DEVELOPMENT OF AN EMI MEASUREMENT SYSTEM Performance Analysis of Bluetooth communication under noise environment

Transcription

1 DEPARTMENT OF TECHNOLOGY AND BUILT ENVIRONMENT DEVELOPMENT OF AN EMI MEASUREMENT SYSTEM Performance Analysis of Bluetooth communication under noise environment Javier Ferrer Coll, Félix Pérez Castelló Sept 2008 Master s Thesis in Electronics/Telecommunications Master s Programme in Electronics/Telecommunications Examiner: Claes Beckman Supervisors: Jose Chilo, Peter Stenumgaard

2 1

3 AKNOWLEDGEMENTS First of all we want to thank the people, who have contributed to this work and supported us during the thesis, specially our supervisors Jose Chilo and Peter Stenumgaard. But we can t forget to be grateful to Claes Beckman and Per Ängskog for all the help offered when we needed it. We have to admit the collaboration of the Centre for RF Measurements Technology of Gävle, where we were working full-time. We would also like to thank our mates from Sätra, who offered us a good year and very good moments, and with their support the work for the thesis went smoother and softer, specially Milena Manceva. And finally we are very grateful for the aid of our family and their trust in us. 2

4 ABSTRACT This thesis is a project carried out at the Centre for RF Measurements Technology of Gävle. The first aim of this work was basically to develop an EMI measurement system, to that purpose, it has been used an EMI Tester receiver, Spectrum Analyzer and a broadband antenna. Tables and graphics are shown to provide the values of the different detectors utilized. Using this measurement system, an interference file was recorded and then inserted in a Bluetooth communication model. The interference file was simulated with Matlab Simulink, to check how the interference affected the communication; the effects of the signal degradation are presented in a graphic. Finally a real Bluetooth communication was established using two Bluetooth modules from Free2Move Company, to prove that the effect of microwave oven interferences produces the increase of transmission time and therefore decrease the Throughput. 3

5 TABLE OF CONTENTS 1 Introduction Goals Theory EMI Measurement Introduction Antennas The Antenna Factor Polarization, Polar Pattern, and Distance Detectors used in EMI measurements Receiver Specifications per CISPR : Bluetooth Introduction Bluetooth Transmission Technology Frequency Hopping Spread Spectrum Radio Characteristics Modulation Characteristics Time Slots Packets Error Correction Bluetooth Version Results EMI Test receiver Vs Spectrum Analyzer Simulink, Bluetooth communication under noise environment simulations Bluetooth 1.0 with different interferences Bluetooth Model version Real Bluetooth communication under noise environment simulations 33 5 Conclusions

6 6 References APPENDIX APPENDIX A: EMI TESTER MEASUREMENTS APPENDIX B: M-FILES USED APPENDIX C: BiLog ANTENNA 30MHz - 2GHz CBL 6112B APPENDIX D: SIMULINK MODELS APPENDIX E: F2M03GLA BLUETOOTH MODULES APPENDIX F: EVALUATION KIT FOR GENERAL PURPOSE BLUETOOTH MODULES DATASHEET APPENDIX G: MODULES AND MICROWAVE OVEN EXPERIMENT

7 LIST OF ABBREVIATIONS 8DPSK (8-Differential Phase Shift Keying) AC (Alternating Current) ACL (Asynchronous Connection-less Link) AF (Antenna Factor) ARQ (Automatic Repeat Request) BER (Bit Error Rate) C/I (Carrier to Interference ratio) CISPR (Comité Internationale Spécial des Perturbations Radioelectrotechnique - International Special Committee on Radio Interference-) CRC (Cyclic Redundancy Code) CW (Continuous Wave) DC (Direct Current) DQPSK (Differential Quadrature Phase Shift Keying) EDR (Enhanced Data Rate) EMC (Electromagnetic Compatibility) EMI (Electromagnetic Interference) EN (European Norm) EUT (Equipment under Test) FCC (Federal Communications Commission-USA-) FEC (Forward Error Correction code) FHSS (Frequency Hopping Spread Spectrum) FM (Frequency Modulation) FSK (Frequency Shift Keying) FTDI (Future Technology Devices International Ltd) GFSK (Gaussian Frequency Shift Keying) GPIB (General Purpose Interface Bus) GSM (Global System for Mobile Communications) HEC (Header Extension Code) 6

8 IEV (International Electrotechnical Vocabulary) IF (Intermediate Frequency) IMT-2000 (International Mobile Telecommunications-2000) ISM (Industrial, Scientific and Medical) MIL-STD (Military Standard) NFC (Near Field Communication) PSK (Phase Shift Keying) RBW (Resolution Bandwidth) RF (Radio Frequency) RFI (Radio Frequency Interference) RMS (Root Mean Square) SCO (Synchronous Connection-Oriented Link) SIG (Special Interest Group) TDD (Time Division Duplex) TTL (Transistor-Transistor Logic) UART (Universal Asynchronous Receiver-Transmitter) UMTS (Universal Mobile Telephone System) USB (Universal Serial Bus) WiFi (Wireless Fidelity) WLAN (Wireless Local Area Network) 7

9 1 INTRODUCTION Nowadays the EMC (electromagnetic compatibility) has a really big importance in the market. EMC studies the unintentional, generation, propagation and reception of electromagnetic energy with reference to the unwanted effects such as energy may induce, this energy is known as EMI (Electromagnetic Interference). All devices need to carry out the standard corresponding for the product which is controlled by international standards. The regulator commission for EMC is CISPR (Comité International Spécial des Perturbations Radioélectriques), wich imposes the rules to be reliable for the market. The thesis is going to show the steps and results of the process to measure the impulses in a laboratory environment. The measures were taken in Centre for RF Measurements Technology of Gävle. EMI measurement system could be used to check the problems that wireless communications can have in an industrial environment and also, to give an idea to stall a new wireless systems in companies and hospitals where sometimes the wireless communications suffers a hard degradation. The system used to measure the Impulses was composed by EMI Tester, Spectrum Analyzer and BiLog Antenna. The first step was to connect, configure and install the software necessary for the measure. Secondly, was to measure the interference included from 30MHz to 2GHz (range of the BiLog Antenna). The EMI Tester software provides a graph of the spectrum in different detectors and a list of the highest peaks. The second part in the thesis was to simulate a Bluetooth communication under noisy environment. The tools Simulink from MatLab were used to simulate a communications with different modulations or systems. These systems were simulated with the points that the EMI Tester measures. A Bluetooth model from Carl Karlsson [1] was utilized and improved. The original model from Carl consists of transmitter, receiver and channel interference; in the new model several noises were inserted in the channel to study the behavior with different interferences as Microwave Oven, GSM, WiFi and EMI measurements. On the other hand, the version of the Bluetooth model provided by Carl Karlsson [1] was 1.0, however the last version of Bluetooth is 2.1 and it was developed in Simulink. The third part of our work is in relation with interferences which are produced in a real Bluetooth communication. By using two Bluetooth models from Free2Move and Microwave Oven a system was carried out, which shows how the Microwave Oven affects the communication of theses Bluetooth modules. These two devices work in ISM-band. This band covers the frequencies around 2.45GHz and it s free access causes many applications to use it. The two modules were configured to transfer a file. The time of the transmission was compared with the measured time by using a Microwave Oven as interference. There are some conclusion about the throughput and how this interference affects the Bluetooth communication. The report is presented in three sections: the theory to understand the basic concepts, the work made where the results are shown, the conclusion and finally some future applications to follow the research started in this thesis. 8

10 2 GOALS The main goal of this thesis was to design a system which is able to measure an Electromagnetic Interferences in industrial and other environments. To develop the EMI measurement system, EMI Tester receiver, Spectrum Analyzer and a broadband antenna were used. The EMI Tester was made to measure especially EMI emissions; on the other hand the Spectrum Analyzer has a general purpose. The results of these measurements are shown with tables and graphs obtaining the interferences found in the margin with regard to the CISPR they have with maximum power levels. Using our system a file with noise has been created within the frequency band and then used in the simulation model. The spectrum obtained by the EMI Tester and Spectrum Analyzer is compared and added to simulated system in Simulink. The model from Simulink consists of a Bluetooth 1.0 communication. The measurements from the EMI Tester and other interferences are added in the channel checking what happens in the reception. The Bluetooth model is improved to version 2.1 by changing the modulator and completing the components necessary for the channel. With regard to interferences and Bluetooth, to check the communication between two real Bluetooth modules, interference was created using a Microwave Oven. The goal of this section is, to prove if the Oven affects the Bluetooth communication. 9

11 3 THEORY This section presents the knowledge which is necessary to understand the obtained results. The first point consists of the explanation of EMI theory, the principal concepts about EMI, definitions and classification of EMI. After that there is an exposition of the antenna used, features, polarization and antenna factor. The detectors utilized to measure are explained and differentiated by how they work and what limits for the different frequencies they have. The second point of the theory exposes the basic concepts of Bluetooth as: modulation, radio characteristics, Frequency Hopping Spread Spectrum and error correction. All of the above mentioned is going to help to understand the behaviour of the results obtained. 3.1 EMI Measurement Introduction EMC (Electromagnetic Compatibility) studies the unintentional, generation, propagation and reception of electromagnetic energy with reference to the unwanted effects such as energy may induce. This energy is known as EMI (Electromagnetic Interference, also called radio frequency interference or RFI). There are some kinds of interferences that can affect the RF, natural interferences, like solar radiation, electrical storm or atmospheric fields, and the artificial interferences made from human. The artificial interferences can be divided in conducted and radiated. The conducted interferences are generated by conductors and the radiated by electromagnetic fields. There are two different radiated emissions, inductive fields, nearby sources, and radiation fields, distant sources. Opposed to conducted interferences the radiated interferences could be found in all space, therefore is difficult to know the direction of radiated fields, so to obtain the interference strength an antenna is needed that could measure the field in all the directions. In our case we are interested in these kind of interference. Another classification for interferences deals about the features of the signal. There are two important categories: narrowband signals and broadband signals. The International Electro technical Vocabulary (IEV) defines a narrowband disturbance as an electromagnetic disturbance, or component thereof, which has a bandwidth less than or equal to that of a particular measuring apparatus, receiver or susceptible device. Consequently, a broadband disturbance is defined as an electromagnetic disturbance which has a bandwidth greater than that of a particular measuring apparatus, receiver or susceptible device. This means that the classification of a signal as narrowband or broadband is determined by the occupied frequency spectrum of the signal under investigation, related to the resolution bandwidth (RBW) of the instrument used for measurement. If the signal spectrum is completely contained in the pass band of the IF filter, it is defined as a narrowband signal. See figure

12 Figure 3.1. Generic definition of narrowband and broadband signals EMI is a temporary unstable signal, an impulsive noise which features, so like narrowband or unpredictable amplitude power are difficult to be measured, but possible when different detectors have been used to obtain the desired results. These detectors will be commented after Antennas The standard that defines the requirements for antennas to be used in EMC measurement is publication 16-1: 1993, Specification for Radio Disturbance and Immunity Measuring Apparatus and Methods, Part 1. [2] CISPR measurements officially require tuned dipoles, but a note in most standards allows the use of broadband antennas where they can be shown to give equivalent results. Then in the reception with broadband antennas must be considerate the Antenna Factor (AF), which will be explained in the next section. In addition, the antenna polarization and the conditions to obtain the maximum gain will be described. Historically, two types of antenna have been used for emissions measurement, biconical and log-periodic. These are electric-field linear polarized and typically cover frequency ranges from 30 to 300 MHz and 300 to 1000 MHz, respectively. Early biconical designs could reach only 200 MHz, but a modification to the structure has removed a resonance between 200 and 300 MHz, allowing that specification to be stretched. The two types can also be combined into one device that will exhibit the characteristics of each, increasing the relevant frequency range. This device is called BiLog antenna. These different types of antenna are shown in the following figure: 11

13 Figure 3.2. Common antenna types are a) BiLog, b) log-periodic, and c) biconical The Antenna Factor Antenna Factor (AF) is perhaps the most widely used device descriptor in the EMC area. However, it is one that is definitely not part of standard antenna terminology. Antenna Factor reflects the use of an antenna as a field measuring device or probe. Succinctly stated, the antenna factor is the factor by which one would multiply the output voltage of a receiving antenna to obtain or recover the incident electric field. Thus, the electric field Antenna Factor is given by: The antenna factor includes losses and mismatch in the antenna and its associated equipment (such as a balun or matching transformer). However, it does not account for the use of an intervening transmission line (such as coaxial cable) to connect the antenna to the receiver. So the losses in the transmission line can be easily accounted for with a multiplicative factor, see the equation. Where C A =e αl is the loss factor of the transmission, α is the attenuation of the cable in nepers/meter and l is the length of the cable in meters. (1) (2) Polarization, Polar Pattern, and Distance When a single antenna factor is specified, an assumption has been made that the antenna will be used under conditions of maximum gain. For the log-periodic antenna, this is in the direction toward which the antenna is pointing, while for the biconical, it is perpendicular to the antennas axis. In all other directions, the response of the antenna falls off and the antenna factor becomes invalid. The polar-pattern response for a dipole is within 1 db of the on-axis value over an azimuth variation of 45 ; for a log-periodic array, the beam is narrower. This is particularly significant when the antenna is used at high frequencies with a height scan from 1 to 4 m and a close-in distance of 3 m. Under such conditions, the antenna will no longer be properly aligned with the EUT (Equipment Under Test), and an error may result. 12

14 Polarization of the antenna refers to the plane of polarization of the electric-field component. CISPR 16 requires that the cross-polarization be better than 20 db, which implies that the design of the antenna must ensure linear polarization. EMC testing requires a fixed and known distance between the antenna and the EUT. In log-periodic or combination antennas, the active element (known as the phase center) shifts with frequency, and so the measuring distance must change. It is therefore a practical necessity to choose a specific point on the antenna boom against which the AF should be calibrated and to mark this permanently on the antenna itself. The inclination of the antenna could be vertical or horizontal. The fields with vertical polarity are greater than the fields with horizontal polarity. Then choice of the inclination must be vertical. The reason why the horizontal component is lower than the vertical component is caused by the effect of the ground plane, so the floor absorbs a part of horizontal component. This is the reason to put the antenna in vertical orientation Detectors used in EMI measurements There are many different types of detectors in use in signal analysis systems. Each has a unique definition as well as differing advantages and disadvantages. And choosing the right one is critical to obtaining valid data. The following points explain different detectors used in EMI measurements, these are Peak, Quasi-Peak, Average and RMS (Root Mean Square). Peak Detector: Initial EMI measurements are made using the peak detector. This mode is much faster than quasi-peak or average modes of detection. Signals are normally displayed on spectrum analyzers or EMC analyzers in peak mode. Since signals measured in peak detection mode always have amplitude values equal to or higher than quasi-peak or average detection modes, it is a very easy process to take a sweep and compare the results to a limit line. The EMC analyzer has an envelope or peak detector in the IF (Intermediate Frequency) chain which has a time constant such that the voltage at the detector output follows the peak value of the IF signal at all times. In other words, the detector can follow the fastest possible changes in the envelope of the IF signal, but not the instantaneous value of the IF sine wave. The peak detector mode calculates the maximum magnitude at each discrete spectral value, thus: Quasi-Peak Detector: Most radiated and conducted limits are based on quasi-peak detection mode. Quasi-peak detectors weigh signals according to their repetition rate. (3) 13

15 As the repetition rate increases, the quasi-peak detector does not have time to discharge as much resulting in a higher voltage output. For CW (Continuous Wave) signals the peak and the quasi-peak are the same. Quasi-peak measurements are much slower by 2 or 3 orders of magnitude compared to using the peak detector. The quasi-peak detector has a charge rate much faster than the discharge rate; therefore the higher the repetition rates of the signal the higher the output of the quasipeak detector. The quasi-peak detector also responds to different amplitude signals in a linear fashion. High amplitude low repetition rate signals could produce the same output as low amplitude high repetition rate signals. See Figure 3.3. Figure 3.3. Quasi-Peak detector response diagram The Quasi-Peak detector mode evaluates the emission according to a physiological disturbance against amplitude-modulation radio. Average detector: The average detector is required for some conducted emissions tests in conjunction with using the quasi-peak detector. Also, radiated emissions measurements above 1 GHz are performed using average detection. The average detector output is always less than or equal to peak detection. Average detection is similar in many aspects to peak detection. The output of the envelope detector is the modulation envelope. Peak detection occurs when the post detection bandwidth is wider than the resolution bandwidth. For average detection to take place, the peak detected signal must pass through a filter whose bandwidth is much less than the resolution bandwidth. The filter averages the higher frequency components, such as noise, at the output of the envelope detector. The average detector mode calculates the mean spectrum from the spectrogram. The formulation for this detector is: (4) RMS Detector: RMS amplitude measurement is the best way to relate AC quantities to DC quantities, or other AC quantities of differing waveform shapes, when dealing with 14

16 measurements of electric power. The RMS detector output is always less than or equal to peak detection, and the value is always the same as or just a little bit larger than the average. The RMS detector mode calculates the RMS value of the magnitude of the spectrogram as follows: (5) As it is described in the equation, to determine RMS value, three mathematical operations are carried out on the function representing the AC waveform; The square of the waveform function (usually a sine wave) is determined, the function resulting from this step is averaged over time and finally the square root of the averaged function. The figure 3.44 below is a comparison between Peak, RMS and Average detectors; it is possible to see the difference of amplitude in each detector depending on the waveform of the signal to detect. RMS = AVG = PK = 2 RMS = Peak AVG = Peak PK = 2 RMS = AVG = 0.5 PK = 2 Figure 3.4. Amplitude difference in each detector used Receiver Specifications per CISPR : Most commercial EMI standards reference CISPR as the standard defining the specifications of EMI receivers where the input, impedance, detector characteristics and IF bandwidth shapes are specified. The current version of CISPR calls out receiver specifications for the frequency range from 9 khz to 18 GHz 15

17 where the realization of the bandwidths, the provision of the required dynamic range for the quasi-peak detector or overload protection is not defined. CISPR is a system specification which defines the response of a receiver to defined input signals. Resolution Bandwidths: Frequency resolution is the ability of an EMI receiver to separate two input signals into distinct responses on the display. Specific resolution bandwidths are called out for measurements in different frequency ranges. In general, receiver IF filters are usually specified by a bandwidth and additional information about its frequency response. CISPR references the 6 db bandwidth values of three IF filters to be used in the frequency range to 2 GHz, these filters, in each range of frequency use, are: 200 Hz (for 9 khz to 150 khz) 9 khz (for 150 khz to 30 MHz) 120 khz (for 30 MHz to 2 GHz) Devices classification and limits: The CISPR classified the devices under testing in two groups depending on their use, class A and class B: Class A: A device which is marketed for use in an industrial application and is not intended for use in the home or residential area. Since the product is being sold to a commercial market, the emissions limits are significantly less stringent than Class B (residential) devices. Products that fall under the category of Class A do not require an official submittal, but simply need a Verification test performed and the data must be keep on hand by the manufacturer. Class B: That one is a device which is marketed for use at home or in a residential area by the customer. Class B devices can require Verification, Certification, or Self Declaration depending on the type of product. Depending on the range of frequency analyzed and the distance from the EUT, the standard CISPR limits the strength of the field. The measured value can t be higher than the limits established and showed in the following table: Frequency 3 Meters 10 Meters 30 MHz to 230 MHz 40 dbµv/m 30 dbµv/m 230 MHz to 1000 MHz 47 dbµv/m 37 dbµv/m Table 3.1. CISPR limits, Quasi Peak detector, class B 16

18 3.2 Bluetooth Introduction The Bluetooth technology has mead a very important step in the world of the communications, allowing the wireless interconnection of different devices. Some years ago the connection among the different peripherals of the computer was carried out by means of cabling, which caused diverse problems, as the complexity of the connection or the excess of cabling that it rebounded in the devices mobility. All these problems have been solved by this technology, allowing more comfortable and quick connections, and also allowing the inclusion of this technology in eventual smaller devices, giving access to mobile devices. From the beginning, Bluetooth technology was intended to hasten the convergence of voice and data to handheld devices, such as cellular telephones and portable computers. Figure 3.5. Capability of Bluetooth connections As the idea grew, the SIG (Special Interest Group) was formed to create a standard for this technology. The original SIG, formed in 1998, consisted of five companies:ericsson, IBM, Intel, Nokia, and Toshiba. And other companies join later: Microsoft, 3Com, Lucent and Motorola. Through the efforts of its developers and the members of the Bluetooth SIG, it is now emerging with features and applications that not only remain true to its original intent, but also provide for broader uses of its technology. Nowadays the Bluetooth SIG has over 10,000 member companies developing, manufacturing, and selling thousands of Bluetooth enabled products worldwide. Now Bluetooth is a norm that defines a wireless global standard of communication, which facilitates the voice and data transmission among different teams by means of a connection for radio frequency. 17

19 3.2.2 Bluetooth Transmission Technology One of the primary advantages of the Bluetooth system is ease of computer vendor product integration. Other key benefits of this technology are low power, long battery life, low cost, low complexity, and wireless connectivity for personal space, peer to peer, cable replacement, and connectivity. The Bluetooth transceiver operates in the globally available 2.4 GHz ISM band (Industrial Scientific Medicine). In most countries around the world the range of this frequency band is MHz. However, several countries have national limitations in this frequency range, and in order to comply, special frequency hopping algorithms have been specified for these countries. Bluetooth uses 79 channels for the communication. The 79 RF channels are spaced 1 MHz apart. The channel is divided into time slots of 625 µs in length. A guard band is used at the lower and upper band edge to comply with out-of-band regulations Frequency Hopping Spread Spectrum The ISM band is occupied by other RF emitters, ranging from WLANs, baby monitors, and cordless phones. Bluetooth is based on a critical technology known as FHSS (Frequency-Hopping Spread Spectrum), applied to combat interference, fading, and to facilitate optional operation at power levels up to 100 mw. FHSS spreads the signal by transmitting a short burst on one frequency and then hops to another frequency for another short burst and so on, figure 3.6. In the FHSS system the carrier frequency of the transmitter hops in accordance with a pseudorandom hopping sequence, unique to each piconet. The frequency-hopping rate is 1600 hops/s for a single slot packet and slightly decreases for multi-slot packets. The transmitter and receiver synchronize to the hop sequence to ensure communication. The average signal strength on any given frequency is relatively low. Hopping also provides enhanced data reception in the presence of interfering signals, like fixed frequency radio networks or microwave ovens. If interference at a specific frequency is experienced, only a portion of the frequency hops will be blocked instead of the whole signal. The unblocked hops make it possible to recover the original data by re-transmitting the-message. Constant interference on a given frequency affects the radio network for only a short time on that specific frequency. 18

20 Figure 3.6. Working mode of FHSS and collision Radio Characteristics Bluetooth devices are divided into three power classes, Class 1, Class 2 and Class 3. The Bluetooth core specification classifies the transmitter equipment as having three classes of radio transmission power, namely 100mW (20dBm), 2.5mW (4dBm) and 1mW (0dBm). With 0dBm power, the communication range may be up to 10 meters while 20dBm transmit power increases the range to100 meters. Above 4dBm, there is power control to transmit appropriate radio power corresponding to the communication distance. The receiver actual sensitivity level is defined as the input level for which a raw BER (Bit Error Rate) of 0.1% is met for 723kbps. The Bluetooth receiver requires an actual sensitivity level of 70 dbm or better. The carrier to interference ratio (C/I) requirement is 11 db for Co-Channel interference. Adjacent interference on 1 MHz channels is 0dB and -30dB on 2MHz channels. Adjacent channels greater than 3 MHz require a signal to interference ratio of -40dB Modulation Characteristics The Bluetooth modulation scheme is GFSK (Gaussian Frequency Shift Keying) with a symbol rate of 1Msym/s and modulation index between The Gaussian-shaped binary FSK modulation minimizes transceiver complexity. Using positive frequency deviation a binary one is represented while a binary zero is represented by a negative frequency deviation, that is the Bluetooth carrier is in 150 khz to transmit 1 and -150 khz to transmit 0. Maximum frequency deviation is between 140 khz and 175 khz. If the frequency change is allowed to occur instantaneously, this can lead to ISI (inter-symbol interference) at the receiver. ISI makes it difficult to interpret what state the bit is trying to represent, this produces the transmission data errors. To reduce the spectral spreading that causes ISI, Bluetooth uses a Gaussian Filter (B T b = 0.5) to slow the transitions between the two frequencies Time Slots A Time Division Duplex (TDD) scheme is used where master and slave alternatively transmit. The baseband burst rate is 1Mbps. A TDD user frequency channel is shared with other users who have time slots allocated at different times. 19

21 Bluetooth allocates one slot at the transmit frequency and one slot on the receive frequency. The master only starts its transmission in even numbered time slots while the slave starts its transmission in odd-numbered time slots. A single packet obtains the RF hop frequency to be used from the current Bluetooth clock value. Multi-slot packet obtains the RF hop frequency to be used for the entire packet from the clock value in the first slot of the packet Packets The packets are broken down into their constituent parts such as access code, packet header, payload header, and payload, figure 3.7. LSB ACCESS CODE HEADER PAYLOAD 72 bits 54 bits bits Figure 3.7. Bluetooth Packet MSB There are currently 14 packet types defined, split into 4 segments; Common Packets, both ACL (Asynchronous Connection-Less) and SCO (Synchronous Connection-Oriented), Single slot, ACL 3 and ACL 4 slot packets. Each packet type has a different level of error correction and protection and different size payloads. The Access code is used to detect the presence of a packet and to address the packet to a specific device. The header packet contains control information associated with the packet such as the address of the Slave for which the packet is intended. Finally, the payload contains the message information. The payload field of all ACL packets is split into the payload header, the payload data and the Cyclic Redundancy Check (CRC) field. Before the payload is sent over the air interface, several bit manipulations are performed in the transmitter to increase reliability and security. An HEC (Header Extension Code) is added to the packet header, the header bits are scrambled with a whitening word, and FEC (Forward Error Correction code) coding is applied. In the receiver, the inverse processes are carried out Error Correction Three data error-correction schemes defined for the baseband controllers are: 1/3, 2/3 rate Forward Error Correction code (FEC), and Automatic Repeat Request (ARQ) scheme. FEC is implemented on the data payload to reduce the number of retransmissions. In a reasonable error-free environment, FEC adds unnecessary overhead, which reduces the throughput. 1/3 FEC uses a simple repetition code that repeats the bit three times. The 2/3 FEC scheme encodes data using a shortened hamming code. In the ARQ scheme packets are transmitted and retransmitted until the transmitting device receives an acknowledgement of a successful reception. Depending on the characteristics of the data that will be transmitted, Bluetooth uses several types of data packets. These packets differ by their payload length and FEC options. The application chooses the packet type to use, depending on the requirements of data rate and degree of error protection. Among various packet types, the ones that are used in broadcasting are asynchronous connectionless packets (ACL). The ACL packets are further classified as DM1, DM3, DM5, DH1, DH3 and DH5, figure 3.8. The difference between the packets DM and DH is that DM has error protection 2/3 FEC and DH doesn t have protection, and the number (1, 3, 5) refers the 20

22 size of the packet. There are two factors that affect packet type selection: one is current bit error rate (BER) of the radio channel (which is also related to the interference level) and the other is effectiveness of the FEC scheme applied in the selected packet type Bluetooth Version 2.1. Figure 3.8. Different kind of packets The main advantage of Bluetooth version 2.1 is the increase of the Basic Rate, due to the use of EDR (Enhanced Data Rate) with different modulation for the payload. For the EDR, PSK (Phase Shift Keying) is used as the modulation scheme. Two variants are specified, π/4-dqpsk (Differential Quadrature Phase Shift Keying) and 8DPSK (Differential Phase Shift Keying). By keeping the symbol rate at a constant 1 Msym/s for all modulation strategies the Basic Rate achieves a maximum of 1 Mbps, the Enhanced Data Rate achieves 2 Mbps maximum for π/4-dqpsk and 3Mbps for 8DPSK. The format of the packet is different, figure 3.9: LSB ACCESS HEADER GUARD SYNC EDR PAYLOAD TRAILER CODE 72 bits 54 bits 5µs 11µs bits Figure 3.9. Bluetooth v.2.1. EDR Packet format The Access Code and Header packet are modulated in FSK, but Synchronism, EDR Payload and Trailer packets are modulated in PSK. The Trailer payload is used for DC compensation. Guard is not a payload is only a security time and Synchronism is a time used to synchronize the communication. MSB 21

23 4 RESULTS 4.1 EMI Test receiver Vs Spectrum Analyzer EMI measurements require a different approach than other types of RF measurements because EMI are unpredictable signals in frequency and power. The test equipment needed to perform measurements is composed by antennas, amplifiers, filters, and the testing device. Two instruments are usually used for EMI testing: EMI Test receivers and Spectrum Analyzers. We needed to compare these two instruments to understand which is better for each measurement. These instruments are often working together as the same thing. The differences between the two devices can start to be explored by the parameters necessary to configure each instrument, there is an article comparing these parameters according to [3]: EMI Tester (R&S EMI Test Receiver 9 khz to 2500 MHz ESPC): Start/Stop Frequency Resolution Bandwidth filter (3 or 6 db) Detectors (Peak, Quasi-Peak, Average) Measurement Time Step size Spectrum Analyzer (R&S FSQ Signal Analyzer): Start/Stop Frequency Resolution Bandwidth filter (3 or 6 db) Detectors (Peak, Quasi-Peak, Average, RMS) Sweep Time Video Bandwidth The EMI Tester from R&S that we use is made for EMI measurements, and it can obtain the EMI measurement directly, then why to use another instrument for these measurements? One of the principal advantages to use the Spectrum Analyzer is the versatility and familiarity of this instrument, since in most laboratories there is one of these devices, and everyone knows its behavior and possibilities. But Spectrum Analyzers measure everything that falls in the pass band of the RF front end. In the case of high amplitude and wide frequency signal could cause overload, and is needed a preselection filter. On the other hand the EMI Tester Receiver doesn t need the preselector filter, because is included in the device. This instrument is recommended to use for EMI applications by standards bodies like CISPR, EN, FCC, MIL-STD and others. Some of the advantages of this instrument are the automatic control of the measurements, like automatic control of RF attenuation, RBW filtering, preselection filtering, preamplification settings, and the step size. The possibility to insert the antenna factor 22

24 and cable loss into the instrument to obtain the correct EMI measurements is another important advantage. The main disadvantage of EMI Testers is that they are slower than analyzers, and have limited use outside of the area of specialization. The measurements have been taken in the Laboratory of the Centre for RF Measurements Technology of Gävle. The devices used in this experiment are: EMI Tester, Spectrum Analyzer, antenna, cable, and computer. The system components are explained below: The EMI Tester utilized is EMI Test Receiver R&S ESPC, with a large frequency range from 150 khz to 1000 MHz, but can be extended from 9 khz to 2500 MHz. The device has Peak, Quasi-Peak and Average detectors and a fast synthesizer with a frequency resolution of 10 and 100 Hz. Is possible add up to 22 transducer factors for the cable effect and antenna factor. The results are showed in a graph directly in the computer by the software of the Tester (ESxS-K1), and the data could be extracted and used in Matlab after some correction with an M-File (APPENDIX B). This device has the option of output a comprehensive test report on a printer or plotter. By means of the Signal Analyzer R&S FSQ is realized the function of the Spectrum Analyzer. The device is a combination of two instruments, with a demodulation and analysis bandwidth that has been enhanced to 120 MHz. It has EMI filters of 6 db bandwidth (200 Hz, 9 khz, and 120 khz). The different detectors that the Analyzer has are Max Peak, Min Peak, Auto Peak (Normal), Sample, RMS, Average, Quasi-Peak. The results of the measurements are obtained, connecting the analyzer with the computer by GPIB, using Matlab and an M-File for store the values of the measurement (APPENDIX B). The antenna used for this measurement is a hybrid Bilog periodic (CBL 6112A), which is a broadband antenna (30 MHz-2GHz), linearly polarized and exhibits an excellent balance and cross polarization performance (APPENDIX C). The antenna has a combination of frequency coverage and high power, handling capability and is the antenna recommended for EMC testing, the following Figure 4.1, shows the Antenna Factor of our antenna. 23

25 30 ANTENNA FACTOR CBL6112A dbµv/m MHz Figure 4.1 Antenna Factor CBL6112A The cable is TS , a Tensolite coaxial cable of 48 inches (1.21 cm). The insertion losses, in the range of the measurements taken, are around db, such as is possible to see in [4]. The insertion losses of the connectors must be included in this factor, in the system are used two connectors, this means that are 0.2 db (0.1 db each connector). The computer needs the software for scan automatically the interferences through the EMI Test Receiver; this program is ESxS-K1 and uses GPIB (General-Purpose Instrumentation Bus) connection for the communication between computer and receiver. To control the Spectrum Analyzer by GPIB, is needed Matlab with GPIB toolbox. The full system is a connection of the EMI Test Receiver, the Signal Analyzer and the computer via GPIB. The Figure 4.2 shows the measurement system, Antenna, EMI Tester and Spectrum Analyzer: 24

26 Figure 4.2. EMI Measurement System, EMI Tester, Spectrum Analyzer and Antenna 25

27 The developed system was tested to measure the noisy environment of the RF laboratory in the Centre for RF Measurements Technology of Gävle. The measurements were taken with the EMI Tester and the Spectrum Analyzer. The scan using the EMI Tester gave us the following graph of the signal spectrum interference. The figure 4.3 represents the power in dbµv/m in each frequency of the range covered by the antenna (30 MHz-2GHz). This scanner uses Peak and Average detectors, the graph of Quasi-Peak and Average detectors could be look up in (APPENDIX A). Figure 4.3. ESPC- Graphic of EMI Tester of different detectors: Peak/Average From the graph values it has designed a table (Table 4.1) showing the principal peaks of the interferences measured with the EMI Tester. The interferences are classified according to the frequency; in each frequency the range where the interference is worst is specified. In the table 4.3 the results of the EMI Tester measurements are obtained with each kind of detector, Peak, Quasi-Peak and Average. It is possible to see the difference of strength field between each detector. The explanation of the columns is the following: Frequency and range: represents the frequency where is located the interference. Peak, Quasi-Peak and Average: show the reading value in the EMI Tester for the different kind of detector. Limit: limit established by the standard CISPR for correct reception of the signal. 26

28 Margin: difference between the measured value, Quasi-Peak measurement, and the limit. If the signal doesn t exceed this value, the interference will not be dangerous for the transmission of the signal. Utilization: is the real utilization, considerate in the Swedish frequency allocation table [5], of the frequency range where is found the interference. Frequenc y (MHz) Range (MHz) Peak (dbµv/m) Quasi- Peak (dbµv/m) Average (dbµv/m) Limit (dbµv/m) Margin (db) (UPL) (DWL) (UPL) (DWL) Utilization FM Sound Analogue, Broadcasting Broadcasting Television GSM GSM GSM (UPL) (DWL) Table 4.1. EMI Tester measurements results UMTS (3G), IMT-2000 With regard to Spectrum Analyzer the steps done are the same, measure the noisy environment of the laboratory. In this case it has to notate that the device has a RMS detector, contrary to EMI Tester. One of the principal disadvantages is the impossibility to add the AF to the instrument, and configure the Spectrum Analyzer like the EMI Tester; this inconvenient is solved adding the reading value to the correction factor, as is showed in the table 4.2. The scan obtained with the Spectrum Analyzer using Peak, Average and RMS detectors is showed in the following figure 4.4: 27

29 Electric field strength [dbµv/m] Peak RMS Average Frequency [MHz] Figure 4.4. Graphic of Signal Analyzer of different detectors: Peak, RMS, Average Viewing the power values of the measure with the Spectrum Analyzer it has constructed a table similar to (Table 4.1). Frequency (MHz) Peak (dbµv/m) RMS (dbµv/m) Average (dbµv/m) Quasi-Peak (dbµv/m) Correction Factor (dbµv/m) Limit (dbµv/m) Margin (db) Utilization ~ ~ FM Sound Analogue, Broadcasting Broadcasting Television GSM GSM GSM Table 4.2. Spectrum Analyzer results UMTS (3G), IMT-2000 If the reading value is added to the correction factor the results in each device are approximately the same. But the EMI Tester reads more power than the Spectrum Analyzer due to the Tester takes more time to evaluate the interference, and the Analyzer show the signal in real time. The results with the EMI Tester are obtained directly to the computer after configure the options of scan; against this, the Spectrum Analyzer measures everything that falls in the range. With the Spectrum Analyzer the signal represented is changeable because the measured signal is from impulsive noise. By means of the EMI Tester the signal is stable but is impossible to see variations immediately, the Tester needs more time to scan the interferences. In this way, the Spectrum Analyzer is better because is 28

30 possible to see the changes of the signal in real time. On contrary, since as the EMI Tester is designed especially for EMI measurements, the Tester set up automatically some important features like the reception filter, and the Spectrum Analyzer must be configured manually. 29

31 4.2 Simulink, Bluetooth communication under noise environment simulations Bluetooth 1.0 with different interferences The interference model is developed in Matlab Simulink and uses the following Simulink libraries: Simulink standard block set, Communications block set, Signal processing block set and Stateflow. The model is composed as separate objects that can be turned off, depending if they are used or not. This gives the user many simulation options and opportunities for how the model is used. The model is based on the model [1]. The changes realized are located in the channel. Firstly it has been added a real interference measurement of a noisy environment instead of microwave oven interference used in [1]. The other change is about the Bluetooth version, since as actually is used the version 2.1. The following figure 4.5. it is possible to see the improvements of the model [1]. The figure shows the new model with the changed blocks. In this model, the noise block has one Sum box to add the new interferences to the microwave oven interferences of [1]. These interferences are: EMI Tester measurements, WiFi and GSM. CHANNEL RECEPTION UU(R,C) FFT Spectrum Scope Rx Before Demodulation [Tx] From7 FDATool z -50 M-FSK [Rx] In1 Out1 u IF Filter Select 1MHz band Move to a symbol boundary M-FSK Demodulator Baseband Power Meter [dbm] -K- Gain TOTAL INTERFERENCE M-FSK Generate 79 possible carriers -39MHz to 39MHz To Frame [Hops] UU(R,C) FFT Spectrum Scope Rx After Modulation [GSM] [MicroWave_Oven] MEASUREMENT SYSTEM [EMI_Tester] 0 [WiFi] Display [Tx] 0 0 From2 Tx Error Rate Calculation Rx BER Data rate1 Data rate2 Throughput (Mbps)1 0 Total Power Interference (dbm) [Rx] From3 Error Rate Calculation Data rate3 Calculate Throughput Th Effective (%) 0 Effective Velocity (Mbps) Figure 4.5. Bluetooth transmission with, Microwave Oven and EMI measurements, WiFi, GSM With regard to the interferences, it was decided to add interferences which affect to Bluetooth communication. WiFi transmission is working in the same band frequency than Bluetooth, then it could be important to count with this interferer signal. 30

32 GSM (Global System for Mobile communications) is working in 900 and 1800 MHz, theoretically this signal doesn t affect the Bluetooth communication, but it is added to the model because is interesting to prove that any signal could be summed as noise. So that the interferences affect the Bluetooth system, every signal is inserted in baseband. As is described in the last section the EMI measurement covers the range from 30 MHz to 2 GHz. In this range, the signal doesn t affect the Bluetooth communication, but it can be simulated in baseband also. The figure 4.6 shows the interference measured by the EMI Tester, which is added to the Bluetooth model: Figure 4.6. EMI measurement interference The channel of the last model sums all of interferences created, GSM, WiFi, Microwave Oven and EMI interferences, this sum produce a total interference that is added to the transmitted signal of Bluetooth, this total interferences has a power which is calculated by a Power Meter showed in the model. The Total Power Interference can be increased by the Gain situated after the sum; this will allow the degradation of the Bluetooth signal, detecting in the receiver a worse Bit Error Rate. The effect of the Total Power Interference on the Bluetooth communication is showed on the figure 4.7. The BER is decreasing at the same time that the Total Power interference increases. How is showed in the features of the Bluetooth modules (APPENDIX E) has a sensibility of 0.1 with regard to the BER, then is easy to see the maximum of Total Power Interference, 60 dbm. When the noise exceed this threshold the communication using this system is not feasible. 31

33 BER Total Power Interference [dbm] Figure 4.7 Graphical of BER against Total Power Interferences (Microwave Oven, EMI measurements, WiFi and GSM signals) of Bluetooth Bluetooth Model version 2.1 The actual Bluetooth version 2.1 is composed by different modulations. As is described in theory, two modulations are used, GFSK and PSK (π/4-dqpsk or 8DPSK). The upper branch, header cost of 126 bits modulated in GFSK, this branch is concatenated with the down branch, total payload bits modulated in PSK. In this case the PSK modulation used for the payload bits is π/4-dqpsk, this provides basic rates to 2Mbs, it could modulate with 8DPSK. After the concatenation the signal is transmitted to the channel like in the other model. The figure of the signal transmitted is in the APPENDIX D, and is possible to see the combination of two modulations, viewing the difference with the transmitted signal in version 1.0 from the model of [1]. In the figure 4.8 is possible see the transmission of this new version 2.1. TRANSMISSION Random Integer CPM Random Integer Generator3 Buffer GFSK Modulation Vert Cat Matrix Concatenation Zero-Order Hold Quantizer -K- Gain [Tx_Signal] Goto1 Random Integer Random Integer Generator1 Buffer1 DQPSK DQPSK Modulator Baseband To Frame From2 [Hops] M-FSK Generate 79 possible carriers -39MHz to 39MHz TX SIGNAL FFT U(R,C) U Figure 4.8. Bluetooth transmission version

34 4.3 Real Bluetooth communication under noise environment simulations This section of the thesis deals about the experiment with real Bluetooth modules in noisy environment. The modules used are from Swedish company Free2move AB. Specifically are the modules F2M03GLA [6] with the evaluation board kit F2M03G-KIT. The Bluetooth module uses Bluetooth v 2.0+EDR, and has a range up to 350 meters (line of sight), but can be improved with another omnidirectional antenna on the circuit board and reach a range up to 1000 meters. Nowadays there are studies with Bluetooth modules that could reach 30 km. With this kind of ranges we can think a lot of possibilities for new applications. F2M03G-KIT has the possibility of easily evaluating the wireless UART (Universal Asynchronous Receiver-Transmitter) firmware as well as customizing it and access to an extensive range of I/O ports. It is possible to connect the evaluation kit to the computer by serial cable (RS232), USB (Universal Serial Bus) or TTL (Transistor- Transistor Logic). USB has the advantage that is self powered and, using serial converter software in the computer (FTDI), is the best way to connect the device. When the module is connected to the computer and detected, the device can be properly set up using the Bluetooth configuration software that come with the product to set connection modes, connect accept settings, security modes, PIN-code, baud rate, etc. For this thesis, only the basic options have been used. Once the modules have been configured, one of the modules will be the master and the other the slave to establish the communication, see the assembly in APPENDIX G. To check the good behavior of the modules, it has been sent information from one module to other by a terminal program, Br@y++ Terminal [7], see main window in figure 4.9. The election of this program was the different configurations and options that it has. The possibilities of Br@y++ Terminal are huge, it can send files or separated characters, and measure the time of transmission, view the graph of received data, change the baud rate, the parity, the handshaking and other useful options. Figure 4.9. Main window of Br@y++ Terminal 33

35 Once the modules are configured and the terminal is working, the next step is send files and measure the time of transmission. Using different Baud Rate for the transmission, it is possible to measure the velocity of the communication, and finally compare the measures with a noisy environment. This case the noisy environment is the presence of microwave interferences. The next graphics (4.10 and 4.11) provide the result of the experiment. The figure 4.10 is for baud rate of bps, and the figure 4.11 is for bps. The baud rate is the number of symbols per second transferred. The graph has three columns, the first column is the theoretical throughput, this means the maximum expected for the modules in ideal conditions (APPENDIX E). The other columns are for the communication with and without interference. The interference used is the microwave oven (APPENDIX G). This experiment checks if there is some effect in the transmission. The effective throughput is obtained dividing the size of the file (bits) by the transmission time (seconds). In this experiment the size of the file is bytes. Figure Throughput in Bluetooth communication with bps 34

36 Figure Throughput in Bluetooth communication with bps The following table compares the Throughput without and with microwave oven, is possible to observe the reduction of the throughput, in other words, the increment of the transmission. Theoretical Throughput (kbps) Throughput (kbps) without interferences Throughput (kbps) with microwave oven Reduced Throughput (%) Table 4.3. Comparative of Throughputs without and with microwave oven interference The resolution of calculation is not too much useful, because the transmission time measured is done with the terminal [7] for communicate the modules and has only 1 second of resolution, with another resolution the measurement could be more reliable. But bearing in mind this is possible to take some conclusions about the measurement of the time transmission in the communication with the two modules sending a file of fixed size. It is possible calculate the Throughput and see the difference with the theory. The measurements were taken first with the two modules without any interference and after this was added the effect of a microwave oven. To increase the effect of the microwave oven interference the door was half-opened, because the microwave oven doors are prepared to isolate the radiation. The transmission time of the file with the microwave oven interfering is greater than without interference; this is due to the retransmission of packets when there is some transmission error. 35

37 The BER of a radio channel is the dominant parameter that affects the effective throughput and reliability of a transmission. To obtain the BER is needed a specific analyzer, a Bluetooth BER Tester. In the other hand is possible relate throughput with BER, but the BER must be a fixed value of the channel behavior; this relation is exposed in reference [8]. One important result is that the throughput stays reasonable up to a threshold BER value, but after that it falls down quite fast, see figure The threshold BER value depends on the packet type. A packet type that has error protection has a higher value for this threshold, which is what was expected. It is showed in the figure the relation with the use of different Bluetooth packets, BER and effective Throughput. Figure Relationship BER, Throughput and packets used 36

38 5 CONCLUSIONS In industrial environment there are a lot of interferences caused by the industrial machinery that emit impulsive interferences. This is an important problem where innovative systems can be added like, wireless communication for the communication between machines and computers. If the industry wants to renew the communication and use wireless technology it must be checked if it is possible to use this technology, because the machines, as a motor, could interfere with the system. In the thesis these interferences are measured in a laboratory of Centre for RF Measurements Technology of Gävle, but these measures can be also taken in an industrial environment. With this information it is possible to simulate the real effect of these interferences in a Bluetooth communication, and assure that a Bluetooth communication is possible in a specific noisy environment. Depending on the environment and the necessrey frequencies to be scanned change of the antenna is needed and maybe the EMI Tester. The measurements in this thesis were limited by the range of the antenna available (30MHz to 2 GHz). Then, the study realized, can be extended using another antenna with a bigger range and scanning different environment as an in industry or hospitals. Other way to continue this thesis could be in the field of simulation. The realized work was focused in Bluetooth communications with interferences, but this can be extended to other technologies. Also, the noise inserted in Simulink was from laboratory environment and could be from anyplace, industry, hospitals, markets, cities, etc. Finally a real case of interferences in a Bluetooth communication has been tested using a terminal to connect and check the difference between, with as well as without interferences. To extend the research with the modules, the measurements could be taken using a BER Tester for Bluetooth which measures the BER and extends the conclusions about the interferences created by the Microwave Oven. In addition, the modules were separated around a half meter where future investigation could measure the throughput and BER between the modules by inserting different obstacles like walls and also increasing the distance or testing the behavior in an environment industry. 37

39 6 REFERENCES [1] Carl Karlsson, Coexistence of IEEE and Bluetooth in industrial applications, Master s Thesis in Electronics/Telecommunications at University of Gävle, Januari [2] The Comite International Special des Perturbations Radioelectriques, [2008/06/05]. [3] Rhode & Schwarz, System Support Center, EMI measurements, Test Receiver vs. Spectrum Analyzer. [4] Cable loss, [2008/05/02] [5] Eva Liljefors, General guidelines of the Swedish National Post and Telecom Agency on the Swedish frequency allocation table, PTSFS 2005:4, Swedish National Post and Telecom Agency, 16 August [6] Free2move AB main page, [2008/03/08] [7] Br@y++ Terminal download, [2008/03/20]. [8] Kaan Dogan, Guray Gurel A., Kerim Kamci, Ibrahim Korpeoglu, A Performance Analysis of Bluetooth Broadcasting Scheme, [9] Schaefer, W., Measurement of Impulsive Signals with a Spectrum Analyzer or EMI Receiver, Electromagnetic Compatibility, EMC International Symposium on Volume 1, Issue, 8-12 August 2005 Page(s): Vol. 1. [10] Jussi Savolainen, Automatic Bluetooth Radio Frequency Measurement System, Elcoteq Design Center Oy, Salo, May [11] Lynch, Jamel Pleasant, Co-Channel Interference In Bluetooth Piconets, Master Thesis in Electrical and Computer Engineering, October [12] Tihany L., Electromagnetic Compatibility in Power Electronics, IEEE Press [13] Clayton R. Paul, Introduction to Electromagnetic Compatibility, Wiley Interscience, January [14] Mark T. Ma, Motohisa Kanda and Myron L. Crawford, A review of electromagnetic compatibility/interference measurement methodologies, Proceedings of the IEEE, Volume 73, No 3, March

40 [15] Zeng, Z.; Alien, B.; Aghvami, A.H., Performance evaluation of a Bluetooth interference canceller in IEEE b wireless networks, Consumer Electronics, IEEE Transactions on, Volume 51, Issue 4, Nov Page(s): [16] Ferro, E.; Potorti, F., Bluetooth and Wi-Fi wireless protocols: a survey and a comparision, Wireless Communications, IEEE Personal Communications, Volume 12, Issue 1, Feb Page(s): [17] Shah, A.; Jalil, A., Investigation and Performance Evaluation of different Bluetooth voice packets against ambient error conditions, Multitopic Conference, INMIC 06. IEEE, Dec Page(s): [18] James McLean, Robert Sutton, Rob Hoffman, Interpreting Antenna Performance Parameters for EMC Applications, Part 3, TDK RF Solutions Inc., 2002,2003. [19] Ling-Jyh Chen, Rohit Kapoor, M. Y. Sanadidi, Mario Gerla, Enhancing Bluetooth TCP Throughput via Packet Type Adaptation, Dept. of Computer Science, UCLA, 2004 IEEE International Conference on Communications, June, 2004, Paris. [20] Eung-in Kim, Jung-Ryun Lee, and Dong-Ho Cho, Throughput Analysis of Data Link Protocol with Adaptive Frame Length in Wireless Networks, AEÜ Int. J. Electron. Commun., 51 (2003) No. 1, 1.8, Ministry of Science and Technology, Korea. [21] Schaefer, W., Narrowband and broadband discrimination with a spectrum analyzer or EMI receiver, Electromagnetic Compatibility, EMC IEEE International Symposium on, Volume 2, August 2006 Page(s): [22] Markus Tengvall, Smart Sensors.Detection of infrasonic waves, Master Thesis project at The Royal Institute of Technology of Stockholm, June [23] EMC Testing, [2008/05/13]. [24] Southwick, R.; Runger, G., A theory to optimize the detection and measurement of EMI signals, Electromagnetic Compatibility, IEEE 1989 National Symposium on May 1989 Page(s): [25] Special Interest Group of Bluetooth, [2008/04/07]. [26] How it Works Bluetooth, [2008/06/04]. [27] Tim Williams, What to Look for in an EMC Antenna, [2008/04/20]. [28] Electronic Theses and Dissertations, [2008/03/18]. 39

41 [29] Peter F. Stenumgaard, Lars-Erik Juhlin, Erling Pettersson, Jenny Skansen, A Novel Method to Identify Pulsed Interference Sources in Radiated Emission Measurements, The Swedish Defense Research Agency, ABB Power Systems, February [30] Joel Galmor, Smart Sensors Evaluation and implementation of wireless standard for remote infrasonic detection, Master of Science Thesis at KTH of Stockholm, [31] Matlab main page, [2008/04/04] [32] Rhode & Schwarz main page, [2008/04/19] [33] Schaefer, W., Signal detection with EMI receivers, Electromagnetic Compatibility, IEEE Syposium on, Volume 2, August 1998 Page(s): vol.2. 40

42 7 APPENDIX APPENDIX A: EMI TESTER MEASUREMENTS Figure A.2. ESPC- Graphic of EMI Tester with different detectors: Quasi-Peak/Average Figure A.3. ESPC- List of peaks with different detectors: QuasiPeak/Average 41

43 APPENDIX B: M-FILES USED The results of the measurements with the EMI Tester can be stored in the computer in format.dat. This format must be transformed to a vector to be represented in Matlab. The m-file used in this case is the next: %Change the file with data results from EMI Tester to the correct vectors >>[s, msg] = replaceinfile(',', '.', 'MEASLAB.DAT'); %M_file to change the characters ASCII for another %ASCII character, in our case change the (,) for (.). %MEASLAB.DAT is the EMI Tester file >>fid = fopen('measlab.dat'); %text in ASCII and numbers. Each format column must be specificate in the order % %s its for cell(ascii) %d8 for (int) %f32(numbers with 32 bits) %u for (natural number) %Now the same order but adding a functionality to eliminates a lines, because its needed to delete the beginners lines >>C = textscan(fid, '%s %d8 %u %f32 %d8 %d8', 'headerlines', 130); >>fclose(fid); >>semilogx(c{3},c{4}) >>AXIS([0 2*10^ ]) %AXIS([XMIN XMAX YMIN YMAX]) sets scaling for the x- and y-axes on the current plot. >>grid In the measurements with the Signal Analyzer the data results are obtained by means of GPIB communication. It is possible to plot the graph through Matlab, now the data mustn t be transformed, all that is needed is a GPIB communication. The instructions to draw the graph of the Signal Analyzer are the followings: %To construct a GPIB object connected to an Nation Instrument % board at index 0 with an instrument at primary address 1: >>g = gpib('ni', 0, 1); % To connect the GPIB object to the instrument: XMA? 200,500 >>fopen(g) >>fprintf(g,'xma? 210,600') >>points=[fscanf(g)]; >>y = str2num(points);%to change the format of the dates recived, Char to double. >>n=length(y); % Ask the length of the vector y >>x=[1:n];%create a vector x with the same length than y >>plot(x,y) %To disconnect the GPIB object from the instrument. >>fclose(g); 42

44 APPENDIX C: 6112B BiLog ANTENNA 30MHz - 2GHz CBL The CBL 6112B is a high gain ultra wideband BiLog antenna for emission and immunity EMC testing. Three Antennas In One The CBL 6112B operates over the unprecedented, wide range 30MHz to 2GHz. It effectively combines the performance of three standard EMC antennas, the Biconical, the Log Periodic and the Waveguide Horn. Considerable savings in the order of 40-50% can be made in expensive test time, plus the added benefit of improved repeatability and reliability by not having to laboriously disconnect and reconnect antennas during testing: Ideal for FCC15, EN/CISPR Compliance emission testing VHF radio approval applications Immunity testing to 300W CW Excellent balance < 1dB No rotational offsets Easily transportable Individual calibration The CBL 6112B is primarily an emission measuring antenna but can handle CW powers up to 300 watts, making it suitable for most immunity measurements requiring fields up to 10V/m, or even greater. The CBL 6112B is linearly polarised and exhibits excellent balance. Technical specifications CBL 6112B: Frequency range: 30MHz - 2GHz Impedance (Nominal): 50 Ω Gain 6dB Typical: 200MHz - 700MHz 8dB Typical: 700MHz - 2GHz Connector: N Female VSWR Average: 2:1 Size L x W x H cm: 153 x 139 x 63 Weight: 4.2kg Max. Power: 300W CW CBL 6112B BiLog mounted on optional tripod CTP6097A 43

45 30 ANTENNA FACTOR CBL6112A dbµv/m MHz 44

46 APPENDIX D: SIMULINK MODELS TRANSMISSION [Tx] Goto1 Random Integer CPM Random Integer Generator Buffer GFSK Modulation [Hops] To Frame INTERFERENCES Out1 [MicroWave_Oven] Microwave Oven Goto5 Out1 [EMI_Tester] EMI Tester Goto3 Tx Signal [WiFi] Transmitter WiFi Goto7 Out1 [GSM] GSM System Goto4 Spectrum Modulated with carrier FFT U(R,C) U Mix to appropriate hop frequency M-FSK Generate 79 possible carriers -39MHz to 39MHz CHANNEL RECEPTION UU(R,C) FFT Spectrum Scope Rx Before Demodulation FDATool z -50 M-FSK In1 Out1 Power Meter [dbm] Gain -K- u Move to a IF Filter symbol Select 1MHz band boundary M-FSK To Frame [Hops] M-FSK Demodulator Baseband UU(R,C) TOTAL INTERFERENCE Generate 79 possible carriers -39MHz to 39MHz [GSM] [MicroWave_Oven] MEASUREMENT SYSTEM [EMI_Tester] [WiFi] HOPS GENERATOR [Tx] From2 Tx Error Rate Calculation Rx BER Data rate1 Data rate2 Random Integer Random Integer Generator1 [Hops] Total Power Interference (dbm) [Rx] From3 Error Rate Calculation Data rate3 Calculate Throughput [Rx] FFT Spectrum Scope Rx After Modulation e+004 Display Throughput (Mbps) Th Effective (%) Effective Velocity (Mbps) Figure D.1. Bluetooth communication with, Microwave Oven and EMI measurements, WiFi, GSM 45

47 TRANSMISSION Random Integer Random Integer Generator3 Buffer Random Integer Random Integer Generator1 Buffer1 Hops Random Integer Random Integer Generator u-39 [Hops] Goto -11 Display1 CPM GFSK Modulation DQPSK DQPSK Modulator Baseband [Tx_Signal] u Vert Cat Matrix Concatenation RECEPTION z -50 Move to a symbol boundary Generate [Hops] 79 possible carriers -39MHz to 39MHz 1 M-FSK To Frame Zero-Order Hold Quantizer To Frame M-FSK From2 [Hops] Generate 79 possible carriers -39MHz to 39MHz M-FSK M-FSK Demodulator Baseband Buffer2 -K- Gain z -5 Delay UU(R,C) FFT TX SIGNAL [Tx_Signal] Goto1 UU(R,C) FFT CARRIERS GENERATED UU(R,C) FFT RX SIGNAL [Rx] Goto2 Figure D.2. Model Bluetooth v.2.1. (GFSK and DQPSK modulations) 46

48 Figure D.3. Bluetooth v.2.1. modulated signal (GFSK and DQPSK) 47

49 APPENDIX E: F2M03GLA BLUETOOTH MODULES Features: Fully qualified end product with Bluetooth v2.0+edr, CE and FCC Low power consumption Integrated high output antenna Transmit power up to +8dBm Class1/ 2/ 3 Configurable Range up to 350m (line of sight) Piconet and Scatternet capability, support for up to 7 slaves Require only few external components Industrial temperature range -40 C to +85 C USB v2.0 compliant Extensive digital and analog I/O interface PCM interface for up to 3 simultaneous voice channels Large external memory for custom applications Support for b/g Co-Existence RoHS compliant Applications: Industrial and domestic appliances Cable replacement Medical systems Automotive applications Stand-alone sensors Embedded systems Cordless headsets Computer peripherals (Mice, Keyboard, USB dongles, etc.) Handheld, laptop and desktop computers Mobile phones 48

50 General Description: F2M03GLA is a Low power embedded Bluetooth v2.0+edr module with built-in high output antenna. The module is a fully Bluetooth compliant device for data and voice communication. With a transmit power of up to +8dBm and receiver sensibility of down to 83dBm combined with low power consumption the F2M03GLA is suitable for the most demanding applications. Developers can easily implement a wireless solution into their product even with limited knowledge in Bluetooth and RF. The module is fully Bluetooth v2.0+edr qualified and it is certified according to CE and FCC, which give fast and easy Plug-and-Go implementation and short time to market. The F2M03GLA comes with an on board highly efficient omnidirectional antenna that simplifies the integration for a developers Bluetooth solution. The high output power combined with the low power consumption makes this module ideal for handheld applications and other battery powered devices. F2M03GLA can be delivered with the exceedingly reliable and powerful easy-to-use Wireless UART firmware implementing the Bluetooth Serial Port Profile (SPP). The following table shows the maximum achieved throughput when streaming data between two connected modules with different configuration and baud rate: Maximum Maximum Direction Baud Rate Throughput (kbit/s Throughput (kbit/s) (throughput mode)) (latency mode) Master to Slave ~57.6 ~57.6 Slave to Master ~57.6 ~57.6 Full duplex ~57.6 ~50.5 Master to Slave ~115.1 ~93.9 Slave to Master ~115.1 ~79.6 Full duplex ~114.5 ~42.0 Master to Slave ~223.1 ~158.0 Slave to Master ~221.4 ~117.7 Full duplex ~172.7 ~86.2 Master to Slave ~228.6 ~206.7 Slave to Master ~222.7 ~154.1 Full duplex ~173.3 ~109.8 Master to Slave ~240.1 ~235.7 Slave to Master ~235.4 ~186.0 Full duplex ~174.7 ~

51 APPENDIX F: EVALUATION KIT FOR GENERAL PURPOSE BLUETOOTH MODULES DATASHEET Key Features: 10 GPIO:s with push buttons and led indication Analog audio interface (mono) (speaker and microphone, 3.5mm socket) Power supply through USB-connector or external adaptor USB-interface (virtual com port or direct HCI access) Programming interface for Free2move s flash utility (parallel port) RS232-interface Pin headers for all digital and analog I/O:s F2M03G-KIT-1 Contains: F2M02BG1 (Evaluation board) USB-cable Serial cable Parallel cable CD (Manuals and software) General Description: F2M03G-KIT is a new evaluation kit for Free2move s general-purpose Bluetooth modules. The evaluation board has extensive I/O functionality for both data and audio streams. The evaluation board is primary made to evaluate the Wireless UART firmware but is also intended to be used for other firmwares and custom made applications. The evaluation board gives the possibility to upgrade the Bluetooth module with new firmware using Free2move s Flash utility. The F2M03G-KIT has both RS232 and USB-interface for data communication. The USB-interface can either act as a direct connection to the module through HCI or it can be used as a virtual com port for computers without a physical com port. All digital and analog I/O can be accessed from pin headers on the evaluation board for external communication. The evaluation board is equipped with an audio codec and amplifier for direct microphone and speaker drive. External audio streams can be connected using the PCM interface. The F2M03G-KIT is delivered with all necessary cables and a CD with manuals and software. The evaluation kit can both be delivered as a single or a double kit. The kit does NOT include any Bluetooth OEM-board. They are ordered separately! 50

52 APPENDIX G: EXPERIMENT MODULES AND MICROWAVE OVEN 51