Energy Conversion and Management 49 (2008) 3746–3754

Contents lists available at ScienceDirect

Energy Conversion and Management

journal homepage: www.elsevier.com/locate/enconman

A microcontroller-based data-acquisition system for meteorological

station monitoring
S. Rosiek, F.J. Batlles *
Dpto. Física Aplicada, Universidad de Almería, 04120 Almería, Spain

a r t i c l e i n f o a b s t r a c t

Article history: This paper presents a study of feasibility of different existing methodologies linked to field’s data acqui-
Received 29 January 2007 sition from remote meteorological stations. The data transmission serves to collect field’s meteorological
Received in revised form 5 October 2007 information, such as temperature, humidity and radiation. In our study the experimental data is regis-
Accepted 25 May 2008
tered in a weather station located about 100 km from the University of Almeria. Various existing tech-
Available online 18 July 2008
niques are studied, especially Radio, GSM (global system of mobile communication) and GPRS (general
packet radio service). In the result of these studies has been designed a system of field’s data acquisition
Keywords:
(herein referred as Meteologger) which we are going to present in this paper. The system is based on an
Data-acquisition system
Microcontroller
ATmega 16 microcontroller, which scans 8 sensors together at any programmable intervals. This paper
Communication system presents the study of the mentioned project, application and some main characteristics of the prototype
Meteorological data system and its program. We attempt to implement the system, and subsequently present the perfor-
mance of tests regarding the mentioned system. To verify its functioning some comparison of this mea-
surement system with two others commercial data-acquisition system (Campbell and Hobo H8) has been
carried out.
Ó 2008 Elsevier Ltd. All rights reserved.

1. Introduction sulted in the installation of many RES power system all around the
world. A disadvantage of RES systems is a fact that the installation
The knowledge of the solar radiation (in a given place) is funda- cost is still high, so their design optimization is desirable. However,
mental for the suitable location of solar systems, both thermal and such an effort requires detailed knowledge of meteorological data
photovoltaic. It seems to us necessary to analyse the complex of of the place where the system will be installed, because the corre-
factors influencing the variability of the solar radiation as; eleva- sponding energy production is highly influenced by the climatic
tion and orientation of the place and its accessibility. The latter is conditions. Thus, is essential to develop techniques that will aid
often limited by nearby elevations. The importance of the variabil- in assessing the available RES potential at the area of interest
ity of the solar incidental radiation in the topography in certain resulting in minimum system cost and maximum operation reli-
applications is fundamental. In this study, the weather stations ability under intermittent energy production conditions. In many
are located in the zone of complex topography and of difficult ac- cases, meteorological data from many different locations is re-
cess. Thus, the utilisation of electronic system to automate mea- quired in order to evaluate models describing the spatial variability
surements will make easier and more efficient the acquisition of of a RES resource, such as the global radiation, across an extended
meteorological data in that kind of location. Today, microproces- geographic region or to fill missing data because of measurements
sors are used to collect meteorological data and most of them are unavailability, leading to the development of data-collection net-
especially designed for solar energy applications [1]. For example, works. Additionally, because of the yearly variation of the climatic
a microprocessor-based system devoted to the collection of solar conditions, statistical processing of a large volume of data available
radiation data was constructed by Thomas et al. [2]. More recently, from past years is required in order to derive accurate models of
Mukaro and Carelse [3,4] in 1999 also built a microcontroller- the RES resources. Thus, the usually applied data organization in
based data-acquisition system for solar and environmental moni- text files is inefficient and the development of automate Database
toring. Some sort of confirmation and justification for our study Management Systems is indispensable. Such systems typically
we found also in the following scientific work. The rapid evolution consist of microcontroller-based unit for recording the signals of
of renewable energy sources (RESs) during the last two decades re- interest, while the collected data are usually transmitted to PC
for storage and further processing. This type of systems is usually
* Corresponding author. Tel.: +34 950 015914; fax: +34 950 015477. installed in geographically isolated areas, while the acquired data
E-mail address: fbatlles@ual.es (F.J. Batlles). must be distributed to the several remote users. It is possible to

0196-8904/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.enconman.2008.05.029
 S. Rosiek, F.J. Batlles / Energy Conversion and Management 49 (2008) 3746–3754 3747

Nomenclature

MBE mean bias error Vin input voltage of the circuit

RMSE root mean square error VREF selected reference voltage
GSM global system of mobile communication EEPROM electrically-erasable programmable read-only memory
GPRS general packet radio service SRAM static random access memory
RES renewable energy sources kB kilobyte
ADC analog-to-digital converter MB megabyte
USART universal synchronous and asynchronous serial receiver
and transmitter

display all of the acquired data to remote users in real-time cause its oneself creation permits to modify it, both physically
through the Internet [5,6]. and programmatic. This system is versatile and easy transportable.
Nowadays, the classical big power plants for the production of Mentioned advantage of our project makes us hopeful that our sys-
electrical power are supported by micro-generators spread along tem will be welcomed with interest also by common users.
the territory. Due to both economical and ethical aspects, the con- The principal purpose of this project is to design a system of
tribution to the generated power provided by renewable sources is data transmission received from the remote stations network, to
rapidly increasing. In particular, photovoltaic (PV) panels were facilitate significantly the reception of data and to diminish the
introduced for domestic power supplies, radio stations, street cost of maintenance of these stations. Firstly we analyze several
signs, water pumps, etc. Therefore, in these situations, the connec- techniques used for collecting and sending/receiving field’s data
tion with the network of power distribution is often unpractical or with the set goal to obtain real-time remote information and then
even useless, the energy is locally produced and managed (stand we describe the field’s data-acquisition system developed to re-
alone plants). Hence this kind of plants are usually difficult to ceive the information from the sensors.
reach, the presence of maintenance operators should be as low Every system of communication must have emitter, receiver
as possible, also considering the moderate value of the produced and means of transmission. The emitter prepares digital informa-
energy. For these reasons, such plants need to be remotely moni- tion for transmission and then transmits the information down
tored and controlled. The monitoring and control system can be (through) the way of transmission. The receiver detects and trans-
functionally divided in three sub-systems: forms the information in order to visualize it, to file it and to ana-
lyze it. Usually optical fibre, radio transmission or cable are used
 The PV sensor cell, providing an estimation of the radiation inci- for transmission [11]. In our case the conjunction of Meteologger
dent on the PV generator. and GSM Modem was chosen as the emitter, the GSM network as
 The data-acquisition system, that measures the plant interesting the means of transmission and another GSM Modem connected
quantities, characterizes the overall performances, and individu- to a computer—as the receiver.
ate possible problems. Nowadays the research in the field of microsystems is progres-
 The transmission system that manages the communication from sively directed towards smart electronic interface, which provides
and to the monitoring and control system. the ability to perform complex operations. Specially designed elec-
tronic interface for specific applications improve the performances
The use of the GSM standard extends the effectiveness of the of the microsystem and provide user-friendly environment for the
system independently wherever the plants are placed, even far control and communication with it [12].
from the electrical distribution network and from the traditional In the present work a microcontroller-based data-acquisition
and wired telecommunication systems. Due to the low cost and system was devised. This system, described in the next sections,
diffusion of the GSM devices, the transmission system is fairly was designed around the 10-bit microcontroller ATmega 16 and
cheap and it is expected to became cheaper and cheaper [7]. will be applied for weather station monitoring. The specific micro-
Automatic data-acquisition systems are currently used for both controller was selected in order to produce a fast and low cost
monitoring system performance and control of its operation. The prototype. The information from the sensors goes to the microcon-
obtained information can be used to evaluate the plant efficiency troller, where it is processed and later send to an external EEPROM
during long periods and to optimize future systems in terms of per- memory and passed to the GSM Modem every 24 h by RS 232
formance and reliability. Several data-acquisition systems have interface. In the case of failure of the data transmission caused
been developed for use in a wide variety of applications, which in- by a problem of the circuit or GSM Modem remains stored in the
clude measuring, acquisition and processing environmental vari- memory which has a capacity to keep it up to 12 days. The number
ables [8], monitoring and evaluating the performance of PV of sensors and the signal that these sensors provide have been ta-
systems, monitoring the status of batteries for water pumping PV ken into account during the process of designing of this circuit.
systems [9], measuring operational parameters of hybrid photovol- Field test and comparison with two different data-acquisition sys-
taic-Diesel system, etc. A common characteristic of the mentioned tems have been performed. Preliminary results obtained during
data-acquisition systems is the use of data loggers or microcontrol- this experiment are presented as well.
lers for measuring and acquiring the signals and transmitting them
to a PC through serial port RS-232. This type of the data-acquisition 2. Experimental data
system allows to measure typical environmental and system vari-
ables of PV systems (DC current, DC voltage, AC current, AC voltage, In this work we use measurements of global radiation, photo-
energy, power, ambient temperature, solar radiation), as well as synthetically activate radiation PAR, temperature of the air and rel-
the I–V curve of the PV plant. The data are transmitted and stored ative humidity, registered in 15 weather stations. These stations
in a computer through different interfaces [10]. Summoning the are located in the north side of the Natural Park of Sierra Nevada,
above examples we wanted to prove that systems like our Meteo- in Huéneja (Granada). Fig. 1 shows the location of the mentioned
logger have very wide applications, due to the needs of user, be- stations. The altitude of the stations oscillates between 1077 and
3748 S. Rosiek, F.J. Batlles / Energy Conversion and Management 49 (2008) 3746–3754

Fig. 1. The locations of the remote stations in Sierra Nevada.

1670 m, where the maximum distance between stations is of the most suitable communication system, according to logistic (avail-
order of 10 km. The global radiation is moderate with piranómetro able means of communication and operation time) and economical
LI-COR 200-SZ the PAR with piranómetro LI-COR 190, where the considerations [7].
period of sampling is of 2.5 min, using HOBO, models: H8 and H8
PRO as system of data acquisition. 3.1. Remote data transmission
The LI-COR sensors provide a very small signal of the order of
microamperes, whereas the inputs of the HOBO data-acquisition The principal systems of remote transmission are: Radio, Satel-
system require a signal inferior of 2.5 V. Therefore, two interfaces lite, Phone and Wireless Network. Fig. 2 shows a general scheme of
between sensors LI-COR and HOBO had been designed. There were the possibilities for field data collection and its transmission to a
used two amplifiers with the factors of amplification of 343 and 98 central computer. The selection of the transmission system has
respectively. The above amplifiers are supplied by a battery of 6 V to be made after considering the following: quality and coverage
and 7.2 A h. The outputs of the amplifiers are connected to a data- of the service, transmission frequency, quantity of transmitted
acquisition system H8 (4–20 mA, 0–2.5 V DC, 8 bit resolution). This information (in kB), type of data-acquisition system, distances be-
system has a non-volatile EEPROM memory of 32 kB and accumu- tween the stations, interferences, power supplies for all devices,
lates the registered measurements every 2.5 min. price of installation and maintenance, license of transmission fre-
quency (Radio), possibilities of future development. Two practical
2.1. Volume of transmitted information aspects which often limit the selection of the transmission system
are the transmission frequency and the cost of the installed instru-
To initiate and read field data from the metrological stations net mentations. Bearing in mind above considerations, we are going to
is a demanding task, considering the great influence of the low and analyse the advantages and disadvantages of some of these
high temperatures in the mountain range, the capacities of the systems.
data-acquisition system of those stations as well as the importance The Radio has the possibility of sending and receiving a huge
of obtaining the field information in real-time. The volume of data amount of information with reduced costs of the transmission,
to be transmitted is a key figure determining the design of the and it is also a good alternative in the case of the lack of telephone
transmission system. Until now the period of registry of the mea- lines. Its main disadvantage is the difficulty in obtaining permis-
surements has amounted to 2.5 min. This interval has been chosen sion for the transmission frequency and the high price of its
primarily due to the limitations of the data-acquisition systems. installation.
Considering the possibility of changes in the period of the mea- The satellite has a great spatial and temporal coverage and it is
surements, the new monthly volume of the data collected from very useful in places without access to telephone lines, however it
all stations has been calculated. Using a time of sampling of 10 s, is a very expensive method. Its main disadvantage is the high cost
the volume of data collected is of the order of 22 MB. For the same of its installation.
period, estimated time of the data collection is about 2 days which As far as wireless communication techniques (WiFi and WiMax)
shows the importance of installing a data transmission system. are concerned, they provide high speed of transmission and limited
coverage rank of the order 200 meters. The overwhelming advan-
3. Methodology tage of the WiMax technology over WiFi lies in the fact of its great-
er coverage, of the order of 50 km. Nowadays use of mobile phones,
In this section we are going to analyze the different systems of especially GSM/GPRS, is being successfully implemented in the
remote transmission in the first place. Many configurations are field of data transmission [13,14]. Its main advantages are the flex-
possible. The first design step consists in the individuation of the ibility and unlimited reach, although the latter is restricted by the
 S. Rosiek, F.J. Batlles / Energy Conversion and Management 49 (2008) 3746–3754 3749

Fig. 2. Possibilities of field data transmission.

cellular phone penetration rate. Another advantage is the low cost The following is the description of systems which will be used
of the amount of transmitted data, as well as the remote control of for remote collecting and transmitting data from the metrological
the equipment and simple power solutions.Having analyzed vari- stations network down to the Work Place.
ous techniques we suggest that the Radio and the Telephone [15]
offer the most significant advantages. 4. Data acquisition developed system (Meteologger)

3.2. Analysis of data transmission and reception system In this section we are going to present the description of field
data-acquisition system, developed to collect and transmit the
In the last decades transmission of metrological and seismic information from the sensors down to the GSM Modem. We have
information from remote databases has been performed by using shown the preference for this type of technology due to the limita-
the Radio. The telephony has not been an interesting alternative tions of the Radio Modem system. Because of the fact that they
because of the price of calls, but lately with GPRS technology, the need the amplifiers to reconstruct and amplify the signal in order
transmission through Telephone has begun to be popular. This to transmit the information at big distances we opt for this partic-
technology opens new possibilities, which have been analysed in ular technology. In the selection of this technology we also state
this work. As it was mentioned in the discussion above, we are that, in the above project, the location of some stations may
going to use wireless communication techniques to conduct field change, as well as the number of the stations may be subject to
data-collection trials from 15 remote stations located in different alternation.
zones in the mountain range, where the conventional networks Fig. 3 shows the transmission scheme. Considering the advanta-
have no signal. In Table 1 we can see numbered meteorological sta- ges of system and the great interest that has been actually arousing
tions together with their geographic coordinates, altitude and cov- in its development, the following equipment is going to be in-
erage of a mobile operator translated into percentage. stalled: Meteologger: it allows to register information and to send
it across GSM/GPRS, GSM/GPRS Modem.

4.1. Meteologger
Table 1
The characteristics of the 15 stations The following is the description of the field data-acquisition
system developed to transmit the information from the sensors
Station no. Coordinate x Coordinate y Altitude (m) Coverage (%)
down to the GSM Modem. Fig. 4 presents the general scheme of
1 502130 4111284 1670 60 the Data Logger. The information from the sensors goes to the
2 502243 4111212 1647 40
3 502532 4111109 1623 20
microcontroller, where it is processed and later send to an external
4 502334 4111674 1562 60 EEPROM memory and it is passed to the GSM Modem every 24 h by
5 502639 4111461 1568 40 RS 232 interface. In the case of the data failure transmission caused
6 502905 4111461 1537 40 by a problem of the circuit or GSM Modem, it remains stored in the
7 503062 4111460 1505 20
8 503325 4111154 1460 80
9 503111 4111107 1447 20 Table 2
10 503539 4112190 1301 80 The accuracy of the different systems of data-acquisition systems
11 503828 4112280 1276 80
12 503658 4112414 1299 80 Reference voltage 2.5 V
13 505529 4114325 1156 100
Hobo_H8 8-bit Meteologger 10-bit Campbell 13-bit
14 508391 4116254 1077 60
15 506280 4118779 1130 100 Accuracy 10 mV 2.5 mV 333 lV
3750 S. Rosiek, F.J. Batlles / Energy Conversion and Management 49 (2008) 3746–3754

The union ‘‘1” is used for sending the programming from a

programmer (circuit designed especially for a microcontroller AT-
MEGA16) down to the microcontroller ‘‘2”. The power source ‘‘3”
provides the necessary voltage to supply the circuit. We consid-
ered the rank of the operative voltage of all circuit’s elements,
that was put on the stabilizer ‘‘4” that stabilizes the voltage from
power supply to the circuit. The signals gathered from sensors ‘‘5”
are converted to their digital form, inside of the microcontroller
‘‘2”. Afterwards these digital signals are stored in 64 kB EEPROM
memory AT24C512 ‘‘6”. The size of the memory was chosen after
considering the quantity of information obtained with the inter-
val of measurement of 10 s as well considering its possible
change. In order to send the information from the microcontroller
Fig. 3. Scheme of field data transmission through the GSM Modem/GPRS.
up to a GSM Modem using the RS 232 interface, it is necessary to
adapt the output voltage from the microcontroller to the input
voltage of the GSM Modem. A MAX 232 Drivers/Receivers had
been used previously for this task ‘‘7”. Finally the Meteologger
had been connected with GSM Modem through a DB9M conector
‘‘8”.
After establishing the system, some performance tests in rela-
tion to clock of reference have been carried out and the necessity
of the higher accuracy in the unit of time generation has been en-
sured. Thus, the second circuit has been designed. It contains the
additional 8 MHz quartz crystal, located under the microcontroller
‘‘2”, Fig. 5b. During these tests a constant deviation of time was
established. It strictly depends on applied oscillator and limited
possibilities of configuration of the ATmega Timer. Using optimal
settings, the time deviation has been reduced to 0.3 s during one
Fig. 4. Flow of the information from the sensors, across the Data Logger, down to a day. The auto-correction was applied to reduce this deviation
GSM Modem/GPRS.
and consists in adding the constant value of 0.3 s every 24 h to a
current time (cf. Fig. 6).
memory. The number of sensors and the signal that these sensors
provide have been taken into account when designing this
circuit. 5. Embedded developed system
Fig. 5a and b show each of the components of the circuit con-
structed, which consists of the following elements: In the following subsections we are going to describe the prin-
cipal rules with relation to the programming of the prototype sys-
 JTAG Interface ‘‘1” tem [16]. The microcontroller and the external EEPROM memory
 Microcontroller ATMEGA16 ‘‘2” are the essential elements for the designed circuit.
 Power supply ‘‘3”
 Stabilizer LM7805 ‘‘4” 5.1. Microcontroller
 Outputs of the sensors ‘‘5”
 External EEPROM memory AT24C512 ‘‘6” In this paper we use an ATMEL high-performance, low-power
 MAX 232 DRIVERS/RECEIVERS ‘‘7” microcontroller ATmega 16 and it is the most important element
 DB 9M connector ‘‘8” of the whole circuit. Continuing, we are going to describe the most

Fig. 5. (a) Aspect of the Meteologger with internal RC oscillator and (b) aspect of the New Meteologger with quartz crystal.
 S. Rosiek, F.J. Batlles / Energy Conversion and Management 49 (2008) 3746–3754 3751

COMPARISON BETWEEN THE METEOLOGGER WITH INTERNAL RC

OSCILATOR AND METEOLOGGER WITH QUARTZ CRYSTAL
12

10
TIME DEVIATION [s]
8 RCOSCILATOR

QUARTZ CRYSTAL
6

QUARTZ CRYSTAL WITH OPTIMAL

4
SETTINGS

0
0 500 1000 1500 2000 2500 3000 3500
TIME [min]

Fig. 6. The comparison between the Meteologger with an internal RC oscillator with Meteologger with quartz crystal.

important elements of the microcontroller, such as analog-to-dig- nal so that its amplitude increases by over 1000 times and the out-
ital converter (ADC) and the universal synchronous and asynchro- put signal is now measured in volts rather than millivolts. The
nous serial receiver and transmitter (USART). amplification is linear and the output is an exact reproduction of
the input, just changed in respect of the amplitude.
5.1.1. Analog-to-digital converter (ADC) The next step is the analog-to-digital conversion. In the basic
The data transmission collects field data, such as temperature, function of analog-to-digital conversion, as shown in Fig. 7, the
humidity and radiation which are typically analogue and are not analog signal must be changed to a digital code so it can be recog-
yet processed to be suitable for direct input to electronic systems. nized by a digital system that processes the information. The
They must be changed to an electrical quantity—a voltage or a cur- microcontroller used in this paper is equipped with a 10-bit Reso-
rent—in order to interface with electronic circuits. Fig. 7 presents lution ADC converter and 8 Multiplexed Single Ended Input Chan-
the basic functions of analogue to digital conversion for the tem- nels. Resolution is the number of bits that the converter (ADC) uses
perature signal input with the amplitude in millivolt. The first step to represent the analog signal. The higher the resolution, the higher
after obtaining the signals from the sensors is conditioning the sig- the number of divisions into which the range is broken, and

Fig. 7. The basic functions of analog-to-digital conversion.

3752 S. Rosiek, F.J. Batlles / Energy Conversion and Management 49 (2008) 3746–3754

therefore, the smaller changes of voltages can be detected. Fig. 7 the memory had been chosen, considering the quantity of informa-
demonstrates the resolution of the hypothetical 10-bit converter, tion obtained with the interval of measurement of 10 s and its pos-
which can resolve 210, or 1024 different levels. This accuracy can sible change.
be maintained because digital quantities are manipulated and pro-
cessed very rapidly, millions of times faster than analog signals 6. Results and discussion
[11].
The result of the conversion for the channel with only one input In this work we have presented the design, implementation and
is expressed in the following equation: test of the data-acquisition system. The system will be used for
ðV IN Þ  1024 reading and storing information from several stations, located in
ADC ¼ ð1Þ zones of difficult access. The designed circuit and its programming,
V REF
offers an advantage because of the possibility of changing the num-
where ADC—the result of the conversion, VIN—the input voltage of ber of placed sensors, as well as changing the frequency of its mea-
the circuit and VREF—the selected reference voltage. surement. The data transmission collects field data, such as global
Making this equation useful the smallest detectable change in and PAR radiation.
the input voltage has been calculated, which is mainly determined In this paragraph we present the results obtained when using
by the resolution of the ADC converter. A 10-bit ADC converter the developed system of data acquisition. To verify its functioning
with the reference voltage of 2.56 V detects a change of 2.5 mV. some field tests and comparison of this measurement system
against two others commercial data-acquisition systems has been
5.1.2. The universal synchronous and asynchronous serial receiver and carried out. The reference system Campbell has been chosen, as it
transmitter (USART) presents the highest precision due to higher ADC resolution (13
Serial communication is a popular means of transmitting data bit). The second system Hobo H8 has been elected with the 8 bit
between a computer and a peripheral device such as a programma- resolution. The Meteologger was mounted outdoors on a horizon-
ble instrument or even another computer. It uses a transmitter to tal surface alongside a Campbell and Hobo H8 (see Fig. 8). The PAR
send data, one bit at a time, over a single communication line to sensor LI-COR 190 was connected to three different systems of data
a receiver. It is very useful when it is necessary to transfer data acquisition. The same tests with the global sensor LI-COR 200-SZ
over long distances. were performed. Global and PAR horizontal radiation readings
The USART has to be initialized before any communication can from these three instruments were taken simultaneously at
take place. The initialization process normally consists of setting 2-min intervals for intensity levels between 20 and 1200 mV. We
the baud rate, setting frame format and enabling the Transmitter have carried out the comparison of three different systems of data
or the Receiver depending on the usage. The information is sent acquisition in mV unity voltage provided by the sensor, in order to
and received in the asynchronous way (UART). eliminate possible errors caused by a factor of its calibration.
The following step is undertaken to specify the setting frame The three used systems acquired voltage provided by the sensor
format. Each transmitted character is packaged in a character with the different precisions due to resolution of each one (see
frame that consists of a single start bit followed by the data bits, Fig. 9). Measuring a linear ramp signal for example gives the three
the optional parity bit, and the stop bit or bits. A start bit signals graphs, as shown in Fig. 9. Fig. 9a shows a 8-bit discrete represen-
the beginning of each character frame. The order of transmission tation of the input ramp signal, measured without artificial noise
of data bits is imposed from least significant bit (LSB) to most sig- added. The quantization steps are the very marked. To increase
nificant bit (MSB). An optional parity bit follows the data bits in the the resolution, the quantization steps need to be reduced Fig. 9b
character frame. This bit is included as a simple means of error and c.
checking. You specify in advance whether the parity of the trans- In this work an Atmel’s AVR microcontroller with an 10-bit res-
mission is to be even or odd. If the parity is chosen to be odd, olution analog-to-digital converter was used. In most cases 10-bit
the transmitter will then set the parity bit in such a way as to make resolution is sufficient. In some cases higher accuracy is desired
an odd number of 1’s among the data bits and the parity bit. The and if it will be necessary special signal processing techniques or
last part of a character frame consists of 1, 1.5, or 2 stop bits. The an external ADC can be used to improve the resolution of the mea-
parity generator calculates the parity bit for the serial frame data. surement. The ADCs reference voltage and the ADCs resolution
When parity bit is enabled, the transmitter control logic inserts the
parity bit between the last data bit and the first stop bit of the
frame that is sent.
The port series is working at the asynchronous mode and the
format of data for transmission consists fundamentally of 11 bits,
that is 1 bit of start, 8 bits of data, 2 bits of stop.

5.2. The external AT24C512 EEPROM memory

When programming it is necessary to plant the system of trans-

mission and storage of the information. The mentioned microcon-
troller consists of a 1 kB SRAM memory and 512 B of EEPROM
memory, which facilitates the organization of the information to
be transmitted. Doing the calculation of the quantity of obtained
information from each station, we can easily estimate the space
occupied in the memory of the microcontroller. Considering that
the design of the mentioned circuit leaves the possibility of using
an external memory, an EEPROM AT24C512 memory has been
used, which we are going to describe in the following paragraph.
The AT24C512 EEPROM memory organized as 65.536 words of
8 bits each has an endurance of a 100,000 write cycles. The size of Fig. 8. Picture of data-acquisition systems used during experiments.
 S. Rosiek, F.J. Batlles / Energy Conversion and Management 49 (2008) 3746–3754 3753

Fig. 9. Enhancing the resolution.

define the ADC step size. The ADC’s reference voltage, VREF, may be tion system: Hobo H8-8 bit, Meteologger-10 bit, Campbell-13 bit
selected to AVCC, an internal 2.56 V reference, or a reference volt- (cf. Table 2). For both cases the deviation is practically null.
age at the AREF pin. A lower VREF provides a higher voltage preci- Figs. 10 and 11 illustrate measurements of global solar radiation
sion but minimizes the dynamic range of the input signal. If the registered by the Meteologger and the Hobo H8 against to the mea-
2.56 V VREF is selected, this will give 2.5 mV accuracy on the con- surements registered by the Campbell, respectively. Fig. 10 shows
version result for Meteologger and the highest input voltage that is that the majority of the experimental points are located over the
measured equals 2.56 V (Eq. (1)). perfect adjust line 1:1, illustrating the minimal dispersion. We dis-
Following we are also going to analyze the results in meaning of tinguish that this behaviour is similar for all voltage range. Fig. 11
the Root Mean Square Error (RMSE) and Mean Bias Error (MBE).
The RMSE informs us about the dispersion of the experimental
data and it is defined as CAMPBELL / METEOLOGGER
900
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ðX estimated  X measured Þ2
P
RMSE ¼ ð2Þ 800
N
METEOLOGGER_GLOBAL

700
where Xestimated—the data measured by the Campbell, Xmeasured—the
data measured by the Meteologger and the Hobo H8. 600
The MBE informs us about the tendency above the underesti- 500
[mV]

mation of experimental data and it is expressed by the following

400
equation: y = 0,9921x + 0,3377
300 R2 = 0,9999
ðX estimated  X measured Þ
P
MBE ¼ ð3Þ 200
N
100
In Tables 3 and 4 we can see the statistical results of about 7500
measurements from the different data-acquisition systems (Meteo- 0
0 100 200 300 400 500 600 700 800 900
logger, Campbell and Hobo H8) for the global and PAR radiation. The
RMSE error caused by the Meteologger against to the Campbell is of CAMPBELL_GLOBAL [mV]
order 1% and it is minor than the RMSE error caused by Hobo H8.
Fig. 10. The correlation between the Global radiation measured by the Meteologger
We suggest that this diminution due to resolution of data-acquisi- and the Campbell.

Table 3
CAMPBELL / HOBO_H8
The correlation between the RMSE and MBE errors of PAR measured by different 900
systems
800
HOBO-H8_GLOBAL [mV]

PAR (mV)
700
Campbell/Meteologger Campbell/Hobo_H8
600
RMSE% 1.02 1.87
MBE% 0.41 0.36 500

400
y = 0,9992x + 3,0112
300 R2 = 0,9987
Table 4 200
The correlation between the RMSE and MBE errors of Global measured by different
100
systems
0
GLOBAL (mV) 0 100 200 300 400 500 600 700 800 900
Campbell/Meteologger Campbell/Hobo_H8 CAMPBELL_GLOBAL [mV]
RMSE% 1.00 1.99
MBE% 0.70 0.69 Fig. 11. The correlation between the Global radiation measured by the Hobo H8
and the Campbell.
3754 S. Rosiek, F.J. Batlles / Energy Conversion and Management 49 (2008) 3746–3754

Table 5
The content of the format of transmitted data

3 BYTES 1 BYTE 3 BYTES 2 BYTES 2 BYTES 2 BYTES 2 BYTES 2 BYTES

Date Number of station Measurement time PAR radiation GLOBAL radiation Humidity Temperature Battery

Table 6 of the developed system against two others commercial data-

Decoding of the received message acquisition systems (Campbell and Hobo H8) has been carried
Code Meaning out. The measurements registered by the Meteologger present
061107 The date of the transmission 2006-11-07 the RMSE error of order 1% and practically the null deviation. We
1 The number of the station 1 suggest that the resolution of the data-acquisition systems is the
143720 The hour of the average 2:37:20 PM essential influence factor for this error.
1538.34 The PAR radiation 1538.34 lmol/s m2 Considering obtained acceptable results, we point out that the
927.90 The Global radiation 927.90 W/m2
future study in this field should focus on the transmission of infor-
73.54 The humidity 73.54%
34.4 The temperature 34.4 °C mation to remote area of University of Almeria.
6.25 The state of the battery 6.25 V
Acknowledgements

illustrates the dispersion for measurements above 200 mV. For This research had been carried out with the help of the project
measurements of PAR radiation we obtained the similar results. ENE2007-67849-C02-02/ALT financed by the Spanish Ministry of
Finally we analyze the content of the format of the information Education.
measured by the Meteologger, that is essential for its future trans-
mission. Table 5 presents the frame format that consists of the
References
specified date of every measurement, the identification of the sta-
tion, time of measurement and the data concerning radiation, [1] Ameur S, Laghrouche M, Adane A. Monitoring a greenhouse using a
humidity, temperature as well as the additional information about microcontroller-based meteorological data-acquisition system. Renew
Energy 2001;24:19–30.
the power of the battery. Taking into account the huge flexibility of
[2] Thomas A, Ganesan A, Mujeeb SA. Solar irradiance monitor. Renew Energy
the prototype system and emphasizing that the programming 1993;3(6–7):599–606.
leaves the possibility of choosing the interval of the measurement, [3] Mukaro R, Carelse XF. A microcontroller-based data acquisition system for
solar radiation and environmental monitoring. IEEE Trans Instrum Meas
the sampling time of which amounted to 2 min, was performed
1999;48:6.
every 6 min on average. It was estimated that the monthly volume [4] Mukaro R, Carelse XF, Olumekor L. First performance analysis of a silicon-cell
of the collected data is equal to 0.5 MB, taking into account that the microcontroller-based solar radiation monitoring system. Solar Energy
size of the individual data frame is of the order of 17 B. In this work 1998;63(5):313–21.
[5] Koutroulis E, Kalaitzakis K. Development of an integrated data-acquisition
we have presented the design, implementation and test of the system for renewable energy sources systems monitoring. Renew Energy
data-acquisition system and we have come to a main conclusion 2003;28:139–52.
of the availability of ordering the information of each station in [6] Papadakis K, Koutroulis E, Kalaitzakis K. A server database system for remote
monitoring and operational evaluation of renewable energy sources plants.
the following format [17]: Renew Energy 2005;30:1649–69.
[7] Gagliarducci M, Lampasi DA, Podesta L. GSM-based monitoring and control of
061107; 1; 143720; 1538:34; 927:90; 73:54; 34:4; 6:25 photovoltaic power generation. Measurement 2007;40(3):314–21.
[8] Santos A, Lopez Barrio C, Garcia Guerra A. Automatic data acquisition system
It is worth mentioning that the commas and points are marked
for a photovoltaic solar plant. Solar Wind Technol 1986;3(4):259.
only for the convenience of reading and they do not appear in the [9] Benghanem M, Arab AH, Mukadam K. Data acquisition system for photovoltaic
transmitted format. In Table 6 we can observe the meanings of water pumps. Renew Energy 1999;17:385–96.
[10] Forero N, Hernandez J, Gordillo G. Development of a monitoring system for a
data fragments.
PV solar plant. Energ Convers Manage 2006;47:2329–36.
[11] Luecke J. Analog and digital circuits for electronic control system applications:
7. Conclusions using the TI MSP430 microcontroller. Oxford: Elsevier Inc.; 2005.
[12] Asimakopoulos P, Kaltsas G, Nassiopoulou G. A microntroller-based interface
circuit for data acquisition and control of a micromechanical thermal flow
In the first phase of the work we have analyzed different sys- sensor. J Phys: Conf Ser 2005;10:301–4.
tems of field data transmission with their advantages and disad- [13] Andersson Ch. GPRS and 3G wireless applications. New York: John Wiley &
Sons, Inc.; 2001.
vantages. It has been observed the technology GSM/GPRS is the [14] Halonen T, Romero J, Melero J. GSM, GPRS and EDGE performance. Evolution
one that offers considerable advantages in transmitting the infor- towards 3G/UMTS. 2nd ed. England: John Wiley & Sons, Ltd.; 2003.
mation at big distances. Favourable conditions, flexibility and the [15] Voskarides SCh, Pattichis CS, Istepanian R, Michaelides C, Schizas CN. Practical
evaluation of GPRS use in a telemedicine system in Cyprus. In: Proc of the 4th
low costs allowed to develop a data-acquisition system for reading
annual IEEE conf on information technology applications in biomedicine, UK;
and storing information from several stations, located in zones of 2003.
difficult access. In the following phase of the work we have focused [16] Barnet RH, Cox S, O’Cull L. Embedded C programming and the Atmel AVR. 2nd
on the development of data-acquisition system. A microcontroller- ed. New York: Delmar Learning; 2006.
[17] Tseng Chwan-Lu, Jiang Joe-Air, Lee Ren-Guey, Lu Fu-Ming, et al. Feasibility
based circuit for data acquisition has been designed, which scans 8 study on application of GSM-SMS technology to field data acquisition. Comp
sensors together at any intervals programmable. The comparison Electron Agr 2006;53:45–59.