Design and Implementation of Low Cost Electronic

Prepaid Energy Meter

Syed Khizar Ali Zaidi I, Hura Masroor I, Syed Rehan Ashraf I and Ahmed Hassan 2
Department of Electronic Engineering
NED University of Engineering and Technology, Karachi, Pakistan.
(skhizarzaidi,hura_masroor,syed_rehan 173,muhammadahmadhassan)@yahoo.com

LINTRODUCTION
Abstract: - Electronic energy meter has got numerous advantages Our meter has outperformed the electromechanical meter not
over the conventional electromechanical meter and due to this; only in functionality & performance but also in cost and
many countries of the world have switched to electronic metering reliability. It is reliable because it is stable towards
system. But unfortunately Pakistan is still deprived of such temperature changes due to its electronic nature. It is also cost
meters. The paper is based on the final year project of the design effective due to the use of ICs and microcontroller which are
& implementation of prepaid electronic energy meter which we quite cheap. Moreover the energy consumption can be read by
are designing in order to eliminate the problems being faced by
the customer through an LCD display that eventually results in
the Pakistani people. By the introduction of prepaid system in
Pakistan the problem of overcharging and over billing and the an easy reading of the energy by the customer. The smart card
trouble being faced by the customers in paying the bills will be system of billing cut the cost of meter reading as no meter
removed all together. Since our meter is electronic in nature, it readers are required.
has got no moving parts and hence the problem of stability &
accuracy due to temperature changes are solved. Our meter is We have broadly classified our project into two main
also tamper resistant which eradicates the chances of the theft of categories
electricity. • Metering system
• Billing system
Keywords: Microcontroller, Smart-Card, Encryption,
Metering engine•

. ------------------------------------------------------~
PIta.I
loHSide
•
,•
•
- I

+sv •••
••
1 " _-_- .4".·

I
1tt • .. ~ " ~

RIII.nllal
•
! ,•
••
Oillet
; ,. ! .

Lln.V_ge
• •own ~.:.., ~>~ ,••
Cllintlonl
is........................~>'
111101~,,
:

iF ". .i.
...A~ :'ED :

.•
'....." , .

1;-"':";.--:.. . v:.::

!
..., , .. // •
Gain :
C8li"-tion •
------------------------------------------------------~
Fig. 1 The Block Diagram of metering system

978-1-4244-2824-3/08/$25.00 ©2008 IEEE

548
 th
Proceedings ofthe 12 IEEE International Multitopic Conference, December 23-24,2008
ILMETERING SYSTEM
vee
A. Functioning ofthe meter 6V

Since our meter is electronic in nature, it accepts voltage 4V

inputs representing local voltage and current in an electrical
power system and converts them to digital using over
sampling AID converters. An on-chip digital processor
continuously computes the product of the two signals, which is
proportional to instantaneous power. A low-cost fixed-
function (embedded) DSP, with on-chip AID converters, for OV
measurement and computation-and an associated w TINE
microcontroller to handle programming tasks and simple
calculations for communications and display. The DSP is u~
< ACTVIE
continuously converting, sampling, and computing ~
instantaneous and average power.
Fig. 2. On Chip Power Supply Monitoring
With the exception of the analog circuitry in the AID
converters and the reference circuit, all other signal processing
(e.g., multiplication and filtering) is carried out in the digital B .Metering Engine
domain. This approach provides superior stability and
accuracy over extremes in environmental conditions and over The metering engine is given the voltage and current inputs
time. We have used LCD display to show the Kilo watt hour and has a voltage reference. Then samplers and quantisers are
consumption. The meter has a novel fault-detection scheme, followed by an ADC section to yield the digitized equivalents
which both warns of fault conditions and allows it to continue of all the inputs. These inputs are then processed using a
accurate billing despite of fault event. It does this by Digital Signal Processor to calculate the various metering
continuously monitoring both the phase and neutral (return) parameters such as powers, energies etc.
currents. A fault is indicated when these currents differ by
more than 12.5%, and billing is continued using the larger of
1. Active Power Calculation
the two currents.

The voltage signals that we finally get from the current and
III.DESCRIPTION OF EACH MODULE
voltage transducers at CHI and CH2 refer fig.3 are passed
through the two ADC's which converts these signals in digital
As in the block diagram, the meter has a power supply, a domain for superior stability over time & temperature. This
metering engine, a processing and communication engine i.e. a signal is then passed through the HPF to remove any DC offset
microcontroller, other add-on modules such as RTC, LCD present in the signal. We have done this in order to ensure
display, communication ports/modules etc. accurate active power calculation. The voltage & current
signals are then multiplied to get the instantaneous power
A. Power Supply Module which is then low pass filtered (LPF) to get the instantaneous
Weare using a power supply of 5V for the metering system. active power.
We have also employed a power supply monitor which brings 2. Digital-To-Frequency Conversion
the metering IC into an inactive state as soon as the supply
The instantaneous active power is passed through the digital to
goes below 4V, 5% tolerance. We have done this in order to
frequency converter. In this block the power is integrated over
ensure the correct device operation at the power up and power
time to get the frequency & this frequency actually
down condition. Our supply also has the immunity towards
corresponds to the average active power. The active power is
false triggering of the device as we have employed hysteresis
then finally converted to energy by counting the number of
monitoring as well.
pulses over a specified time period.

549
 Proceedings ofthe 12th IEEE International Multitopic Conference, December 23-24, 2008
Fl D. RTC and other add-on modules
F2

These are attached as slaves to the processing and

communication section for various input/output functions. In
our meter, most if not all of this is implemented inside the
microprocessor, such as the Real Time Clock (RTC),
temperature sensor, memory and analog to digital converters.
INSTANTANEOUS ACTIV POWER SIGNAL
POWER SIGNAL-P(t)
Vxl IV.THEFT RESISTANT ENERGY METER
VxI
p(t)=I(t)XV(t)
--- WERE, 2
............--........
Our electronic energy meter is designed as a tamper resistant
V(t)=v cos (Wt) energy meter by monitoring current in both the phase and
I(t)=I cos (Wt) return path. In our energy meter, we are monitoring both phase
TIME P(t)=(VxI)12(1+c:os(2wt» and neutral and are calculating real average power based on
the larger of the two currents thus eliminating most pilfering
schemes that tamper with the current path through the energy
Fig. 3. Signal Processing Block Diagram meter. A fault is indicated when the phase and neutral currents
. - - ~ differ by more than 12.5%. Billing is continued using the
microcontroller and using the controller for counting purpose. larger of the two currents.
The output of the controller will provide us with the energy by
the following formula.
Jre ~ JX1MYpqu1i<ml to tre avaagefia:p.m;y, is
AverngeFIaf.U£Y=AverngePCMTer=Coonter (I)
Tnrrr
Jre eJrJgyCCflillOUl~anintegmtimIXIiOOis

Cf
JlJ1JL
Fig. 5. Neutral Missing system

V.BILLING SYSTEM
fA1.1.J2
Our billing is done through a prepayment system that is using
smart card technology for the billing purpose. The basic
Fig. 4. Interfacing To Micro controller For Energy Metering modules in our billing system are

C. Processing and communication section (microcontroller) • Encoder

• Smart card
• Encryption
This section has the responsibility of calculating the various
derived quantities from the digital values generated by the
metering engine. This also has the responsibility of A. Encoder
communication using various protocols and interface with
other add-on modules connected as slaves to it. The metering In our prepayment system, the customer has to buy the credit
IC, LCD display and the smart cart chip are all attached with using smart card technology. This is done by charging the
the micro controller. smart card. In this system, the smart card will be charged by
any service provider shop in which smart card reader writer
has been installed. So the shopkeeper will enter the amount

550
 Proceedings ofthe 12th IEEE International Multitopic Conference, December 23-24,2008
say 200 ~pees to charge ~he card. This amount will finally go The purpose of Rb is to convert the current coming from CT to
to the mlcrocontroller U?It. SO how the MCU will understand the voltage and send this voltage to the metering IC. Its value
that the ente~ed amount IS 200 :upees? Ac.tual~y there is a code is selected keeping in view the maximum current i.e. 40A. To
for very declI~al number that IS entere~ In. bInary format a~d achieve the best results 233.3 mV i.e. half of the rms input
the MCU basIcally understands only t?IS bInary format. So In voltage applied to VI is selected. If2500:1 is the CT tum ratio
order to convert the entered amount Into the format or code then
which is understood by the MCU, the encoder is needed.
40A rms/2500 turns = 16m A rms (3)
B. Smart Card Rb= 233.3m V/16m Arms = 14.58 n (4)

This resistance monitors the load current similarly another

The most significant module of our billing system is the smart
resistance is inserted in CT2, which monitors the neutral
card that has been used for writing and reading purpose. We
current.
have developed a separate programmer for reading/writing
purpose. The units (say 200) that have been told by the
customer to the shop keeper to be loaded on the card are
C. CT Design
written on the card by the help of smart card writer. Similarly
The current transformers for this design should be selected to
when the customer inserts this card on the meter, the units that
maximize the use of the dynamic range on Channels VIA and
have been loaded on the card is read by the controller in the
VIB. There are several important considerations when picking
meter by the help of the smart card reader. Actually when the
a current transformer for the application. First, a CT that is
card is inserted in the meter, the units enter into the memory of
selected must be linear over the dynamic range of interest. The
the microcontroller and are used for energy consumption. Now
burden resistance should be as small as possible to preserve
as the energy is consumed, the units start decreasing which is
linearity over a large current range. Therefore a burden
displayed on the LCD. As soon as 75% of the energy is
resistance of less than 15 n is used for this application.
consumed, an LED gets on indicating the user that the units
are about to get finished. If the user fails to recharge the card,
Second, the CT introduces a phase shift between the primary
and all the units are consumed, a solid state relay will operate
and secondary current. This phase shift should be kept to a
and cut off the supply of electricity. If after this, the user
minimum. In this design, the phase of the voltage channel is
recharges the card, then again the connection is resumed by
shifted to match the phase shift in the current channel caused
the relay and electricity supply is provided.
by the current transformers.
C. Encryption

VII.CALIBRATION MODES OF THE METER

In order to prevent the unauthorized use of electricity by the
customer, we have added encryption feature in our meter.
Through the use of encryption, we have converted our original A. Normal Mode
information (the units to be charged) into a code word that
cannot be easily interpreted by the user. We have employed
In this mode, the current is sensed by 2 CT'S and this current
RSA algorithm for encrypting our meter. The public key is
is converted to voltage by a burden resistor (Rb). An
given to the user and the code can only be decrypted by the
attenuation network made up of series resistors is employed to
private key which is only known to the service provider.
attenuate the input voltage to the desired value.

VI.DESIGN ISSUES
Calibration
In this mode, the signal to be calibrated is the line voltage.
A. Power supply Design
First the load current is set to lb. Then we have used the
attenuation network of resistors to achieve the desired
In order to design the supply, it is important to investigate
frequency at output CF.
what amount of current the meter will draw. The supply
current for the metering IC is less than 4mA. The counter
CF = 220*5mA*0.027777*32 = 0.9777 (5)
drive requires a maximum of lmA average. An additional
current is needed for the LED, which is 4mA, or less. An extra
We have made jumpers by shorting out two solder pads. This
milliamp is allowed for the voltage regulation quiescent
approach is preferred over the use of trim pots because of
current, so the total power needed for supply is 9mA.
reliability over time and if CF increases when any jumper is
closed, it should be opened again.
B. Burden resistor calculation (Rb)

551
 Proceedings ofthe 12th IEEE International Multitopic Conference, December 23-24, 2008

240VAC
R A A R
'V
'. }IISCAL
R R R R

J
J J J
._-_._.-
., ........~.'.'..'::.:'.'::..~','~',','::::. .."._.-""' - '

R R R
"',.JIJa.,,''''.;;:'.,,,
364mV ... ,&A .. """,..... ,.......

J J J

Fig. 6. Line Attenuation Network

J J J J J
B. Ground r'ault Mode
This is the mode where the neutral is either disconnected or Fig. 8. MISCAL Attenuation Network
grounded. If this is the condition then current carried by CT I
will be different from that of CT2. But we have designed our C. Missing Neutral Mode
meter in such a way that it will take the larger of the two
currents for power calculation. In this mode, the ground is completely disconnected & the
current flows only in the phase i.e. here the line voltage is o.
Calibration
Calibration
The first step is to check the two CT'S if they are carrying
equal current or not. For calibration, VIB is disconnected, For the calibration in this mode we've employed a MISCAL
with Ib on VIA and note CF. Then we disconnect VIA with Ib attenuation network which is approximately the same network
on V IB and again note CF. We have then matched the later as that employed for normal mode. The only difference is that
CF with the former one by varying R26 & R27 as shown in here we've to calibrate MISCAL voltage instead of line
figure 7. voltage by shorting the jumpers to achieve the required CF as
shown in figure 8.

233mV
VlA REFERENCES

[1]. Zhiqun Chen, Java Car Technology For Smart Cards - Architecture
And Programmers Guide, Addison-Wesley, 2000.
[2]. W. Rankl, W.Effing , "Smart Card Hand Book", John Wiley And Son,
1997.
233mV VlB [3]. Micro Chip Data Book, Second Edition, July 1992.
[4]. Muhammad Ali mazidi , The 8051 Microcontroller and Embedded
Systems, Second Edition,

Fig. 7 Current Input Network

VI.TEST RESULTS
Load
Current Cos Load Frequency %
Voltage (V) (KWh)
(I) <I> (KW) Calculated Observed ERROR
Observed
1 212 7.37 1 1.56244 0.04340 0.04761 1.713 9%

2 210 3.47 1 0.7287 0.02024 0.02222 0.79992 9.11%

3 211 0.95 1 0.20045 0.005568 0.006211 0.22359 11.54%

4 209 5.40 1 1.1286 0.03135 0.03393 1.22148 8.2296%

%ERROR- <Energy measured by IC - True Energy) XIOO

True Energy

552