor that supervises the power supply
(VDD). At power-up, until VDD reaches 2 V ± 10%, the chip is
in an inactive state. As VDD crosses this threshold, the power
supply monitor keeps the chip in this inactive state for an
additional 26 ms, allowing VDD to achieve 3.3 V − 10%, the
minimum recommended supply voltage. Because the PM0 and
PM1 pins have internal pull-up resistors and the external micro-
processor keeps them high, the ADE7854/ADE7858/ADE7868/
ADE7878 always power-up in sleep mode (PSM3). Then, an
external circuit (that is, a microprocessor) sets the PM1 pin to a
low level, allowing the ADE78xx to enter normal mode (PSM0).
The passage from PSM3 mode, in which most of the internal
circuitry is turned off, to PSM0 mode, in which all functionality
is enabled, is accomplished in less than 40 ms (see Figure 24 for
details).
If PSM0 mode is the only desired power mode, the PM1 pin can
be set low permanently by using a direct connection to ground.
The PM0 pin can remain open because the internal pull-up
resistor ensures that its state is high.
When the ADE7854/ADE7858/ADE7868/ADE7878 enter PSM0
mode, the I2C port is the active serial port. If the SPI port is
used, then the SS/HSA pin must be toggled three times, high to
low. This action selects the SPI port for further use. If I2C is the
active serial port, Bit 1 (I2C_LOCK) of the CONFIG2 register
must be set to 1 to lock it in. From this moment, the ADE78xx
ignores spurious toggling of the SS/HSA pin, and an eventual
switch to use the SPI port is no longer possible. Likewise, if SPI
is the active serial port, any write to the CONFIG2 register locks
the port, at which time a switch to use the I2C port is no longer
possible. Only a power-down or by setting the RESET pin low
can the ADE7854/ADE7858/ADE7868/ADE7878 be reset to use
the I2C port. Once locked, the serial port choice is maintained
when the ADE78xx changes PSMx power modes.
Immediately after entering PSM0, the ADE7854/ADE7858/
ADE7868/ADE7878 set all registers to their default values,
including the CONFIG2 and LPOILVL registers.
The ADE7854/ADE7858/ADE7868/ADE7878 signals the end of
the transition period by triggering the IRQ1 interrupt pin low and
setting Bit 15 (RSTDONE) in the STATUS1 register to 1. This
bit is 0 during the transition period and becomes 1 when the
transition ends. The status bit is cleared and the IRQ1 pin is
returned high by writing the STATUS1 register with the corres-
ponding bit set to 1. Because the RSTDONE is an unmaskable
interrupt, Bit 15 (RSTDONE) in the STATUS1 register must be
cancelled for the IRQ1 pin to return high. It is recommended to
wait until the IRQ1 pin goes low before accessing the STATUS1
register to test the state of the RSTDONE bit. At this point, as a
good programming practice, it is also recommended to cancel
all other status flags in the STATUS1 and STATUS0 registers by
writing the corresponding bits with 1.
Initially, the DSP is in idle mode, which means it does not
execute any instruction. This is the moment to initialize all
ADE78xx registers. The last register in the queue must be
written three times to ensure the register has been initialized.
Then, enable the data memory RAM protection and write
0x0001 into the run register to start the DSP (see the Digital
Signal Processor section for details on data memory RAM
protection and the run register).
If the supply voltage, VDD, drops lower than 2 V ± 10%, the
ADE7854/ADE7858/ADE7868/ADE7878 enter an inactive state,
which means that no measurements or computations are executed.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 25 of 96
HARDWARE RESET
The ADE7854/ADE7858/ADE7868/ADE7878 each has a
RESET pin. If the ADE7854, ADE7858, ADE7868, or ADE7878
is in PSM0 mode and the RESET pin is set low, then the
ADE78xx enters the hardware reset state. The ADE78xx must
be in PSM0 mode for a hardware reset to be considered. Setting
the RESET pin low while the ADE78xx is in PSM1, PSM2, and
PSM3 modes does not have any effect.
If the ADE7854, ADE7858, ADE7868, or ADE7878 is in PSM0
mode and the RESET pin is toggled from high to low and then
back to high after at least 10 µs, all the registers are set to their
default values, including the CONFIG2 and LPOILVL registers.
The ADE78xx signals the end of the transition period by triggering
the IRQ1 interrupt pin low and setting Bit 15 (RSTDONE) in the
STATUS1 register to 1. This bit is 0 during the transition period
and becomes 1 when the transition ends. The status bit is cleared
and the IRQ1 pin is returned high by writing to the STATUS1
register with the corresponding bit set to 1.
After a hardware reset, the DSP is in idle mode, which means it
does not execute any instruction.
Because the I2C port is the default serial port of the ADE7854/
ADE7858/ADE7868/ADE7878, it becomes active after a reset
state. If SPI is the port used by the external microprocessor, the
procedure to enable it must be repeated immediately after the
RESET pin is toggled back to high (see the
section for details).
Serial Interfaces
At this point, it is recommended to initialize all of the ADE78xx
registers, enable data memory RAM protection, and then write
0x0001 into the run register to start the DSP. See the Digital
Signal Processor section for details on data memory RAM
protection and the run register.
SOFTWARE RESET FUNCTIONALITY
Bit 7 (SWRST) in the CONFIG register manages the software
reset functionality in PSM0 mode. The default value of this bit is 0.
If this bit is set to 1, then the ADE7854/ADE7858/ADE7868/
ADE7878 enter the software reset state. In this state, almost all
internal registers are set to their default values. In addition, the
choice of which serial port, I2C or SPI, is in use remains unchanged
if the lock-in procedure has been executed previously (see the
Serial Interfaces for details). The registers that maintain their
values despite the SWRST bit being set to 1 are the CONFIG2
and LPOILVL registers. When the software reset ends, Bit 7
(SWRST) in the CONFIG register is cleared to 0, the IRQ1
interrupt pin is set low, and Bit 15 (RSTDONE) in the STATUS1
register is set to 1. This bit is 0 during the transition period and
becomes 1 when the transition ends. The status bit is cleared and
the IRQ1 pin is set back high by writing to the STATUS1 register
with the corresponding bit set to 1.
After a software reset ends, the DSP is in idle mode, which
means it does not execute any instruction. It is recommended
to initialize all the ADE7854/ADE7858/ADE7868/ADE7878
registers and then enable the data memory RAM protection and
write 0x0001 into the run register to start the DSP (see the
Digital Signal Processor section for details on data memory
RAM protection and the run register).
Software reset functionality is not available in PSM1, PSM2, or
PSM3 mode.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 26 of 96
THEORY OF OPERATION
ANALOG INPUTS
The ADE7868/ADE7878 have seven analog inputs forming current
and voltage channels. The ADE7854/ADE7858 have six analog
inputs, not offering the neutral current. The current channels
consist of four pairs of fully differential voltage inputs: IAP and
IAN, IBP and IBN, ICP and ICN, and INP and INN. These vol-
tage input pairs have a maximum differential signal of ±0.5 V. In
addition, the maximum signal level on analog inputs for the
IxP/IxN pair is ±0.5 V with respect to AGND. The maximum
common-mode signal allowed on the inputs is ±25 mV. Figure 25
presents a schematic of the input for the current channels and
their relation to the maximum common-mode voltage.
IAP, IBP,
ICP, OR INP
IAN, IBN,
ICN, OR INN
V
CM
V
2
V
1
+
500m
V
V
CM
V
1
+ V
2
DIFFERENTIAL INPUT
V
1
+ V
2
= 500mV MAX PEAK
COMMON MODE
V
CM
= ±25mV MAX
500m
V
08510-010
Figure 25. Maximum Input Level, Current Channels, Gain = 1
All inputs have a programmable gain amplifier (PGA) with a
possible gain selection of 1, 2, 4, 8, or 16. The gain of IA, IB, and
IC inputs is set in Bits[2:0] (PGA1[2:0]) of the gain register. For
the ADE7868 and ADE7878 only, the gain of the IN input is set
in Bits[5:3] (PGA2[2:0]) of the gain register; thus, a different gain
from the IA, IB, or IC inputs is possible. See Table 44 for details
on the gain register.
The voltage channel has three single-ended voltage inputs: VAP,
VBP, and VCP. These single-ended voltage inputs have a maximum
input voltage of ±0.5 V with respect to VN. In addition, the max-
imum signal level on analog inputs for VxP and VN is ±0.5 V
with respect to AGND. The maximum common-mode signal
allowed on the inputs is ±25 mV. Figure 26 presents a schematic
of the voltage channels inputs and their relation to the maximum
common-mode voltage.
VAP, VBP,
OR VCP
VN
V
CM
V
1
+
500m
V
V
CM
V
1
DIFFERENTIAL INPUT
V
1
+ V
2
= 500mV MAX PEAK
COMMON MODE
V
CM
= ±25mV MAX
500m
V
08510-012
Figure 26. Maximum Input Level, Voltage Channels, Gain = 1
All inputs have a programmable gain with a possible gain
selection of 1, 2, 4, 8, or 16. To set the gain, use Bits[8:6]
(PGA3[2:0]) in the gain register (see Table 44).
Figure 27 shows how the gain selection from the gain register
works in both current and voltage channels.
K × V
IN
GAIN
SELECTION
IxN, VN
IxP, VyP
V
IN
NOTES
1. x = A, B, C, N
y = A, B, C.
08510-011
Figure 27. PGA in Current and Voltage Channels
ANALOG-TO-DIGITAL CONVERSION
The ADE7868/ADE7878 have seven sigma-delta (Σ-) analog-
to-digital converters (ADCs), and the ADE7854/ADE7858 have
six Σ- ADCs. In PSM0 mode, all ADCs are active. In PSM1
mode, only the ADCs that measure the Phase A, Phase B, and
Phase C currents are active. The ADCs that measure the neutral
current and the A, B, and C phase voltages are turned off. In
PSM2 and PSM3 modes, the ADCs are powered down to
minimize power consumption.
For simplicity, the block diagram in Figure 28 shows a first-
order Σ- ADC. The converter is composed of the Σ- modulator
and the digital low-pass filter.
24
DIGITAL
LOW-PASS
FILTER
R
C
+
CLKIN/16
INTEGRATOR
V
REF
1-BIT DAC
LATCHED
COMPARATOR
ANALOG
LOW-PASS FILTER
.....10100101.....
+
08510-013
Figure 28. First-Order
Σ
-∆ ADC
A Σ- modulator converts the input signal into a continuous
serial stream of 1s and 0s at a rate determined by the sampling
clock. In the ADE7854/ADE7858/ADE7868/ADE7878, the
sampling clock is equal to 1.024 MHz (CLKIN/16). The 1-bit
DAC in the feedback loop is driven by the serial data stream.
The DAC output is subtracted from the input signal. If the loop
gain is high enough, the average value of the DAC output (and,
therefore, the bit stream) can approach that of the input signal
level. For any given input value in a single sampling interval, the
data from the 1-bit ADC is virtually meaningless. Only when a
large number of samples are averaged is a meaningful result
obtained. This averaging is carried out in the second part of the
ADC, the digital low-pass filter. By averaging a large number of
bits from the modulator, the low-pass filter can produce 24-bit
data-words that are proportional to the input signal level.
The Σ- converter uses two techniques to achieve high resolu-
tion from what is essentially a 1-bit conversion technique. The
first is oversampling. Oversampling means that the signal is
sampled at a rate (frequency) that is many times higher than
the bandwidth of interest. For example, the sampling rate in
the ADE7854/ADE7858/ADE7868/ADE7878 is 1.024 MHz,
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 27 of 96
and the bandwidth of interest is 40 Hz to 2 kHz. Oversampling
has the effect of spreading the quantization noise (noise due to
sampling) over a wider bandwidth. With the noise spread more
thinly over a wider bandwidth, the quantization noise in the band
of interest is lowered, as shown in Figure 29. However, oversam-
pling alone is not efficient enough to improve the signal-to-noise
ratio (SNR) in the band of interest. For example, an oversampling
factor of 4 is required just to increase the SNR by a mere 6 dB
(1 bit). To keep the oversampling ratio at a reasonable level, it is
possible to shape the quantization noise so that the majority of
the noise lies at the higher frequencies. In the Σ- modulator,
the noise is shaped by the integrator, which has a high-pass-type
response for the quantization noise. This is the second technique
used to achieve high resolution. The result is that most of the
noise is at the higher frequencies where it can be removed by
the digital low-pass filter. This noise shaping is shown in Figure 29.
NOISE
SIGNAL
NOISE
SIGNAL
0 2 4 512
FREQUENCY (kHz)
HIGH RESOLUTION
OUTPUT FROM
DIGITAL LPF
1024
0 2 4 512
FREQUENCY (kHz)
1024
DIGITAL FILTER
SHAPED NOISE
ANTIALIAS FILTER
(RC)
SAMPLING
FREQUENCY
08510-014
Figure 29. Noise Reduction Due to Oversampling and
Noise Shaping in the Analog Modulator
Antialiasing Filter
Figure 28 also shows an analog low-pass filter (RC) on the input
to the ADC. This filter is placed outside the ADE7854/ADE7858/
ADE7868/ADE7878, and its role is to prevent aliasing. Aliasing
is an artifact of all sampled systems as shown in Figure 30. Aliasing
means that frequency components in the input signal to the
ADC, which are higher than half the sampling rate of the ADC,
appear in the sampled signal at a frequency below half the
sampling rate. Frequency components above half the sampling
frequency (also known as the Nyquist frequency, that is, 512 kHz)
are imaged or folded back down below 512 kHz. This happens
with all ADCs regardless of the architecture. In the example shown,
only frequencies near the sampling frequency, that is, 1.024 MHz,
move into the band of interest for metering, that is, 40 Hz to
2 kHz. To attenuate the high frequency (near 1.024 MHz) noise
and prevent the distortion of the band of interest, a low-pass
filer (LPF) must be introduced. For conventional current
sensors, it is recommended to use one RC filter with a corner
frequency of 5 kHz for the attenuation to be sufficiently high at
the sampling frequency of 1.024 MHz. The 20 dB per decade
attenuation of this filter is usually sufficient to eliminate the
effects of aliasing for conventional current sensors. However, for a
di/dt sensor such as a Rogowski coil, the sensor has a 20 dB per
decade gain. This neutralizes the 20 dB per decade attenuation
produced by the LPF. Therefore, when using a di/dt sensor, take
care to offset the 20 dB per decade gain. One simple approach is
to cascade one additional RC filter, thereby producing a −40 dB
per decade attenuation.
A
LIASING EFFECTS SAMPLING
FREQUENCY
IMAGE
FREQUENCIES
0 2 4 512
FREQUENCY (kHz)
1024
08510-015
Figure 30. Aliasing Effects
ADC Transfer Function
All ADCs in the ADE7854/ADE7858/ADE7868/ADE7878 are
designed to produce the same 24-bit signed output code for the
same input signal level. With a full-scale input signal of 0.5 V
and an internal reference of 1.2 V, the ADC output code is nomi-
nally 5,928,256 (0x5A7540). The code from the ADC can vary
between 0x800000 (−8,388,608) and 0x7FFFFF (+8,388,607);
this is equivalent to an input signal level of ±0.707 V. However,
for specified performance, do not exceed the nominal range of
±0.5 V; ADC performance is guaranteed only for input signals
lower than ±0.5 V.
CURRENT CHANNEL ADC
Figure 31 shows the ADC and signal processing path for
Input IA of the current channels (it is the same for IB and IC).
The ADC outputs are signed twos complement 24-bit data-
words and are available at a rate of 8 kSPS (thousand samples
per second). With the specified full-scale analog input signal
of ±0.5 V, the ADC produces its maximum output code value.
Figure 31 shows a full-scale voltage signal applied to the differ-
ential inputs (IAP and IAN). The ADC output swings between
−5,928,256 (0xA58AC0) and +5,928,256 (0x5A7540). The
input, IN, corresponds to the neutral current of a 3-phase
system (available in the ADE7868 and ADE7878 only). If no
neutral line is present, connect this input to AGND. The
datapath of the neutral current is similar to the path of the
phase currents as shown in Figure 32.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 28 of 96
AIGAIN[23:0]
DSP
REFERENCE HPFDIS
[23:0] DIGITAL
INTEGRATOR
INTEN BIT
CONFIG[0]
TOTAL/FUNDAMENTAL
ACTIVE AND REACTIVE
POWER CALCULATION
CURRENT PEAK,
OVERCURRENT
DETECT
IAWV WAVEFORM
SAMPLE REGISTER
CURRENT RMS (IRMS)
CALCULATION
HPF
ADC
PGA1
IAP
ZX DETECTION
PGA1 BITS
GAIN[2:0]
×1, ×2, ×4, ×8, ×16
V
IN
V
IN
IAN
ANALOG INPUT RANGE ANALOG OUTPUT RANGE
0x5A7540 =
+5,928,256
ZX SIGN
A
L
DATA RANGE
0V
0xA58AC0 =
–5,928,256
0x5A7540 =
+5,928,256
CURRENT CHANNE L
DATA RANGE AFTER
INTEGRATION
0V
0xA58AC0 =
–5,928,256
0x5A7540 =
+5,928,256
CURRENT CHANNE L
DATA RANGE
0V
0xA58AC0 =
–5,928,256
+0.5V/GAIN
–0.5V/GAIN
0V
LPF1
08510-019
Figure 31. Current Channel Signal Path
NIGAIN[23:0]
DSP
REFERENCE
DIGITAL
INTEGRATOR
INTEN BIT
CONFIG[0]
INWV WAVEFORM
SAMPLE REGISTER
CURRENT RMS (IRMS)
CALCULATION
HPF
ADC
PGA2
INP
PGA2 BITS
GAIN[5:3]
×1, ×2, ×4, ×8, ×16
V
IN
INN
08510-120
Figure 32. Neutral Current Signal Path (ADE7868, ADE7878 Only)
Current Waveform Gain Registers
There is a multiplier in the signal path of each phase and
neutral current. The current waveform can be changed by
±100% by writing a corresponding twos complement number
to the 24-bit signed current waveform gain registers (AIGAIN,
BIGAIN, CIGAIN, and NIGAIN). For example, if 0x400000 is
written to those registers, the ADC output is scaled up by 50%.
To scale the input by −50%, write 0xC00000 to the registers.
Equation 3 describes mathematically the function of the current
waveform gain registers.
Current Waveform =
+× 23
2
1RegisterGainCurrentofContent
OutputADC (3)
Changing the content of the AIGAIN, BIGAIN, CIGAIN, or
INGAIN registers affects all calculations based on its current;
that is, it affects the corresponding phase active/reactive/
apparent energy and current rms calculation. In addition,
waveform samples scale accordingly.
Note that the serial ports of the ADE7854, ADE7858, ADE7868,
and/or ADE7878 work on 32-, 16-, or 8-bit words, and the DSP
works on 28 bits. The 24-bit AIGAIN, BIGAIN, CIGAIN, and
NIGAIN registers are accessed as 32-bit registers with the four
most significant bits (MSBs) padded with 0s and sign extended
to 28 bits. See Figure 33 for details.
31 28 27 24 23 0
24-BIT NUMBER0000
BITS[27:24] ARE
EQUAL TO BIT 23
BIT 23 IS A SIGN BIT
08510-016
Figure 33. 24-Bit xIGAIN Transmitted as 32-Bit Words
Current Channel HPF
The ADC outputs can contain a dc offset. This offset can create
errors in power and rms calculations. High-pass filters (HPFs)
are placed in the signal path of the phase and neutral currents
and of the phase voltages. If enabled, the HPF eliminates any dc
offset on the current channel. All filters are implemented in the
DSP and, by default, they are all enabled: the 24-bit HPFDIS
register is cleared to 0x00000000. All filters are disabled by
setting HPFDIS to any nonzero value.
As stated in the Current Waveform Gain Registers section, the
serial ports of the ADE78xx work on 32-, 16-, or 8-bit words.
The HPFDIS register is accessed as a 32-bit register with eight
MSBs padded with 0s. See Figure 34 for details.
31 24 23 0
24-BIT NUMBER0000 0000
08510-017
Figure 34. 24-Bit HPFDIS Register Transmitted as 32-Bit Word
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 29 of 96
Current Channel Sampling
The waveform samples of the current channel are taken at the
output of HPF and stored in the 24-bit signed registers, IAWV,
IBWV, ICWV, and INWV (ADE7868 and ADE7878 only) at a
rate of 8 kSPS. All power and rms calculations remain uninter-
rupted during this process. Bit 17 (DREADY) in the STATUS0
register is set when the IAWV, IBWV, ICWV, and INWV registers
are available to be read using the I2C or SPI serial port. Setting
Bit 17 (DREADY) in the MASK0 register enables an interrupt
to be set when the DREADY flag is set. See the Digital Signal
Processor section for more details on Bit DREADY.
As stated in the Current Waveform Gain Registers section, the
serial ports of the ADE78xx work on 32-, 16-, or 8-bit words.
When the IAWV, IBWV, ICWV, and INWV 24-bit signed
registers are read from the ADE78xx (INWV is available on
ADE7868/ADE7878 only), they are transmitted sign extended
to 32 bits. See Figure 35 for details.
31 24 23 22 0
24-BIT SIGNED NUMBER
BITS[31:24] ARE
EQUAL TO BIT 23
BIT 23 IS A SIGN BIT
08510-018
Figure 35. 24-Bit IxWV Register Transmitted as 32-Bit Signed Word
The ADE7854/ADE7858/ADE7868/ADE7878 devices each
contain a high speed data capture (HSDC) port that is specially
designed to provide fast access to the waveform sample registers.
See the HSDC Interface section for more details.
di/dt CURRENT SENSOR AND DIGITAL INTEGRATOR
The di/dt sensor detects changes in the magnetic field caused by
the ac current. Figure 36 shows the principle of a di/dt current
sensor.
MAGNETIC FIELD CREATED BY CURRENT
(DIRECTLY PROPORTIONAL TO CURRENT)
+ EMF (ELECTROMOTIVE FORCE)
– INDUCED BY CHANGES IN
MAGNETIC FLUX DENSITY (di/dt)
08510-020
Figure 36. Principle of a di/dt Current Sensor
The flux density of a magnetic field induced by a current is
directly proportional to the magnitude of the current. The
changes in the magnetic flux density passing through a conductor
loop generate an electromotive force (EMF) between the two
ends of the loop. The EMF is a voltage signal that is propor-
tional to the di/dt of the current. The voltage output from the
di/dt current sensor is determined by the mutual inductance
between the current carrying conductor and the di/dt sensor.
Due to the di/dt sensor, the current signal needs to be filtered
before it can be used for power measurement. On each phase and
neutral current datapath, there is a built-in digital integrator to
recover the current signal from the di/dt sensor. The digital inte-
grator is disabled by default when the ADE78xx is powered up
and after a reset. Setting Bit 0 (INTEN) of the CONFIG register
turns on the integrator. Figure 37 and Figure 38 show the
magnitude and phase response of the digital integrator.
Note that the integrator has a −20 dB/dec attenuation and
approximately −90° phase shift. When combined with a di/dt
sensor, the resulting magnitude and phase response should be a
flat gain over the frequency band of interest. However, the di/dt
sensor has a 20 dB/dec gain associated with it and generates sig-
nificant high frequency noise. An antialiasing filter of at least
the second order is needed to avoid noise aliasing back in the
band of interest when the ADC is sampling (see the Antialiasing
Filter section).
50
0
–50
–100
–50
0
MAGNITUDE (dB)PHASE (Degrees)
0 500 1000 1500 2000 2500 3000 3500 4000
FREQUENCY (Hz)
0.01 0.1 1 10 100 1000
FREQUENCY (Hz)
08510-116
Figure 37. Combined Gain and Phase Response of the
Digital Integrator
The DICOEFF 24-bit signed register is used in the digital
integrator algorithm. At power-up or after a reset, its value is
0x000000. Before turning on the integrator, this register must be
initialized with 0xFFF8000. DICOEFF is not used when the
integrator is turned off and can remain at 0x000000 in that case.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 30 of 96
–30
15
–20
–25
–89.96
–89.97
–89.98
–89.99
MAGNITUDE (dB)PHASE (Degrees)
30 35 40 45 50 55 60 65 70
FREQUENCY (Hz)
30 35 40 45 50 55 60 65 70
FREQUENCY (Hz)
0
8510-101
Figure 38. Combined Gain and Phase Response of the
Digital Integrator (40 Hz to 70 Hz)
As stated in the Current Waveform Gain Registers section, the
serial ports of the ADE78xx work on 32-, 16-, or 8-bit words.
Similar to the registers shown in Figure 33, the DICOEFF 24-bit
signed register is accessed as a 32-bit register with four MSBs
padded with 0s and sign extended to 28 bits, which practically
means it is transmitted equal to 0xFFF8000.
When the digital integrator is switched off, the ADE78xx can
be used directly with a conventional current sensor, such as
a current transformer (CT).
VOLTAGE CHANNEL ADC
Figure 39 shows the ADC and signal processing chain for
Input VA in the voltage channel. The VB and VC channels
have similar processing chains. The ADC outputs are signed
twos complement 24-bit words and are available at a rate of
8 kSPS. With the specified full-scale analog input signal of
±0.5 V, the ADC produces its maximum output code value.
Figure 39 shows a full-scale voltage signal being applied to the
differential inputs (VA and VN). The ADC output swings
between −5,928,256 (0xA58AC0) and +5,928,256 (0x5A7540).
AVGAIN[23:0]
REFERENCE HPFDIS
[23:0]
DSP
TOTAL/FUNDAMENTAL
ACTIVE AND REACTIVE
POWER CALCULATION
V
OLT
A
GE PEAK,
OVERVOLTAGE,
SAG DETECT
VAWV WAVEFORM
SAMPLE REGISTER
CURRENT RMS (VRMS)
CALCULATION
HPF
ADC
PGA3
VAP
ZX DETECTION
PGA3 BITS
GAIN[8:6]
×1, ×2, ×4, ×8, ×16
V
IN
V
IN
VN
ANALOG INPUT RANGE ANALOG OUTPUT RANGE
0x5A7540 =
+5,928,256
ZX SIGNAL
DATA RANGE
0V
0xA58AC0 =
–5,928,256
0x5A7540 =
+5,928,256
VOLTAGE CHANNEL
DATA RANGE
0V
0xA58AC0 =
–5,928,256
+0.5V/GAIN
–0.5V/GAIN
0V
LPF1
08510-025
Figure 39. Voltage Channel Datapath
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 31 of 96
Voltage Waveform Gain Registers
There is a multiplier in the signal path of each phase voltage.
The voltage waveform can be changed by ±100% by writing
a corresponding twos complement number to the 24-bit signed
current waveform gain registers (AVGAIN, BVGAIN, and
CVGAIN). For example, if 0x400000 is written to those registers,
the ADC output is scaled up by 50%. To scale the input by −50%,
write 0xC00000 to the registers. Equation 4 describes mathe-
matically the function of the current waveform gain registers.
Voltage Waveform =
+× 23
2
1RegisterGainVoltageofContent
OutputADC (4)
Changing the content of the AVGAIN, BVGAIN, and CVGAIN
registers affects all calculations based on its voltage; that is, it
affects the corresponding phase active/reactive/apparent energy
and voltage rms calculation. In addition, waveform samples are
scaled accordingly.
As stated in the Current Waveform Gain Registers section, the
serial ports of the ADE78xx work on 32-, 16-, or 8-bit words,
and the DSP works on 28 bits. As presented in Figure 33, the
AVGAIN, BVGAIN, and CVGAIN registers are accessed as
32-bit registers with four MSBs padded with 0s and sign
extended to 28 bits.
Voltage Channel HPF
As explained in the Current Channel HPF section, the ADC
outputs can contain a dc offset that can create errors in power
and rms calculations. HPFs are placed in the signal path of the
phase voltages, similar to the ones in the current channels. The
HPFDIS register can enable or disable the filters. See the
Current Channel HPF section for more details.
Voltage Channel Sampling
The waveform samples of the voltage channel are taken at the
output of HPF and stored into VAWV, VBWV, and VCWV
24-bit signed registers at a rate of 8 kSPS. All power and rms
calculations remain uninterrupted during this process. Bit 17
(DREADY) in the STATUS0 register is set when the VAWV,
VBWV, and VCWV registers are available to be read using the
I2C or SPI serial port. Setting Bit 17 (DREADY) in the MASK0
register enables an interrupt to be set when the DREADY flag is
set. See the Digital Signal Processor section for more details on
Bit DREADY.
As stated in the Current Waveform Gain Registers section, the
serial ports of the ADE78xx work on 32-, 16-, or 8-bit words.
Similar to registers presented in Figure 35, the VAWV, VBWV,
and VCWV 24-bit signed registers are transmitted sign
extended to 32 bits.
The ADE7854/ADE7858/ADE7868/ADE7878 each contain an
HSDC port especially designed to provide fast access to the
waveform sample registers. See the HSDC Interface section for
more details.
CHANGING PHASE VOLTAGE DATAPATH
The ADE7854/ADE7858/ADE7868/ADE7878 can direct one
phase voltage input to the computational datapath of another
phase. For example, Phase A voltage can be introduced in the
Phase B computational datapath, which means all powers
computed by the ADE78xx in Phase B are based on Phase A
voltage and Phase B current.
Bits[9:8] (VTOIA[1:0]) of the CONFIG register manage the
Phase A voltage measured at the VA pin. If VTOIA[1:0] = 00
(default value), the voltage is directed to the Phase A computa-
tional datapath. If VTOIA[1:0] = 01, the voltage is directed to
the Phase B path. If VTOIA[1:0] = 10, the voltage is directed to the
Phase C path. If VTOIA[1:0] = 11, the ADE7878 behaves as if
VTOIA[1:0] = 00.
Bits[11:10] (VTOIB[1:0]) of the CONFIG register manage the
Phase B voltage measured at the VB pin. If VTOIB[1:0] = 00
(default value), the voltage is directed to the Phase B computa-
tional datapath. If VTOIB[1:0] = 01, the voltage is directed to
the Phase C path. If VTOIB[1:0] = 10, the voltage is directed to
the Phase A path. If VTOIB[1:0] = 11, the ADE78xx behaves
as if VTOIB[1:0] = 00.
Bits[13:12] (VTOIC[1:0]) of the CONFIG register manage the
Phase C voltage measured at the VC pin. If VTOIC[1:0] = 00
(default value), the voltage is directed to Phase C computational
datapath, if VTOIC[1:0] = 01, the voltage is directed to the
Phase A path. If VTOIC[1:0] = 10, the voltage is directed to the
Phase B path. If VTOIC[1:0] = 11, the ADE78xx behaves as if
VTOIC[1:0] = 00.
I
A
VA
IB
VB
IC
VC
PHASE A
COMPUTATIONAL
DATAPATH
PHASE B
COMPUTATIONAL
DATAPATH
PHASE C
COMPUTATIONAL
DATAPATH
VTOIA[1:0] = 01,
PHASE A VOLTAGE
DIRECTED
TO PHASE B
VTOIB[1:0] = 01,
PHASE B VOLTAGE
DIRECTED
TO PHASE C
VTOIC[1:0] = 01,
PHASE C VOLTAGE
DIRECTED
TO PHASE A
CPHCAL
BPHCAL
APHCAL
08510-026
Figure 40. Phase Voltages Used in Different Datapaths
Figure 40 presents the case in which Phase A voltage is used in
the Phase B datapath, Phase B voltage is used in the Phase C
datapath, and Phase C voltage is used in the Phase A datapath.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 32 of 96
POWER QUALITY MEASUREMENTS
Zero-Crossing Detection
The ADE7854/ADE7858/ADE7868/ADE7878 have a zero-
crossing (ZX) detection circuit on the phase current and voltage
channels. The neutral current datapath does not contain a zero-
crossing detection circuit. Zero-crossing events are used as a
time base for various power quality measurements and in the
calibration process.
The output of LPF1 is used to generate zero crossing events.
The low-pass filter is intended to eliminate all harmonics of
50 Hz and 60 Hz systems, and to help identify the zero-crossing
events on the fundamental components of both current and
voltage channels.
The digital filter has a pole at 80 Hz and is clocked at 256 kHz.
As a result, there is a phase lag between the analog input signal
(one of IA, IB, IC, VA, VB, and VC) and the output of LPF1.
The error in ZX detection is 0.0703° for 50 Hz systems (0.0843°
for 60 Hz systems). The phase lag response of LPF1 results in a
time delay of approximately 31.4° or 1.74 ms (at 50 Hz) between
its input and output. The overall delay between the zero crossing
on the analog inputs and ZX detection obtained after LPF1 is
about 39.6° or 2.2 ms (at 50 Hz). The ADC and HPF introduce
the additional delay. The LPF1 cannot be disabled to assure a
good resolution of the ZX detection. Figure 41 shows how the
zero-crossing signal is detected.
xIGAIN[23:0] OR
xVGAIN[23:0]
REFERENCE HPFDIS
[23:0]
DSP
HPF
PGA ADC
IA, IB, IC,
OR
VA, VB, VC ZX
DETECTION
LPF1
IA, IB, IC,
OR VA, VB, VC
39.6° OR 2.2ms @ 50Hz
1
0.855
0V ZX ZX
ZX
ZX
LPF1 OUTPUT
0
8510-027
Figure 41. Zero-Crossing Detection on Voltage and Current Channels
To provide further protection from noise, input signals to the
voltage channel with amplitude lower than 10% of full scale do
not generate zero-crossing events at all. The Current Channel ZX
detection circuit is active for all input signals independent of their
amplitudes.
The ADE7854/ADE7858/ADE7868/ADE7878 contain six zero-
crossing detection circuits, one for each phase voltage and
current channel. Each circuit drives one flag in the STATUS1
register. If a circuit placed in the Phase A voltage channel
detects one zero-crossing event, Bit 9 (ZXVA) in the STATUS1
register is set to 1.
Similarly, the Phase B voltage circuit drives Bit 10 (ZXVB), the
Phase C voltage circuit drives Bit 11 (ZXVC), and circuits placed
in the current channel drive Bit 12 (ZXIA), Bit 13 (ZXIB), and
Bit 14 (ZXIC) in the STATUS1 register. If a ZX detection bit is
set in the MASK1 register, the IRQ1 interrupt pin is driven low
and the corresponding status flag is set to 1. The status bit is
cleared and the IRQ1 pin is set to high by writing to the STATUS1
register with the status bit set to 1.
Zero-Crossing Timeout
Every zero-crossing detection circuit has an associated timeout
register. This register is loaded with the value written into the
16-bit ZXTOUT register and is decremented (1 LSB) every
62.5 µs (16 kHz clock). The register is reset to the ZXTOUT
value every time a zero crossing is detected. The default value of
this register is 0xFFFF. If the timeout register decrements to 0
before a zero crossing is detected, one of Bits[8:3] of the
STATUS1 register is set to 1. Bit 3 (ZXTOVA), Bit 4 (ZXTOVB),
and Bit 5 (ZXTOVC) in the STATUS1 register refer to Phase A,
Phase B, and Phase C of the voltage channel; Bit 6 (ZXTOIA),
Bit 7 (ZXTOIB), and Bit 8 (ZXTOIC) in the STATUS1 register
refer to Phase A, Phase B, and Phase C of the current channel.
If a ZXTOIx or ZXTOVx bit is set in the MASK1 register, the
IRQ1 interrupt pin is driven low when the corresponding status bit
is set to 1. The status bit is cleared and the IRQ1 pin is returned to
high by writing to the STATUS1 register with the status bit set to 1.
The resolution of the ZXOUT register is 62.5 µs (16 kHz clock)
per LSB. Thus, the maximum timeout period for an interrupt is
4.096 sec: 216/16 kHz.
Figure 42 shows the mechanism of the zero-crossing timeout
detection when the voltage or the current signal stays at a fixed
dc level for more than 62.5 µs × ZXTOUT µs.
16-BIT INTERNAL
REGISTER VALUE
ZXTOUT
VOLTAGE
OR
CURRENT
SIGNAL
IRQ1 INTERRUPT PIN
ZXZOxy FLAG IN
S
TATUS1[31:0], x = V, A
y = A, B, C
0V
08510-028
Figure 42. Zero-Crossing Timeout Detection
Phase Sequence Detection
The ADE7854/ADE7858/ADE7868/ADE7878 have on-chip
phase sequence error detection circuits. This detection works
on phase voltages and considers only the zero crossings
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 33 of 96
determined by their negative-to-positive transitions. The regular
succession of these zero-crossing events is Phase A followed by
Phase B followed by Phase C (see Figure 44). If the sequence of
zero-crossing events is, instead, Phase A followed by Phase C
followed by Phase B, then Bit 19 (SEQERR) in the STATUS1
register is set.
If Bit 19 (SEQERR) in the MASK1 register is set to 1 and a
phase sequence error event is triggered, the IRQ1 interrupt pin
is driven low. The status bit is cleared and the IRQ1 pin is set
high by writing to the STATUS1 register with the Status Bit 19
(SEQERR) set to 1.
The phase sequence error detection circuit is functional only
when the ADE78xx is connected in a 3-phase, 4-wire, three voltage
sensors configuration (Bits[5:4], CONSEL[1:0] in the ACCMODE
register, set to 00). In all other configurations, only two voltage
sensors are used; therefore, it is not recommended to use the
detection circuit. In these cases, use the time intervals between
phase voltages to analyze the phase sequence (see the Time
Interval Between Phases section for details).
Figure 43 presents the case in which Phase A voltage is not
followed by Phase B voltage but by Phase C voltage. Every time
a negative-to-positive zero crossing occurs, Bit 19 (SEQERR) in
the STATUS1 register is set to 1 because such zero crossings on
Phase C, Phase B, or Phase A cannot come after zero crossings
from Phase A, Phase C, or respectively, Phase B zero crossings.
ZX BZX C
PHASE C PHASE BPHASE A
A, B, C PHASE
VOLTAGES AFTER
LPF1
BIT 19 (SEQERR) IN
STATUS1 REGISTE
R
IRQ1
ZX A
STATUS1[19] SET TO 1 STATUS1[19] CANCELLED
BY A WRITE TO THE
STATUS1 REGISTER WITH
SEQERR BIT SET
08510-029
Figure 43. SEQERR Bit Set to 1 When Phase A Voltage Is Followed by
Phase C Voltage
Once a phase sequence error has been detected, the time
measurement between various phase voltages (see the Time
Interval Between Phases section) can help to identify which
phase voltage should be considered with another phase current
in the computational datapath. Bits[9:8] (VTOIA[1:0]), Bits[11:10]
(VTOIB[1:0]), and Bits[13:12] (VTOIC[1:0]) in the CONFIG
register can be used to direct one phase voltage to the datapath
of another phase. See the Changing Phase Voltage Datapath
section for details.
Time Interval Between Phases
The ADE7854/ADE7858/ADE7868/ADE7878 have the capa-
bility to measure the time delay between phase voltages, between
phase currents, or between voltages and currents of the same
phase. The negative-to-positive transitions identified by the zero-
crossing detection circuit are used as start and stop measuring
points. Only one set of such measurements is available at one time,
based on Bits[10:9] (ANGLESEL[1:0]) in the COMPMODE
register.
ZX CZX B
PHASE B PHASE CPHASE A
ZX A
0
8510-030
Figure 44. Regular Succession of Phase A, Phase B, and Phase C
When the ANGLESEL[1:0] bits are set to 00, the default value,
the delays between voltages and currents on the same phase are
measured. The delay between Phase A voltage and Phase A
current is stored in the 16-bit unsigned ANGLE0 register (see
Figure 45 for details). In a similar way, the delays between voltages
and currents on Phase B and Phase C are stored in the ANGLE1
and ANGLE2 registers, respectively.
PHASE A
CURRENT
ANGLE0
PHASE A
VOLTAGE
08510-031
Figure 45. Delay Between Phase A Voltage and Phase A Current Is
Stored in the ANGLE0 Register
When the ANGLESEL[1:0] bits are set to 01, the delays between
phase voltages are measured. The delay between Phase A voltage
and Phase C voltage is stored into the ANGLE0 register. The
delay between Phase B voltage and Phase C voltage is stored in
the ANGLE1 register, and the delay between Phase A voltage
and Phase B voltage is stored in the ANGLE2 register (see
Figure 46 for details).
When the ANGLESEL[1:0] bits are set to 10, the delays between
phase currents are measured. Similar to delays between phase
voltages, the delay between Phase A and Phase C currents is stored
into the ANGLE0 register, the delay between Phase B and Phase C
currents is stored in the ANGLE1 register, and the delay between
Phase A and Phase B currents is stored into the ANGLE2
register (see Figure 46 for details).
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 34 of 96
PHASE B PHASE CPHASE A
ANGLE2
ANGLE0
ANGLE1
0
8510-032
Figure 46. Delays Between Phase Voltages (Currents)
The ANGLE0, ANGLE1, and ANGLE2 registers are 16-bit
unsigned registers with 1 LSB corresponding to 3.90625 s
(256 kHz clock), which means a resolution of 0.0703° (360° ×
50 Hz/256 kHz) for 50 Hz systems and 0.0843° (360° × 60 Hz/
256 kHz) for 60 Hz systems. The delays between phase voltages
or phase currents are used to characterize how balanced the
load is. The delays between phase voltages and currents are
used to compute the power factor on each phase as shown in
the following Equation 5:
cosφx = cos
×
×kHz256
360 LINE
f
ANGLEx
o
(5)
where fLINE = 50 Hz or 60 Hz.
Period Measurement
The ADE7854/ADE7858/ADE7868/ADE7878 provide the
period measurement of the line in the voltage channel. Bits[1:0]
(PERSEL[1:0]) in the MMODE register select the phase voltage
used for this measurement. The period register is a 16-bit
unsigned register and updates every line period. Because of the
LPF1 filter (see Figure 41), a settling time of 30 ms to 40 ms is
associated with this filter before the measurement is stable.
The period measurement has a resolution of 3.90625 s/LSB
(256 kHz clock), which represents 0.0195% (50 Hz/256 kHz)
when the line frequency is 50 Hz and 0.0234% (60 Hz/256 kHz)
when the line frequency is 60 Hz. The value of the period register
for 50 Hz networks is approximately 5120 (256 kHz/50 Hz) and
for 60 Hz networks is approximately 4267 (256 kHz/60 Hz). The
length of the register enables the measurement of line frequencies
as low as 3.9 Hz (256 kHz/216). The period register is stable at
±1 LSB when the line is established and the measurement does
not change.
The following expressions can be used to compute the line
period and frequency using the period register:
[]
sec
3E256
1+
=0]PERIOD[15:
TL (6)
]Hz[
1
3E256
+
=0]PERIOD[15:
fL (7)
Phase Voltage Sag Detection
The ADE7854/ADE7858/ADE7868/ADE7878 can be pro-
grammed to detect when the absolute value of any phase voltage
drops below a certain peak value for a number of half-line cycles.
The phase where this event takes place is identified in Bits[14:12]
(VSPHASE[x]) of the PHSTATUS register. This condition is
illustrated in Figure 47.
PHASE A VOLTAGE
BIT 16 (SAG) IN
STATUS1[31:0]
VSPHASE[0] =
PHSTATUS[12]
IRQ1 PIN
FULL SCALE
SAGLVL[23:0]
FULL SCALE
SAGLVL[23:0]
SAGCYC[7:0] = 0x4
PHASE B VOLTAGE
VSPHASE[1] =
PHSTATUS[13]
STATUS[16] AND
PHSTATUS[13]
SET TO 1
08510-033
STATUS1[16] AND
PHSTATUS[12]
CANCELLED BY A
WRITE TO
STATUS1[31:0]
WITH SAG BIT SET
SAGCYC[7:0] = 0x4
Figure 47. SAG Detection
Figure 47 shows Phase A voltage falling below a threshold that
is set in the SAG level register (SAGLVL) for four half-line cycles
(SAGCYC = 4). When Bit 16 (SAG) in the STATUS1 register is set
to 1 to indicate the condition, Bit VSPHASE[0] in the PHSTATUS
register is also set to 1 because the event happened on Phase A
Bit 16 (SAG) in the STATUS1 register. All Bits[14:12] (VSPHASE[2],
VSPHASE[1], and VSPHASE[0]) of the PHSTATUS register (not
just the VSPHASE[0] bit) are erased by writing the STATUS1
register with the SAG bit set to 1.
The SAGCYC register represents the number of half-line cycles
the phase voltage must remain below the level indicated in the
SAGLVL register to trigger a SAG condition; 0 is not a valid
number for SAGCYC. For example, when the SAG cycle
(SAGCYC[7:0]) contains 0x07, the SAG flag in the STATUS1
register is set at the end of the seventh half line cycle for which
the line voltage falls below the threshold. If Bit 16 (SAG) in
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 35 of 96
MASK1 is set, the IRQ1 interrupt pin is driven low in case of a
SAG event in the same moment the Status Bit 16 (SAG) in
STATUS1 register is set to 1. The SAG status bit in the STATUS1
register and all Bits[14:12] (VSPHASE[2], VSPHASE[1], and
VSPHASE[0]]) of the PHSTATUS register are cleared, and the
IRQ1 pin is returned to high by writing to the STATUS1
register with the status bit set to 1.
When the Phase B voltage falls below the indicated threshold
into the SAGLVL register for two line cycles, Bit VSPHASE[1]
in the PHSTATUS register is set to 1, and Bit VSPHASE[0] is
cleared to 0. Simultaneously, Bit 16 (SAG) in the STATUS1 register
is set to 1 to indicate the condition.
Note that the internal zero-crossing counter is always active. By
setting the SAGLVL register, the first SAG detection result is,
therefore, not executed across a full SAGCYC period. Writing to
the SAGCYC register when the SAGLVL register is already initia-
lized resets the zero-crossing counter, thus ensuring that the first
SAG detection result is obtained across a full SAGCYC period.
The recommended procedure to manage SAG events is the
following:
1. Enable SAG interrupts in the MASK1 register by setting
Bit 16 (SAG) to 1.
2. When a SAG event happens, the IRQ1 interrupt pin goes
low and Bit 16 (SAG) in the STATUS1 is set to 1.
3. The STATUS1 register is read with Bit 16 (SAG) set to 1.
4. The PHSTATUS register is read, identifying on which
phase or phases a SAG event happened.
5. The STATUS1 register is written with Bit 16 (SAG) set to 1.
Immediately, the SAG bit and all Bits[14:12] (VSPHASE[2],
VSPHASE[1], and VSPHASE[0]) of the PHSTATUS register
are erased.
SAG Level Set
The content of the SAGLVL[23:0] SAG level register is compared
to the absolute value of the output from HPF. Writing 5,928,256
(0x5A7540) to the SAGLVL register, puts the SAG detection
level at full scale (see the Voltage Channel ADC section), thus;
the SAG event is triggered continuously. Writing 0x00 or 0x01
puts the SAG detection level at 0, therefore, the SAG event is
never triggered.
As stated in the Current Waveform Gain Registers section, the
serial ports of the ADE78xx work on 32-, 16-, or 8-bit words.
Similar to the register presented in Figure 34, the SAGLVL
register is accessed as a 32-bit register with eight MSBs padded
with 0s.
Peak Detection
The ADE7854/ADE7858/ADE7868/ADE7878 record the
maximum absolute values reached by the voltage and current
channels over a certain number of half-line cycles and stores
them into the less significant 24 bits of the VPEAK and IPEAK
32-bit registers.
The PEAKCYC register contains the number of half-line cycles
used as a time base for the measurement. The circuit uses the
zero-crossing points identified by the zero-crossing detection
circuit. Bits[4:2] (PEAKSEL[2:0]) in the MMODE register select
the phases upon which the peak measurement is performed. Bit 2
selects Phase A, Bit 3 selects Phase B, and Bit 4 selects Phase C.
Selecting more than one phase to monitor the peak values
decreases proportionally the measurement period indicated in
the PEAKCYC register because zero crossings from more
phases are involved in the process. When a new peak value is
determined, one of Bits[26:24] (IPPHASE[2:0] or VPPHASE[2:0])
in the IPEAK and VPEAK registers is set to 1, identifying the
phase that triggered the peak detection event. For example, if a
peak value has been identified on Phase A current, Bit 24
(IPPHASE[0]) in the IPEAK register is set to 1. If next time a
new peak value is measured on Phase B, Bit 24 (IPPHASE[0])
of the IPEAK register is cleared to 0, and Bit 25 (IPPHASE[1])
of the IPEAK register is set to 1. Figure 48 shows the composition
of the IPEAK and VPEAK registers.
PEAK DETECTED
ON PHASE C
00000
31 27 26 25 24 23 0
24 BIT UNSIGNED NUMBER
PEAK DETECTED
ON PHASE A
IPPHASE/VPPHASE BITS
PEAK DETECTED
ON PHASE B
0
8510-034
Figure 48. Composition of IPEAK[31:0] and VPEAK[31:0] Registers
PHASE A
CURRENT
PHASE B
CURRENT
BIT 24
OF IPEAK
BIT 25
OF IPEAK
PEAK VALUE WRITTEN INTO
IPEAK AT THE END OF FIRST
PEAKCYC PERIOD
END OF FIRST
PEAKCYC = 16 PERIOD
BIT 24 OF IPEAK
CLEARED TO 0 AT
THE END OF SECOND
PEAKCYC PERIOD
BIT 25 OF IPEAK
SET TO 1 AT THE
END OF SECOND
PEAKCYC PERIOD
END OF SECOND
PEAKCYC = 16 PERIOD
PEAK VALUE WRITTEN INTO
IPEAK AT THE END OF SECOND
PEAKCYC PERIOD
08510-035
Figure 49. Peak Level Detection
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 36 of 96
Figure 49 shows how the ADE78xx records the peak value on the
current channel when measurements on Phase A and Phase B are
enabled (Bit PEAKSEL[2:0] in the MMODE register are 011).
PEAKCYC is set to 16, meaning that the peak measurement
cycle is four line periods. The maximum absolute value of Phase A
is the greatest during the first four line periods (PEAKCYC = 16),
so the maximum absolute value is written into the less signifi-
cant 24 bits of the IPEAK register, and Bit 24 (IPPHASE[0]) of
the IPEAK register is set to 1 at the end of the period. This bit
remains at 1 for the duration of the second PEAKCYC period of
four line cycles. The maximum absolute value of Phase B is the
greatest during the second PEAKCYC period; therefore, the
maximum absolute value is written into the less significant
24 bits of the IPEAK register, and Bit 25 (IPPHASE[1]) in the
IPEAK register is set to 1 at the end of the period.
At the end of the peak detection period in the current channel,
Bit 23 (PKI) in the STATUS1 register is set to 1. If Bit 23 (PKI)
in the MASK1 register is set, the IRQ1 interrupt pin is driven low
at the end of PEAKCYC period and Status Bit 23 (PKI) in the
STATUS1 register is set to 1. In a similar way, at the end of the
peak detection period in the voltage channel, Bit 24 (PKV) in the
STATUS1 register is set to 1. If Bit 24 (PKV) in the MASK1
register is set, the IRQ1 interrupt pin is driven low at the end of
PEAKCYC period and Status Bit 24 (PKV) in the STATUS1
register is set to 1. To find the phase that triggered the interrupt,
one of either the IPEAK or VPEAK registers is read immediately
after reading the STATUS1 register. Next, the status bits are
cleared, and the IRQ1 pin is set to high by writing to the
STATUS1 register with the status bit set to 1.
Note that the internal zero-crossing counter is always active. By
setting Bits[4:2] (PEAKSEL[2:0]) in the MMODE register, the
first peak detection result is, therefore, not executed across a full
PEAKCYC period. Writing to the PEAKCYC register when the
PEAKSEL[2:0] bits are set resets the zero-crossing counter,
thereby ensuring that the first peak detection result is obtained
across a full PEAKCYC period.
Overvoltage and Overcurrent Detection
The ADE7854/ADE7858/ADE7868/ADE7878 detect when the
instantaneous absolute value measured on the voltage and
current channels becomes greater than the thresholds set in the
OVLVL and OILVL 24-bit unsigned registers. If Bit 18 (OV) in
the MASK1 register is set, the IRQ1 interrupt pin is driven low
in case of an overvoltage event. There are two status flags set
when the IRQ1 interrupt pin is driven low: Bit 18 (OV) in the
STATUS1 register and one of Bits[11:9] (OVPHASE[2:0]) in the
PHSTATUS register to identify the phase that generated the
overvoltage. The Status Bit 18 (OV) in the STATUS1 register
and all Bits[11:9] (OVPHASE[2:0]) in the PHSTATUS register
are cleared, and the IRQ1 pin is set to high by writing to the
STATUS1 register with the status bit set to 1. presents
overvoltage detection in Phase A voltage.
Figure 50
OVLVL[23:0]
BIT 18 (OV) OF
STATUS1
BIT 9 (OVPHASE)
OF PHSTATUS
PHASE
A
VOLTAGE CHANNEL
OVERVOLTAGE
DETECTED
STATUS1[18] AND
PHSTATUS[9]
CANCELLED BY A
WRITE OF STATUS1
WITH OV BIT SET.
0
8510-036
Figure 50. Overvoltage Detection
Whenever the absolute instantaneous value of the voltage goes
above the threshold from the OVLVL register, Bit 18 (OV) in
the STATUS1 register and Bit 9 (OVPHASE[0]) in the PHSTATUS
register are set to 1. Bit 18 (OV) of the STATUS1 register and
Bit 9 (OVPHASE[0]) in the PHSTATUS register are cancelled
when the STATUS1 register is written with Bit 18 (OV) set to 1.
The recommended procedure to manage overvoltage events is
the following:
1. Enable OV interrupts in the MASK1 register by setting
Bit 18 (OV) to 1.
2. When an overvoltage event happens, the IRQ1 interrupt
pin goes low.
3. The STATUS1 register is read with Bit 18 (OV) set to 1.
4. The PHSTATUS register is read, identifying on which
phase or phases an overvoltage event happened.
5. The STATUS1 register is written with Bit 18 (OV) set to 1.
In this moment, Bit OV is erased and also all Bits[11:9]
(OVPHASE[2:0]) of the PHSTATUS register.
In case of an overcurrent event, if Bit 17 (OI) in the MASK1
register is set, the IRQ1 interrupt pin is driven low. Immediately,
Bit 17 (OI) in the STATUS1 register and one of Bits[5:3]
(OIPHASE[2:0]) in the PHSTATUS register, which identify
the phase that generated the interrupt, are set. To find the
phase that triggered the interrupt, the PHSTATUS register
is read immediately after reading the STATUS1 register. Next,
Status Bit 17 (OI) in the STATUS1 register and Bits[5:3]
(OIPHASE[2:0]) in the PHSTATUS register are cleared and the
IRQ1 pin is set to high by writing to the STATUS1 register with
the status bit set to 1. The process is similar with overvoltage
detection.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 37 of 96
Overvoltage and Overcurrent Level Set
The content of the overvoltage (OVLVL), and overcurrent,
(OILVL) 24-bit unsigned registers is compared to the absolute
value of the voltage and current channels. The maximum value of
these registers is the maximum value of the HPF outputs:
+5,928,256 (0x5A7540). When the OVLVL or OILVL register is
equal to this value, the overvoltage or overcurrent conditions
are never detected. Writing 0x0 to these registers signifies the
overvoltage or overcurrent conditions are continuously detected,
and the corresponding interrupts are permanently triggered.
As stated in the Current Waveform Gain Registers section, the
serial ports of the ADE78xx work on 32-, 16-, or 8-bit words.
Similar to the register presented in Figure 34, OILVL and
OVLVL registers are accessed as 32-bit registers with the eight
MSBs padded with 0s.
Neutral Current Mismatch—ADE7868, ADE7878
Neutral current mismatch is available in the ADE7868 and
ADE7878 only. In 3-phase systems, the neutral current is equal
to the algebraic sum of the phase currents
IN(t) = IA(t) + IB(t) + IC(t)
If there is a mismatch between these two quantities, then a
tamper situation may have occurred in the system.
The ADE7868/ADE7878 compute the sum of the phase
currents adding the content of the IAWV, IBWV, and ICWV
registers, and storing the result into the ISUM 28-bit signed
register: ISUM(t) = IA(t) + IB(t) + IC(t). ISUM is computed every
125 µs (8 kHz frequency), the rate at which the current samples
are available, and Bit 17 (DREADY) in the STATUS0 register is
used to signal when the ISUM register can be read. See the
Digital Signal Processor section for more details on Bit DREADY.
To recover ISUM(t) value from the ISUM register, use the
following expression:
FS
M
AX
SUM I
ADC
ISUM[27:0]
tI ×=)(
where:
ADCMAX = 5,928,256, the ADC output when the input is at full
scale.
IFS is the full-scale ADC phase current.
The ADE7868/ADE7878 compute the difference between the
absolute values of ISUM and the neutral current from the
INWV register, take its absolute value and compare it against
the ISUMLVL threshold. If ISUMLVLINWVISUM ,
then it is assumed that the neutral current is equal to the sum
of the phase currents, and the system functions correctly. If
ISUMLVLINWVISUM > , then a tamper situation may
have occurred, and Bit 20 (MISMTCH) in the STATUS1 register
is set to 1. An interrupt attached to the flag can be enabled by
setting Bit 20 (MISMTCH) in the MASK1 register. If enabled,
the IRQ1 pin is set low when Status Bit MISMTCH is set to 1.
The status bit is cleared and the IRQ1 pin is set back to high by
writing to the STATUS1 register with Bit 20 (MISMTCH) set to 1.
If ISUMLVLINWVISUM , then MISMTCH = 0
If ISUMLVLINWVISUM > , then MISMTCH = 1
ISUMLVL, the positive threshold used in the process, is a 24-bit
signed register. Because it is used in a comparison with an
absolute value, always set ISUMLVL as a positive number,
somewhere between 0x00000 and 0x7FFFFF. ISUMLVL uses
the same scale of the current ADCs outputs, so writing
+5,928,256 (0x5A7540) to the ISUMLVL register puts the
mismatch detection level at full scale; see the Current Channel
ADC section for details. Writing 0x000000, the default value, or
a negative value, signifies the MISMTCH event is always triggered.
The right value for the application should be written into the
ISUMLVL register after power-up or after a hardware/software
reset to avoid continuously triggering MISMTCH events.
As stated in the Current Waveform Gain Registers section, the
serial ports of the ADE7868/ADE7878 work on 32-, 16-, or 8-bit
words and the DSP works on 28 bits. As presented in Figure 51,
ISUM, the 28-bit signed register, is accessed as a 32-bit register
with the four most significant bits padded with 0s.
31 28 27
BIT 27 IS A SIGN BIT
0
28-BIT SIGNED NUMBER0000
08510-250
Figure 51. The ISUM[27:0] Register is Transmitted As a 32-Bit Word
Similar to the registers presented in Figure 33, the ISUMLVL
register is accessed as a 32-bit register with four most significant
bits padded with 0s and sign extended to 28 bits.
PHASE COMPENSATION
As described in the Current Channel ADC and Voltage Channel
ADC sections, the datapath for both current and voltages is the
same. The phase error between current and voltage signals
introduced by the ADE7854/ADE7858/ADE7868/ADE7878 is
negligible. However, the ADE7854/ADE7858/ADE7868/ADE7878
must work with transducers that may have inherent phase
errors. For example, a current transformer (CT) with a phase
error of 0.1° to 3° is not uncommon. These phase errors can
vary from part to part, and they must be corrected to perform
accurate power calculations.
The errors associated with phase mismatch are particularly
noticeable at low power factors. The ADE78xx provides a means
of digitally calibrating these small phase errors. The ADE78xx
allows a small time delay or time advance to be introduced into
the signal processing chain to compensate for the small phase
errors.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 38 of 96
(8)
The phase calibration registers (APHCAL, BPHCAL, and
CPHCAL) are 10-bit registers that can vary the time advance
in the voltage channel signal path from −374.0 µs to +61.5 s.
Negative values written to the PHCAL registers represent a time
advance whereas positive values represent a time delay. One LSB
is equivalent to 0.976 µs of time delay or time advance (clock
rate of 1.024 MHz). With a line frequency of 60 Hz, this gives
a phase resolution of 0.0211° (360° × 60 Hz/1.024 MHz) at the
fundamental. This corresponds to a total correction range of
−8.079° to +1.329° at 60 Hz. At 50 Hz, the correction range is
−6.732° to +1.107° and the resolution is 0.0176° (360° × 50 Hz/
1.024 MHz).
Given a phase error of x degrees, measured using the phase
voltage as the reference, the corresponding LSBs are computed
dividing x by the phase resolution (0.0211°/LSB for 60 Hz and
0.0176°/LSB for 50 Hz). Results between −383 and +63 only are
acceptable; numbers outside this range are not accepted. If the
result is negative, the absolute value is written into the PHCAL
registers. If the result is positive, 512 is added to the result
before writing it into xPHCAL.
APHCAL,
BPHCAL, or
CPHCAL =
>+
0,512
_
0,
_
x
resolutionphase
x
x
resolutionphase
x
Figure 53 illustrates how the phase compensation is used to remove
x =1° phase lead in IA of the current channel from the external
current transducer (equivalent of 55.5 µs for 50 Hz systems). To
cancel the lead (1°) in the current channel of Phase A, a phase
lead must be introduced into the corresponding voltage channel.
Using Equation 8, APHCAL is 57 least significant bits, rounded
up from 56.8. The phase lead is achieved by introducing a time
delay of 55.73 µs into the Phase A current.
As stated in the Current Waveform Gain Registers section, the
serial ports of the ADE785xx work on 32-, 16-, or 8-bit words.
As shown in Figure 52, APHCAL, BPHCAL, and CPHCAL
10-bit registers are accessed as 16-bit registers with the six MSBs
padded with 0s.
0000 00
15 10 9 0
xPHCAL
08510-038
Figure 52. xPHCAL Registers Communicated As 16-Bit Registers
PHASE
CALIBRATION
APHCAL = 57
ADC
PGA3
VAP
VA
VN
ADC
PGA1
I
A
P
IA
IAN
PHASE COMPENSATION
ACHIEVED DELAYING
IA BY 56µs
50Hz
VA
IA
VA
IA
08510-039
Figure 53. Phase Calibration on Voltage Channels
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 39 of 96
REFERENCE CIRCUIT
The nominal reference voltage at the REFIN/OUT pin is 1.2 ±
0.075% V. This is the reference voltage used for the ADCs in
the ADE7854/ADE7858/ADE7868/ADE7878. The REFIN/OUT
pin can be overdriven by an external source, for example, an
external 1.2 V reference. The voltage of the ADE78xx reference
drifts slightly with temperature; see the Specifications section for
the temperature coefficient specification (in ppm/°C). The value of
the temperature drift varies from part to part. Because the refer-
ence is used for all ADCs, any x% drift in the reference results in
a 2x% deviation of the meter accuracy. The reference drift resulting
from temperature changes is usually very small and typically
much smaller than the drift of other components on a meter.
Alternatively, the meter can be calibrated at multiple temperatures.
If Bit 0 (EXTREFEN) in the CONFIG2 register is cleared to 0 (the
default value), the ADE7854/ADE7858/ADE7868/ADE7878 use
the internal voltage reference. If the bit is set to 1, the external
voltage reference is used. Set the CONFIG2 register during the
PSM0 mode. Its value is maintained during the PSM1, PSM2, and
PSM3 power modes.
DIGITAL SIGNAL PROCESSOR
The ADE7854/ADE7858/ADE7868/ADE7878 contain a fixed
function digital signal processor (DSP) that computes all powers
and rms values. It contains program memory ROM and data
memory RAM.
The program used for the power and rms computations is
stored in the program memory ROM and the processor executes
it every 8 kHz. The end of the computations is signaled by
setting Bit 17 (DREADY) to 1 in the STATUS0 register. An
interrupt attached to this flag can be enabled by setting Bit 17
(DREADY) in the MASK0 register. If enabled, the IRQ0 pin is
set low and Status Bit DREADY is set to 1 at the end of the
computations. The status bit is cleared and the IRQ0 pin is set
to high by writing to the STATUS0 register with Bit 17 (DREADY)
set to 1.
The registers used by the DSP are located in the data memory
RAM, at addresses between 0x4380 and 0x43BE. The width of
this memory is 28 bits. Within the DSP core, the DSP contains a
two stage pipeline. This means that when a single register needs
to be initialized, two more writes are required to ensure the
value has been written into RAM, and if two or more registers
need to be initialized, the last register must be written two more
times to ensure the value has been written into RAM.
As explained in the Power-Up Procedure section, at power-up
or after a hardware or software reset, the DSP is in idle mode.
No instruction is executed. All the registers located in the data
memory RAM are initialized at 0, their default values, and they
can be read/written without any restriction. The run register,
used to start and stop the DSP, is cleared to 0x0000. The run
register needs to be written with 0x0001 for the DSP to start
code execution. It is recommended to first initialize all ADE78xx
registers located in the data memory RAM with their desired
values. Next, write the last register in the queue two additional
times to flush the pipeline, and then write the run register with
0x0001. In this way, the DSP starts the computations from a
desired configuration.
To protect the integrity of the data stored in the data memory
RAM of the DSP (addresses between 0x4380 and 0x43BE), a
write protection mechanism is available. By default, the
protection is disabled and registers placed between 0x4380 and
0x43BE can be written without restriction. When the protection
is enabled, no writes to these registers is allowed. Registers can
be always read, without restriction, independent of the write
protection state.
To enable the protection, write 0xAD to an internal 8-bit
register located at Address 0xE7FE, followed by a write of 0x80
to an internal 8-bit register located at Address 0xE7E3.
It is recommended to enable the write protection before starting
the DSP. If any data memory RAM based register needs to be
changed, simply disable the protection, change the value and
then re-enable the protection. There is no need to stop the DSP
to change these registers.
To disable the protection, write 0xAD to an internal 8-bit
register located at Address 0xE7FE, followed by a write of 0x00
to an internal 8-bit register located at Address 0xE7E3.
The recommended procedure to initialize the registers located
in the data memory RAM is as follows:
Initialize all registers. Write the last register in the queue
three times to ensure its value was written into the RAM.
Initialize all of the other registers of the ADE7854/ADE7858/
ADE7868/ADE7878 here as well.
Enable the write protection by writing 0xAD to an internal
8-bit register located at Address 0xE7FE, followed by a write of
0x80 to an internal 8-bit register located at Address 0xE7E3.
Read back all data memory RAM registers to ensure they
were initialized with the desired values.
In the remote case that one or more registers are not initia-
lized correctly, disable the protection by writing 0xAD to
an internal 8-bit register located at Address 0xE7FE, followed
by a write of 0x00 to an internal 8-bit register located at
Address 0xE7E3. Reinitialize the registers. Write the last
register in the queue three times. Enable the write protec-
tion by writing 0xAD to an internal 8-bit register located
at Address 0xE7FE, followed by a write of 0x80 to an internal
8-bit register located at Address 0xE7E3..
Start the DSP by setting run = 1.
There is no obvious reason to stop the DSP if the ADE78xx is
maintained in PSM0 normal mode. All ADE78xx registers,
including ones located in the data memory RAM, can be
modified without stopping the DSP. However, to stop the DSP,
0x0000 has to be written into run register. To restart the DSP,
one of the following procedures must be followed:
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 40 of 96
If the ADE7854/ADE7858/ADE7868/ADE7878 registers
located in the data memory RAM have not been modified,
write 0x0001 into the run register to start the DSP.
If the ADE7854/ADE7858/ADE7868/ADE7878 registers
located in the data memory RAM have to be modified, first
execute a software or a hardware reset, initialize all
ADE7854/ADE7858/ADE7868/ADE7878 registers at
desired values, enable the write protection, and then write
0x0001 into the run register to start the DSP.
As mentioned in the Power Management section, when the
ADE7854/ADE7858/ADE7868/ADE7878 switch out of PSM0
power mode, it is recommended to stop the DSP by writing
0x0000 into the run register (see Table 11 and Table 12 for
the recommended actions when changing power modes).
ROOT MEAN SQUARE MEASUREMENT
Root mean square (rms) is a measurement of the magnitude of
an ac signal. Its definition can be both practical and mathematical.
Defined practically, the rms value assigned to an ac signal is the
amount of dc required to produce an equivalent amount of
power in the load. Mathematically, the rms value of a conti-
nuous signal f(t) is defined as
()
dttf
t
rmsF
=t
0
2
1 (9)
For time sampling signals, rms calculation involves squaring the
signal, taking the average, and obtaining the square root.
[]
=
=N
N
nf
N
rmsF
1
2
1 (10)
Equation 10 implies that for signals containing harmonics, the
rms calculation contains the contribution of all harmonics, not
only the fundamental. The ADE78xx uses two different methods
to calculate rms values. The first one is very accurate and is active
only in PSM0 mode. The second one is less accurate, uses the
estimation of the mean absolute value (mav) measurement, is
active in PSM0 and PSM1 modes, and is available for the
ADE7868 and ADE7878 only.
The first method is to low-pass filter the square of the input
signal (LPF) and take the square root of the result (see Figure 54).
(
k
k
kγtωkFtf +=
=
sin2)(
1
)
)
(11)
Then
()
(
=
=
=
+×+××+
++=
mk
mk
m
k
m
k
kkk
kk
γtωmγtωkFF
γtωkFFtf
1,
1
2
1
22
sinsin22
)2cos()(
(12)
After the LPF and the execution of the square root, the rms
value of f(t) is obtained by
=
=
1
2
k
k
FF (13)
The rms calculation based on this method is simultaneously
processed on all seven analog input channels. Each result is
available in the 24-bit registers: AIRMS, BIRMS, CIRMS,
AVRMS, BVRMS, CVRMS, and NIRMS (NIRMS is available
on the ADE7868 and ADE7878 only).
The second method computes the absolute value of the input
signal and then filters it to extract its dc component. It computes
the absolute mean value of the input. If the input signal in
Equation 12 has a fundamental component only, its average
value is
××××= T
T
1
T
1
DC dttωFdttωF
T
F
2
2
0
)sin(2)sin(2
1
1
DC F
π
F××= 2
2
The calculation based on this method is simultaneously processed
only on the three phase currents. Each result is available in the
20-bit registers, which are available on the AE7868 and ADE7878
only: AIMAV, BMAV, and CMAV. Note that the proportionality
between mav and rms values is maintained for the fundamental
components only. If harmonics are present in the current channel,
the mean absolute value is no longer proportional to rms.
Current RMS Calculation
This section presents the first approach to compute the rms
values of all phase and neutral currents.
Figure 54 shows the detail of the signal processing chain for the
rms calculation on one of the phases of the current channel.
The current channel rms value is processed from the samples
used in the current channel. The current rms values are signed
24-bit values and they are stored into the AIRMS, BIRMS, CIRMS,
and NIRMS (ADE7868/ADE7878 only) registers. The update
rate of the current rms measurement is 8 kHz.
With the specified full-scale analog input signal of 0.5 V, the
ADC produces an output code that is approximately ±5,928,256.
The equivalent rms value of a full-scale sinusoidal signal is
4,191,910 (0x3FF6A6), independent of the line frequency. If
the integrator is enabled, that is, when Bit 0 (INTEN) in the
CONFIG register is set to 1, the equivalent rms value of a full-
scale sinusoidal signal at 50 Hz is 4,191,910 (0x3FF6A6) and at
60 Hz is 3,493,258 (0x354D8A).
The accuracy of the current rms is typically 0.1% error from
the full-scale input down to 1/1000 of the full-scale input when
PGA = 1. Additionally, this measurement has a bandwidth of
2 kHz. It is recommended to read the rms registers synchronous
to the voltage zero crossings to ensure stability. The IRQ1 inter-
rupt can be used to indicate when a zero crossing has occurred
(see the section). shows the settling time for
the I rms measurement, which is the time it takes for the rms
register to reflect the value at the input to the current channel
when starting from 0.
Interrupts Table 13
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 41 of 96
Table 13. Settling Time for I rms Measurement
Integrator Status 50 Hz Input signals
As stated in the Current Waveform Gain Registers section, the
serial ports of the ADE78xx work on 32-, 16-, or 8-bit words.
Similar to the register presented in Figure 34, the AIRMS,
BIRMS, CIRMS, and NIRMS (ADE7868/ADE7878 only) 24-bit
signed registers are accessed as 32-bit registers with the eight
MSBs padded with 0s.
60 Hz Input signals
Integrator Off 440 ms 440 ms
Integrator On 550 ms 500 ms
0xA58AC0 =
–5,928,256
0x5A7540 =
5,928,256
0V
CURRENT SIGNAL FROM
HPF OR INTEGRATOR
(IF ENABLED) LPF
x
2
2
7
xIRMSOS[23:0]
xIRMS[23:0]
08510-040
Figure 54. Current RMS Signal Processing
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 42 of 96
Current RMS Offset Compensation
The ADE7854/ADE7858/ADE7868/ADE7878 incorporate a
current rms offset compensation register for each phase:
AIRMSOS, BIRMSOS, CIRMSOS registers, and the NIRMSOS
register for ADE7878 and ADE7868 only. These are 24-bit
signed registers that are used to remove offsets in the current
rms calculations. An offset can exist in the rms calculation due
to input noises that are integrated in the dc component of I2(t).
The current rms offset compensation register is added to the
squared current rms before the square root is executed. Assuming
that the maximum value from the current rms calculation is
4,191,400 with full-scale ac inputs (50 Hz), one LSB of the current
rms offset represents 0.00037% (( 12841912+/4191 − 1) × 100)
of the rms measurement at 60 dB down from full scale. Conduct
offset calibration at low current; avoid using currents equal to
zero for this purpose.
IRMSOSrmsIrmsI ×+= 128
2
0 (14)
where I rms0 is the rms measurement without offset correction.
As stated in the Current Waveform Gain Registers section, the
serial ports of the ADE78xx work on 32-, 16-, or 8-bit words
and the DSP works on 28 bits. Similar to the register presented
in Figure 33, the AIRMSOS, BIRMSOS, CIRMSOS, and
NIRMSOS (ADE7868/ADE7878 only) 24-bit signed registers
are accessed as 32-bit registers with four MSBs padded with 0s
and sign extended to 28 bits.
Current Mean Absolute Value Calculation—ADE7868
and ADE7878 Only
This section presents the second approach to estimate the rms
values of all phase currents using the mean absolute value (mav)
method. This approach is used in PSM1 mode, which is available
to the ADE7868 and ADE7878 only, to allow energy accumu-
lation based on current rms values when the missing neutral
case demonstrates to be a tamper attack. This datapath is active
also in PSM0 mode to allow for its gain calibration. The gain is
used in the external microprocessor during PSM1 mode. The
mav value of the neutral current is not computed using this
method. Figure 55 shows the details of the signal processing
chain for the mav calculation on one of the phases of the current
channel.
CURRENT SIGNAL
COMING FROM ADC xIMAV[23:0]
HPF HPF
|X|
08510-251
Figure 55. Current MAV Signal Processing for PSM1 Mode
The current channel mav value is processed from the samples
used in the current channel waveform sampling mode. The
samples are passed through a high-pass filter to eliminate the
eventual dc offsets introduced by the ADCs and the absolute
values are computed. The outputs of this block are then filtered
to obtain the average. The current mav values are unsigned 20-bit
values and they are stored in the AIMAV, BIMAV, and CIMAV
registers. The update rate of this mav measurement is 8 kHz.
207000
207500
208000
208500
209000
209500
210000
210500
211000
211500
212000
45 50 55
FREQUENCY (Hz)
LSB
60 65
08510-252
Figure 56. xIMAV Register Values at Full Scale, 45 Hz to 65 Hz Line
Frequencies
The mav values of full-scale sinusoidal signals of 50 Hz and
60 Hz are 209,686 and 210,921, respectively. As seen in Figure 56,
there is a 1.25% variation between the mav estimate at 45 Hz
and the one at 65 Hz for full-scale sinusoidal inputs. The accuracy
of the current mav is typically 0.5% error from the full-scale
input down to 1/100 of the full-scale input. Additionally, this
measurement has a bandwidth of 2 kHz. The settling time for
the current mav measurement, that is the time it takes for the
mav register to reflect the value at the input to the current
channel within 0.5% error, is 500 ms.
As stated in the Current Waveform Gain Registers section, the
serial ports of the ADE7868/ADE7878 work on 32-, 16-, or 8-
bit words. As presented in Figure 57, the AIMAV, BIMAV, and
CIMAV 20-bit unsigned registers are accessed as 32-bit registers
with the 12 MSBs padded with 0s.
31 20 19 0
20-BIT UNSIGNED NUMBER0000 0000 0000
08510-253
Figure 57. xIMAV Registers Transmitted as 32-Bit Registers
Current MAV Gain and Offset Compensation
The current rms values stored in the AIMAV, BIMAV, and
CIMAV registers can be calibrated using gain and offset
coefficients corresponding to each phase. It is recommended to
calculate the gains in PSM0 mode by supplying the ADE7868/
ADE7878 with nominal currents. The offsets can be estimated
by supplying the ADE7868/ADE7878 with low currents, usually
equal to the minimum value at which the accuracy is required.
Every time the external microcontroller reads the AIMAV,
BIMAV, and CIMAV registers, it uses these coefficients stored
in its memory to correct them.
Voltage Channel RMS Calculation
Figure 58 shows the detail of the signal processing chain for the
rms calculation on one of the phases of the voltage channel. The
voltage channel rms value is processed from the samples used in
the voltage channel. The voltage rms values are signed 24-bit
values and they are stored into the Registers AVRMS, BVRMS, and
CVRMS. The update rate of the current rms measurement is 8 kHz.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 43 of 96
With the specified full-scale analog input signal of 0.5 V, the
ADC produces an output code that is approximately ±5,928,256.
The equivalent rms value of a full-scale sinusoidal signal is
4,191,910 (0x3FF6A6), independent of the line frequency.
The settling time for the V rms measurement is 440 ms for both
50 Hz and 60 Hz input signals. The V rms measurement is the
time it takes for the rms register to reflect the value at the input
to the voltage channel when starting from 0.
As stated in the Current Waveform Gain Registers section, the
serial ports of the ADE78xx work on 32-, 16-, or 8-bit words.
Similar to the register presented in Figure 34, the AVRMS,
BVRMS, and CVRMS 24-bit signed registers are accessed as
32-bit registers with the eight MSBs padded with 0s.
The accuracy of the voltage rms is typically 0.1% error from the
full-scale input down to 1/1000 of the full-scale input. Additionally,
this measurement has a bandwidth of 2 kHz. It is recommended
to read the rms registers synchronous to the voltage zero crossings
to ensure stability. The IRQ1 interrupt can be used to indicate
when a zero crossing has occurred (see the section). Interrupts
0xA58AC0 =
–5,928,256
0x5A7540 =
5,928,256
0V
VOLTAGE SIGNAL
FROM HPF LPF
x
2
2
7
xVRMSOS[23:0]
xVRMS[23:0]
08510-041
Figure 58. Voltage RMS Signal Processing
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 44 of 96
Voltage RMS Offset Compensation
The ADE78xx incorporates voltage rms offset compensation
registers for each phase: AVRMSOS, BVRMSOS, and CVRMSOS.
These are 24-bit signed registers used to remove offsets in the
voltage rms calculations. An offset can exist in the rms calculation
due to input noises that are integrated in the dc component of
V2(t). The voltage rms offset compensation register is added to the
squared voltage rms before the square root is executed. Assuming
that the maximum value from the voltage rms calculation is
4,191,400 with full-scale ac inputs (50 Hz), one LSB of the current
rms offset represents 0.00037% (( 12841912+/4191 − 1) × 100)
of the rms measurement at 60 dB down from full scale. Conduct
offset calibration at low current; avoid using voltages equal to zero
for this purpose.
VRMSOSrmsVrmsV ×+= 128
2
0 (15)
where V rms0 is the rms measurement without offset correction.
As stated in the Current Waveform Gain Registers section, the
serial ports of the ADE78xx work on 32-, 16-, or 8-bit words
and the DSP works on 28 bits. Similar to registers presented in
Figure 33, the AVRMSOS, BVRMSOS, and CVRMSOS 24-bit
registers are accessed as 32-bit registers with the four most
significant bits padded with 0s and sign extended to 28 bits.
ACTIVE POWER CALCULATION
The ADE7854/ADE7858/ADE7868/ADE7878 compute the
total active power on every phase. Total active power considers
in its calculation all fundamental and harmonic components of
the voltages and currents. In addition, the ADE7878 computes
the fundamental active power, the power determined only by
the fundamental components of the voltages and currents.
Total Active Power Calculation
Electrical power is defined as the rate of energy flow from source
to load, and it is given by the product of the voltage and current
waveforms. The resulting waveform is called the instantaneous
power signal, and it is equal to the rate of energy flow at every
instant of time. The unit of power is the watt or joules/sec. If an
ac system is supplied by a voltage, v(t), and consumes the current,
i(t), and each of them contains harmonics, then
sin2)(
1
=
=
k
k
Vtv (kωt + φk) (16)
()
k
k
kγtωkIti +=
=
sin2)(
1
where:
Vk, Ik are rms voltage and current, respectively, of each
harmonic.
φk, γk are the phase delays of each harmonic.
The instantaneous power in an ac system is
p(t) = v(t) × i(t) = cos(φkγk) − cos(2kωt + φk + γk) +
{cos[(km)ωt + φkγm] – cos[(k + m)ωt + φk + γm]}
=1
k
kk IV
=1
k
kk IV
=
mk
mk
m
kIV
1,
(17)
The average power over an integral number of line cycles (n) is
given by the expression in Equation 18.
P =
()
=
=
1
0
1
kkk
nT
IVdttp
nT cos(φkγk) (18)
where:
T is the line cycle period.
P is referred to as the total active or total real power.
Note that the total active power is equal to the dc component of
the instantaneous power signal p(t) in Equation 17, that is,
=1k
kk IV cos(φkγk)
This is the expression used to calculate the total active power in
the ADE78xx for each phase. The expression of fundamental active
power is obtained from Equation 18 with k = 1, as follows:
FP = V1I1 cos(φ1γ1) (19)
Figure 59 shows how the ADE78xx computes the total active
power on each phase. First, it multiplies the current and voltage
signals in each phase. Next, it extracts the dc component of the
instantaneous power signal in each phase (A, B, and C) using
LPF2, the low-pass filter.
AVGAIN
HPFDIS
[23:0]
HPF
LPF
AIGAIN
HPFDIS
[23:0]
DIGITAL
INTEGRATOR
HPF
V
A
I
A
AWGAIN AWATTOS
AWATT
2
4
APHCAL
INSTANTANEOUS
PHASE A ACTIVE
POWER
08510-145
DIGITAL SIGNAL PROCESSOR
Figure 59. Total Active Power Datapath
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 45 of 96
If the phase currents and voltages contain only the fundamental
component, are in phase (that is φ1 = γ1 = 0), and they correspond
to full-scale ADC inputs, then multiplying them results in an
instantaneous power signal that has a dc component, V1 × I1,
and a sinusoidal component, V1 × I1 cos(2ωt); Figure 60 shows
the corresponding waveforms.
INSTANTANEOU
S
POWER SIGNAL
INSTANTANEOUS
ACTIVE POWER
SIGNAL: V rms × I rms
p(t)= V rms × I rms – V rms × I rms × cos(2ωt)
0x3FED4D6
67,032,278
V rms × I rms
0x1FF6A6B =
33,516,139
0x000 0000
i(t) = I rmsin(ωt)
v(t) = 2×V rmsin(ωt)
08510-043
Figure 60. Active Power Calculation
Because LPF2 does not have an ideal brick wall frequency
response (see Figure 61), the active power signal has some
ripple due to the instantaneous power signal. This ripple is
sinusoidal and has a frequency equal to twice the line frequency.
Because the ripple is sinusoidal in nature, it is removed when
the active power signal is integrated over time to calculate the
energy.
0
–5
–10
–15
–20
–25
0.1 13
FREQUENCY (Hz)
MAGNITUDE (dB)
10
08510-103
Figure 61. Frequency Response of the LPF Used
to Filter Instantaneous Power in Each Phase
The ADE7854/ADE7858/ADE7868/ADE7878 store the
instantaneous total phase active powers into the AWATT,
BWATT, and CWATT registers. Their expression is
=
××=
1kFS
k
FS
k
I
I
U
U
xWATT cos(φk – γk) × PMAX × 4
2
1 (20)
where:
UFS, IFS are the rms values of the phase voltage and current when
the ADC inputs are at full scale.
PMAX = 33,516,139; it is the instantaneous power computed
when the ADC inputs are at full scale and in phase.
The xWATT[23:0] waveform registers can be accessed using
various serial ports. Refer to the Waveform Sampling Mode
section for more details.
Fundamental Active Power Calculation—ADE7878 Only
The ADE7878 computes the fundamental active power using
a proprietary algorithm that requires some initializations function
of the frequency of the network and its nominal voltage measured
in the voltage channel. Bit 14 (SELFREQ) in the COMPMODE
register must be set according to the frequency of the network in
which the ADE7878 is connected. If the network frequency is
50 Hz, clear this bit to 0 (the default value). If the network fre-
quency is 60 Hz, set this bit to 1. In addition, initialize the VLEVEL
24-bit signed register with a positive value based on the
following expression:
520,491×=
n
FS
U
U
VLEVEL (21)
where:
UFS is the rms value of the phase voltages when the ADC inputs
are at full scale.
Un is the rms nominal value of the phase voltage.
As stated in the Current Waveform Gain Registers section, the
serial ports of the ADE7878 work on 32-, 16-, or 8-bit words
and the DSP works on 28 bits. Similar to the registers presented
in Figure 33, the VLEVEL 24-bit signed register is accessed as a
32-bit register with four most significant bits padded with 0s
and sign extended to 28 bits.
Table 14 presents the settling time for the fundamental active
power measurement.
Table 14. Settling Time for Fundamental Active Power
Input Signals
63% Full Scale 100% Full Scale
375 ms 875 ms
Active Power Gain Calibration
Note that the average active power result from the LPF2 output
in each phase can be scaled by ±100% by writing to the phases
watt gain 24-bit register (AWGAIN, BWGAIN, CWGAIN,
AFWGAIN, BFWGAIN, or CFWGAIN). The xWGAIN
registers are placed in each phase of the total active power
datapath, and the xFWGAIN (available for the ADE7878 only)
registers are placed in each phase of the fundamental active
power datapath. The watt gain registers are twos complement,
signed registers and have a resolution of 2−23/LSB. Equation 22
describes mathematically the function of the watt gain registers.
+×
=
23
2
12 gisterReGainWatt
OutputLPF
DataPowe
r
Average
(22)
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 46 of 96
The output is scaled by −50% by writing 0xC00000 to the watt
gain registers, and it is increased by +50% by writing 0x400000
to them. These registers are used to calibrate the active power
(or energy) calculation in the ADE7854/ADE7858/ADE7868/
ADE7878 for each phase.
As stated in the Current Waveform Gain Registers section, the
serial ports of the ADE7854/ADE7858/ADE7868/ADE7878
work on 32-, 16-, or 8-bit words, and the DSP works on 28 bits.
Similar to registers presented in Figure 33, AWGAIN, BWGAIN,
CWGAIN, AFWGAIN, BFWGAIN, and CFWGAIN 24-bit
signed registers are accessed as 32-bit registers with the four
MSBs padded with 0s and sign extended to 28 bits.
Active Power Offset Calibration
The ADE7854/ADE7858/ADE7868/ADE7878 incorporate a
watt offset 24-bit register on each phase and on each active
power. The AWATTOS, BWATTOS, and CWATTOS registers
compensate the offsets in the total active power calculations,
and the AFWATTOS, BFWATTOS, and CFWATTOS registers
compensate offsets in the fundamental active power calculations.
These are signed twos complement, 24-bit registers that are
used to remove offsets in the active power calculations. An
offset can exist in the power calculation due to crosstalk between
channels on the PCB or in the chip itself. One LSB in the active
power offset register is equivalent to 1 LSB in the active power
multiplier output. With full-scale current and voltage inputs,
the LPF2 output is PMAX = 33,516,139. At −80 dB down from
the full scale (active power scaled down 104 times), one LSB of
the active power offset register represents 0.0298% of PMAX.
As stated in the Current Waveform Gain Registers section, the
serial ports of the ADE7854/ADE7858/ADE7868/ADE7878
work on 32-, 16-, or 8-bit words and the DSP works on 28 bits.
Similar to registers presented in Figure 33, the AWATTOS,
BWATTOS, CWATTOS, AFWATTOS, BFWATTOS, and
CFWATTOS 24-bit signed registers are accessed as 32-bit
registers with the four MSBs padded with 0s and sign extended
to 28 bits.
Sign of Active Power Calculation
The average active power is a signed calculation. If the phase
difference between the current and voltage waveform is more
than 90°, the average power becomes negative. Negative power
indicates that energy is being injected back on the grid. The
ADE78xx has sign detection circuitry for total active power
calculations. It can monitor the total active powers or the
fundamental active powers. As described in the Active Energy
Calculation section, the active energy accumulation is performed
in two stages. Every time a sign change is detected in the energy
accumulation at the end of the first stage, that is, after the energy
accumulated into the internal accumulator reaches the WTHR
register threshold, a dedicated interrupt is triggered. The sign of
each phase active power can be read in the PHSIGN register.
Bit 6 (REVAPSEL) in the ACCMODE register sets the type of
active power being monitored. When REVAPSEL is 0, the
default value, the total active power is monitored. When
REVAPSEL is 1, the fundamental active power is monitored.
Bits[8:6] (REVAPC, REVAPB, and REVAPA, respectively) in the
STATUS0 register are set when a sign change occurs in the
power selected by Bit 6 (REVAPSEL) in the ACCMODE
register.
Bits[2:0] (CWSIGN, BWSIGN, and AWSIGN, respectively) in
the PHSIGN register are set simultaneously with the REVAPC,
REVAPB, and REVAPA bits. They indicate the sign of the power.
When they are 0, the corresponding power is positive. When
they are 1, the corresponding power is negative.
Bit REVAPx of STATUS0 and Bit xWSIGN in the PHSIGN
register refer to the total active power of Phase x, the power type
being selected by Bit 6 (REVAPSEL) in the ACCMODE register.
Interrupts attached to Bits[8:6] (REVAPC, REVAPB, and REVAPA,
respectively) in the STATUS0 register can be enabled by setting
Bits[8:6] in the MASK0 register. If enabled, the IRQ0 pin is set
low, and the status bit is set to 1 whenever a change of sign occurs.
To find the phase that triggered the interrupt, the PHSIGN register
is read immediately after reading the STATUS0 register. Next, the
status bit is cleared and the IRQ0 pin is returned to high by writing
to the STATUS0 register with the corresponding bit set to 1.
Active Energy Calculation
As previously stated, power is defined as the rate of energy flow.
This relationship can be expressed mathematically as
d
t
dEnergy
Power = (23)
Conversely, energy is given as the integral of power, as follows:
()
dttpEnergy
= (24)
Total and fundamental active energy accumulations are always
signed operations. Negative energy is subtracted from the active
energy contents.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 47 of 96
AVGAIN
HPFDIS
[23:0]
HPF
AIGAIN
HPFDIS
[23:0]
DIGITAL
INTEGRATOR
REVAPA BIT IN
STATUS0[31:0]
HPF
VA
IA
AWGAIN AWATTOS
APHCAL ACCUMULATOR
2
4
WTHR[47:0]
AWATT
AWATTHR[31:0]
32-BIT
REGISTER
LPF2
08510-044
DIGITAL SIGNAL PROCESSOR
Figure 62. Total Active Energy Accumulation
The ADE7854/ADE7858/ADE7868/ADE7878 achieve the
integration of the active power signal in two stages (see Figure 62).
The process is identical for both total and fundamental active
powers. The first stage is accomplished inside the DSP: every
125 µs (8 kHz frequency) the instantaneous phase total or funda-
mental active power is accumulated into an internal register.
When a threshold is reached, a pulse is generated at the processor
port, and the threshold is subtracted from the internal register.
The sign of the energy in this moment is considered the sign of
the active power (see Sign of Active Power Calculation section
for details). The second stage is done outside the DSP and consists
of accumulating the pulses generated by the processor into internal
32-bit accumulation registers. The content of these registers is
transferred to watt-hour registers, xWATTHR and xFWATTHR,
when these registers are accessed.
WTHR[47:0]
1 DSP PULSE = 1LSB OF WATTHR[31:0]
ACTIVE POWER
A
CCUMULATION
IN DSP
DSP
GENERATED
PULSES
08510-045
Figure 63. Active Power Accumulation Inside the DSP
Figure 63 explains this process. The WTHR 48-bit signed register
contains the threshold. It is introduced by the user and is common
for all phase total active and fundamental powers. Its value
depends on how much energy is assigned to one LSB of watt-
hour registers. Supposing a derivative of wh [10n wh], n as an
integer, is desired as one LSB of the xWATTHR register. Then
WTHR is computed using the following expression:
FSFS
n
S
IU
fPMAX
WTHR ×
×××
=103600 (25)
where:
PMAX = 33,516,139 = 0x1FF6A6B as the instantaneous power
computed when the ADC inputs are at full scale.
fS = 8 kHz, the frequency with which the DSP computes the
instantaneous power.
UFS, IFS are the rms values of phase voltages and currents when
the ADC inputs are at full scale.
The maximum value that can be written on WTHR is 247 − 1.
The minimum value is 0x0, but it is recommended to write a
number equal to or greater than PMAX. Never use negative
numbers.
WTHR is a 48-bit register. As stated in the Current Waveform
Gain Registers section, the serial ports of the ADE7854/ADE7858/
ADE7868/ADE7878 work on 32-, 16-, or 8-bit words. As shown
in Figure 64, the WTHR register is accessed as two 32-bit
registers (WTHR1 and WTHR0), each having eight MSBs
padded with 0s.
WTHR[47:0]
47 24
31 24 23 0 31 24 23 0
23 0
WTHR1[31:0] WTHR0[31:0]
0000 0000 24 BIT SIGNED NUMBER 0000 0000 24 BIT SIGNED NUMBER
08510-046
Figure 64. WTHR[47:0] Communicated As Two 32-Bit Registers
This discrete time accumulation or summation is equivalent to
integration in continuous time following the description in
Equation 26.
()
()
×==
=
0
0T
Lim
n
TnTpdttpEnergy (26)
where:
n is the discrete time sample number.
T is the sample period.
In the ADE7854/ADE7858/ADE7868/ADE7878, the total phase
active powers are accumulated in the AWATTHR, BWATTHR, and
CWATTHR 32-bit signed registers, and the fundamental phase
active powers are accumulated in AFWATTHR, BFWATTHR, and
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 48 of 96
CFWATTHR 32-bit signed registers. The active energy register
content can roll over to full-scale negative (0x80000000) and
continue increasing in value when the active power is positive.
Conversely, if the active power is negative, the energy register
underflows to full-scale positive (0x7FFFFFFF) and continues
decreasing in value.
Bit 0 (AEHF) in the STATUS0 register is set when Bit 30 of
one of the xWATTHR registers changes, signifying one of these
registers is half full. If the active power is positive, the watt-hour
register becomes half full when it increments from 0x3FFF FFFF to
0x4000 0000. If the active power is negative, the watt-hour
register becomes half full when it decrements from 0xC000
0000 to 0xBFFF FFFF. Similarly, Bit 1 (FAEHF) in STATUS0
register, is set when Bit 30 of one of the xFWATTHR registers
changes, signifying one of these registers is half full.
Setting Bits[1:0] in the MASK0 register enable the FAEHF and
AEHF interrupts, respectively. If enabled, the IRQ0 pin is set
low and the status bit is set to 1 whenever one of the energy
registers, xWATTHR (for the AEHF interrupt) or xFWATTHR
(for the FAEHF interrupt), become half full. The status bit is
cleared and the IRQ0 pin is set to logic high by writing to the
STATUS0 register with the corresponding bit set to 1.
Setting Bit 6 (RSTREAD) of the LCYCMODE register enables a
read-with-reset for all watt-hour accumulation registers, that is,
the registers are reset to 0 after a read operation.
Integration Time Under Steady Load
The discrete time sample period (T) for the accumulation register
is 125 µs (8 kHz frequency). With full-scale sinusoidal signals
on the analog inputs and the watt gain registers set to 0x00000, the
average word value from each LPF2 is PMAX = 33,516,139 =
0x1FF6A6B. If the WTHR register threshold is set at the PMAX
level, this means the DSP generates a pulse that is added at watt-
hour registers every 125 µs.
The maximum value that can be stored in the watt-hour
accumulation register before it overflows is 231 − 1 or
0x7FFFFFFF. The integration time is calculated as
Time = 0x7FFF,FFFF × 125 s = 74 hr 33 min 55 sec (27)
Energy Accumulation Modes
The active power accumulated in each watt-hour accumulation
32-bit register (AWATTHR, BWATTHR, CWATTHR,
AF WAT TH R, B FWAT T H R, a nd CF WAT T H R) d e pend s on t h e
configuration of Bit 5 and Bit 4 (CONSEL bits) in the ACCMODE
register. The various configurations are described in Table 15.
Table 15. Inputs to Watt-Hour Accumulation Registers
CONSEL AWATTHR BWATTHR CWATTHR
00 VA × IA VB × IB VC × IC
01 VA × IA 0 VC × IC
10 VA × IA VB × IB VC × IC
VB = −VA − VC
11 VA × IA VB × IB VC × IC
VB = −VA
Depending on the polyphase meter service, choose the appro-
priate formula to calculate the active energy. The American
ANSI C12.10 standard defines the different configurations of
the meter. Table 16 describes which mode to choose in these
various configurations.
Table 16. Meter Form Configuration
ANSI Meter Form Configuration CONSEL
5S/13S 3-wire delta 01
6S/14S 4-wire wye 10
8S/15S 4-wire delta 11
9S/16S 4-wire wye 00
Bits[1:0] (WATTACC[1:0]) in the ACCMODE register determine
how the CF frequency output can be generated as a function of
the total and fundamental active powers. Whereas the watt-hour
accumulation registers accumulate the active power in a signed
format, the frequency output can be generated in signed mode
or in absolute mode as a function of the WATTACC[1:0] bits.
See the Energy-to-Frequency Conversion section for details.
Line Cycle Active Energy Accumulation Mode
In line cycle energy accumulation mode, the energy accumula-
tion is synchronized to the voltage channel zero crossings such
that active energy is accumulated over an integral number of
half line cycles. The advantage of summing the active energy
over an integer number of line cycles is that the sinusoidal compo-
nent in the active energy is reduced to 0. This eliminates any
ripple in the energy calculation and allows the energy to be
accumulated accurately over a shorter time. By using the line
cycle energy accumulation mode, the energy calibration can be
greatly simplified, and the time required to calibrate the meter
can be significantly reduced. In line cycle energy accumulation
mode, the ADE7854/ADE7858/ADE7868/ADE7878 transfer the
active energy accumulated in the 32-bit internal accumulation
registers into the xWATHHR or xFWATTHR registers after an
integral number of line cycles, as shown in Figure 65. The
number of half line cycles is specified in the LINECYC register.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 49 of 96
ZERO-
CROSSING
DETECTION
(PHASE A)
ZERO-
CROSSING
DETECTION
(PHASE B)
CALIBRATION
CONTROL
ZERO-
CROSSING
DETECTION
(PHASE C)
LINECYC[15:0]
AWATTHR[31:0]
ZXSEL[0] IN
LCYCMODE[7:0]
ZXSEL[1] IN
LCYCMODE[7:0]
ZXSEL[2] IN
LCYCMODE[7:0]
OUTPUT
FROM LPF2
AWGAIN
AWATTOS
ACCUMUL ATOR
WTHR[47:0]
32-BIT
REGISTER
08510-147
Figure 65. Line Cycle Active Energy Accumulation Mode
The line cycle energy accumulation mode is activated by setting
Bit 0 (LWATT) in the LCYCMODE register. The energy accu-
mulation over an integer number of half line cycles is written
to the watt-hour accumulation registers after LINECYC number
of half line cycles is detected. When using the line cycle
accumulation mode, the Bit 6 (RSTREAD) of the LCYCMODE
register should be set to Logic 0 because the read with reset of
watt-hour registers is not available in this mode.
Phase A, Phase B, and Phase C zero crossings are, respectively,
included when counting the number of half line cycles by setting
Bits[5:3] (ZXSEL[x]) in the LCYCMODE register. Any combi-
nation of the zero crossings from all three phases can be used
for counting the zero crossing. Select only one phase at a time
for inclusion in the zero crossings count during calibration.
The number of zero crossings is specified by the LINECYC 16-bit
unsigned register. The ADE78xx can accumulate active power
for up to 65,535 combined zero crossings. Note that the internal
zero-crossing counter is always active. By setting Bit 0 (LWATT)
in the LCYCMODE register, the first energy accumulation
result is, therefore, incorrect. Writing to the LINECYC register
when the LWATT bit is set resets the zero-crossing counter, thus
ensuring that the first energy accumulation result is accurate.
At the end of an energy calibration cycle, Bit 5 (LENERGY) in
the STATUS0 register is set. If the corresponding mask bit in
the MASK0 interrupt mask register is enabled, the IRQ0 pin
also goes active low. The status bit is cleared and the IRQ0 pin is
set to high again by writing to the STATUS0 register with the
corresponding bit set to 1.
Because the active power is integrated on an integer number of
half-line cycles in this mode, the sinusoidal components are
reduced to 0, eliminating any ripple in the energy calculation.
Therefore, total energy accumulated using the line cycle
accumulation mode is
()
=
+==
1kkk
nTt
t
IVnTdttpe cos(φk – γk) (28)
where nT is the accumulation time.
Note that line cycle active energy accumulation uses the same
signal path as the active energy accumulation. The LSB size of
these two methods is equivalent.
REACTIVE POWER CALCULATION—ADE7858,
ADE7868, ADE7878 ONLY
The ADE7858/ADE7868/ADE7878 can compute the total reactive
power on every phase. Total reactive power integrates all funda-
mental and harmonic components of the voltages and currents.
The ADE7878 also computes the fundamental reactive power,
the power determined only by the fundamental components of
the voltages and currents.
A load that contains a reactive element (inductor or capacitor)
produces a phase difference between the applied ac voltage and
the resulting current. The power associated with reactive elements
is called reactive power, and its unit is VAR. Reactive power is
defined as the product of the voltage and current waveforms when
all harmonic components of one of these signals are phase
shifted by 90°.
Equation 31 gives an expression for the instantaneous reactive
power signal in an ac system when the phase of the current
channel is shifted by +90°.
=
=
1
2)(
k
k
Vtv sin(kωt + φk) (29)
()
k
kkγtωkIti +=
=
sin2)(
1
(30)
++=
=2
sin2)('
1
π
γtωkIti k
k
k
where iʹ(t) is the current waveform with all harmonic
components phase shifted by 90°.
Next, the instantaneous reactive power, q(t), can be expressed as
q(t) = v(t) × iʹ(t) (31)
=
×=
1
2)(
k
kk IVtq sin(kωt + φk) × sin(kωt + γk + 2
π) +
× 2sin(kωt + φk) × sin(mωt + γm +
=
mk
mk
m
kIV
1, 2
π)
Note that q(t) can be rewritten as
=
=
1
)(
k
kk IVtq {cos(φkγk 2
π) − cos(2 kωt + φk + γk + 2
π)} +
=
mk
mk
m
kIV
1, {cos[(km)ωt + φkγk 2
π]} (32)
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 50 of 96
The average total reactive power over an integral number of line
cycles (n) is given by the expression in Equation 33.
()
=
== nT
kkk IVdttq
nT
Q
01
1cos(φkγk 2
π) (33)
=
=
1k
kk IVQ sin(φkγk)
where:
T is the period of the line cycle.
Q is referred to as the total reactive power. Note that the total
reactive power is equal to the dc component of the instantaneous
reactive power signal q(t) in Equation 32, that is,
=1k
kk IV sin(φkγk)
This is the relationship used to calculate the total reactive power
in the ADE7858/ADE7868/ADE7878 for each phase. The
instantaneous reactive power signal, q(t), is generated by multi-
plying each harmonic of the voltage signals by the 90° phase-
shifted corresponding harmonic of the current in each phase.
The ADE7858/ADE7868/ADE7878 store the instantaneous
total phase reactive powers into the AVAR, BVAR, and CVAR
registers. Their expression is
=
××=
1kFS
k
FS
k
I
I
U
U
xVAR sin(φkγk) × PMAX × 4
2
1 (34)
where:
UFS, IFS are the rms values of the phase voltage and current when
the ADC inputs are at full scale.
PMAX = 33,516,139, the instantaneous power computed when
the ADC inputs are at full scale and in phase.
The xVAR waveform registers can be accessed using various
serial ports. Refer to the Waveform Sampling Mode section for
more details.
The expression of fundamental reactive power is obtained from
Equation 33 with k = 1, as follows:
FQ = V1I1 sin(φ1γ1)
The ADE7878 computes the fundamental reactive power using
a proprietary algorithm that requires some initialization function
of the frequency of the network and its nominal voltage measured
in the voltage channel. These initializations are introduced in
the Active Power Calculation section and are common for both
fundamental active and reactive powers.
Table 17 presents the settling time for the fundamental reactive
power measurement, which is the time it takes the power to
reflect the value at the input of the ADE7878.
Table 17. Settling Time for Fundamental Reactive Power
Input Signals
63% Full Scale 100% Full Scale
375 ms 875 ms
Reactive Power Gain Calibration
The average reactive power from the LPF output in each phase can
be scaled by ±100% by writing to one of the phases VAR gain 24-bit
register (AVARGAIN, BVARGAIN, CVARGAIN, AFVARGAIN,
BFVARGAIN, or CFVARGAIN). The xVARGAIN registers are
placed in each phase of the total reactive power datapath. The
xFVARGAIN registers are placed in each phase of the fundamental
reactive power datapath. The xVARGAIN registers are twos com-
plement signed registers and have a resolution of 2−23/LSB. The
function of the xVARGAIN registers is expressed by
+×
=
23
2
12 gisterRexVARGAIN
OutputLPF
PowerReactiveAverage
(35)
The output is scaled by –50% by writing 0xC00000 to the
xVARGAIN registers and increased by +50% by writing
0x400000 to them. These registers can be used to calibrate the
reactive power (or energy) gain in the ADE78xx for each phase.
As stated in the Current Waveform Gain Registers section, the
serial ports of the ADE7858/ADE7868/ADE7878 work on 32-,
16-, or 8-bit words and the DSP works on 28 bits. Similar to
registers presented in Figure 33, the AVARGAIN, BVARGAIN,
CVARGAIN, AFVARGAIN, BFVARGAIN, and CFVARGAIN
24-bit signed registers are accessed as 32-bit registers with the
four MSBs padded with 0s and sign extended to 28 bits.
Reactive Power Offset Calibration
The ADE7858/ADE7868/ADE7878 provide a reactive power
offset register on each phase and on each reactive power. AVAROS,
BVAROS, and CVAROS registers compensate the offsets in the
total reactive power calculations, whereas AFVAROS, BFVAROS,
and CFVAROS registers compensate offsets in the fundamental
reactive power calculations. These are signed twos complement,
24-bit registers that are used to remove offsets in the reactive
power calculations. An offset can exist in the power calculation
due to crosstalk between channels on the PCB or in the chip
itself. The offset resolution of the registers is the same as for the
active power offset registers (see the Active Power Offset
Calibration section).
As stated in the Current Waveform Gain Registers section, the
serial ports of the ADE7858/ADE7868/ADE7878 work on 32-,
16-, or 8-bit words and the DSP works on 28 bits. Similar to the
registers presented in Figure 33, the AVAROS, BVAROS, and
CVAROS 24-bit signed registers are accessed as 32-bit registers
with the four MSBs padded with 0s and sign extended to 28 bits.
Sign of Reactive Power Calculation
Note that the reactive power is a signed calculation. Table 18
summarizes the relationship between the phase difference between
the voltage and the current and the sign of the resulting reactive
power calculation.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 51 of 96
The ADE7858/ADE7868/ADE7878 have sign detection circuitry
for reactive power calculations that can monitor the total reactive
powers or the fundamental reactive powers. As described in the
Reactive Energy Calculation section, the reactive energy accu-
mulation is executed in two stages. Every time a sign change is
detected in the energy accumulation at the end of the first stage,
that is, after the energy accumulated into the internal accumulator
reaches the VARTHR register threshold, a dedicated interrupt is
triggered. The sign of each phase reactive power can be read in
the PHSIGN register. Bit 7 (REVRPSEL) in the ACCMODE
register sets the type of reactive power being monitored. When
REVRPSEL is 0, the default value, the total reactive power is
monitored. When REVRPSEL is 1, then the fundamental
reactive power is monitored.
Bits[12:10] (REVRPC, REVRPB, and REVRPA, respectively)
in the STATUS0 register are set when a sign change occurs in
the power selected by Bit 7 (REVRPSEL) in the ACCMODE
register.
Bits[6:4] (CVARSIGN, BVARSIGN, and AVARSIGN, respectively)
in the PHSIGN register are set simultaneously with the REVRPC,
REVRPB, and REVRPA bits. They indicate the sign of the reactive
power. When they are 0, the reactive power is positive. When
they are 1, the reactive power is negative.
Bit REVRPx of the STATUS0 register and Bit xVARSIGN in the
PHSIGN register refer to the reactive power of Phase x, the
power type being selected by Bit REVRPSEL in ACCMODE
register.
Setting Bits[12:10] in the MASK0 register enables the REVRPC,
REVRPB, and REVRPA interrupts, respectively. If enabled, the
IRQ0 pin is set low and the status bit is set to 1 whenever a change
of sign occurs. To find the phase that triggered the interrupt,
the PHSIGN register is read immediately after reading the
STATUS0 register. Next, the status bit is cleared and the IRQ0
pin is set to high by writing to the STATUS0 register with the
corresponding bit set to 1.
Table 18. Sign of Reactive Power Calculation
Φ1 Integrator Sign of Reactive Power
Between 0 to +180 Off Positive
Between −180 to 0 Off Negative
Between 0 to +180 On Positive
Between −180 to 0 On Negative
1 Φ is defined as the phase angle of the voltage signal minus the current
signal; that is, Φ is positive if the load is inductive and negative if the load is
capacitive.
Reactive Energy Calculation
Reactive energy is defined as the integral of reactive power.
Reactive Energy = ∫q(t)dt (36)
Both total and fundamental reactive energy accumulations are
always a signed operation. Negative energy is subtracted from
the reactive energy contents.
Similar to active power, the ADE7858/ADE7868/ADE7878
achieve the integration of the reactive power signal in two
stages (see Figure 66). The process is identical for both total and
fundamental active powers.
The first stage is conducted inside the DSP: every 125 µs
(8 kHz frequency), the instantaneous phase total reactive
or fundamental power is accumulated into an internal
register. When a threshold is reached, a pulse is generated
at the processor port and the threshold is subtracted from
the internal register. The sign of the energy in this moment
is considered the sign of the reactive power (see the Sign of
Reactive Power Calculation section for details).
The second stage is performed outside the DSP and consists
in accumulating the pulses generated by the processor into
internal 32-bit accumulation registers. The content of these
registers is transferred to the var-hour registers (xVARHR and
xFVARHR) when these registers are accessed. AVARHR,
BVARHR, CVARHR, AFWATTHR, BFWATTHR, and
CFWATTHR represent phase fundamental reactive powers.
Figure 63 from the Active Energy Calculation section explains
this process. The VARTHR 48-bit signed register contains the
threshold and it is introduced by the user. It is introduced by the
user and is common for both total and fundamental phase reactive
powers. Its value depends on how much energy is assigned to
one LSB of var-hour registers. Supposing a derivative of a volt
ampere reactive hour (varh) at [10n varh] where n is an integer,
is desired as one LSB of the VARHR register. Then, the VARTHR
register can be computed using the following equation:
FSFS
n
s
IU
fPMAX
VARTHR ×
×××
=103600
where:
PMAX = 33,516,139 = 0x1FF6A6B, the instantaneous power
computed when the ADC inputs are at full scale.
fS = 8 kHz, the frequency with which the DSP computes the
instantaneous power.
UFS, IFS are the rms values of phase voltages and currents when
the ADC inputs are at full scale.
The maximum value that may be written on the VARTHR
register is 247 − 1. The minimum value is 0x0, but it is
recommended to write a number equal to or greater than
PMAX. Never use negative numbers.
VARTHR is a 48-bit register. As previously stated in the Vo lt ag e
Waveform Gain Registers section, the serial ports of the ADE7858/
ADE7868/ADE7878 work on 32-, 16-, or 8-bit words. Similar to
the WTHR register shown in Figure 64, VARTHR is accessed as
two 32-bit registers (VARTHR1 and VARTHR0), each having eight
MSBs padded with 0s.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 52 of 96
This discrete time accumulation or summation is equivalent to
integration in continuous time following the expression in
Equation 37:
Bit 2 (REHF) in the STATUS0 register is set when Bit 30 of
one of the xVARHR registers changes, signifying one of these
registers is half full. If the reactive power is positive, the var-hour
register becomes half full when it increments from 0x3FFF FFFF
to 0x4000 0000. If the reactive power is negative, the var-hour
register becomes half full when it decrements from 0xC000 0000
to 0xBFFF FFFF. Analogously, Bit 3 (FREHF) in the STATUS0
register is set when Bit 30 of one of the xFVARHR registers
changes, signifying one of these registers is half full.
()
()
×==
=
0
0T
Lim
n
TnTqdttqergyReactiveEn (37)
where:
n is the discrete time sample number.
T is the sample period.
Setting Bits[3:2] in the MASK0 register enable the FREHF and
REHF interrupts, respectively. If enabled, the IRQ0 pin is set
low and the status bit is set to 1 whenever one of the energy
registers, xVARHR (for REHF interrupt) or xFVARHR (for
FREHF interrupt), becomes half full. The status bit is cleared
and the IRQ0 pin is set to high by writing to the STATUS0
register with the corresponding bit set to 1.
On the ADE7858/ADE7868/ADE7878, the total phase reactive
powers are accumulated in the AVARHR, BVARHR, and
CVARHR 32-bit signed registers. The fundamental phase reactive
powers are accumulated in the AFVARHR, BFVARHR, and
CFVARHR 32-bit signed registers. The reactive energy register
content can roll over to full-scale negative (0x80000000) and
continue increasing in value when the reactive power is positive.
Conversely, if the reactive power is negative, the energy register
underflows to full-scale positive (0x7FFFFFFF) and continues
to decrease in value.
Setting Bit 6 (RSTREAD) of the LCYCMODE register enables a
read-with-reset for all var-hour accumulation registers, that is,
the registers are reset to 0 after a read operation.
AVGAIN
HPFDIS
[23:0]
HPF
AIGAIN
HPFDIS
[23:0]
DIGITAL
INTEGRATOR
REVRPA BIT IN
STATUS0[31:0]
HPF
VA
IA
AVARGAIN
AVAROS
APHCAL ACCUMULATOR
VARTHR[47:0]
AVARHR[31:0]
32-BIT
REGISTER
TOTAL
REACTIVE
POWER
ALGORITHM
08510-245
DIGITAL SIGNAL PROCESSOR 2
4
AVAR
Figure 66. Total Reactive Energy Accumulation
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 53 of 96
Integration Time Under A Steady Load
The discrete time sample period (T) for the accumulation register
is 125 µs (8 kHz frequency). With full-scale pure sinusoidal signals
on the analog inputs and a 90° phase difference between the vol-
tage and the current signal (the largest possible reactive power),
the average word value representing the reactive power is PMAX =
33,516,139 = 0x1FF6A6B. If the VARTHR threshold is set at the
PMAX level, this means the DSP generates a pulse that is added
at the var-hour registers every 125 µs.
The maximum value that can be stored in the var-hour
accumulation register before it overflows is 231 − 1 or
0x7FFFFFFF. The integration time is calculated as
Time = 0x7FFF,FFFF × 125 s = 74 hr 33 min 55 sec (38)
Energy Accumulation Modes
The reactive power accumulated in each var-hour accumulation
32-bit register (AVARHR, BVARHR, CVARHR, AFVARHR,
BFVARHR, and CFVARHR) depends on the configuration of
Bits[5:4] (CONSEL[1:0]) in the ACCMODE register, in correlation
with the watt-hour registers. The different configurations are
described in Table 19. Note that IA/IB/IC are the phase-shifted
current waveforms.
Table 19. Inputs to Var-Hour Accumulation Registers
CONSEL[1:0]
AVARHR,
AFVARHR
BVARHR,
BFVARHR
CVARHR,
CFVARHR
00 VA × IA VB × IB’ VC × IC’
01 VA × IA 0 VC × IC’
10 VA × IA VB × IB’ VC × IC’
VB = −VA − VC
11 VA × IA VB × IB’ VC × IC’
VB = −VA
Bits[3:2] (VARACC[1:0]) in the ACCMODE register determine
how CF frequency output can be a generated function of the total
active and fundamental powers. While the var-hour accumulation
registers accumulate the reactive power in a signed format, the
frequency output can be generated in either the signed mode or the
sign adjusted mode function of VARACC[1:0]. See the Energy-to-
Frequency Conversion section for details.
Line Cycle Reactive Energy Accumulation Mode
As mentioned in the Line Cycle Active Energy Accumulation
Mode section, in line cycle energy accumulation mode, the
energy accumulation can be synchronized to the voltage
channel zero crossings so that reactive energy can be accu-
mulated over an integral number of half line cycles.
In this mode, the ADE7858/ADE7868/ADE7878 transfer the
reactive energy accumulated in the 32-bit internal accumulation
registers into the xVARHR or xFVARHR registers after an
integral number of line cycles, as shown in Figure 67. The
number of half line cycles is specified in the LINECYC register.
The line cycle reactive energy accumulation mode is activated by
setting Bit 1 (LVAR) in the LCYCMODE register. The total reactive
energy accumulated over an integer number of half line cycles
or zero crossings is available in the var-hour accumulation registers
after the number of zero crossings specified in the LINECYC reg-
ister is detected. When using the line cycle accumulation mode,
Bit 6 (RSTREAD) of the LCYCMODE register should be set to
Logic 0 because a read with the reset of var-hour registers is not
available in this mode.
ZERO-
CROSSING
DETECTION
(PHASE A)
ZERO-
CROSSING
DETECTION
(PHASE B)
CALIBRATION
CONTROL
ZERO-
CROSSING
DETECTION
(PHASE C)
LINECYC[15:0]
AVARHR[31:0]
ZXSEL[0] IN
LCYCMODE[7:0]
ZXSEL[1] IN
LCYCMODE[7:0]
ZXSEL[2] IN
LCYCMODE[7:0]
OUTPUT
FROM
TOTAL
REACTIVE
POWER
ALGORITHM
AVARGAI N
AVAROS
ACCUMUL ATOR
VARTHR[47:0]
32-BIT
REGISTER
08510-146
Figure 67. Line Cycle Reactive Energy Accumulation Mode
Phase A, Phase B, and Phase C zero crossings are, respectively,
included when counting the number of half line cycles by setting
Bits[5:3] (ZXSEL[x]) in the LCYCMODE register. Any combi-
nation of the zero crossings from all three phases can be used
for counting the zero crossing. Select only one phase at a time
for inclusion in the zero-crossings count during calibration.
For details on setting the LINECYC register and the Bit 5
(LENERGY) in the MASK0 interrupt mask register associated
with the line cycle accumulation mode, see the Line Cycle
Active Energy Accumulation Mode section.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 54 of 96
The ADE7854/ADE7858/ADE7868/ADE7878 store the instan-
taneous phase apparent powers into the AVA, BVA, and CVA
registers. Their expression is
APPARENT POWER CALCULATION
Apparent power is defined as the maximum power that can be
delivered to a load. One way to obtain the apparent power is by
multiplying the voltage rms value by the current rms value (also
called the arithmetic apparent power) 4
2
1
×××= PMAX
I
I
U
U
xVA
FSFS
(40)
S = V rms × I rms (39) where:
U, I are the rms values of the phase voltage and current.
where:
UFS, IFS are the rms values of the phase voltage and current when
the ADC inputs are at full scale.
S is the apparent power.
V rms and I rms are the rms voltage and current, respectively.
PMAX = 33,516,139, the instantaneous power computed when
the ADC inputs are at full scale and in phase.
The ADE7854/ADE7858/ADE7868/ADE7878 compute the
arithmetic apparent power on each phase. Figure 68 illustrates
the signal processing in each phase for the calculation of the
apparent power in the ADE78xx. Because V rms and I rms con-
tain all harmonic information, the apparent power computed by
the ADE78xx is total apparent power. The ADE7878 does not
compute fundamental apparent power because it does not measure
the rms values of the fundamental voltages and currents.
The xVA[23:0] waveform registers may be accessed using
various serial ports. Refer to the Waveform Sampling Mo de
section for more details.
The ADE7854/ADE7858/ADE7868/ADE7878 can compute the
apparent power in an alternative way by multiplying the phase
rms current by an rms voltage introduced externally. See the
Apparent Power Calculation Using VNOM section for details.
AVAHR[31:0]
AVAGAIN
DIGITAL SIGNAL PROCESSOR
AIRMS
ACCUMULATOR
AVRMS
VATHR[47:0]AVA
2
4
32-BIT REGISTER
08510-048
Figure 68. Apparent Power Data Flow and Apparent Energy Accumulation
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 55 of 96
Apparent Power Gain Calibration
The average apparent power result in each phase can be scaled
by ±100% by writing to one of the phases VAGAIN 24-bit registers
(AVAGAIN, BVAGAIN, or CVAG AIN). The VAGAIN registers
are twos complement, signed registers and have a resolution of
2−23/LSB. The function of the xVAGAIN registers is expressed
mathematically as
+××
=
23
2
1RegisterVAGAIN
rmsIrmsV
PowerApparentAverage
(41)
The output is scaled by –50% by writing 0xC00000 to the
xVAGAIN registers, and it is increased by +50% by writing
0x400000 to them. These registers calibrate the apparent power
(or energy) calculation in the ADE7854/ADE7858/ADE7868/
ADE7878 for each phase.
As previously stated in the Current Waveform Gain Registers
section, the serial ports of the ADE78xx work on 32-, 16-, or 8-bit
words and the DSP works on 28 bits. Similar to registers presented
in Figure 33, the AVAGAIN, BVAGAIN, and CVAGAIN 24-bit
registers are accessed as 32-bit registers with the four MSBs
padded with 0s and sign extended to 28 bits.
Apparent Power Offset Calibration
Each rms measurement includes an offset compensation register
to calibrate and eliminate the dc component in the rms value
(see the Root Mean Square Measurement section). The voltage
and current rms values are multiplied together in the apparent
power signal processing. As no additional offsets are created in
the multiplication of the rms values, there is no specific offset
compensation in the apparent power signal processing. The offset
compensation of the apparent power measurement in each phase is
accomplished by calibrating each individual rms measurement.
Apparent Power Calculation Using VNOM
The ADE7854/ADE7858/ADE7868/ADE7878 can compute the
apparent power by multiplying the phase rms current by an rms
voltage introduced externally in the VNOM 24-bit signed register.
When one of Bits[13:11] (VNOMCEN, VNOMBEN, or
VNOMAEN) in the COMPMODE register is set to 1, the
apparent power in the corresponding phase (Phase x for
VNOMxEN) is computed in this way. When the VNOMxEN
bits are cleared to 0, the default value, then the arithmetic
apparent power is computed.
The VNOM register contains a number determined by U, the
desired rms voltage, and , the rms value of the phase vol-
tage when the ADC inputs are at full scale:
FS
U
910,191,4×=
FS
U
U
VNOM (42)
where U is the nominal phase rms voltage.
As stated in the Current Waveform Gain Registers, the serial
ports of the ADE78xx work on 32-, 16-, or 8-bit words. Similar
to the register presented in Figure 34, the VNOM 24-bit signed
register is accessed as a 32-bit register with the eight MSBs
padded with 0s.
Apparent Energy Calculation
Apparent energy is defined as the integral of apparent power.
Apparent Energy = ∫s(t)dt (43)
Similar to active and reactive powers, the ADE7854/ADE7858/
ADE7868/ADE7878 achieve the integration of the apparent power
signal in two stages (see Figure 68). The first stage is conducted
inside the DSP: every 125 µs (8 kHz frequency), the instanta-
neous phase apparent power is accumulated into an internal
register. When a threshold is reached, a pulse is generated at the
processor port and the threshold is subtracted from the internal
register. The second stage is conducted outside the DSP and
consists of accumulating the pulses generated by the processor
into internal 32-bit accumulation registers. The content of these
registers is transferred to the VA-hour registers, xVAHR, when
these registers are accessed. Figure 63 from the Active Energy
Calculation section illustrates this process. The VATHR 48-bit
register contains the threshold. Its value depends on how much
energy is assigned to one LSB of the VA-hour registers. When a
derivative of apparent energy (VAh) of [10n VAh], where n is an
integer, is desired as one LSB of the xVAHR register; then, the
xVATHR register can be computed using the following equation:
FSFS
n
s
IU
fPMAX
VATHR ×
×××
=103600
where:
PMAX = 33,516,139 = 0x1FF6A6B, the instantaneous power
computed when the ADC inputs are at full scale.
fS = 8 kHz, the frequency with which the DSP computes the
instantaneous power.
UFS, IFS are the rms values of phase voltages and currents when
the ADC inputs are at full scale.
VATHR is a 48-bit register. As previously stated in the Current
Waveform Gain Registers section, the serial ports of the ADE7854/
ADE7858/ADE7868/ADE7878 work on 32-, 16-, or 8-bit words.
Similar to the WTHR register presented in Figure 64, the VATHR
register is accessed as two 32-bit registers (VATHR1 and VATHR0),
each having eight MSBs padded with 0s.
This discrete time accumulation or summation is equivalent to
integration in continuous time following the description in
Equation 44.
()
()
×==
=
0
0T
Lim
n
TnTsdttsergyApparentEn (44)
where:
n is the discrete time sample number.
T is the sample period.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 56 of 96
In the ADE7854/ADE7858/ADE7868/ADE7878, the phase
apparent powers are accumulated in the AVAHR, BVAHR, and
CVAHR 32-bit signed registers. The apparent energy register
content can roll over to full-scale negative (0x80000000) and
continue increasing in value when the apparent power is posi-
tive. Conversely, if because of offset compensation in the rms
datapath, the apparent power is negative, the energy register
underflows to full-scale positive (0x7FFFFFFF) and continues
to decrease in value.
The maximum value that can be stored in the xVAHR
accumulation register before it overflows is 231 − 1 or
0x7FFFFFFF. The integration time is calculated as
Time = 0x7FFF,FFFF × 125 s = 74 hr 33 min 55 sec (45)
Energy Accumulation Mode
The apparent power accumulated in each accumulation register
depends on the configuration of Bits[5:4] (CONSEL[1:0]) in the
ACCMODE register. The various configurations are described
in Table 20.
Bit 4 (VAEHF) in the STATUS0 register is set when Bit 30 of one of
the xVAHR registers changes, signifying one of these registers is
half full. As the apparent power is always positive and the xVAHR
registers are signed, the VA-hour registers become half full when
they increment from 0x3FFFFFFF to 0x4000 0000. Interrupts
attached to Bit VAEHF in the STATUS0 register can be enabled by
setting Bit 4 in the MASK0 register. If enabled, the IRQ0 pin is set
low and the status bit is set to 1 whenever one of the Energy
Registers xVAHR becomes half full. The status bit is cleared and
the IRQ0 pin is set to high by writing to the STATUS0 register
with the corresponding bit set to 1.
Table 20. Inputs to VA-Hour Accumulation Registers
CONSEL[1:0] AVAHR BVAHR CVAHR
00 AVRMS × AIRMS BVRMS × BIRMS CVRMS × CIRMS
01 AVRMS × AIRMS 0 CVRMS × CIRMS
10 AVRMS × AIRMS BVRMS × BIRMS CVRMS × CIRMS
VB = −VA − VC
11 AVRMS × AIRMS BVRMS × BIRMS CVRMS × CIRMS
VB = −VA
Line Cycle Apparent Energy Accumulation Mode
As described in the Line Cycle Active Energy Accumulation
Mode section, in line cycle energy accumulation mode, the
energy accumulation can be synchronized to the voltage channel
zero crossings allowing apparent energy to be accumulated over an
integral number of half line cycles. In this mode, the ADE7854/
ADE7858/ADE7868/ADE7878 transfer the apparent energy
accumulated in the 32-bit internal accumulation registers into
the xVAHR registers after an integral number of line cycles, as
shown in Figure 69. The number of half line cycles is specified
in the LINECYC register.
Setting Bit 6 (RSTREAD) of the LCYCMODE register enables
a read-with-reset for all xVAHR accumulation registers, that is,
the registers are reset to 0 after a read operation.
Integration Time Under Steady Load
The discrete time sample period for the accumulation register is
125 µs (8 kHz frequency). With full-scale pure sinusoidal signals
on the analog inputs, the average word value representing the
apparent power is PMAX. If the VATHR threshold register is set
at the PMAX level, this means the DSP generates a pulse that
is added at the xVAHR registers every 125 µs.
ZERO-
CROSSING
DETECTION
(PHASE A)
ZERO-
CROSSING
DETECTION
(PHASE B)
CALIBRATION
CONTROL
ZERO-
CROSSING
DETECTION
(PHASE C)
LINECYC[15:0]
AVAHR[31:0]
ZXSEL[0] IN
LCYCMODE[7:0]
ZXSEL[1] IN
LCYCMODE[7:0]
ZXSEL[2] IN
LCYCMODE[7:0]
AVAGAINAIRMS
AVRMS
ACCUMUL ATOR
VAHR[47:0]
32-BIT
REGISTER
08510-049
Figure 69. Line Cycle Apparent Energy Accumulation Mode
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 57 of 96
The line cycle apparent energy accumulation mode is activated
by setting Bit 2 (LVA) in the LCYCMODE register. The apparent
energy accumulated over an integer number of zero crossings is
written to the xVAHR accumulation registers after the number
of zero crossings specified in LINECYC register is detected. When
using the line cycle accumulation mode, set Bit 6 (RSTREAD) of
the LCYCMODE register to Logic 0 because a read with the reset
of xVAHR registers is not available in this mode.
Phase A, Phase B, and Phase C zero crossings are, respectively,
included when counting the number of half line cycles by setting
Bits[5:3] (ZXSEL[x]) in the LCYCMODE register. Any combi-
nation of the zero crossings from all three phases can be used
for counting the zero crossing. Select only one phase at a time
for inclusion in the zero-crossings count during calibration.
For details on setting the LINECYC register and Bit 5 (LENERGY)
in the MASK0 interrupt mask register associated with the line
cycle accumulation mode, see the Line Cycle Active Energy
Accumulation Mode section.
WAVEFORM SAMPLING MODE
The waveform samples of the current and voltage waveform,
the active, reactive, and apparent power outputs are stored
every 125 µs (8 kHz rate) into 24-bit signed registers that can be
accessed through various serial ports of the ADE7854/ADE7858/
ADE7868/ADE7878. Table 21 provides a list of registers and their
descriptions.
Table 21. Waveform Registers List
Register Description
IAWV Phase A current
VAWV Phase A voltage
IBWV Phase B current
VBWV Phase B voltage
ICWV Phase C current
VCWV Phase C voltage
INWV Neutral current, available in the ADE7868
and ADE7878 only
AVA Phase A apparent power
BVA Phase B apparent power
CVA Phase C apparent power
AWATT Phase A active power
BWATT Phase B active power
CWATT Phase C active power
AVAR Phase A reactive power
BVAR Phase B reactive power
CVAR Phase C reactive power
Bit 17 (DREADY) in the STATUS0 register can be used to
signal when the registers listed in Table 21 can be read using
I2C or SPI serial ports. An interrupt attached to the flag can be
enabled by setting Bit 17 (DREADY) in the MASK0 register.
See the Digital Signal Processor section for more details on
Bit DREADY.
The ADE7854/ADE7858/ADE7868/ADE7878 contain a high
speed data capture (HSDC) port that is specially designed to
provide fast access to the waveform sample registers. Read the
HSDC Interface section for more details.
As stated in the Current Waveform Gain Registers section, the
serial ports of the ADE7854/ADE7858/ADE7868/ADE7878
work on 32-, 16-, or 8-bit words. All registers listed in Table 21
are transmitted signed extended from 24 bits to 32 bits (see
Figure 35).
ENERGY-TO-FREQUENCY CONVERSION
The ADE7854/ADE7858/ADE7868/ADE7878 provide three
frequency output pins: CF1, CF2, and CF3. The CF3 pin is
multiplexed with the HSCLK pin of the HSDC interface. When
HSDC is enabled, the CF3 functionality is disabled at the pin.
CF1 and CF2 pins are always available. After initial calibration
at manufacturing, the manufacturer or end customer verifies
the energy meter calibration. One convenient way to verify the
meter calibration is to provide an output frequency propor-
tional to the active, reactive, or apparent powers under steady
load conditions. This output frequency can provide a simple,
single-wire, optically isolated interface to external calibration
equipment. Figure 70 illustrates the energy-to-frequency
conversion in the ADE7854/ADE7858/ADE7868/ADE7878.
The DSP computes the instantaneous values of all phase powers:
total active, fundamental active, total reactive, fundamental
reactive, and apparent. The process in which the energy is sign
accumulated in various xWATTHR, xVARHR, and xVAHR
registers has already been described in the energy calculation
sections: Active Energy Calculation, Reactive Energy Calculation,
and Apparent Energy Calculation. In the energy-to-frequency
conversion process, the instantaneous powers generate signals
at the frequency output pins (CF1, CF2, and CF3). One digital-
to-frequency converter is used for every CFx pin. Every converter
sums certain phase powers and generates a signal proportional
to the sum. Two sets of bits decide what powers are converted.
First, Bits[2:0] (TERMSEL1[2:0]), Bits[5:3] (TERMSEL2[2:0]),
and Bits[8:6] (TERMSEL3[2:0]) of the COMPMODE register
decide which phases, or which combination of phases, are added.
The TERMSEL1 bits refer to the CF1 pin, the TERMSEL2 bits
refer to the CF2 pin, and the TERMSEL3 bits refer to the CF3
pin. The TERMSELx[0] bits manage Phase A. When set to 1,
Phase A power is included in the sum of powers at the CFx
converter. When cleared to 0, Phase A power is not included.
The TERMSELx[1] bits manage Phase B, and the TERMSELx[2]
bits manage Phase C. Setting all TERMSELx bits to 1 means all
3-phase powers are added at the CFx converter. Clearing all
TERMSELx bits to 0 means no phase power is added and no
CF pulse is generated.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 58 of 96
Second, Bits[2:0] (CF1SEL[2:0]), Bits[5:3] (CF2SEL[2:0]), and
Bits[8:6] (CF3SEL[2:0]) in the CFMODE register decide what
type of power is used at the inputs of the CF1, CF2, and CF3
converters, respectively. Table 22 shows the values that CFxSEL
can have: total active, total reactive (available in the ADE7858,
ADE7868, and ADE7878 only), apparent, fundamental active
(available in the ADE7878 only), or fundamental reactive
(available in the ADE7878 only) powers.
Table 22. CFxSEL Bits Description
CFxSEL Description
Registers
Latched When
CFxLATCH = 1
000 CFx signal proportional to the
sum of total phase active
powers
AWATTHR,
BWATTHR,
CWATTHR
001 CFx signal proportional to the
sum of total phase reactive
powers (ADE7858/ADE7868/
ADE7878)
AVARHR, BVARHR,
CVARHR
010 CFx signal proportional to the
sum of phase apparent powers
AVAHR, BVAHR,
CVAHR
011 CFx signal proportional to the
sum of fundamental phase
active powers (ADE7878 only)
AFWATTHR,
BFWATTHR,
CFWATTHR
100 CFx signal proportional to the
sum of fundamental phase
reactive powers (ADE7878 only)
AFVARHR,
BFVARHR,
CFVARHR
101 to
111
Reserved
08510-050
DIGITAL SIGNAL
PROCESSOR
INSTANTANEOUS
PHASE A ACTIVE
POWER
INSTANTANEOUS
PHASE B ACTIVE
POWER
INSTANTANEOUS
PHASE C ACTIVE
POWER
TERMSELx BITS IN
COMPMODE
WTHR[47:0]
FREQUENCY
DIVIDER
2
7
2
7
VA
VAR
FWATT
1
FVAR
1
WATT
CFxSEL BITS
IN CFMODE
REVPSUMx BIT OF
STATUS0[31:0]
CFxDEN
ACCUMULATOR
CFx PULSE
OUTPUT
1
FWATT AND FVAR FOR ADE7878 ONLY.
Figure 70. Energy-to-Frequency Conversion
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 59 of 96
By default, the TERMSELx bits are all 1 and the CF1SEL bits are
000, the CF2SEL bits are 001, and the CF3SEL bits are 010. This
means that by default, the CF1 digital-to-frequency converter
produces signals proportional to the sum of all 3-phase total
active powers, CF2 produces signals proportional to total
reactive powers, and CF3 produces signals proportional to
apparent powers.
Similar to the energy accumulation process, the energy-to-
frequency conversion is accomplished in two stages. In the first
stage, the instantaneous phase powers obtained from the DSP at
the 8 kHz rate are shifted left by seven bits and then accumulate
into an internal register at a 1 MHz rate. When a threshold is
reached, a pulse is generated and the threshold is subtracted
from the internal register. The sign of the energy in this moment
is considered the sign of the sum of phase powers (see the Sign
of Sum-of-Phase Powers in the CFx Datapath section for details).
The threshold is the same threshold used in various active,
reactive, and apparent energy accumulators in the DSP, such
as the WTHR, VARTHR, or VATHR registers, except for being
shifted left by seven bits. The advantage of accumulating the
instantaneous powers at the 1 MHz rate is that the ripple at the
CFx pins is greatly diminished.
The second stage consists of the frequency divider by the
CFxDEN 16-bit unsigned registers. The values of CFxDEN
depend on the meter constant (MC), measured in impulses/kWh
and how much energy is assigned to one LSB of various energy
registers: xWATTHR, xVARHR, and so forth. Supposing a deri-
vative of wh [10n wh], n a positive or negative integer, is desired
as one LSB of xWATTHR register. Then, CFxDEN is as follows:
n
MC
CFxDEN 10]imp/kwh[
103
×
= (46)
The derivative of wh must be chosen in such a way to obtain a
CFxDEN register content greater than 1. If CFxDEN = 1, then
the CFx pin stays active low for only 1 µs, therefore, avoid this
number. The frequency converter cannot accommodate fractional
results; the result of the division must be rounded to the nearest
integer. If CFxDEN is set equal to 0, then the ADE78xx considers it
to be equal to 1.
The pulse output for all digital-to-frequency converters stays
low for 80 ms if the pulse period is larger than 160 ms (6.25 Hz). If
the pulse period is smaller than 160 ms and CFxDEN is an even
number, the duty cycle of the pulse output is exactly 50%. If the
pulse period is smaller than 160 ms and CFxDEN is an odd
number, the duty cycle of the pulse output is
(1+1/CFxDEN) × 50%
The pulse output is active low and preferably connected to an
LED, as shown in Figure 71.
V
DD
CFx PIN
08510-051
Figure 71. CFx Pin Recommended Connection
Bits[11:9] (CF3DIS, CF2DIS, and CF1DIS) of the CFMODE
register decide if the frequency converter output is generated
at the CF3, CF2, or CF1 pin. When Bit CFxDIS is set to 1 (the
default value), the CFx pin is disabled and the pin stays high.
When Bit CFxDIS is cleared to 0, the corresponding CFx pin
output generates an active low signal.
Bits[16:14] (CF3, CF2, CF1) in the Interrupt Mask Register MASK0
manage the CF3, CF2, and CF1 related interrupts. When the
CFx bits are set, whenever a high-to-low transition at the corres-
ponding frequency converter output occurs, an interrupt IRQ0
is triggered and a status bit in the STATUS0 register is set to 1.
The interrupt is available even if the CFx output is not enabled
by the CFxDIS bits in the CFMODE register.
Synchronizing Energy Registers with CFx Outputs
The ADE7854/ADE7858/ADE7868/ADE7878 contain a feature
that allows synchronizing the content of phase energy accu-
mulation registers with the generation of a CFx pulse. When
a high-to-low transition at one frequency converter output
occurs, the content of all internal phase energy registers that
relate to the power being output at CFx pin is latched into hour
registers and then resets to 0. See Table 22 for the list of registers
that are latched based on the CFxSEL[2:0] bits in the CFMODE
register. All 3-phase registers are latched independent of the
TERMSELx bits of the COMPMODE register. The process is
shown in Figure 72 for CF1SEL[2:0] = 010 (apparent powers
contribute at the CF1 pin) and CFCYC = 2.
The CFCYC 8-bit unsigned register contains the number of high to
low transitions at the frequency converter output between two
consecutive latches. Avoid writing a new value into the CFCYC
register during a high-to-low transition at any CFx pin.
CFCYC = 2
AVAHR, BVAHR,
CVAHR LATCHED
ENERGY REGISTERS
RESET
AVAHR, BVAHR,
CVAHR LATCHED
ENERGY REGISTERS
RESET
CF1 PULSE
BASED ON
PHASE A AND
PHASE B
APPARENT
POWERS
08510-052
Figure 72. Synchronizing AVAHR and BVAHR with CF1
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 60 of 96
Bits[14:12] (CF3LATCH, CF2LATCH, and CF1LATCH) of the
CFMODE register enable this process when set to 1. When
cleared to 0, the default state, no latch occurs. The process is
available even if the CFx output is not enabled by the CFxDIS
bits in the CFMODE register.
CF Outputs for Various Accumulation Modes
Bits[1:0] (WATTACC[1:0]) in the ACCMODE register deter-
mine the accumulation modes of the total active and fundamental
powers when signals proportional to the active powers are chosen
at the CFx pins (the CFxSEL[2:0] bits in the CFMODE register
equal 000 or 011). When WATTACC[1:0] = 00 (the default value),
the active powers are sign accumulated before entering the energy-
to-frequency converter. Figure 73 shows how signed active power
accumulation works. In this mode, the CFx pulses synchronize
perfectly with the active energy accumulated in xWATTHR regis-
ters because the powers are sign accumulated in both data paths.
NEGPOSPOS
APNOLOAD
SIGN = POSITIVE
NEG
NO-LOAD
THRESHOLD
ACTIVE POWER
NO-LOAD
THRESHOLD
A
CTIVE ENER
G
Y
REVAPx BIT
IN STATUS0
xWSIGN BIT
IN PHSIGN
08510-053
Figure 73. Active Power Signed Accumulation Mode
When WATTACC[1:0] = 11, the active powers are accumulated
in absolute mode. When the powers are negative, they change
sign and accumulate together with the positive power. Figure 74
shows how absolute active power accumulation works. Note
that in this mode, the xWATTHR registers continue to accumulate
active powers in signed mode, even if the CFx pulses are gener-
ated based on the absolute accumulation mode.
Bits[3:2] (VARACC[1:0]) in the ACCMODE register determine the
accumulation modes of the total and fundamental reactive powers
when signals proportional to the reactive powers are chosen at the
CFx pins (the CFxSEL[2:0] bits in the CFMODE register equal
001 or 100). When VARACC[1:0] = 00, the default value, the
reactive powers are sign accumulated before entering the
energy-to-frequency converter. Figure 75 shows how signed
reactive power accumulation works. In this mode, the CFx
pulses synchronize perfectly with the reactive energy accumu
lated in the xVARHR registers because the powers are sign
accumulated in both datapaths.
NEGPOSPOS
APNOLOAD
SIGN = POSITIVE
NEG
NO-LOAD
THRESHOLD
ACTIVE POWER
NO-LOAD
THRESHOLD
A
CTIVE ENERGY
REVAPx BIT
IN STATUS0
xWSIGN BIT
IN PHSIGN
08510-054
Figure 74. Active Power Absolute Accumulation Mode
NEGPOSPOS
VARNOLOAD
SIGN = POSITIVE
NEG
NO-LOAD
THRESHOLD
REACTIVE
POWER
NO-LOAD
THRESHOLD
REACTIVE
ENERGY
REVRPx BIT
IN STATUS0
x
VARSIGN BIT
IN PHSIGN
08510-153
Figure 75. Reactive Power Signed Accumulation Mode
When VARACC[1:0] = 10, the reactive powers are accumulated
depending on the sign of the corresponding active power. If the
active power is positive, the reactive power is accumulated as is.
If the active power is negative, the sign of the reactive power is
changed for accumulation. Figure 76 shows how the sign adjusted
reactive power accumulation mode works. In this mode, the
xVARHR registers continue to accumulate reactive powers in
signed mode, even if the CFx pulses are generated based on the
sign adjusted accumulation mode.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 61 of 96
POSPOS
VARNOLOAD
SIGN = POSITIVE
NEG
NO-LOAD
THRESHOLD
NO-LOAD
THRESHOLD
NO-LOAD
THRESHOLD
REACTIVE
POWER
ACTIVE
POWER
REACTIVE
ENERGY
REVRPx BIT
IN STATUS0
x
VARSIGN BIT
IN PHSIGN
08510-155
Figure 76. Reactive Power Accumulation in Sign Adjusted Mode
Sign of Sum-of-Phase Powers in the CFx Datapath
The ADE7854/ADE7858/ADE7868/ADE7878 have sign detection
circuitry for the sum of phase powers that are used in the CFx
datapath. As seen in the beginning of the Energy-to-Frequency
Conversion section, the energy accumulation in the CFx data-
path is executed in two stages. Every time a sign change is detected
in the energy accumulation at the end of the first stage, that is,
after the energy accumulated into the accumulator reaches one
of the WTHR, VARTHR, or VATHR thresholds, a dedicated
interrupt can be triggered synchronously with the corresponding
CFx pulse. The sign of each sum can be read in the PHSIGN
register.
Bit 18, Bit 13, and Bit 9 (REVPSUM3, REVPSUM2, and
REVPSUM1, respectively) of the STATUS0 register are set
to 1 when a sign change of the sum of powers in CF3, CF2,
or CF1 datapaths occurs. To correlate these events with the
pulses generated at the CFx pins, after a sign change occurs,
Bit REVPSUM3, Bit REVPSUM2, and Bit REVPSUM1 are set
in the same moment in which a high-to-low transition at the
CF3, CF2, and CF1 pin, respectively, occurs.
Bit 8, Bit 7, and Bit 3 (SUM3SIGN, SUM2SIGN, and SUM1SIGN,
respectively) of the PHSIGN register are set in the same moment
with Bit REVPSUM3, Bit REVPSUM2, and Bit EVPSUM1 and
indicate the sign of the sum of phase powers. When cleared to
0, the sum is positive. When set to 1, the sum is negative.
Interrupts attached to Bit 18, Bit 13, and Bit 9 (REVPSUM3,
REVPSUM2, and REVPSUM1, respectively) in the STATUS0
register are enabled by setting Bit 18, Bit 13, and Bit 9 in the
MASK0 register. If enabled, the IRQ0 pin is set low, and the
status bit is set to 1 whenever a change of sign occurs. To find
the phase that triggered the interrupt, the PHSIGN register is
read immediately after reading the STATUS0 register. Next, the
status bit is cleared, and the IRQ0 pin is set high again by writing
to the STATUS0 register with the corresponding bit set to 1.
NO LOAD CONDITION
The no load condition is defined in metering equipment standards
as occurring when the voltage is applied to the meter and no cur-
rent flows in the current circuit. To eliminate any creep effects in
the meter, the ADE7854/ADE7858/ADE7868/ADE7878 contain
three separate no load detection circuits: one related to the total
active and reactive powers (ADE7858/ADE7868/ADE7878
only), one related to the fundamental active and reactive powers
(ADE7878 only), and one related to the apparent powers.
No Load Detection Based On Total Active, Reactive
Powers
This no load condition is triggered when the absolute values of
both phase total active and reactive powers are less than or equal
to positive thresholds indicated in the respective APNOLOAD
and VARNOLOAD signed 24-bit registers. In this case, the total
active and reactive energies of that phase are not accumulated
and no CFx pulses are generated based on these energies. The
APNOLOAD register represents the positive no load level of
active power relative to PMAX, the maximum active power
obtained when full-scale voltages and currents are provided at
ADC inputs. The VARNOLOAD register represents the positive
no load level of reactive power relative to PMAX. The expres-
sion used to compute APNOLOAD signed 24-bit value is
PMAX
I
I
U
U
APNOLOAD
FS
NOLOAD
FS
n××= (47)
where:
PMAX = 33,516,139 = 0x1FF6A6B, the instantaneous power
computed when the ADC inputs are at full scale.
UFS, IFS are the rms values of phase voltages and currents when
the ADC inputs are at full scale.
Un is the nominal rms value of phase voltage.
INOLOAD is the minimum rms value of phase current the meter
starts measuring.
The VARNOLOAD register usually contains the same value as
the APNOLOAD register. When APNOLOAD and VARNOLOAD
are set to negative values, the no load detection circuit is disabled.
Note that the ADE7854 measures only the total active powers.
To ensure good functionality of the ADE7854 no-load circuit,
set the VARNOLOAD register at 0x800000.
As previously stated in the Current Waveform Gain Registers
section, the serial ports of the ADE78xx work on 32-, 16-, or
8-bit words and the DSP works on 28 bits. APNOLOAD and
VARNOLOAD 24-bit signed registers are accessed as 32-bit
registers with the four MSBs padded with 0s and sign extended
to 28 bits. See Figure 33 for details.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 62 of 96
Bit 0 (NLOAD) in the STATUS1 register is set when this no
load condition in one of the three phases is triggered. Bits[2:0]
(NLPHASE[2:0]) in the PHNOLOAD register indicate the state
of all phases relative to a no load condition and are set simulta-
neously with Bit NLOAD in the STATUS1 register. NLPHASE[0]
indicates the state of Phase A, NLPHASE[1] indicates the state
of Phase B, and NLPHASE[2] indicates the state of Phase C.
When Bit NLPHASE[x] is cleared to 0, it means the phase is out
of a no load condition. When set to 1, it means the phase is in a
no load condition.
An interrupt attached to Bit 0 (NLOAD) in the STATUS1
register can be enabled by setting Bit 0 in the MASK1 register.
If enabled, the IRQ1 pin is set to low, and the status bit is set
to 1 whenever one of three phases enters or exits this no load
condition. To find the phase that triggered the interrupt, the
PHNOLOAD register is read immediately after reading the
STATUS1 register. Next, the status bit is cleared, and the IRQ1
pin is set to high by writing to the STATUS1 register with the
corresponding bit set to 1.
No Load Detection Based on Fundamental Active and
Reactive Powers—ADE7878 Only
This no load condition (available on the ADE7878 only) is
triggered when the absolute values of both phase fundamental
active and reactive powers are less than or equal to the respective
APNOLOAD and VARNOLOAD positive thresholds. In this
case, the fundamental active and reactive energies of that phase
are not accumulated, and no CFx pulses are generated based on
these energies. APNOLOAD and VARNOLOAD are the same
no load thresholds set for the total active and reactive powers.
When APNOLOAD and VARNOLOAD are set to negative
values, this no load detection circuit is disabled.
Bit 1 (FNLOAD) in the STATUS1 register is set when this no
load condition in one of the three phases is triggered. Bits[5:3]
(FNLPHASE[2:0]) in the PHNOLOAD register indicate the
state of all phases relative to a no load condition and are set
simultaneously with Bit FNLOAD in the STATUS1 register.
FNLPHASE[0] indicates the state of Phase A, FNLPHASE[1]
indicates the state of Phase B, and FNLPHASE[2] indicates the
state of Phase C. When Bit FNLPHASE[x] is cleared to 0, it
means the phase is out of the no load condition. When set to 1,
it means the phase is in a no load condition.
An interrupt attached to the Bit 1 (FNLOAD) in the STATUS1
register can be enabled by setting Bit 1 in the MASK1 register. If
enabled, the IRQ1 pin is set low and the status bit is set to 1
whenever one of three phases enters or exits this no load
condition. To find the phase that triggered the interrupt, the
PHNOLOAD register is read immediately after reading the
STATUS1 register. Then the status bit is cleared and the IRQ1
pin is set back high by writing to the STATUS1 register with the
corresponding bit set to 1.
No Load Detection Based on Apparent Power
This no load condition is triggered when the absolute value
of phase apparent power is less than or equal to the threshold
indicated in the VANOLOAD 24-bit signed register. In this
case, the apparent energy of that phase is not accumulated
and no CFx pulses are generated based on this energy. The
VANOLOAD register represents the positive no load level
of apparent power relative to PMAX, the maximum apparent
power obtained when full-scale voltages and currents are
provided at the ADC inputs. The expression used to compute
the VANOLOAD signed 24-bit value is
PMAX
I
I
U
U
VANOLOAD
FS
NOLOAD
FS
n××= (48)
where:
PMAX = 33,516,139 = 0x1FF6A6B, the instantaneous apparent
power computed when the ADC inputs are at full scale.
UFS, IFS are the rms values of phase voltages and currents when
the ADC inputs are at full scale.
Un is the nominal rms value of phase voltage.
INOLOAD is the minimum rms value of phase current the meter
starts measuring.
When the VANOLOAD register is set to negative values, the no
load detection circuit is disabled.
As stated in the Current Waveform Gain Registers section, the
serial ports of the ADE7854/ADE7858/ADE7868/ADE7878
work on 32-, 16-, or 8-bit words and the DSP works on 28 bits.
Similar to the registers presented in Figure 33, the VANOLOAD
24-bit signed register is accessed as a 32-bit register with the
four MSBs padded with 0s and sign extended to 28 bits.
Bit 2 (VANLOAD) in the STATUS1 register is set when this no
load condition in one of the three phases is triggered. Bits[8:6]
(VANLPHASE[2:0]) in the PHNOLOAD register indicate the
state of all phases relative to a no load condition and they are set
simultaneously with Bit VANLOAD in the STATUS1 register:
Bit VANLPHASE[0] indicates the state of Phase A.
Bit VANLPHASE[1] indicates the state of Phase B.
Bit VANLPHASE[2] indicates the state of Phase C.
When Bit VANLPHASE[x] is cleared to 0, it means the phase is
out of no load condition. When set to 1, it means the phase is in
no load condition.
An interrupt attached to Bit 2 (VANLOAD) in the STATUS1
register is enabled by setting Bit 2 in the MASK1 register. If
enabled, the IRQ1 pin is set low and the status bit is set to 1
whenever one of three phases enters or exits this no load
condition. To find the phase that triggered the interrupt, the
PHNOLOAD register is read immediately after reading the
STATUS1 register. Next, the status bit is cleared, and the IRQ1
pin is set to high by writing to the STATUS1 register with the
corresponding bit set to 1.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 63 of 96
CHECKSUM REGISTER bi(0) = 1, i = 0, 1, 2, …, 31, the initial state of the bits that form
the CRC. Bit b0 is the least significant bit, and Bit b31 is the most
significant.
The ADE7854/ADE7858/ADE7868/ADE7878 have a checksum
32-bit register, CHECKSUM, that ensures certain very important
configuration registers maintain their desired value during
Normal Power Mode PSM0.
gi, i = 0, 1, 2, …, 31 are the coefficients of the generating
polynomial defined by the IEEE802.3 standard as follows:
G(x) = x32 + x26 + x23 + x22 + x16 + x12 + x11 + x10 + x8 + x7 +
x5 + x4 + x2 + x + 1 (49)
The registers covered by this register are MASK0, MASK1,
COMPMODE, gain, CFMODE, CF1DEN, CF2DEN, CF3DEN,
CONFIG, MMODE, ACCMODE, LCYCMODE, HSDC_CFG,
and another six 8-bit reserved internal registers that always have
default values. The ADE78xx computes the cyclic redundancy
check (CRC) based on the IEEE802.3 standard. The registers
are introduced one-by-one into a linear feedback shift register
(LFSR) based generator starting with the least significant bit (as
shown in Figure 77). The 32-bit result is written in the
CHECKSUM register. After power-up or a hardware/software
reset, the CRC is computed on the default values of the registers
giving the results presented in the Table 23.
g0 = g1 = g2 = g4 = g5 = g7 = 1
g8 = g10 = g11 = g12 = g16 = g22 = g26 = g31 = 1 (50)
All of the other gi coefficients are equal to 0.
FB(j) = aj – 1 XOR b31(j – 1) (51)
b0(j) = FB(j) AND g0 (52)
bi(j) = FB(j) AND gi XOR bi − 1(j – 1), i = 1, 2, 3, ..., 31 (53)
Equation 51, Equation 52, and Equation 53 must be repeated for
j = 1, 2, …, 256. The value written into the CHECKSUM register
contains the Bit bi(256), i = 0, 1, …, 31. The value of the CRC,
after the bits from the reserved internal register have passed
through LFSR, is obtained at Step j = 48 and is presented in the
Table 23.
Table 23. Default Values of CHECKSUM and of Internal
Registers CRC
Part No.
Default Value of
CHECKSUM
CRC of Internal
Registers
Two different approaches can be followed in using the CHECK-
SUM register. One is to compute the CRC based on the relations
(47) to (51) and then compare the value against the CHECKSUM
register. Another is to periodically read the CHECKSUM register.
If two consecutive readings differ, it can be assumed that one of
the registers has changed value and therefore, the ADE7854,
ADE7858, ADE7868, or ADE7878 has changed configuration.
The recommended response is to initiate a hardware/software
reset that sets the values of all registers to the default, including
the reserved ones, and then reinitialize the configuration registers.
ADE7854 0x2689B124 0x3A7ABC72
ADE7858 0xD6744F93 0x3E7D0FC1
ADE7868 0x93D774E6 0x23F7C7B1
ADE7878 0x33666787 0x2D32A389
Figure 78 shows how the LFSR works. The MASK0, MASK1,
COMPMODE, gain, CFMODE, CF1DEN, CF2DEN, CF3DEN,
CONFIG, MMODE, ACCMODE, LCYCMODE, and HSDC_CFG
registers, and the six 8-bit reserved internal registers form the
bits [a255, a254,…, a0] used by LFSR. Bit a0 is the least significant
bit of the first internal register to enter LFSR; Bit a255 is the most
significant bit of the MASK0 register, the last register to enter
LFSR. The formulas that govern LFSR are as follows:
31 0 0 15 0 15 0 01531
255 248 240 232 224 216
7070707 0
0
07 07
40 32 24 16 8 7
MASK0 MASK1 COMPMODE CFMODEGAIN INTERNAL
REGISTER
INTERNAL
REGISTER
INTERNAL
REGISTER
INTERNAL
REGISTER
INTERNAL
REGISTER
INTERNAL
REGISTER
LFSR
GENERATOR
08510-055
Figure 77. CHECKSUM Register Calculation
b
0
LFSR
FB
g
0
g
1
g
2
g
31
b
1
g
3
b
2
b
31
a
255
,a
254
,....,a
2
,a
1
,a
0
08510-056
Figure 78. LFSR Generator Used in CHECKSUM Register Calculation
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 64 of 96
INTERRUPTS
The ADE7854/ADE7858/ADE7868/ADE7878 have two interrupt
pins, IRQ0 and IRQ1. Each of the pins is managed by a 32-bit
interrupt mask register, MASK0 and MASK1, respectively. To
enable an interrupt, a bit in the MASKx register must be set to
1. To disable it, the bit must be cleared to 0. Two 32-bit status
registers, STATUS0 and STATUS1, are associated with the inter-
rupts. When an interrupt event occurs in the ADE78xx, the
corresponding flag in the interrupt status register is set to a Logic 1
(see and ). If the mask bit for this interrupt in
the interrupt mask register is Logic 1, then the
Table 37 Table 38
IRQx logic output
goes active low. The flag bits in the interrupt status register are set
irrespective of the state of the mask bits. To determine the source
of the interrupt, the MCU should perform a read of the corres-
ponding STATUSx register and identify which bit is set to 1. To
erase the flag in the status register, write back to the STATUSx
register with the flag set to 1. After an interrupt pin goes low, the
status register is read and the source of the interrupt is identified.
Then, the status register is written back without any change to
clear the status flag to 0. The IRQx pin remains low until the
status flag is cancelled.
By default, all interrupts are disabled. However, the RSTDONE
interrupt is an exception. This interrupt can never be masked
(disabled) and, therefore, Bit 15 (RSTDONE) in the MASK1
register does not have any functionality. The IRQ1 pin always
goes low, and Bit 15 (RSTDONE) in the STATUS1 register is set
to 1 whenever a power-up or a hardware/software reset process
ends. To cancel the status flag, the STATUS1 register has to be
written with Bit 15 (RSTDONE) set to 1.
Certain interrupts are used in conjunction with other status
registers. The following bits in the MASK1 register work in
conjunction with the status bits in the PHNOLOAD register:
Bit 0 (NLOAD)
Bit1 (FNLOAD), available in the ADE7878 only
Bit 2 (VANLOAD)
The following bits in the MASK1 register work with the status bits
in the PHSTATUS register:
Bit 16, (SAG)
Bit 17 (OI)
Bit 18 (OV)
The following bits in the MASK1 register work with the status bits
in the IPEAK and VPEAK registers, respectively:
Bit 23 (PKI)
Bit 24 (PKV)
The following bits in the MASK0 register work with the status bits
in the PHSIGN register:
Bits[6:8] (REVAPx)
Bits[10:12] (REVRPx), available in the ADE7858,
ADE7868, and ADE7878 only
Bit 9, Bit 13, and Bit 18 (REVPSUMx)
When the STATUSx register is read and one of these bits is set
to 1, the status register associated with the bit is immediately
read to identify the phase that triggered the interrupt and only
at that time can the STATUSx register be written back with the bit
set to 1.
Using the Interrupts with an MCU
Figure 79 shows a timing diagram that illustrates a suggested
implementation of the ADE7854/ADE7858/ADE7868/ADE7878
interrupt management using an MCU. At Time t1, the IRQx pin
goes active low indicating that one or more interrupt events
have occurred in the ADE78xx, at which point the following
steps should be taken:
1. Tie the IRQx pin to a negative-edge-triggered external
interrupt on the MCU.
2. On detection of the negative edge, configure the MCU to
start executing its interrupt service routine (ISR).
3. On entering the ISR, disable all interrupts using the global
interrupt mask bit. At this point, the MCU external interrupt
flag can be cleared to capture interrupt events that occur
during the current ISR.
4. When the MCU interrupt flag is cleared, a read from
STATUSx, the interrupt status register, is carried out. The
interrupt status register content is used to determine the
source of the interrupt(s) and, hence, the appropriate
action to be taken.
5. The same STATUSx content is written back into the
ADE78xx to clear the status flag(s) and reset the IRQx line
to logic high (t2).
If a subsequent interrupt event occurs during the ISR (t3), that
event is recorded by the MCU external interrupt flag being set
again.
On returning from the ISR, the global interrupt mask bit is
cleared (same instruction cycle) and the external interrupt flag
uses the MCU to jump to its ISR once again. This ensures that
the MCU does not miss any external interrupts.
Figure 80 shows a recommended timing diagram when the
status bits in the STATUSx registers work in conjunction with
bits in other registers. When the IRQx pin goes active low, the
STATUSx register is read, and if one of these bits is 1, a second
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 65 of 96
JUMP
TO ISR
GLOBAL
INTERRUPT
MASK
CLEAR MCU
INTERRUPT
FLAG
READ
STATUSx JUMP
TO ISR
WRITE
BACK
STATUSx
ISR ACTION
(BASED ON STATUSx CONTENTS)
ISR RETURN
GLOBAL INTERRUPT
MASK RESET
MCU
INTERRUPT
FLAG SET
PROGRAM
S
EQUENCE
t
1
t
2
t
3
IRQx
0
8510-057
Figure 79. Interrupt Management
JUMP
TO ISR
GLOBAL
INTERRUPT
MASK
CLEAR MCU
INTERRUPT
FLAG
READ
STATUSx READ
PHx
JUMP
TO ISR
WRITE
BACK
STATUSx
ISR ACTION
(BASED ON STATUSx CONTENTS)
ISR RETURN
GLOBAL INTERRUPT
MASK RESET
MCU
INTERRUPT
FLAG SET
PROGRAM
SEQUENCE
t
1
t
2
t
3
IRQx
08510-058
Figure 80. Interrupt Management when PHSTATUS, IPEAK, VPEAK, or PHSIGN Registers are Involved
status register is read immediately to identify the phase that
triggered the interrupt. The name, PHx, in Figure 80 denotes
one of the PHSTATUS, IPEAK, VPEAK, or PHSIGN registers.
Then, STATUSx is written back to clear the status flag(s).
SERIAL INTERFACES
The ADE7854/ADE7858/ADE7868/ADE7878 have three serial
port interfaces: one fully licensed I2C interface, one serial
peripheral interface (SPI), and one high speed data capture port
(HSDC). As the SPI pins are multiplexed with some of the pins
of the I2C and HSDC ports, the ADE78xx accepts two confi-
gurations: one using the SPI port only and one using the I2C
port in conjunction with the HSDC port.
Serial Interface Choice
After reset, the HSDC port is always disabled. Choose between
the I2C and SPI ports by manipulating the SS/HSA pin after
power-up or after a hardware reset. If the SS/HSA pin is kept
high, then the ADE7854/ADE7858/ADE7868/ADE7878 use the
I2C port until a new hardware reset is executed. If the SS/HSA
pin is toggled high to low three times after power-up or after a
hardware reset, the ADE7854/ADE7858/ADE7868/ADE7878
use the SPI port until a new hardware reset is executed. This
manipulation of the SS/HSA pin can be accomplished in two
ways. First, use the SS/HSA pin of the master device (that is, the
microcontroller) as a regular I/O pin and toggle it three times.
Second, execute three SPI write operations to a location in the
address space that is not allocated to a specific ADE78xx register
(for example 0xEBFF, where eight bit writes can be executed).
These writes allow the SS/HSA pin to toggle three times. See the
section for details on the write protocol
involved.
SPI Write Operation
After the serial port choice is completed, it needs to be locked.
Consequently, the active port remains in use until a hardware
reset is executed in PSM0 normal mode or until a power-down.
If I2C is the active serial port, Bit 1 (I2C_LOCK) of the CONFIG2
register must be set to 1 to lock it in. From this moment, the
ADE7854/ADE7858/ADE7868/ADE7878 ignore spurious
toggling of the SS pin and an eventual switch into using the SPI
port is no longer possible. If the SPI is the active serial port, any
write to the CONFIG2 register locks the port. From this moment,
a switch into using the I2C port is no longer possible. Once locked,
the serial port choice is maintained when the ADE78xx changes
PSMx power modes.
The functionality of the ADE78xx is accessible via several on-
chip registers. The contents of these registers can be updated or
read using either the I2C or SPI interfaces. The HSDC port provides
the state of up to 16 registers representing instantaneous values of
phase voltages and neutral currents, and active, reactive, and
apparent powers.
I2C-Compatible Interface
The ADE7854/ADE7858/ADE7868/ADE7878 supports a fully
licensed I2C interface. The I2C interface is implemented as a full
hardware slave. SDA is the data I/O pin, and SCL is the serial
clock. These two pins are shared with the MOSI and SCLK pins
of the on-chip SPI interface. The maximum serial clock frequency
supported by this interface is 400 kHz.
The two pins used for data transfer, SDA and SCL, are confi-
gured in a wire-AND’ed format that allows arbitration in a
multimaster system.
The transfer sequence of an I2C system consists of a master device
initiating a transfer by generating a start condition while the bus
is idle. The master transmits the address of the slave device and
the direction of the data transfer in the initial address transfer. If
the slave acknowledges, the data transfer is initiated. This con-
tinues until the master issues a stop condition, and the bus
becomes idle.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 66 of 96
I2C Write Operation
The write operation using the I2C interface of the ADE7854/
ADE7858/ADE7868/ADE7878 initiate when the master generates
a start condition and consists in one byte representing the
address of the ADE78xx followed by the 16-bit address of the
target register and by the value of the register.
The most significant seven bits of the address byte constitute
the address of the ADE7854/ADE7858/ADE7868/ADE7878
and they are equal to 0111000b. Bit 0 of the address byte is a
read/write bit. Because this is a write operation, it has to be
cleared to 0; therefore, the first byte of the write operation is
0x70. After every byte is received, the ADE7854/ADE7858/
ADE7868/ADE7878 generate an acknowledge. As registers can
have 8, 16, or 32 bits, after the last bit of the register is transmitted
and the ADE78xx acknowledges the transfer, the master gene-
rates a stop condition. The addresses and the register content
are sent with the most significant bit first. See for
details of the I2C write operation.
Figure 81
ACKNOWLEDGE
GENERATED BY
ADE78xx
START
STOP
S
A
C
K
A
C
K
A
C
K
A
C
K
A
C
K
A
C
K
A
C
K
S0
15
SLAVE ADDRESS
MS 8 BITS OF
REGISTER ADDRESS
LS 8 BITS OF
REGISTER ADDRESS
BYTE 3 (MS)
OF REGISTER BYTE 2 OF REGISTER BYTE 1 OF REGISTER BYTE 0 (LS) OF
REGISTER
87 031 1615 87 0 07
1110000
08510-059
Figure 81. I2C Write Operation of a 32-Bit Register
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 67 of 96
I2C Read Operation
The read operation using the I2C interface of the ADE7854/
ADE7858/ADE7868/ADE7878 is accomplished in two stages.
The first stage sets the pointer to the address of the register. The
second stage reads the content of the register.
As seen in Figure 82, the first stage initiates when the master
generates a start condition and consists in one byte representing
the address of the ADE7854/ADE7858/ADE7868/ADE7878
followed by the 16-bit address of the target register. The ADE78xx
acknowledges every byte received. The address byte is similar to
the address byte of a write operation and is equal to 0x70 (see
the I2C Write Operation section for details). After the last byte
of the register address has been sent and acknowledged by the
ADE7854/ADE7858/ADE7868/ADE7878, the second stage
begins with the master generating a new start condition followed
by an address byte. The most significant seven bits of this address
byte constitute the address of the ADE78xx, and they are equal
to 0111000b. Bit 0 of the address byte is a read/write bit. Because
this is a read operation, it must be set to 1; thus, the first byte of
the read operation is 0x71. After this byte is received, the ADE78xx
generates an acknowledge. Then, the ADE78xx sends the value
of the register, and after every eight bits are received, the master
generates an acknowledge. All the bytes are sent with the most
significant bit first. Because registers can have 8, 16, or 32 bits,
after the last bit of the register is received, the master does not
acknowledge the transfer but generates a stop condition.
ACKNOWLEDGE
GENERATED BY
ADE78xx
ACKNOWLEDGE
GENERATED BY
MASTER
START
S
A
C
K
A
C
K
A
C
K
0
15
SLAVE ADDRESS MSB 8 BITS OF
REGISTER ADDRESS
LSB 8 BITS OF
REGISTER ADDRESS
87 0
1110000
START
STOP
S
A
C
K
A
C
K
A
C
K
A
C
K
S0
SLAVE ADDRESS BYTE 3 (MSB)
OF REGISTER
BYTE 2 OF
REGISTER
BYTE 1 OF
REGISTER
BYTE 0 (LSB)
OF REGISTER
31 16 15 8 7007
1110001
ACKNOWLEDGE
GENERATED BY
ADE78xx
N
O
A
C
K
08510-060
Figure 82. I2C Read Operation of a 32-Bit Register
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 68 of 96
SPI-Compatible Interface
The SPI of the ADE7854/ADE7858/ADE7868/ADE7878 is
always a slave of the communication and consists of four pins
(with dual functions): SCLK/SCL, MOSI/SDA, MISO/HSD, and
SS/HSA. The functions used in the SPI-compatible interface are
SCLK, MOSI, MISO, and SS. The serial clock for a data transfer
is applied at the SCLK logic input. This logic input has a Schmitt
trigger input structure that allows the use of slow rising (and
falling) clock edges. All data transfer operations synchronize
to the serial clock. Data shifts into the ADE78xx at the MOSI
logic input on the falling edge of SCLK and the ADE78xx
samples it on the rising edge of SCLK. Data shifts out of the
ADE7854/ ADE7858/ADE7868/ADE7878 at the MISO logic
output on a falling edge of SCLK and can be sampled by the
master device on the raising edge of SCLK. The most significant
bit of the word is shifted in and out first. The maximum serial
clock frequency supported by this interface is 2.5 MHz. MISO
stays in high impedance when no data is transmitted from the
ADE7854/ADE7858/ADE7868/ADE7878. See for
details of the connection between the ADE78xx SPI and a
master device containing an SPI interface.
Figure 83
The SS logic input is the chip select input. This input is used
when multiple devices share the serial bus. Drive the SS input
low for the entire data transfer operation. Bringing SS high
during a data transfer operation aborts the transfer and places
the serial bus in a high impedance state. A new transfer can
then be initiated by returning the SS logic input to low. However,
because aborting a data transfer before completion leaves the
accessed register in a state that cannot be guaranteed, every
time a register is written, its value should be verified by reading
it back. The protocol is similar to the protocol used in I2C
interface.
MOSI/SDA
MISO/HSD
SCLK/SCL
ADE78xx
MOSI
MISO
SCK
SPI DEVICE
SS/HSA SS
08510-061
Figure 83. Connecting ADE78xx SPI with an SPI Device
SPI Read Operation
The read operation using the SPI interface of the ADE7854/
ADE7858/ADE7868/ADE7878 initiate when the master sets the
SS/HSA pin low and begins sending one byte, representing the
address of the ADE7854/ADE7858/ADE7868/ADE7878, on the
MOSI line. The master sets data on the MOSI line starting with
the first high-to-low transition of SCLK. The SPI of the ADE78xx
samples data on the low-to-high transitions of SCLK. The most
significant seven bits of the address byte can have any value, but
as a good programming practice, they should be different from
0111000b, the seven bits used in the I2C protocol. Bit 0 (read/write)
of the address byte must be 1 for a read operation. Next, the
master sends the 16-bit address of the register that is read. After
the ADE78xx receives the last bit of address of the register on a
low-to-high transition of SCLK, it begins to transmit its contents
on the MISO line when the next SCLK high-to-low transition
occurs; thus, the master can sample the data on a low-to-high
SCLK transition. After the master receives the last bit, it sets the
SS and SCLK lines high and the communication ends. The data
lines, MOSI and MISO, go into a high impedance state. See
for details of the SPI read operation. Figure 84
10
15 14
SCLK
MOSI
MISO
10
31 30 1 0
000000
REGISTER VALUE
REGISTER ADDRESS
SS
08510-062
Figure 84. SPI Read Operation of a 32-Bit Register
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 69 of 96
SPI Write Operation
The write operation using the SPI interface of the ADE78xx
initiates when the master sets the SS/HSA pin low and begins
sending one byte representing the address of the ADE7854/
ADE7858/ADE7868/ADE7878 on the MOSI line. The master
sets data on the MOSI line starting with the first high-to-low
transition of SCLK. The SPI of the ADE78xx samples data on
the low-to-high transitions of SCLK. The most significant seven
bits of the address byte can have any value, but as a good pro-
gramming practice, they should be different from 0111000b, the
seven bits used in the I2C protocol. Bit 0 (read/write) of the
address byte must be 0 for a write operation. Next, the master
sends both the 16-bit address of the register that is written and
the 32-, 16-, or 8-bit value of that register without losing any
SCLK cycle. After the last bit is transmitted, the master sets the
SS and SCLK lines high at the end of the SCLK cycle and the
communication ends. The data lines, MOSI and MISO, go into
a high impedance state. See for details of the SPI write
operation.
Figure 85
0
15 14
SCLK
MOSI
103130 10
00 0 0000
REGISTER ADDRESS REGISTER VALUE
SS
08510-063
Figure 85. SPI Write Operation of a 32-Bit Register
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 70 of 96
HSDC Interface
The high speed data capture (HSDC) interface is disabled after
default. It can be used only if the ADE7854/ADE7858/ADE7868/
ADE7878 is configured with an I2C interface. The SPI interface
of the ADE7854/ADE7858/ADE7868/ADE7878 cannot be used
simultaneously with HSDC.
Bit 6 (HSDCEN) in the CONFIG register activates HSDC when
set to 1. If Bit HSDCEN is cleared to 0, the default value, the
HSDC interface is disabled. Setting Bit HSDCEN to 1 when SPI
is in use does not have any effect. HSDC is an interface for
sending to an external device (usually a microprocessor or a
DSP) up to sixteen 32-bit words. The words represent the
instantaneous values of the phase currents and voltages, neutral
current, and active, reactive, and apparent powers. The registers
being transmitted include IAWV, VAWV, IBWV, VBWV, ICWV,
VCWV, AVA, INWV, BVA, CVA, AWATT, BWATT, CWATT,
AVAR, BVAR, and CVAR. All are 24-bit registers that are sign
extended to 32-bits (see Figure 35 for details). In the case of
ADE7854 and ADE7858, the INWV register is not available. In
its place, the HSDC transmits one 32-bit word always equal
to 0. In addition, the AVAR, BVAR, and CVAR registers are not
available in the ADE7854. In their place, the HSDC transmits
three 32-bit words that are always equal to 0.
HSDC can be interfaced with SPI or similar interfaces. HSDC is
always a master of the communication and consists of three
pins: HSA, HSD, and HSCLK. HSA represents the select signal.
It stays active low or high when a word is transmitted and it is
usually connected to the select pin of the slave. HSD sends data
to the slave and it is usually connected to the data input pin of
the slave. HSCLK is the serial clock line that is generated by the
ADE7854/ADE7858/ADE7868/ADE7878 and it is usually con-
nected to the serial clock input of the slave. Figure 86 shows the
connections between the ADE78xx HSDC and slave devices
containing an SPI interface.
08510-064
MISO/HSD
CF3/HSCLK
ADE78xx
MISO
SCK
SPI DEVICE
SS/HSA SS
Figure 86. Connecting the ADE78xx HSDC with an SPI
The HSDC communication is managed by the HSDC_CFG
register (see Table 53). It is recommended to set the HSDC_CFG
register to the desired value before enabling the port using Bit 6
(HSDCEN) in the CONFIG register. In this way, the state of
various pins belonging to the HSDC port do not take levels incon-
sistent with the desired HSDC behavior. After a hardware reset
or after power-up, the MISO/HSD and SS/HSA pins are set high.
Bit 0 (HCLK) in the HSDC_CFG register determines the serial
clock frequency of the HSDC communication. When HCLK is
0 (the default value), the clock frequency is 8 MHz. When HCLK
is 1, the clock frequency is 4 MHz. A bit of data is transmitted
for every HSCLK high-to-low transition. The slave device that
receives data from HSDC samples the HSD line on the low-to-
high transition of HSCLK.
The words can be transmitted as 32-bit packages or as 8-bit
packages. When Bit 1 (HSIZE) in the HSDC_CFG register is 0 (the
default value), the words are transmitted as 32-bit packages. When
Bit HSIZE is 1, the registers are transmitted as 8-bit packages. The
HSDC interface transmits the words MSB first.
Bit 2 (HGAP) introduces a gap of seven HSCLK cycles between
packages when Bit 2 (HGAP) is set to 1. When Bit HGAP is cleared
to 0 (the default value), no gap is introduced between packages
and the communication time is shortest. In this case, HSIZE
does not have any influence on the communication and a data
bit is placed on the HSD line with every HSCLK high-to-low
transition.
Bits[4:3] (HXFER[1:0]) decide how many words are transmitted.
When HXFER[1:0] is 00, the default value, then all 16 words are
transmitted. When HXFER[1:0] is 01, only the words representing
the instantaneous values of phase and neutral currents and phase
voltages are transmitted in the following order: IAWV, VAWV,
IBWV, VBWV, ICWV, VCWV, and one 32-bit word that is always
equal to INWV. When HXFER[1:0] is 10, only the instantaneous
values of phase powers are transmitted in the following order:
AVA, BVA, CVA, AWATT, BWATT, CWATT, AVAR, BVAR, and
CVAR. The value, 11, for HXFER[1:0] is reserved and writing it is
equivalent to writing 00, the default value.
Bit 5 (HSAPOL) determines the polarity of HSA function of the
SS/HSA pin during communication. When HSAPOL is 0 (the
default value), HSA is active low during the communication.
This means that HSA stays high when no communication is in
progress. When the communication starts, HSA goes low and
stays low until the communication ends. Then it goes back to
high. When HSAPOL is 1, the HSA function of the SS/HSA pin
is active high during the communication. This means that HSA
stays low when no communication is in progress. When the
communication starts, HSA goes high and stays high until the
communication ends; then, it goes back to low.
Bits[7:6] of the HSDC_CFG register are reserved. Any value
written into these bits does not have any consequence on HSDC
behavior.
Figure 87 shows the HSDC transfer protocol for HGAP = 0,
HXFER[1:0] = 00 and HSAPOL = 0. Note that the HSDC
interface sets a data bit on the HSD line every HSCLK high-
to-low transition and the value of Bit HSIZE is irrelevant.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 71 of 96
Figure 88 shows the HSDC transfer protocol for HSIZE = 0,
HGAP = 1, HXFER[1:0] = 00, and HSAPOL = 0. Note that the
HSDC interface introduces a seven-HSCLK cycles gap between
every 32-bit word.
Figure 89 shows the HSDC transfer protocol for HSIZE = 1,
HGAP = 1, HXFER[1:0] = 00, and HSAPOL = 0. Note that the
HSDC interface introduces a seven-HSCLK cycles gap between
every 8-bit word.
See Table 53 for the HSDC_CFG register and descriptions for
the HCLK, HSIZE, HGAP, HXFER[1:0], and HSAPOL bits.
Table 24 lists the time it takes to execute an HSDC data transfer
for all HSDC_CFG register settings. For some settings, the
transfer time is less than 125 s (8 kHz), the waveform sample
registers update rate. This means the HSDC port transmits data
every sampling cycle. For settings in which the transfer time is
greater than 125 s, the HSDC port transmits data only in the
first of two consecutive 8 kHz sampling cycles. This means it
transmits registers at an effective rate of 4 kHz.
Table 24. Communication Times for Various HSDC Settings
HXFER[1:0] HGAP HSIZE1 HCLK Communication Time (μs)
00 0 N/A 0 64
00 0 N/A 1 128
00 1 0 0
77.125
00 1 0 1
154.25
00 1 1 0
119.25
00 1 1 1
238.25
01 0 N/A 0
28
01 0 N/A 1
56
01 1 0 0
33.25
01 1 0 1
66.5
01 1 1 0
51.625
01 1 1 1
103.25
10 0 N/A 0
36
10 0 N/A 1
72
10 1 0 0
43
10 1 0 1
86
10 1 1 0
66.625
10 1 1 1 133.25
1 N/A means not applicable.
HSCLK
HSD
HSA
0
31 0
31 0
31 0
31
IAVW (32 BITS) VAWV (32 BITS) IBWV (32 BITS) CVAR (32 BITS)
08510-066
Figure 87. HSDC Communication for HGAP = 0, HXFER[1:0] = 00, and HSAPOL = 0; HSIZE Is Irrelevant
HSCLK
HSDATA
HSACTIVE
0
31 0
31 0
31
IAVW (32 BITS)
7 HCLK CYCLES
VAWV (32 BITS) IBWV (32 BITS)
7 HCLK CYCLES
0
31
CVAR (32 BITS)
08510-067
Figure 88. HSDC Communication for HSIZE = 0, HGAP = 1, HXFER[1:0] = 00, and HSAPOL = 0
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 72 of 96
HSCLK
HSDATA
HSACTIVE
24
31 16
23 8
15 0
7
IAVW (BYTE 3)
7 HCLK CYCLES
IAWV (BYTE 2) IAWV (BYTE 1) CVAR (BYTE 0)
7 HCLK CYCLES
08510-068
Figure 89. HSDC Communication for HSIZE = 1, HGAP = 1, HXFER[1:0] = 00, and HSAPOL = 0
ADE7878 EVALUATION BOARD
An evaluation board built upon the ADE7878 configuration
supports all ADE7854, ADE7858, ADE7868, and ADE7878
components. Visit www.analog.com/ADE7878 for details.
DIE VERSION
The register named version identifies the version of the die. It is
an 8-bit, read-only version register located at Address 0xE707.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 73 of 96
SILICON ANOMALY
This anomaly list describes the known issues with the ADE7854, ADE7858, ADE7868, and ADE7878 silicon identified by the version
register (Address 0xE707) being equal to 2 and to 4.
Analog Devices, Inc., is committed, through future silicon revisions, to continuously improve silicon functionality. Analog Devices tries
to ensure that these future silicon revisions remain compatible with your present software/systems by implementing the recommended
workarounds outlined here.
ADE7854/ADE7858/ADE7868/ADE7878 FUNCTIONALITY ISSUES
Silicon Revision
Identifier Chip Marking Silicon Status Anomaly Sheet No. of Reported Issues
Version = 2 ADE7854ACPZ Released Rev. A 4 (er001, er002, er003, er004)
ADE7858ACPZ
ADE7868ACPZ
ADE7878ACPZ
Version = 4 ADE7854ACPZ Released Rev. B 1 (er005)
ADE7858ACPZ
ADE7868ACPZ
ADE7878ACPZ
FUNCTIONALITY ISSUES
Table 25. Offset RMS Registers Cannot be Set to Negative Values [er001, Version = 2 Silicon]
Background When the AIRMSOS, AVRMSOS, BIRMSOS, BVRMSOS, CIRMSOS, CVRMSOS, and NIRMSOS registers are set to a negative
value, for sufficiently small inputs, the argument of the square root used in the rms data path may become negative. In
this case, the corresponding AIRMS, AVRMS, BIRMS, BVRMS, CIRMS, or CVRMS rms register is automatically set to 0.
Issue Negative values for the AIRMSOS, AVRMSOS, BIRMSOS, BVRMSOS, CIRMSOS, CVRMSOS, and NIRMSOS registers are not
supported in the silicon version identified by the version register being equal to 2.
Workaround Do not use negative values for the AIRMSOS, AVRMSOS, BIRMSOS, BVRMSOS, CIRMSOS, CVRMSOS, and NIRMSOS
registers.
If further details on this issue are required, please use the following website to submit your query:
www.analog.com/en/content/technical_support_page/fca.html.
Related Issues None.
Table 26. Values Written to the CF1DEN, CF2DEN, CF3DEN, SAGLVL, and ZXTOUT Registers May Not Be Immediately Used By
ADE7854, ADE7858, ADE7868, ADE7878 [er002, Version = 2 Silicon]
Background Usually, the CF1DEN, CF2DEN, CF3DEN, SAGLVL, and ZXTOUT registers initialize immediately after power-up or after a
hardware/software reset. After the RUN register is set to 1, the energy-to-frequency converter (for CF1DEN, CF2DEN, and CF3DEN), the
phase voltage sag detector (for SAGLVL), and the zero-crossing timeout circuit (for ZXTOUT) use these values immediately.
Issue After the CF1DEN register is initialized with a new value after power-up or a hardware/software reset, the new value may be
delayed and, therefore, not immediately available for use by the energy-to-frequency converter. It is, however, used by the
converter after the first high-to-low transition is triggered at t the CF1 pin using the CF1DEN default value (0x0).
CF2DEN and CF3DEN registers present similar behavior at the CF2 and CF3 pins, respectively. CF1DEN, CF2DEN and CF3DEN
above behavior has been corrected in Version = 4 silicon.
After the SAGLVL register is initialized with a new value after power-up or a hardware or software reset, the new value may be
delayed and not available for immediate use by the phase voltage sag detector. However, it is used by the detector after at least
one phase voltage rises above 10% of the full-scale input at the phase voltage ADCs.
After the ZXTOUT register is initialized with a new value after power-up or a hardware or software reset, the new value may be
delayed and not available for immediate use by the zero-crossing timeout circuit. However, the circuit does use the new value
after at least one phase voltage rises above 10% of the full-scale input at the phase voltage ADCs.
Workaround If the behavior outlined in the Issue row does not conflict with the meter specification, then the new values of the CF1DEN,
CF2DEN, CF3DEN, SAGLVL, and ZXTOUT registers may be written one time only.
If the behavior is not acceptable, write the new value into the CF1DEN, CF2DEN, and CF3DEN registers eight consecutive times.
This ensures the probability of the new value not being considered immediately by the energy-to-frequency converter becomes
lower than 0.2 ppm.
Usually, at least one of the phase voltages is greater than 10% of full scale after power-up or after a hardware/software reset. If
this cannot be guaranteed, then the SAGLVL and ZXTOUT registers should also be written eight consecutive times to reduce the
probability of not being considered immediately by the phase voltage sag detector and zero-crossing timeout circuit.
Related Issues None.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 74 of 96
Table 27. The Read-Only RMS Registers May Show the Wrong Value [er003, Version = 2 Silicon]
Background The read-only rms registers (AVRMS, BVRMS, CVRMS, AIRMS, BIRMS, CIRMS, and NIRMS) can be read without restrictions at
any time.
Issue The fixed function DSP of ADE7854, ADE7858, ADE7868, and ADE7878 computes all the powers and rms values in a loop
with a period of 125 µs (8 kHz frequency). If two rms registers are accessed (read) consecutively, the value of the second
register may be corrupted. Consequently, the apparent power computed during that 125 µs cycle is also corrupted. The
rms calculation recovers in the next 125 µs cycle, and all the rms and apparent power values compute correctly.
The issue appears independent of the communication type, SPI or I2C, when the time between the start of two
consecutive rms readings is lower than 265 µs. The issue affects only the rms registers; all of the other registers of
ADE7854, ADE7858, ADE7868, and ADE7878 can be accessed without any restrictions.
Workaround The rms registers can be read one at a time with at least 265 µs between the start of the readings. DREADY interrupt at the
IRQ0 pin can be used to time one rms register reading every three consecutive DREADY interrupts. This ensures 375 µs
between the start of the rms readings.
Alternatively, the rms registers can be read interleaved with readings of other registers that are not affected by this
restriction as long as the time between the start of two consecutive rms register readings is 265 s.
Related Issues None.
Table 28. To Obtain Best Accuracy Performance, Internal Setting Must Be Changed [er004, Version = 2 Silicon]
Background Internal default settings provide best accuracy performance for ADE7854, ADE7858, ADE7868, and ADE7878.
Issue It was found that if a different setting is used, the accuracy performance can be improved.
Workaround To enable a new setting for this internal register, execute two consecutive 8-bit register write operations:
The first write operation: 0xAD is written to Address 0xE7FE.
The second write operation: 0x01 is written to Address 0xE7E2.
The write operations must be executed consecutively without any other read/write operation in between. As a
verification that the value was captured correctly, a simple 8-bit read of Address 0xE7E2 should show the 0x01 value.
Related Issues None.
Table 29. Values Written to the SAGLVL and ZXTOUT Registers May Not Be Immediately Used by ADE7854, ADE7858,
ADE7868, and ADE7878 [er005, Version = 4 Silicon]
Background Usually, the SAGLVL and ZXTOUT registers initialize immediately after power-up or after a hardware/software reset. After
the run register is set to 1, the phase voltage sag detector (for SAGLVL), and the zero-crossing timeout circuit (for ZXTOUT)
use these values immediately.
Issue After the SAGLVL register is initialized with a new value after power-up or a hardware or software reset, the new value
may be delayed and not available for immediate use by the phase voltage sag detector. However, it is used by the
detector after at least one phase voltage rises above 10% of the full-scale input at the phase voltage ADCs.
After the ZXTOUT register is initialized with a new value after power-up or a hardware or software reset, the new value
may be delayed and not available for immediate use by the zero-crossing timeout circuit. However, the circuit does use
the new value after at least one phase voltage rises above 10% of the full-scale input at the phase voltage ADCs.
Workaround Usually, at least one of the phase voltages is greater than 10% of full scale after power-up or after a hardware/software
reset. If this cannot be guaranteed, then the SAGLVL and ZXTOUT registers should be written eight consecutive times to
reduce the probability of not being considered immediately by the phase voltage sag detector and zero-crossing timeout
circuit below 0.2 ppm.
Related Issues None.
SECTION 1. ADE7854/ADE7858/ADE7868/ADE7878 FUNCTIONALITY ISSUES
Reference
Number Description Status
er001 Offset rms registers cannot be set to negative values. Identified
er002 Values written to the CF1DEN, CF2DEN, CF2DEN, SAGLVL, and ZXTOUT registers may not be immediately
used by ADE7854, ADE7858, ADE7868, and ADE7878.
Identified
er003 The read-only rms registers may show the wrong value. Identified
er004 To obtain best accuracy performance, internal setting must be changed. Identified
er005 Values written to the SAGLVL and ZXTOUT registers may not be immediately used by ADE7854, ADE7858,
ADE7868, and ADE7878.
Identified
This completes the Silicon Anomaly section.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 75 of 96
REGISTERS LIST
Table 30. Registers List Located in DSP Data Memory RAM
Address
Register
Name R/W1
Bit
Length
Bit Length During
Communication2 Type 3
Default
Value Description
0x4380 AIGAIN R/W 24 32 ZPSE S 0x000000 Phase A current gain adjust.
0x4381 AVGAIN R/W 24 32 ZPSE S 0x000000 Phase A voltage gain adjust.
0x4382 BIGAIN R/W 24 32 ZPSE S 0x000000 Phase B current gain adjust.
0x4383 BVGAIN R/W 24 32 ZPSE S 0x000000 Phase B voltage gain adjust.
0x4384 CIGAIN R/W 24 32 ZPSE S 0x000000 Phase C current gain adjust.
0x4385 CVGAIN R/W 24 32 ZPSE S 0x000000 Phase C voltage gain adjust.
0x4386 NIGAIN R/W 24 32 ZPSE S 0x000000 Neutral current rms offset (ADE7868 and
ADE7878 only).
0x4387 AIRMSOS R/W 24 32 ZPSE S 0x000000 Phase A current rms offset.
0x4388 AVRMSOS R/W 24 32 ZPSE S 0x000000 Phase A voltage rms offset.
0x4389 BIRMSOS R/W 24 32 ZPSE S 0x000000 Phase B current rms offset.
0x438A BVRMSOS R/W 24 32 ZPSE S 0x000000 Phase B voltage rms offset.
0x438B CIRMSOS R/W 24 32 ZPSE S 0x000000 Phase C current rms offset.
0x438C CVRMSOS R/W 24 32 ZPSE S 0x000000 Phase C voltage rms offset.
0x438D NIRMSOS R/W 24 32 ZPSE S 0x000000
Neutral current rms offset (ADE7868 and
ADE7878 only).
0x438E AVAGAIN R/W 24 32 ZPSE S 0x000000 Phase A apparent power gain adjust.
0x438F BVAGAIN R/W 24 32 ZPSE S 0x000000 Phase B apparent power gain adjust.
0x4390 CVAGAIN R/W 24 32 ZPSE S 0x000000 Phase C apparent power gain adjust.
0x4391 AWGAIN R/W 24 32 ZPSE S 0x000000 Phase A total active power gain adjust.
0x4392 AWATTOS R/W 24 32 ZPSE S 0x000000 Phase A total active power offset adjust.
0x4393 BWGAIN R/W 24 32 ZPSE S 0x000000 Phase B total active power gain adjust.
0x4394 BWATTOS R/W 24 32 ZPSE S 0x000000 Phase B total active power offset adjust.
0x4395 CWGAIN R/W 24 32 ZPSE S 0x000000 Phase C total active power gain adjust.
0x4396 CWATTOS R/W 24 32 ZPSE S 0x000000 Phase C total active power offset adjust.
0x4397 AVARGAIN R/W 24 32 ZPSE S 0x000000
Phase A total reactive power gain adjust
(ADE7858, ADE7868, ADE7878 only).
0x4398 AVAROS R/W 24 32 ZPSE S 0x000000
Phase A total reactive power offset adjust
(ADE7858, ADE7868, ADE7878 only).
0x4399 BVARGAIN R/W 24 32 ZPSE S 0x000000
Phase B total reactive power gain adjust
(ADE7858, ADE7868, ADE7878 only).
0x439A BVAROS R/W 24 32 ZPSE S 0x000000
Phase B total reactive power offset adjust
(ADE7858, ADE7868, and ADE7878 only).
0x439B CVARGAIN R/W 24 32 ZPSE S 0x000000
Phase C total reactive power gain adjust
(ADE7858, ADE7868, and ADE7878 only).
0x439C CVAROS R/W 24 32 ZPSE S 0x000000
Phase C total reactive power offset adjust
(ADE7858, ADE7868, and ADE7878 only).
0x439D AFWGAIN R/W 24 32 ZPSE S 0x000000 Phase A fundamental active power gain
adjust. Location reserved for ADE7854,
ADE7858, and ADE7868.
0x439E AFWATTOS R/W 24 32 ZPSE S 0x000000 Phase A fundamental active power offset
adjust. Location reserved for ADE7854,
ADE7858, and ADE7868.
0x439F BFWGAIN R/W 24 32 ZPSE S 0x000000
Phase B fundamental active power gain
adjust (ADE7878 only).
0x43A0 BFWATTOS R/W 24 32 ZPSE S 0x000000
Phase B fundamental active power offset
adjust (ADE7878 only).
0x43A1 CFWGAIN R/W 24 32 ZPSE S 0x000000 Phase C fundamental active power gain
adjust.
0x43A2 CFWATTOS R/W 24 32 ZPSE S 0x000000 Phase C fundamental active power offset
adjust (ADE7878 only).
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 76 of 96
Address
Register
Name R/W1
Bit
Length
Bit Length During
Communication2 Type 3
Default
Value Description
0x43A3 AFVARGAIN R/W 24 32 ZPSE S 0x000000 Phase A fundamental reactive power gain
adjust (ADE7878 only).
0x43A4 AFVAROS R/W 24 32 ZPSE S 0x000000
Phase A fundamental reactive power
offset adjust (ADE7878 only).
0x43A5 BFVARGAIN R/W 24 32 ZPSE S 0x000000
Phase B fundamental reactive power gain
adjust (ADE7878 only).
0x43A6 BFVAROS R/W 24 32 ZPSE S 0x000000 Phase B fundamental reactive power
offset adjust (ADE7878 only).
0x43A7 CFVARGAIN, R/W 24 32 ZPSE S 0x000000
Phase C fundamental reactive power gain
adjust (ADE7878 only).
0x43A8 CFVAROS R/W 24 32 ZPSE S 0x000000 Phase C fundamental reactive power
offset adjust (ADE7878 only).
0x43A9 VATHR1 R/W 24 32 ZP U 0x000000
Most significant 24 bits of VATHR[47:0]
threshold used in phase apparent power
datapath.
0x43AA VATHR0 R/W 24 32 ZP U 0x000000
Less significant 24 bits of VATHR[47:0]
threshold used in phase apparent power
datapath.
0x43AB WTHR1 R/W 24 32 ZP U 0x000000
Most significant 24 bits of WTHR[47:0]
threshold used in phase total/fundamental
active power datapath.
0x43AC WTHR0 R/W 24 32 ZP U 0x000000
Less significant 24 bits of WTHR[47:0]
threshold used in phase total/fundamental
active power datapath.
0x43AD VARTHR1 R/W 24 32 ZP U 0x000000
Most significant 24 bits of VARTHR[47:0]
threshold used in phase total/fundamental
reactive power datapath (ADE7858,
ADE7868, ADE7878 only).
0x43AE VARTHR0 R/W 24 32 ZP U 0x000000
Less significant 24 bits of VARTHR[47:0]
threshold used in phase total/fundamental
reactive power datapath (ADE7858,
ADE7868, ADE7878 only).
0x43AF Reserved N/A4 N/A4 N/A4 N/A4 0x000000 This memory location should be kept at
0x000000 for proper operation.
0x43B0 VANOLOAD R/W 24 32 ZPSE S 0x0000000
No load threshold in the apparent power
datapath.
0x43B1 APNOLOAD R/W 24 32 ZPSE S 0x0000000
No load threshold in the total/fundamental
active power datapath.
0x43B2 VARNOLOAD R/W 24 32 ZPSE S 0x0000000
No load threshold in the total/fundamental
reactive power datapath. Location
reserved for ADE7854.
0x43B3 VLEVEL R/W 24 32 ZPSE S 0x000000 Register used in the algorithm that
computes the fundamental active and
reactive powers (ADE7878 only).
0x43B4 Reserved N/A4 N/A4 N/A4 N/A4 0x000000 This location should not be written for
proper operation.
0x43B5 DICOEFF R/W 24 32 ZPSE S 0x0000000
Register used in the digital integrator
algorithm. If the integrator is turned on, it
must be set at 0xFF8000. In practice, it is
transmitted as 0xFFF8000.
0x43B6 HPFDIS R/W 24 32 ZP U 0x000000
Disables/enables the HPF in the current
datapath (see Table 34).
0x43B7 Reserved N/A4 N/A4 N/A4 N/A4 0x000000 This memory location should be kept at
0x000000 for proper operation.
0x43B8 ISUMLVL R/W 24 32 ZPSE S 0x000000 Threshold used in comparison between
the sum of phase currents and the neutral
current (ADE7868 and ADE7878 only).
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 77 of 96
Address
Register
Name R/W1
Bit
Length
Bit Length During
Communication2 Type 3
Default
Value Description
0x43B9 to
0x43BE
Reserved N/A4 N/A4 N/A4 N/A4 0x000000 These memory locations should be kept
at 0x000000 for proper operation.
0x43BF ISUM R 28 32 ZP S N/A4 Sum of IAWV, IBWV, and ICWV registers
(ADE7868 and ADE7878 only).
0x43C0 AIRMS R 24 32 ZP S N/A4 Phase A current rms value.
0x43C1 AVRMS R 24 32 ZP S N/A4 Phase A voltage rms value.
0x43C2 BIRMS R 24 32 ZP S N/A4 Phase B current rms value.
0x43C3 BVRMS R 24 32 ZP S N/A4 Phase B voltage rms value.
0x43C4 CIRMS R 24 32 ZP S N/A4 Phase C current rms value.
0x43C5 CVRMS R 24 32 ZP S N/A4 Phase C voltage rms value.
0x43C6 NIRMS R 24 32 ZP S N/A4 Neutral current rms value (ADE7868 and
ADE7878 only).
0x43C7 to
0x43FF
Reserved N/A4 N/A4 N/A4 N/A4 N/A4 These memory locations should not be
written for proper operation.
1 R is read, and W is write.
2 32 ZPSE = 24-bit signed register that is transmitted as a 32-bit word with four MSBs padded with 0s and sign extended to 28 bits. Whereas 32 ZP = 28- or 24-bit signed
or unsigned register that is transmitted as a 32-bit word with four MSBs or eight MSBs, respectively, padded with 0s.
3 U is unsigned register, and S is signed register in twos complement format.
4 N/A means not applicable.
Table 31. Internal DSP Memory RAM Registers
Address
Register
Name R/W1
Bit
Length
Bit Length
During
Communication Type2
Default
Value Description
0xE203 Reserved R/W 16 16 U 0x0000 This memory location should not be written for
proper operation.
0xE228 Run R/W 16 16 U 0x0000 Run register starts and stops the DSP. See the
Digital Signal Processor section for more details.
1 R is read, and W is write.
2 U is unsigned register, and S is signed register in twos complement format.
Table 32. Billable Registers
Address
Register
Name R/W1, 2
Bit
Length2
Bit Length
During
Communication2 Type2, 3
Default
Value Description
0xE400 AWATTHR R 32 32 S 0x00000000 Phase A total active energy accumulation.
0xE401 BWATTHR R 32 32 S 0x00000000 Phase B total active energy accumulation.
0xE402 CWATTHR R 32 32 S 0x00000000 Phase C total active energy accumulation.
0xE403 AFWATTHR R 32 32 S 0x00000000 Phase A fundamental active energy
accumulation (ADE7878 only).
0xE404 BFWATTHR R 32 32 S 0x00000000
Phase B fundamental active energy
accumulation (ADE7878 only).
0xE405 CFWATTHR R 32 32 S 0x00000000
Phase C fundamental active energy
accumulation (ADE7878 only).
0xE406 AVARHR R 32 32 S 0x00000000
Phase A total reactive energy accumulation
(ADE7858, ADE7868, and ADE7878 only).
0xE407 BVARHR R 32 32 S 0x00000000
Phase B total reactive energy accumulation
(ADE7858, ADE7868, and ADE7878 only).
0xE408 CVARHR R 32 32 S 0x00000000
Phase C total reactive energy accumulation
(ADE7858, ADE7868, and ADE7878 only).
0xE409 AFVARHR R 32 32 S 0x00000000 Phase A fundamental reactive energy
accumulation (ADE7878 only).
0xE40A BFVARHR R 32 32 S 0x00000000
Phase B fundamental reactive energy
accumulation (ADE7878 only).
0xE40B CFVARHR R 32 32 S 0x00000000
Phase C fundamental reactive energy
accumulation (ADE7878 only).
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 78 of 96
Address
Register
Name R/W1, 2
Bit
Length2
Bit Length
During
Communication2 Type2, 3
Default
Value Description
0xE40C AVAHR R 32 32 S 0x00000000 Phase A apparent energy accumulation.
0xE40D BVAHR R 32 32 S 0x00000000 Phase B apparent energy accumulation.
0xE40E CVAHR R 32 32 S 0x00000000 Phase C apparent energy accumulation.
1 R is read, and W is write.
2 N/A is not applicable.
3 U is unsigned register, and S is signed register in twos complement format.
Table 33. Configuration and Power Quality Registers
Address
Register
Name R/W1
Bit
Length
Bit Length
During
Communication2 Type 3
Default
Value 4 Description
0xE500 IPEAK R 32 32 U N/A Current peak register. See Figure 48
and Table 35 for details about its
composition.
0xE501 VPEAK R 32 32 U N/A Voltage peak register. See Figure 48
and Table 36 for details about its
composition.
0xE502 STATUS0 R/W 32 32 U N/A Interrupt Status Register 0. See Table 37.
0xE503 STATUS1 R/W 32 32 U N/A Interrupt Status Register 1. See Table 38.
0xE504 AIMAV R 20 32 ZP U N/A Phase A current mean absolute value
computed during PSM0 and PSM1
modes (ADE7868 and ADE7878 only).
0xE505 BIMAV R 20 32 ZP U N/A Phase B current mean absolute value
computed during PSM0 and PSM1
modes (ADE7868 and ADE7878 only).
0xE506 CIMAV R 20 32 ZP U N/A Phase C current mean absolute value
computed during PSM0 and PSM1
modes (ADE7868 and ADE7878 only).
0xE507 OILVL R/W 24 32 ZP U 0xFFFFFF Overcurrent threshold.
0xE508 OVLVL R/W 24 32 ZP U 0xFFFFFF Overvoltage threshold.
0xE509 SAGLVL R/W 24 32 ZP U 0x000000 Voltage SAG level threshold.
0xE50A MASK0 R/W 32 32 U 0x00000000 Interrupt Enable Register 0. See Table 39.
0xE50B MASK1 R/W 32 32 U 0x00000000 Interrupt Enable Register 1. See Table 40.
0xE50C IAWV R 24 32 SE S N/A Instantaneous value of Phase A current.
0xE50D IBWV R 24 32 SE S N/A Instantaneous value of Phase B current.
0xE50E ICWV R 24 32 SE S N/A Instantaneous value of Phase C current.
0xE50F INWV R 24 32 SE S N/A Instantaneous value of neutral current
(ADE7868 and ADE7878 only).
0xE510 VAWV R 24 32 SE S N/A Instantaneous value of Phase A voltage.
0xE511 VBWV R 24 32 SE S N/A Instantaneous value of Phase B voltage.
0xE512 VCWV R 24 32 SE S N/A Instantaneous value of Phase C voltage.
0xE513 AWATT R 24 32 SE S N/A Instantaneous value of Phase A total
active power.
0xE514 BWATT R 24 32 SE S N/A Instantaneous value of Phase B total
active power.
0xE515 CWATT R 24 32 SE S N/A Instantaneous value of Phase C total
active power.
0xE516 AVAR R 24 32 SE S N/A Instantaneous value of Phase A total
reactive power (ADE7858, ADE7868,
and ADE7878 only).
0xE517 BVAR R 24 32 SE S N/A Instantaneous value of Phase B total
reactive power (ADE7858, ADE7868,
and ADE7878 only).
0xE518 CVAR R 24 32 SE S N/A Instantaneous value of Phase C total
reactive power (ADE7858, ADE7868,
and ADE7878 only).
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 79 of 96
Address
Register
Name R/W1
Bit
Length
Bit Length
During
Communication2 Type 3
Default
Value 4 Description
0xE519 AVA R 24 32 SE S N/A Instantaneous value of Phase A
apparent power.
0xE51A BVA R 24 32 SE S N/A Instantaneous value of Phase B
apparent power.
0xE51B CVA R 24 32 SE S N/A Instantaneous value of Phase C
apparent power.
0xE51F CHECKSUM R 32 32 U 0x33666787
Checksum verification. See the
Checksum Register section for details.
0xE520 VNOM R/W 24 32 ZP S 0x000000
Nominal phase voltage rms used in the
alternative computation of the
apparent power.
0xE521 to
0xE52E
Reserved These addresses should not be written
for proper operation.
0xE600 PHSTATUS R 16 16 U N/A Phase peak register. See Table 41.
0xE601 ANGLE0 R 16 16 U N/A Time Delay 0. See the Time Interval
Between Phases section for details.
0xE602 ANGLE1 R 16 16 U N/A Time Delay 1. See the Time Interval
Between Phases section for details.
0xE603 ANGLE2 R 16 16 U N/A Time Delay 2. See the Time Interval
Between Phases section for details.
0xE604 to
0xE606
Reserved These addresses should not be written
for proper operation.
0xE607 PERIOD R 16 16 U N/A Network line period.
0xE608 PHNOLOAD R 16 16 U N/A Phase no load register. See Table 42.
0xE609 to
0xE60B
Reserved These addresses should not be written
for proper operation.
0xE60C LINECYC R/W 16 16 U 0xFFFF Line cycle accumulation mode count.
0xE60D ZXTOUT R/W 16 16 U 0xFFFF Zero-crossing timeout count.
0xE60E COMPMODE R/W 16 16 U 0x01FF Computation-mode register. See
Table 43.
0xE60F Gain R/W 16 16 U 0x0000 PGA gains at ADC inputs. See Table 44.
0xE610 CFMODE R/W 16 16 U 0x0E88 CFx configuration register. See Table 45.
0xE611 CF1DEN R/W 16 16 U 0x0000 CF1 denominator.
0xE612 CF2DEN R/W 16 16 U 0x0000 CF2 denominator.
0xE613 CF3DEN R/W 16 16 U 0x0000 CF3 denominator.
0xE614 APHCAL R/W 10 16 ZP U 0x0000 Phase calibration of Phase A. See
Table 46.
0xE615 BPHCAL R/W 10 16 ZP U 0x0000 Phase calibration of Phase B. See Table 46.
0xE616 CPHCAL R/W 10 16 ZP U 0x0000 Phase calibration of Phase C. See Table 46.
0xE617 PHSIGN R 16 16 U N/A Power sign register. See Table 47.
0xE618 CONFIG R/W 16 16 U 0x0000 ADE7878 configuration register. See
Table 48.
0xE700 MMODE R/W 8 8 U 0x1C Measurement mode register.
See Table 49.
0xE701 ACCMODE R/W 8 8 U 0x00 Accumulation mode register.
See Table 50.
0xE702 LCYCMODE R/W 8 8 U 0x78 Line accumulation mode behavior. See
Table 52.
0xE703 PEAKCYC R/W 8 8 U 0x00 Peak detection half line cycles.
0xE704 SAGCYC R/W 8 8 U 0x00 SAG detection half line cycles.
0xE705 CFCYC R/W 8 8 U 0x01 Number of CF pulses between two
consecutive energy latches. See the
Synchronizing Energy Registers with
CFx Outputs section.
0xE706 HSDC_CFG R/W 8 8 U 0x00 HSDC configuration register. See Table 53.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E | Page 80 of 96
Address
Register
Name R/W1
Bit
Length
Bit Length
During
Communication2 Type 3
Default
Value 4 Description
0xE707 Version R 8 8 U Version of die.
0xEBFF Reserved 8 8 This address can be used in manipulating
the SS/HSA pin when SPI is chosen as
the active port. See the
section for details.
Serial Interfaces
0xEC00 LPOILVL R/W 8 8 U 0x07 Overcurrent threshold used during
PSM2 mode (ADE7868 and ADE7878
only). See Table 54 in which the register
is detailed.
0xEC01 CONFIG2 R/W 8 8 U 0x00 Configuration register used during
PSM1 mode. See Table 55.
1 R is read, and W is write.
2 32 ZP = 24- or 20-bit signed or unsigned register that is transmitted as a 32-bit word with 8 or 12 MSBs, respectively, padded with 0s. 32 SE = 24-bit signed register that
is transmitted as a 32-bit word sign extended to 32 bits. 16 ZP = 10-bit unsigned register that is transmitted as a 16-bit word with six MSBs padded with 0s.
3 U is unsigned register, and S is signed register in twos complement format.
4 N/A is not applicable.
Table 34. HPFDIS Register (Address 0x43B6)
Bit
Location
Default
Value Description
23:0 00000000
When HPFDIS = 0x00000000, then all high-pass filters in voltage and current channels are enabled. When the
register is set to any nonzero value, all high-pass filters are disabled.
Table 35. IPEAK Register (Address 0xE500)
Bit Location Bit Mnemonic Default Value Description
23:0 IPEAKVAL[23:0] 0 These bits contain the peak value determined in the current channel.
24 IPPHASE[0] 0 When this bit is set to 1, Phase A current generated IPEAKVAL[23:0] value.
25 IPPHASE[1] 0 When this bit is set to 1, Phase B current generated IPEAKVAL[23:0] value.
26 IPPHASE[2] 0 When this bit is set to 1, Phase C current generated IPEAKVAL[23:0] value.
31:27 00000 These bits are always 0.
Table 36. VPEAK Register (Address 0xE501)
Bit Location Bit Mnemonic Default Value Description
23:0 VPEAKVAL[23:0] 0 These bits contain the peak value determined in the voltage channel.
24 VPPHASE[0] 0 When this bit is set to 1, Phase A voltage generated VPEAKVAL[23:0] value.
25 VPPHASE[1] 0 When this bit is set to 1, Phase B voltage generated VPEAKVAL[23:0] value.
26 VPPHASE[2] 0 When this bit is set to 1, Phase C voltage generated VPEAKVAL[23:0] value.
31:27 00000 These bits are always 0.
Table 37. STATUS0 Register (Address 0xE502)
Bit
Location Bit Mnemonic Default Value Description
0 AEHF 0 When this bit is set to 1, it indicates that Bit 30 of any one of the total active energy
registers (AWATTHR, BWATTHR, or CWATTHR) has changed.
1 FAEHF 0 When this bit is set to 1, it indicates that Bit 30 of any one of the fundamental active
energy registers, FWATTHR, BFWATTHR, or CFWATTHR, has changed. This bit is always 0
for ADE7854, ADE7858, and ADE7868.
2 REHF 0 When this bit is set to 1, it indicates that Bit 30 of any one of the total reactive energy
registers (AVARHR, BVARHR, or CVARHR) has changed. This bit is always 0 for ADE7854.
3 FREHF 0 When this bit is set to 1, it indicates that Bit 30 of any one of the fundamental reactive
energy registers, AFVARHR, BFVARHR, or CFVARHR, has changed. This bit is always 0 for
ADE7854, ADE7858, and ADE7868.
4 VAEHF 0 When this bit is set to 1, it indicates that Bit 30 of any one of the apparent energy
registers (AVAHR, BVAHR, or CVAHR) has changed.
5 LENERGY 0 When this bit is set to 1, in line energy accumulation mode, it indicates the end of an
integration over an integer number of half line cycles set in the LINECYC register.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E| Page 81 of 96
Bit
Location Bit Mnemonic Default Value Description
6 REVAPA 0 When this bit is set to 1, it indicates that the Phase A active power identified by Bit 6
(REVAPSEL) in the ACCMODE register (total or fundamental) has changed sign. The sign
itself is indicated in Bit 0 (AWSIGN) of the PHSIGN register (see Table 47).
7 REVAPB 0 When this bit is set to 1, it indicates that the Phase B active power identified by Bit 6
(REVAPSEL) in the ACCMODE register (total or fundamental) has changed sign. The sign
itself is indicated in Bit 1 (BWSIGN) of the PHSIGN register (see Table 47).
8 REVAPC 0 When this bit is set to 1, it indicates that the Phase C active power identified by Bit 6
(REVAPSEL) in the ACCMODE register (total or fundamental) has changed sign. The sign
itself is indicated in Bit 2 (CWSIGN) of the PHSIGN register (see Table 47).
9 REVPSUM1 0 When this bit is set to 1, it indicates that the sum of all phase powers in the CF1 datapath
has changed sign. The sign itself is indicated in Bit 3 (SUM1SIGN) of the PHSIGN register
(see Table 47).
10 REVRPA 0 When this bit is set to 1, it indicates that the Phase A reactive power identified by Bit 7
(REVRPSEL) in the ACCMODE register (total or fundamental) has changed sign. The sign
itself is indicated in Bit 4 (AVARSIGN) of the PHSIGN register (see Table 47). This bit is
always 0 for ADE7854.
11 REVRPB 0 When this bit is set to 1, it indicates that the Phase B reactive power identified by Bit 7
(REVRPSEL) in the ACCMODE register (total or fundamental) has changed sign. The sign
itself is indicated in Bit 5 (BVARSIGN) of the PHSIGN register (see Table 47). This bit is
always 0 for ADE7854.
12 REVRPC 0 When this bit is set to 1, it indicates that the Phase C reactive power identified by Bit 7
(REVRPSEL) in the ACCMODE register (total or fundamental) has changed sign. The sign
itself is indicated in Bit 6 (CVARSIGN) of the PHSIGN register (see Table 47). This bit is
always 0 for ADE7854.
13 REVPSUM2 0 When this bit is set to 1, it indicates that the sum of all phase powers in the CF2 datapath
has changed sign. The sign itself is indicated in Bit 7 (SUM2SIGN) of the PHSIGN register
(see Table 47).
14 CF1 When this bit is set to 1, it indicates a high to low transition has occurred at CF1 pin; that
is, an active low pulse has been generated. The bit is set even if the CF1 output is disabled
by setting Bit 9 (CF1DIS) to 1 in the CFMODE register. The type of power used at the CF1
pin is determined by Bits[2:0] (CF1SEL[2:0]) in the CFMODE register (see Table 45).
15 CF2 When this bit is set to 1, it indicates a high-to-low transition has occurred at the CF2 pin;
that is, an active low pulse has been generated. The bit is set even if the CF2 output is
disabled by setting Bit 10 (CF2DIS) to 1 in the CFMODE register. The type of power used at
the CF2 pin is determined by Bits[5:3] (CF2SEL[2:0]) in the CFMODE register (see Table 45).
16 CF3 When this bit is set to 1, it indicates a high-to-low transition has occurred at CF3 pin; that
is, an active low pulse has been generated. The bit is set even if the CF3 output is disabled
by setting Bit 11 (CF3DIS) to 1 in the CFMODE register. The type of power used at the CF3
pin is determined by Bits[8:6] (CF3SEL[2:0]) in the CFMODE register (see Table 45).
17 DREADY 0 When this bit is set to 1, it indicates that all periodical (at 8 kHz rate) DSP computations
have finished.
18 REVPSUM3 0 When this bit is set to 1, it indicates that the sum of all phase powers in the CF3 datapath
has changed sign. The sign itself is indicated in Bit 8 (SUM3SIGN) of the PHSIGN register
(see Table 47).
31:19 Reserved 0 0000 0000 0000 Reserved. These bits are always 0.
Table 38. STATUS1 Register (Address 0xE503)
Bit
Location Bit Mnemonic Default Value Description
0 NLOAD 0 When this bit is set to 1, it indicates that at least one phase entered no load condition based
on total active and reactive powers. The phase is indicated in Bits[2:0] (NLPHASE[x]) in the
PHNOLOAD register (see Table 42).
1 FNLOAD 0 When this bit is set to 1, it indicates that at least one phase entered no load condition based
on fundamental active and reactive powers. The phase is indicated in Bits[5:3] (FNLPHASE[x])
in PHNOLOAD register (see Table 42 in which this register is described). This bit is always 0
for ADE7854, ADE7858, and ADE7868.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E| Page 82 of 96
Bit
Location Bit Mnemonic Default Value Description
2 VANLOAD 0 When this bit is set to 1, it indicates that at least one phase entered no load condition based
on apparent power. The phase is indicated in Bits[8:6] (VANLPHASE[x]) in the PHNOLOAD
register (see Table 42).
3 ZXTOVA 0 When this bit is set to 1, it indicates a zero crossing on Phase A voltage is missing.
4 ZXTOVB 0 When this bit is set to 1, it indicates a zero crossing on Phase B voltage is missing.
5 ZXTOVC 0 When this bit is set to 1, it indicates a zero crossing on Phase C voltage is missing.
6 ZXTOIA 0 When this bit is set to 1, it indicates a zero crossing on Phase A current is missing.
7 ZXTOIB 0 When this bit is set to 1, it indicates a zero crossing on Phase B current is missing.
8 ZXTOIC 0 When this bit is set to 1, it indicates a zero crossing on Phase C current is missing.
9 ZXVA 0 When this bit is set to 1, it indicates a zero crossing has been detected on Phase A voltage.
10 ZXVB 0 When this bit is set to 1, it indicates a zero crossing has been detected on Phase B voltage.
11 ZXVC 0 When this bit is set to 1, it indicates a zero crossing has been detected on Phase C voltage.
12 ZXIA 0 When this bit is set to 1, it indicates a zero crossing has been detected on Phase A current.
13 ZXIB 0 When this bit is set to 1, it indicates a zero crossing has been detected on Phase B current.
14 ZXIC 0 When this bit is set to 1, it indicates a zero crossing has been detected on Phase C current.
15 RSTDONE 1 In case of a software reset command, Bit 7 (SWRST) is set to 1 in the CONFIG register, or a
transition from PSM1, PSM2, or PSM3 to PSM0, or a hardware reset, this bit is set to 1 at the
end of the transition process and after all registers changed value to default. The IRQ1 pin
goes low to signal this moment because this interrupt cannot be disabled.
16 SAG 0 When this bit is set to 1, it indicates a SAG event has occurred on one of the phases indicated
by Bits[14:12] (VSPHASE[x]) in the PHSTATUS register (see Table 41).
17 OI 0 When this bit is set to 1, it indicates an overcurrent event has occurred on one of the phases
indicated by Bits[5:3] (OIPHASE[x]) in the PHSTATUS register (see Table 41).
18 OV 0 When this bit is set to 1, it indicates an overvoltage event has occurred on one of the phases
indicated by Bits[11:9] (OVPHASE[x]) in the PHSTATUS register (see Table 41).
19 SEQERR 0 When this bit is set to 1, it indicates a negative-to-positive zero crossing on Phase A voltage
was not followed by a negative-to-positive zero crossing on Phase B voltage but by a
negative-to-positive zero crossing on Phase C voltage.
20 MISMTCH 0 When this bit is set to 1, it indicates ISUMLVLINWVISUM > , where ISUMLVL is
indicated in the ISUMLVL register. This bit is always 0 for ADE7854 and ADE7858.
21 Reserved 1 Reserved. This bit is always set to 1.
22 Reserved 0 Reserved. This bit is always set to 0.
23 PKI 0 When this bit is set to 1, it indicates that the period used to detect the peak value in the
current channel has ended. The IPEAK register contains the peak value and the phase where
the peak has been detected (see Table 35).
24 PKV 0 When this bit is set to 1, it indicates that the period used to detect the peak value in the
voltage channel has ended. VPEAK register contains the peak value and the phase where the
peak has been detected (see Table 36).
31:25 Reserved 000 0000 Reserved. These bits are always 0.
Table 39. MASK0 Register (Address 0xE50A)
Bit
Location Bit Mnemonic Default Value Description
0 AEHF 0 When this bit is set to 1, it enables an interrupt when Bit 30 of any one of the total active
energy registers (AWATTHR, BWATTHR, or CWATTHR) changes.
1 FAEHF 0 When this bit is set to 1, it enables an interrupt when Bit 30 of any one of the fundamental
active energy registers (AFWATTHR, BFWATTHR, or CFWATTHR) changes. Setting this bit to1
does not have any consequence for ADE7854, ADE7858, and ADE7868.
2 REHF 0 When this bit is set to 1, it enables an interrupt when Bit 30 of any one of the total reactive
energy registers (AVARHR, BVARHR, CVARHR) changes. Setting this bit to1 does not have any
consequence for ADE7854.
3 FREHF 0 When this bit is set to 1, it enables an interrupt when Bit 30 of any one of the fundamental
reactive energy registers (AFVARHR, BFVARHR, or CFVARHR) changes. Setting this bit to1
does not have any consequence for ADE7854, ADE7858, and ADE7868.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E| Page 83 of 96
Bit
Location Bit Mnemonic Default Value Description
4 VAEHF 0 When this bit is set to 1, it enables an interrupt when Bit 30 of any one of the apparent
energy registers (AVAHR, BVAHR, or CVAHR) changes.
5 LENERGY 0 When this bit is set to 1, in line energy accumulation mode, it enables an interrupt at the end
of an integration over an integer number of half line cycles set in the LINECYC register.
6 REVAPA 0 When this bit is set to 1, it enables an interrupt when the Phase A active power identified by
Bit 6 (REVAPSEL) in the ACCMODE register (total or fundamental) changes sign.
7 REVAPB 0 When this bit is set to 1, it enables an interrupt when the Phase B active power identified by
Bit 6 (REVAPSEL) in the ACCMODE register (total or fundamental) changes sign.
8 REVAPC 0 When this bit is set to 1, it enables an interrupt when the Phase C active power identified by
Bit 6 (REVAPSEL) in the ACCMODE register (total or fundamental) changes sign.
9 REVPSUM1 0 When this bit is set to 1, it enables an interrupt when the sum of all phase powers in the CF1
datapath changes sign.
10 REVRPA 0 When this bit is set to 1, it enables an interrupt when the Phase A reactive power identified
by Bit 7 (REVRPSEL) in the ACCMODE register (total or fundamental) changes sign. Setting
this bit to1 does not have any consequence for ADE7854.
11 REVRPB 0 When this bit is set to 1, it enables an interrupt when the Phase B reactive power identified
by Bit 7 (REVRPSEL) in the ACCMODE register (total or fundamental) changes sign. Setting
this bit to1 does not have any consequence for ADE7854.
12 REVRPC 0 When this bit is set to 1, it enables an interrupt when the Phase C reactive power identified
by Bit 7 (REVRPSEL) in the ACCMODE register (total or fundamental) changes sign. Setting
this bit to1 does not have any consequence for ADE7854.
13 REVPSUM2 0 When this bit is set to 1, it enables an interrupt when the sum of all phase powers in the CF2
datapath changes sign.
14 CF1 When this bit is set to 1, it enables an interrupt when a high-to-low transition occurs at the
CF1 pin, that is, an active low pulse is generated. The interrupt can be enabled even if the
CF1 output is disabled by setting Bit 9 (CF1DIS) to 1 in the CFMODE register. The type of
power used at the CF1 pin is determined by Bits[2:0] (CF1SEL[2:0]) in the CFMODE register
(see Table 45).
15 CF2 When this bit is set to 1, it enables an interrupt when a high-to-low transition occurs at CF2
pin, that is, an active low pulse is generated. The interrupt may be enabled even if the CF2
output is disabled by setting Bit 10 (CF2DIS) to 1 in the CFMODE register. The type of power
used at the CF2 pin is determined by Bits[5:3] (CF2SEL[2:0]) in the CFMODE register (see Table 45).
16 CF3 When this bit is set to 1, it enables an interrupt when a high to low transition occurs at CF3
pin, that is, an active low pulse is generated. The interrupt may be enabled even if the CF3
output is disabled by setting Bit 11 (CF3DIS) to 1 in the CFMODE register. The type of power
used at the CF3 pin is determined by Bits[8:6] (CF3SEL[2:0]) in the CFMODE register (see Table 45).
17 DREADY 0 When this bit is set to 1, it enables an interrupt when all periodical (at 8 kHz rate) DSP
computations finish.
18 REVPSUM3 0 When this bit is set to 1, it enables an interrupt when the sum of all phase powers in the CF3
datapath changes sign.
31:19 Reserved 00 0000 0000
0000
Reserved. These bits do not manage any functionality.
Table 40. MASK1 Register (Address 0xE50B)
Bit
Location Bit Mnemonic Default Value Description
0 NLOAD 0 When this bit is set to 1, it enables an interrupt when at least one phase enters no load
condition based on total active and reactive powers.
1 FNLOAD 0 When this bit is set to 1, it enables an interrupt when at least one phase enters no load
condition based on fundamental active and reactive powers. Setting this bit to 1 does not
have any consequence for ADE7854, ADE7858, and ADE7868.
2 VANLOAD 0 When this bit is set to 1, it enables an interrupt when at least one phase enters no load
condition based on apparent power.
3 ZXTOVA 0 When this bit is set to 1, it enables an interrupt when a zero crossing on Phase A voltage is
missing.
4 ZXTOVB 0 When this bit is set to 1, it enables an interrupt when a zero crossing on Phase B voltage is
missing.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E| Page 84 of 96
Bit
Location Bit Mnemonic Default Value Description
5 ZXTOVC 0 When this bit is set to 1, it enables an interrupt when a zero crossing on Phase C voltage is
missing.
6 ZXTOIA 0 When this bit is set to 1, it enables an interrupt when a zero crossing on Phase A current is
missing.
7 ZXTOIB 0 When this bit is set to 1, it enables an interrupt when a zero crossing on Phase B current is
missing.
8 ZXTOIC 0 When this bit is set to 1, it enables an interrupt when a zero crossing on Phase C current is
missing.
9 ZXVA 0 When this bit is set to 1, it enables an interrupt when a zero crossing is detected on Phase A
voltage.
10 ZXVB 0 When this bit is set to 1, it enables an interrupt when a zero crossing is detected on Phase B
voltage.
11 ZXVC 0 When this bit is set to 1, it enables an interrupt when a zero crossing is detected on Phase C
voltage.
12 ZXIA 0 When this bit is set to 1, it enables an interrupt when a zero crossing is detected on Phase A
current.
13 ZXIB 0 When this bit is set to 1, it enables an interrupt when a zero crossing is detected on Phase B
current.
14 ZXIC 0 When this bit is set to 1, it enables an interrupt when a zero crossing is detected on Phase C
current.
15 RSTDONE 0 Because the RSTDONE interrupt cannot be disabled, this bit does not have any functionality
attached. It can be set to 1 or cleared to 0 without having any effect.
16 SAG 0 When this bit is set to 1, it enables an interrupt when a SAG event occurs on one of the
phases indicated by Bits[14:12] (VSPHASE[x]) in the PHSTATUS register (see Table 41).
17 OI 0 When this bit is set to 1, it enables an interrupt when an overcurrent event occurs on one of
the phases indicated by Bits[5:3] (OIPHASE[x]) in the PHSTATUS register (see Table 41).
18 OV 0 When this bit is set to 1, it enables an interrupt when an overvoltage event occurs on one of
the phases indicated by Bits[11:9] (OVPHASE[x]) in the PHSTATUS register (see Table 41).
19 SEQERR 0 When this bit is set to 1, it enables an interrupt when a negative-to-positive zero crossing on
Phase A voltage is not followed by a negative-to-positive zero crossing on Phase B voltage,
but by a negative-to-positive zero crossing on Phase C voltage.
20 MISMTCH 0 When this bit is set to 1, it enables an interrupt when ISUMLVLINWVISUM > is
greater than the value indicated in ISUMLVL register. Setting this bit to1 does not have any
consequence for ADE7854 and ADE7858.
22:21 Reserved 00 Reserved. These bits do not manage any functionality.
23 PKI 0 When this bit is set to 1, it enables an interrupt when the period used to detect the peak
value in the current channel has ended.
24 PKV 0 When this bit is set to 1, it enables an interrupt when the period used to detect the peak
value in the voltage channel has ended.
31:25 Reserved 000 0000 Reserved. These bits do not manage any functionality.
Table 41. PHSTATUS Register (Address 0xE600)
Bit
Location Bit Mnemonic Default Value Description
2:0 Reserved 000 Reserved. These bits are always 0.
3 OIPHASE[0] 0 When this bit is set to 1, Phase A current generates Bit 17 (OI) in the STATUS1 register.
4 OIPHASE[1] 0 When this bit is set to 1, Phase B current generates Bit 17 (OI) in the STATUS1 register.
5 OIPHASE[2] 0 When this bit is set to 1, Phase C current generates Bit 17 (OI) in the STATUS1 register.
8:6 Reserved 000 Reserved. These bits are always 0.
9 OVPHASE[0] 0 When this bit is set to 1, Phase A voltage generates Bit 18 (OV) in the STATUS1 register.
10 OVPHASE[1] 0 When this bit is set to 1, Phase B voltage generates Bit 18 (OV) in the STATUS1 register.
11 OVPHASE[2] 0 When this bit is set to 1, Phase C voltage generates Bit 18 (OV) in the STATUS1 register.
12 VSPHASE[0] 0 When this bit is set to 1, Phase A voltage generates Bit 16 (SAG) in the STATUS1 register.
13 VSPHASE[1] 0 When this bit is set to 1, Phase B voltage generates Bit 16 (SAG) in the STATUS1 register.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E| Page 85 of 96
Bit
Location Bit Mnemonic Default Value Description
14 VSPHASE[2] 0 When this bit is set to 1, Phase C voltage generates Bit16 (SAG) in the STATUS1 register.
15 Reserved 0 Reserved. This bit is always 0.
Table 42. PHNOLOAD Register (Address 0xE608)
Bit
Location Bit Mnemonic Default Value Description
0 NLPHASE[0] 0 0: Phase A is out of no load condition based on total active/reactive powers.
1: Phase A is in no load condition based on total active/reactive powers. Bit set together with
Bit 0 (NLOAD) in the STATUS1 register.
The ADE7854 no load condition is based only on the total active powers.
1 NLPHASE[1] 0 0: Phase B is out of no load condition based on total active/reactive powers.
1: Phase B is in no load condition based on total active/reactive powers. Bit set together with
Bit 0 (NLOAD) in the STATUS1 register.
The ADE7854 no load condition is based only on the total active powers.
2 NLPHASE[2] 0 0: Phase C is out of no load condition based on total active/reactive powers.
1: Phase C is in no load condition based on total active/reactive powers. Bit set together with
Bit 0 (NLOAD) in the STATUS1 register.
The ADE7854 no load condition is based only on the total active powers.
3 FNLPHASE[0] 0 0: Phase A is out of no load condition based on fundamental active/reactive powers. This bit
is always 0 for ADE7854, ADE7858, and ADE7868.
1: Phase A is in no load condition based on fundamental active/reactive powers. This bit is
set together with Bit 1 (FNLOAD) in STATUS1.
4 FNLPHASE[1] 0 0: Phase B is out of no load condition based on fundamental active/reactive powers. This bit
is always 0 for ADE7854, ADE7858, and ADE7868.
1: Phase B is in no load condition based on fundamental active/reactive powers. This bit is
set together with Bit 1 (FNLOAD) in STATUS1.
5 FNLPHASE[2] 0 0: Phase C is out of no load condition based on fundamental active/reactive powers. This bit
is always 0 for ADE7854, ADE7858, and ADE7868.
1: Phase C is in no load condition based on fundamental active/reactive powers. This bit is
set together with Bit 1 (FNLOAD) in STATUS1.
6 VANLPHASE[0] 0 0: Phase A is out of no load condition based on apparent power.
1: Phase A is in no load condition based on apparent power. Bit set together with Bit 2
(VANLOAD) in the STATUS1 register.
7 VANLPHASE[1] 0 0: Phase B is out of no load condition based on apparent power.
1: Phase B is in no load condition based on apparent power. Bit set together with Bit 2
(VANLOAD) in the STATUS1 register.
8 VANLPHASE[2] 0 0: Phase C is out of no load condition based on apparent power.
1: Phase C is in no load condition based on apparent power. Bit set together with Bit 2
(VANLOAD) in the STATUS1 register.
15:9 Reserved 000 0000 Reserved. These bits are always 0.
Table 43. COMPMODE Register (Address 0xE60E)
Bit
Location Bit Mnemonic Default Value Description
0 TERMSEL1[0] 1 Setting all TERMSEL1[2:0] to 1 signifies the sum of all three phases is included in the CF1
output. Phase A is included in the CF1 outputs calculations.
1 TERMSEL1[1] 1 Phase B is included in the CF1 outputs calculations.
2 TERMSEL1[2] 1 Phase C is included in the CF1 outputs calculations.
3 TERMSEL2[0] 1 Setting all TERMSEL2[2:0] to 1 signifies the sum of all three phases is included in the CF2
output. Phase A is included in the CF2 outputs calculations.
4 TERMSEL2[1] 1 Phase B is included in the CF2 outputs calculations.
5 TERMSEL2[2] 1 Phase C is included in the CF2 outputs calculations.
6 TERMSEL3[0] 1 Setting all TERMSEL3[2:0] to 1 signifies the sum of all three phases is included in the CF3
output. Phase A is included in the CF3 outputs calculations.
7 TERMSEL3[1] 1 Phase B is included in the CF3 outputs calculations.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E| Page 86 of 96
Bit
Location Bit Mnemonic Default Value Description
8 TERMSEL3[2] 1 Phase C is included in the CF3 outputs calculations.
10:9 ANGLESEL[1:0] 00 00: the angles between phase voltages and phase currents are measured.
01: the angles between phase voltages are measured.
10: the angles between phase currents are measured.
11: no angles are measured.
11 VNOMAEN 0 When this bit is 0, the apparent power on Phase A is computed regularly.
When this bit is 1, the apparent power on Phase A is computed using VNOM register instead
of regular measured rms phase voltage.
12 VNOMBEN 0 When this bit is 0, the apparent power on Phase B is computed regularly.
When this bit is 1, the apparent power on Phase B is computed using VNOM register instead
of regular measured rms phase voltage.
13 VNOMCEN 0 When this bit is 0, the apparent power on Phase C is computed regularly.
When this bit is 1, the apparent power on Phase C is computed using VNOM register instead
of regular measured rms phase voltage.
14 SELFREQ 0 When the ADE7878 is connected to 50 Hz networks, this bit should be cleared to 0 (default
value). When the ADE7878 is connected to 60 Hz networks, this bit should be set to 1. This
bit does not have any consequence for ADE7854, ADE7858, and ADE7868.
15 Reserved 0 This bit is 0 by default and it does not manage any functionality.
Table 44. Gain Register (Address 0xE60F)
Bit
Location Bit Mnemonic Default Value Description
Phase currents gain selection.
000: gain = 1.
001: gain = 2.
010: gain = 4.
011: gain = 8.
100: gain = 16.
2:0 PGA1[2:0] 000
101, 110, 111: reserved. When set, the ADE7854/ADE7858/ADE7868/ADE7878 behave
like PGA1[2:0] = 000.
5:3 PGA2[2:0] 000 Neutral current gain selection.
000: gain = 1. These bits are always 000 for ADE7854 and ADE7858.
001: gain = 2.
010: gain = 4.
011: gain = 8.
100: gain = 16.
101, 110, 111: reserved. When set, the ADE7868/ADE7878 behave like PGA2[2:0] =
000.
Phase voltages gain selection.
000: gain = 1.
001: gain = 2.
010: gain = 4.
011: gain = 8.
100: gain = 16.
8:6 PGA3[2:0] 000
101, 110, 111: reserved. When set, the ADE7854/ADE7858/ADE7868/ADE7878 behave
like PGA3[2:0] = 000.
15:9 Reserved 000 0000 Reserved. These bits do not manage any functionality.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E| Page 87 of 96
Table 45. CFMODE Register (Address 0xE610)
Bit
Location Bit Mnemonic Default Value Description
000: the CF1 frequency is proportional to the sum of total active powers on each phase
identified by Bits[2:0] (TERMSEL1[x]) in the COMPMODE register.
001: the CF1 frequency is proportional to the sum of total reactive powers on each phase
identified by Bits[2:0] (TERMSEL1[x]) in the COMPMODE register. This condition does not
have any consequence for the ADE7854.
010: the CF1 frequency is proportional to the sum of apparent powers on each phase
identified by Bits[2:0] (TERMSEL1[x]) in the COMPMODE register.
011: the CF1 frequency is proportional to the sum of fundamental active powers on each
phase identified by Bits[2:0] (TERMSEL1[x]) in the COMPMODE register. This condition does
not have any consequence for the ADE7854, ADE7858, and ADE7868.
100: the CF1 frequency is proportional to the sum of fundamental reactive powers on each
phase identified by Bits[2:0] (TERMSEL1[x]) in the COMPMODE register. This condition does
not have any consequence for the ADE7854, ADE7858, and ADE7868.
2:0 CF1SEL[2:0] 000
101, 110, 111: reserved. When set, the CF1 signal is not generated.
000: the CF2 frequency is proportional to the sum of total active powers on each phase
identified by Bits[5:3] (TERMSEL2[x]) in the COMPMODE register.
001: the CF2 frequency is proportional to the sum of total reactive powers on each phase
identified by Bits[5:3] (TERMSEL2[x]) in the COMPMODE register. This condition does not
have any consequence for the ADE7854.
010: the CF2 frequency is proportional to the sum of apparent powers on each phase
identified by Bits[5:3] (TERMSEL2[x]) in the COMPMODE register.
011: the CF2 frequency is proportional to the sum of fundamental active powers on each
phase identified by Bits[5:3] (TERMSEL2[x]) in the COMPMODE register. This condition does
not have any consequence for the ADE7854, ADE7858, and ADE7868.
100: the CF2 frequency is proportional to the sum of fundamental reactive powers on each
phase identified by Bits[5:3] (TERMSEL2[x]) in the COMPMODE register. This condition does
not have any consequence for the ADE7854, ADE7858, and ADE7868.
5:3 CF2SEL[2:0] 001
101,110,111: reserved. When set, the CF2 signal is not generated.
000: the CF3 frequency is proportional to the sum of total active powers on each phase
identified by Bits[8:6] (TERMSEL3[x]) in the COMPMODE register.
001: the CF3 frequency is proportional to the sum of total reactive powers on each phase
identified by Bits[8:6] (TERMSEL3[x]) in the COMPMODE register. This condition does not
have any consequence for the ADE7854.
010: the CF3 frequency is proportional to the sum of apparent powers on each phase
identified by Bits[8:6] (TERMSEL3[x]) in the COMPMODE register.
011: CF3 frequency is proportional to the sum of fundamental active powers on each phase
identified by Bits[8:6] (TERMSEL3[x]) in the COMPMODE register. This condition does not
have any consequence for the ADE7854, ADE7858, and ADE7868.
100: CF3 frequency is proportional to the sum of fundamental reactive powers on each
phase identified by Bits[8:6] (TERMSEL3[x]) in the COMPMODE register. This condition does
not have any consequence for the ADE7854, ADE7858, and ADE7868.
8:6 CF3SEL[2:0] 010
101,110,111: reserved. When set, the CF3 signal is not generated.
9 CF1DIS 1 When this bit is set to 1, the CF1 output is disabled. The respective digital to frequency
converter remains enabled even if CF1DIS = 1.
When this bit is set to 0, the CF1 output is enabled.
10 CF2DIS 1 When this bit is set to 1, the CF2 output is disabled. The respective digital to frequency
converter remains enabled even if CF2DIS = 1.
When this bit is set to 0, the CF2 output is enabled.
11 CF3DIS 1 When this bit is set to 1, the CF3 output is disabled. The respective digital to frequency
converter remains enabled even if CF3DIS = 1.
When this bit is set to 0, the CF3 output is enabled.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E| Page 88 of 96
Bit
Location Bit Mnemonic Default Value Description
12 CF1LATCH 0 When this bit is set to 1, the content of the corresponding energy registers is latched when a
CF1 pulse is generated. See the Synchronizing Energy Registers with CFx Outputs section.
13 CF2LATCH 0 When this bit is set to 1, the content of the corresponding energy registers is latched when a
CF2 pulse is generated. See the Synchronizing Energy Registers with CFx Outputs section.
14 CF3LATCH 0 When this bit is set to 1, the content of the corresponding energy registers is latched when a
CF3 pulse is generated. See the Synchronizing Energy Registers with CFx Outputs section.
15 Reserved 0 Reserved. This bit does not manage any functionality.
Table 46. APHCAL, BPHCAL, CPHCAL Registers (Address 0xE614, Address 0xE615, Address 0xE616)
Bit
Location Bit Mnemonic Default Value Description
If current channel compensation is necessary, these bits can vary only between 0 and 383.
If voltage channel compensation is necessary, these bits can vary only between 512 and 575.
If the PHCALVAL bits are set with numbers between 384 and 511, the compensation behaves
like PHCALVAL set between 256 and 383.
9:0 PHCALVAL 0000000000
If the PHCALVAL bits are set with numbers between 576 and 1023, the compensation
behaves like PHCALVAL bits set between 384 and 511.
15:10 Reserved 000000 Reserved. These bits do not manage any functionality.
Table 47. PHSIGN Register (Address 0xE617)
Bit
Location Bit Mnemonic Default Value Description
0: if the active power identified by Bit 6 (REVAPSEL) in the ACCMODE register (total of
fundamental) on Phase A is positive.
0 AWSIGN 0
1: if the active power identified by Bit 6 (REVAPSEL) in the ACCMODE register (total of
fundamental) on Phase A is negative.
1 BWSIGN 0 0: if the active power identified by Bit 6 (REVAPSEL) in the ACCMODE register (total of
fundamental) on Phase B is positive.
1: if the active power identified by Bit 6 (REVAPSEL) in the ACCMODE register (total of
fundamental) on Phase B is negative.
2 CWSIGN 0 0: if the active power identified by Bit 6 (REVAPSEL) in the ACCMODE register (total of
fundamental) on Phase C is positive.
1: if the active power identified by Bit 6 (REVAPSEL) bit in the ACCMODE register (total of
fundamental) on Phase C is negative.
3 SUM1SIGN 0 0: if the sum of all phase powers in the CF1 datapath is positive.
1: if the sum of all phase powers in the CF1 datapath is negative. Phase powers in the CF1
datapath are identified by Bits[2:0] (TERMSEL1[x]) of the COMPMODE register and by
Bits[2:0] (CF1SEL[x]) of the CFMODE register.
4 AVARSIGN 0 0: if the reactive power identified by Bit 7 (REVRPSEL) in the ACCMODE register (total of
fundamental) on Phase A is positive. This bit is always 0 for ADE7854.
1: if the reactive power identified by Bit 7 (REVRPSEL) in the ACCMODE register (total of
fundamental) on Phase A is negative.
5 BVARSIGN 0 0: if the reactive power identified by Bit 7 (REVRPSEL) in the ACCMODE register (total of
fundamental) on Phase B is positive. This bit is always 0 for ADE7854.
1: if the reactive power identified by Bit 7 (REVRPSEL) in the ACCMODE register (total of
fundamental) on Phase B is negative.
6 CVARSIGN 0 0: if the reactive power identified by Bit 7 (REVRPSEL) in the ACCMODE register (total of
fundamental) on Phase C is positive. This bit is always 0 for ADE7854.
1: if the reactive power identified by Bit 7 (REVRPSEL) in the ACCMODE register (total of
fundamental) on Phase C is negative.
7 SUM2SIGN 0 0: if the sum of all phase powers in the CF2 datapath is positive.
1: if the sum of all phase powers in the CF2 datapath is negative. Phase powers in the CF2
datapath are identified by Bits[5:3] (TERMSEL2[x]) of the COMPMODE register and by
Bits[5:3] (CF2SEL[x]) of the CFMODE register.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E| Page 89 of 96
Bit
Location Bit Mnemonic Default Value Description
8 SUM3SIGN 0 0: if the sum of all phase powers in the CF3 datapath is positive.
1: if the sum of all phase powers in the CF3 datapath is negative. Phase powers in the CF3
datapath are identified by Bits[8:6] (TERMSEL3[x]) of the COMPMODE register and by
Bits[8:6] (CF3SEL[x]) of the CFMODE register.
15:9 Reserved 000 0000 Reserved. These bits are always 0.
Table 48. CONFIG Register (Address 0xE618)
Bit
Location Bit Mnemonic Default Value Description
0 INTEN 0 Integrator enable. When this bit is set to 1, the internal digital integrator is enabled for use in
meters utilizing Rogowski coils on all 3-phase and neutral current inputs.
When this bit is cleared to 0, the internal digital integrator is disabled.
2:1 Reserved 00 Reserved. These bits do not manage any functionality.
3 SWAP 0 When this bit is set to 1, the voltage channel outputs are swapped with the current channel
outputs. Thus, the current channel information is present in the voltage channel registers
and vice versa.
4 MOD1SHORT 0 When this bit is set to 1, the voltage channel ADCs behave as if the voltage inputs were put
to ground.
5 MOD2SHORT 0 When this bit is set to 1, the current channel ADCs behave as if the voltage inputs were put
to ground.
6 HSDCEN 0 When this bit is set to 1, the HSDC serial port is enabled and HSCLK functionality is chosen at
CF3/HSCLK pin.
When this bit is cleared to 0, HSDC is disabled and CF3 functionality is chosen at CF3/HSCLK pin.
7 SWRST 0 When this bit is set to 1, a software reset is initiated.
These bits decide what phase voltage is considered together with Phase A current in the
power path.
00 = Phase A voltage.
01 = Phase B voltage.
10 = Phase C voltage.
9:8 VTOIA[1:0] 00
11 = reserved. When set, the ADE7854/ADE7858/ADE7868/ADE7878 behave like VTOIA[1:0] = 00.
These bits decide what phase voltage is considered together with Phase B current in the
power path.
00 = Phase B voltage.
01 = Phase C voltage.
10 = Phase A voltage.
11:10 VTOIB[1:0] 00
11 = reserved. When set, the ADE7854/ADE7858/ADE7868/ADE7878 behave like VTOIB[1:0] = 00.
These bits decide what phase voltage is considered together with Phase C current in the
power path.
00 = Phase C voltage.
01 = Phase A voltage.
10 = Phase B voltage.
13:12 VTOIC[1:0] 00
11 = reserved. When set, the ADE7854/ADE7858/ADE7868/ADE7878 behave like VTOIC[1:0] = 00.
15:14 Reserved 0 Reserved. These bits do not manage any functionality.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E| Page 90 of 96
Table 49. MMODE Register (Address 0xE700)
Bit
Location Bit Mnemonic Default Value Description
00: Phase A selected as the source of the voltage line period measurement.
01: Phase B selected as the source of the voltage line period measurement.
10: Phase C selected as the source of the voltage line period measurement.
1:0 PERSEL[1:0] 00
11: reserved. When set, the ADE7854/ADE7858/ADE7868/ADE7878 behave like PERSEL[1:0] = 00.
2 PEAKSEL[0] 1 PEAKSEL[2:0] bits can all be set to 1 simultaneously to allow peak detection on all three
phases simultaneously. If more than one PEAKSEL[2:0] bits are set to 1, then the peak
measurement period indicated in the PEAKCYC register decreases accordingly because zero
crossings are detected on more than one phase.
When this bit is set to 1, Phase A is selected for the voltage and current peak registers.
3 PEAKSEL[1] 1 When this bit is set to 1, Phase B is selected for the voltage and current peak registers.
4 PEAKSEL[2] 1 When this bit is set to 1, Phase C is selected for the voltage and current peak registers.
7:5 Reserved 000 Reserved. These bits do not manage any functionality.
Table 50. ACCMODE Register (Address 0xE701)
Bit
Location Bit Mnemonic Default Value Description
00: signed accumulation mode of the total and fundamental active powers. Fundamental
active powers are available in the ADE7878.
01: reserved. When set, the device behaves like WATTACC[1:0] = 00.
10: reserved. When set, the device behaves like WATTACC[1:0] = 00.
1:0 WATTACC[1:0] 00
11: absolute accumulation mode of the total and fundamental active powers.
00: signed accumulation of the total and fundamental reactive powers. Total reactive powers
are available in the ADE7858, ADE7868, and ADE7878. Fundamental reactive powers are
available in the ADE7878. These bits are always 00 for the ADE7854.
01: reserved. When set, the device behaves like VARACC[1:0] = 00.
10: the total and fundamental reactive powers are accumulated, depending on the sign of
the total and fundamental active power: if the active power is positive, the reactive power is
accumulated as is, whereas if the active power is negative, the reactive power is accumulated
with reversed sign.
3:2 VARACC[1:0] 00
11: reserved. When set, the device behave like VARACC[1:0] = 00.
These bits select the inputs to the energy accumulation registers. IA, IB’, and IC are IA, IB, and
IC shifted respectively by −90°. See Table 51.
00: 3-phase four wires with three voltage sensors.
01: 3-phase three wires delta connection.
10: 3-phase four wires with two voltage sensors.
5:4 CONSEL[1:0] 00
11: 3-phase four wires delta connection.
6 REVAPSEL 0 0: The total active power on each phase is used to trigger a bit in the STATUS0 register as
follows: on Phase A triggers Bit 6 (REVAPA), on Phase B triggers Bit 7 (REVAPB), and on
Phase C triggers Bit 8 (REVAPC). This bit is always 0 for the ADE7854, ADE7858, and ADE7868.
1: The fundamental active power on each phase is used to trigger a bit in the STATUS0
register as follows: on Phase A triggers Bit 6 (REVAPA), on Phase B triggers Bit 7 (REVAPB),
and on Phase C triggers Bit 8 (REVAPC).
7 REVRPSEL 0 0: The total active power on each phase is used to trigger a bit in the STATUS0 register as
follows: on Phase A triggers Bit 10 (REVRPA), on Phase B triggers Bit 11 (REVRPB), and on
Phase C triggers Bit 12 (REVRPC). This bit is always 0 for the ADE7854, ADE7858, and
ADE7868.
1: The fundamental active power on each phase is used to trigger a bit in the STATUS0
register as follows: on Phase A triggers Bit 10 (REVRPA), on Phase B triggers Bit 11 (REVRPB),
and on Phase C triggers Bit 12 (REVRPC).
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E| Page 91 of 96
Table 51. CONSEL[1:0] Bits in Energy Registers
Energy Registers CONSEL[1:0] = 00 CONSEL[1:0] = 01 CONSEL[1:0] = 10 CONSEL[1:0] = 11
AWATTHR, AFWATTHR VA × IA VA × IA VA × IA VA × IA
BWATTHR, BFWATTHR VB × IB 0 VB = −VA − VC VB = −VA
VB × IB VB × IB
CWATTHR, CFWATTHR VC × IC VC × IC VC × IC VC × IC
AVARHR, AFVARHR VA × IA’ VA × IA VA × IA’ VA × IA’
BVARHR, BFVARHR VB × IB’ 0 VB = −VA − VC VB = −VA
VB × IB’ VB × IB’
CVARHR, CFVARHR VC × IC’ VC × IC VC × IC’ VC × IC’
AVAHR VA rms × IA rms VA rms × IA rms VA rms × IA rms VA rms × IA rms
BVAHR VB rms × IB rms 0 VB rms × IB rms VB rms × IB rms
CVAHR VC rms × IC rms VC rms × IC rms VC rms × IC rms VC rms × IC rms
Table 52. LCYCMODE Register (Address 0xE702)
Bit
Location Bit Mnemonic Default Value Description
0 LWATT 0 0: the watt-hour accumulation registers (AWATTHR, BWATTHR, CWATTHR, AFWATTHR,
BFWATTHR, and CFWATTHR) are placed in regular accumulation mode.
1: the watt-hour accumulation registers (AWATTHR, BWATTHR, CWATTHR, AFWATTHR,
BFWATTHR, and CFWATTHR) are placed into line cycle accumulation mode.
1 LVAR 0 0: the var-hour accumulation registers (AVARHR, BVARHR, and CVARHR) are placed in regular
accumulation mode. This bit is always 0 for the ADE7854.
1: the var-hour accumulation registers (AVARHR, BVARHR, and CVARHR) are placed into line-
cycle accumulation mode.
2 LVA 0 0: the VA-hour accumulation registers (AVAHR, BVAHR, and CVAHR) are placed in regular
accumulation mode.
1: the VA-hour accumulation registers (AVAHR, BVAHR, and CVAHR) are placed into line-cycle
accumulation mode.
3 ZXSEL[0] 1 0: Phase A is not selected for zero-crossings counts in the line cycle accumulation mode.
1: Phase A is selected for zero-crossings counts in the line cycle accumulation mode. If more
than one phase is selected for zero-crossing detection, the accumulation time is shortened
accordingly.
4 ZXSEL[1] 1 0: Phase B is not selected for zero-crossings counts in the line cycle accumulation mode.
1: Phase B is selected for zero-crossings counts in the line cycle accumulation mode.
5 ZXSEL[2] 1 0: Phase C is not selected for zero-crossings counts in the line cycle accumulation mode.
1: Phase C is selected for zero-crossings counts in the line cycle accumulation mode.
6 RSTREAD 1 0: read-with-reset of all energy registers is disabled. Clear this bit to 0 when Bits[2:0] (LWATT,
LVAR, and LVA) are set to 1.
1: enables read-with-reset of all xWATTHR, xVARHR, xVAHR, xFWATTHR, and xFVARHR
registers. This means a read of those registers resets them to 0.
7 Reserved 0 Reserved. This bit does not manage any functionality.
Table 53. HSDC_CFG Register (Address 0xE706)
Bit
Location Bit Mnemonic Default Value Description
0 HCLK 0 0: HSCLK is 8 MHz.
1: HSCLK is 4 MHz.
1 HSIZE 0 0: HSDC transmits the 32-bit registers in 32-bit packages, most significant bit first.
1: HSDC transmits the 32-bit registers in 8-bit packages, most significant bit first.
2 HGAP 0 0: no gap is introduced between packages.
1: a gap of seven HCLK cycles is introduced between packages.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E| Page 92 of 96
Bit
Location Bit Mnemonic Default Value Description
4:3 HXFER[1:0] 00 00 = for ADE7854, HSDC transmits sixteen 32-bit words in the following order: IAWV, VAWV,
IBWV, VBWV, ICWV, and VCWV, one 32-bit word equal to 0, AVA, BVA, CVA, AWATT, BWATT,
and CWATT, three 32-bit words equal to 0. For ADE7858, HSDC transmits sixteen 32-bit
words in the following order: IAWV, VAWV, IBWV, VBWV, ICWV, and VCWV, one 32-bit word
equal to 0, AVA, BVA, CVA, AWATT, BWATT, CWATT, AVAR, BVAR, and CVAR. For the ADE7868
and ADE7878, HSDC transmits sixteen 32-bit words in the following order: IAWV, VAWV, IBWV,
VBWV, ICWV, VCWV, INWV, AVA, BVA, CVA, AWATT, BWATT, CWATT, AVAR, BVAR, and CVAR.
01 = for the ADE7854 and ADE7858, HSDC transmits six instantaneous values of currents
and voltages: IAWV, VAWV, IBWV, VBWV, ICWV, and VCWV, and one 32-bit word equal to 0.
For the ADE7868 and ADE7878, HSDC transmits seven instantaneous values of currents and
voltages: IAWV, VAWV, IBWV, VBWV, ICWV, VCWV, and INWV.
10 = for the ADE7854, HSDC transmits six instantaneous values of phase powers: AVA, BVA,
CVA, AWATT, BWATT, and CWATT and three 32-bit words equal to 0. For the ADE7858,
ADE7868, and ADE7878, HSDC transmits nine instantaneous values of phase powers: AVA,
BVA, CVA, AWATT, BWATT, CWATT, AVAR, BVAR, and CVAR.
11 = reserved. If set, the ADE7854/ADE7858/ADE7868/ADE7878 behave as if HXFER[1:0] = 00.
5 HSAPOL 0 0: SS/HSA output pin is active low.
1: SS/HSA output pin is active high.
7:6 Reserved 00 Reserved. These bits do not manage any functionality.
Table 54. LPOILVL Register (Address 0xEC00)1
Bit Location Bit Mnemonic Default Value Description
2:0 LPOIL[2:0] 111 Threshold is put at a value corresponding to full scale multiplied by LPOIL/8.
7:3 LPLINE[4:0] 00000 The measurement period is (LPLINE + 1)/50 seconds.
1 The LPOILVL register is available only for the ADE7868 and ADE7878; it is reserved for ADE7854 and ADE7858.
Table 55. CONFIG2 Register (Address 0xEC01)
Bit
Location Bit Mnemonic Default Value Description
0 EXTREFEN 0 When this bit is 0, it signifies that the internal voltage reference is used in the ADCs.
When this bit is 1, an external reference is connected to the Pin 17 REFIN/OUT.
1 I2C_LOCK 0 When this bit is 0, the SS/HSA pin can be toggled three times to activate the SPI port. If I2C is
the active serial port, this bit must be set to 1 to lock it in. From this moment on, spurious
toggling of the SS/HSA pin and an eventual switch into using the SPI port is no longer possible. If
SPI is the active serial port, any write to CONFIG2 register locks the port. From this moment
on, a switch into using I2C port is no longer possible. Once locked, the serial port choice is
maintained when the ADE7854/ADE7858/ADE7868/ADE7878 change PSMx power modes.
7:2 Reserved 0 Reserved. These bits do not manage any functionality.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E| Page 93 of 96
OUTLINE DIMENSIONS
111808-A
0.50
BSC
BOTTOM VIEWTOP VIEW
PIN 1
INDICATOR
EXPOSED
PAD
PIN 1
INDICATOR
SEATING
PLANE
0.05 MAX
0.02 NOM
0.20 REF
COPLANARITY
0.08
0.30
0.23
0.18
6.10
6.00 SQ
5.90
0.80
0.75
0.70
FOR PROPER CONNECTION OF
THE EXPOSED PAD, REFER TO
THE PIN CONFIGURATION AND
FUNCTION DESCRIPTIONS
SECTION OF THIS DATA SHEET.
0.45
0.40
0.35
0.25 MIN
4.45
4.30 SQ
4.25
COMPLIANT TO JEDEC STANDARDS MO-220-WJJD.
40
1
11
20
21
30
31
10
Figure 90. 40-Lead Lead Frame Chip Scale Package [LFCSP_WQ]
6 mm x 6 mm Body, Very Very Thin Quad
(CP-40-10)
Dimensions shown in millimeters
ORDERING GUIDE
Model1 Temperature Range Package Description Package Option
ADE7854ACPZ −40°C to +85°C 40-Lead LFCSP_WQ CP-40-10
ADE7854ACPZ-RL −40°C to +85°C 40-Lead LFCSP_WQ, 13” Tape and Reel CP-40-10
ADE7858ACPZ −40°C to +85°C 40-Lead LFCSP_WQ CP-40-10
ADE7858ACPZ-RL −40°C to +85°C 40-Lead LFCSP_WQ, 13” Tape and Reel CP-40-10
ADE7868ACPZ −40°C to +85°C 40-Lead LFCSP_WQ CP-40-10
ADE7868ACPZ-RL −40°C to +85°C 40-Lead LFCSP_WQ, 13” Tape and Reel CP-40-10
ADE7878ACPZ −40°C to +85°C 40-Lead LFCSP_WQ CP-40-10
ADE7878ACPZ-RL −40°C to +85°C 40-Lead LFCSP_WQ, 13” Tape and Reel CP-40-10
1 Z = RoHS Compliant Part.
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E| Page 94 of 96
NOTES
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E| Page 95 of 96
NOTES
ADE7854/ADE7858/ADE7868/ADE7878
Rev. E| Page 96 of 96
NOTES
I2C refers to a communications protocol originally developed by Philips Semiconductors (now NXP Semiconductors).
©2010–2011 Analog Devices, Inc. All rights reserved. Trademarks and
registered trademarks are the property of their respective owners.
D08510-0-4/11(E)