L-Band DAB Eureka 147 Field Trials and Coverage Measurements in Urban Areas M. M.Vélez (jtpveelm@bi.ehu.es), P. Angueira, D. de la Vega, J.L. Ordiales, A Arrinda. University of the Basque Country Bilbao Engineering College Alda. Urkijo S/N 48013 BILBAO, Spain Abstract- This paper presents the field trials and the coverage measurements carried out in urban areas in the first experimental terrestrial DAB (T-DAB) network at L-Band in Spain. The trials focused on obtaining the parameters that characterize the propagation channel for different urban environments. The results will be discussed taking into consideration recommended parameters and values given by international organizations (ITU-R, ETSI) as well as the ones obtained in former experimental networks worldwide. I. INTRODUCTION The Eureka-147 DAB system [1] has been recommended as an international standard for terrestrial and satellite digital sound broadcasting to vehicular, portable and fixed receivers in the frequency range 30-3000 MHz [2],[3]. Some L-Band (1452-1492 MHz) test transmission pilot projects have been successfully carried out worldwide so far [4],[5],[6],[7]. As a result, a few countries have already set up terrestrial networks broadcasting operational services in this band. In Spain this band is likely to allocate local services in the future. Digital Radio in Spain began with pilot stations in 1998 using VHF Band-III (174 MHz - 240 MHz). During year 2000 an experimental network was placed to test all the parameters of the system. This network included one L-Band transmitter in the city of Madrid to evaluate the L-Band propagation channel in urban areas and to survey the potential problems that will arise when operational networks have to be implemented. In the second half of January 2001 some field trials and coverage measurements, including both vehicular and fixed reception, were made in the city of Madrid in order to test the L-Band T-DAB signal behavior. The measurements presented here try to obtain the main parameters that characterize the urban reception for network planning purposes. This paper will explain the field trials, describing the measuring equipment and techniques applied. Further on, the set of measurements will be summarized on different curves and tables, comparing the results with the recommendations given by ITU-R [4] and ETSI [8]. Finally and considering all the results obtained, some conclusions will be proposed as well as some guidelines for future works. II. OBJECTIVES The main targets of the work presented here are: Evaluation of the propagation radio channel for T-DAB networks at L-Band in urban areas. Evaluation of the characteristics for vehicular and fixed reception in different environments. Validation of the results recommended by the ITU-R and the ETSI. Definition of the parameters and values needed for L-Band T-DAB planning in urban areas The DAB signal parameters evaluated have been the received power, the Pseudo-channel BER[9], CRC errors, frequency spectrum and subjective audio quality. The urban channel parameters to be obtained have been the minimum Carrier to Noise ratio (C/N) and the minimum field strength needed for an optimal coverage, defined as the reception without audible errors that correspond to a netber less than 10-4. The standard deviation of the field strength levels in large areas has been also calculated. This value corresponds to the signal slow fading.
III. EXPERIMENTAL NETWORK The L-Band network installed in the city Madrid consisted of a low power transmitter (300W EPR) with good line of sight coverage over a wide urban area. This transmitter was intentionally used in order to test the real viability of such low power equipment for local services. The transmission mode selected for the trials was DAB Mode II [1]. The main characteristics of the transmission system are presented in Table 1. As the network consisted of a single transmitter it has not been possible to test the behavior of Single Frequency Networks (SFN) in urban environments. Table 1. Transmission system parameters Operating Frequency 1454.672 MHz (LB) Channel Bandwidth 1.536 MHz Polarization Linear vertical Signal modulation COFDM Number of carriers 384 Carriers modulation D-QPSK Frame duration 24 ms Gross bitrate 2.3 Mbit/s The transmitter broadcasted a multiplex of 6 audio programs with different bitrates and error protection levels. All the tests presented were made on the audio program with 192 Kbit/s and error protection level 3 (0.5 average Code Rate in the convolutional encoder [1]). IV. MEASUREMENT CAMPAIGN A. Measurement Equipment The field trials have been developed using a specially equipped mobile unit (Fig. 1). Two different measurement systems have been used. Both systems used a 2dBi gain omnidirectional receiving antenna placed on the van roof. The receiving antenna height is 2.2m. In the first measurement system (referred to as MS-1 in this paper) the DAB signal is fed to three different ways by means of a power divider. Each way is connected to the power meter, the spectrum analyzer and the professional DAB receiver respectively. The main drawback of this system is the high attenuation, which yields to a high minimum field strength value. Another system (referred to as MS-2 in this paper) has been used. In this case the DAB antenna feeds a low noise amplifier that reduces the noise figure and decreases the minimum field strength required. ANTENNA DAB RECEIVER PERSONAL COMPUTER MS-1 MS-2 POWER DIVIDER SPECTRUM ANALYZER LOW NOISE AMPLIFIER MS-2 POWER METER Control & Data adquisition Figure 1. Mobile unit equipment GPS Table 2 shows the comparison between the L- Band receivers proposed by the ITU-R [4] and ETSI [8] and the ones used in these trials (MS-1, MS-2) For comparison purposes, the minimum C/N corresponds to a simulated Rayleigh channel (14 db) [8]. (In this paper a different value for this parameter will be later proposed). Table 2. Comparison of L-Band receiver systems ITU-R ETSI ETSI (amp) MS-1 MS-2 (amp) Fr 3 3.5 6 6 6 Fa 2.3 1.5 Fs 4 2.9 Gamp 10 20 L(max) 1 2.5 4 5.8 6.8 Ga 0 1.1-0.9 2 2 C/N-rf 14 14 14 14 14 FS(min) 45.1 47 47 52 43.1 Fr : Receiver noise figure (db) Fa : Amplifier noise figure (db) Fs: System noise figure (including amplifier and cable loss) (db) Gamp : Amplifier gain (db) L(max) : Loss in antenna circuit (max) (db) Ga : Antenna gain (dbi) C/N-rf : C/N (assuming Rayleigh fading) (db) FS (min) : Minimum Field Strength (dbµv/m)
The ITU-R has defined a typical receiver system for DAB satellite networks with minimum field strength of 45.1 dbµv/m. The ETSI considers two different systems to achieve the minimum field strength of 47 dbµv/m proposed for network planning. The ITU-R receiver and the ETSI receiver without amplifier are quite unrealistic for terrestrial reception because they lead to good noise figures, which may not be easy to provide in a mass-produced set. The MS-2 system, in spite of being very useful for the field trials due to his low minimum field requirement, can t be either considered a standard receiver system because the professional amplifier used has a noise figure very low if compared with the ones used in commercial sets. It s very important to note that the receiver system characteristics influence the minimum field strength requirements. As a consequence, receiver features have to be carefully defined when comparing different measurement campaigns [5],[6],[7]. Therefore, this article will focus on C/N values rather than on field strength values. B. Measurement Techniques and Data Processing The measurement techniques and data processing have varied during the trials depending upon the reception type (fixed or mobile) and the target parameter being surveyed. (1) Fixed reception The fixed reception measurements have been collected in 64 urban locations. According to [4], two different urban scenarios have been considered: Dense urban areas, in which most of the buildings have more than four stories (43 locations) Urban areas, with buildings up to four stories, but with some open space between (21 locations). For fixed reception, the system carried out 25 power measurements per second and recorded the averaged Pc-BER and CRC-errors once every 1.2 seconds (50 DAB frames). The instantaneous signal spectrum was recorded twice a second. An average time of 3 minutes was set for recording data at each location. During the measurement time the subjective audio quality was also recorded with the rest of the parameters. (2) Mobile reception To obtain the minimum C/N needed for mobile reception, 19 routes with a total of 121 Km in urban areas have been tested. The routes were selected to cross the limit of the transmitter coverage area in order to obtain the threshold value. These routes were planned in urban and dense urban areas with different traffic density including some high-speed urban roads. Whenever possible it was tried to collect data at different vehicle speeds. For mobile reception the system carried out 25 power measurements per second and recorded the averaged Pc-BER and CRC-errors once every 1.2 seconds (50 DAB frames). Subjective audio quality was also recorded along the whole route. (3) Signal strength local-area variations In order to obtain the field strength local-area variations, a total of 180 Km of urban routes with correct signal reception were tested. In this case, the system was limited by the measurement speed of the power meter (25 measurements per second). With this restriction the study was limited to the slow fading component. Fast fading and combined fading analysis should have required extremely low vehicle speeds [10]. The recorded data has been processed in areas of 500m length. These 500m length areas are considered as large areas for network planning [4]. (4) Coverage target The edge of service target for an adequate audio quality is a BER value after Viterbi (also named netber) less than 10-4 [8]. The receiver used provided the Pseudo-channel BER (Pc-BER), which is related to the channel-ber (BER before Viterbi) [9]. The Pc-BER is a valid parameter to find the error free reception target, enabling coverage measurements using on-air signals. Despite that, the Pc-BER value that provides the netber equal to 10-4 depends on the travelling speed and the nature of the terrain. As a consequence the Pc-BER threshold has to be measured for each receiving scenario considered. V. RESULTS AND DISCUSSION All the subjective audio tests, in fixed and mobile reception, confirmed that the audio failures were directly related to the existence of Scale Factors CRC
(SF-CRC) errors. The scale factors are the values that normalize the audio samples [1] and the correct reception of these values is checked by these CRC codes. Therefore, the SF-CRC error free reception condition will be the reference for detecting system threshold values. As previously known, if compared to MS-1, the MS-2 behaved significantly better, due to the low noise amplifier. A. Fixed Reception The standard deviation of the averaged spectrum gives an idea of the multipath level at a location. This parameter varied between 0.85 db and 6.1 db, with a 3.2 db mean value. The lowest values correspond to urban locations were the line of sight to the transmitter is unobstructed. In those locations the multipath level caused by buildings and other objects is negligible if compared to the direct ray level. The highest values correspond to locations where the direct ray from transmitter is shadowed and signal is received by way of scattered rays. Using the mean as a threshold value, 39 locations present a spectrum standard deviation lower than 3.2 db and 25 locations higher than 3.2 db. The first set is considered as low multipath condition and the second one as strong multipath condition. Figures 2 and 3 group the results of the mean Pc- BER and the mean C/N values for the low and the strong multipath locations, respectively. Mean Pc-BER 1.00E+00 1.00E-01 1.00E-02 1.00E-03 1.00E-04 1.00E-05 1.00E-06 0.00 10.00 20.00 30.00 40.00 50.00 Mean C/N (db) Figure 2. Pc-BER vs. C/N for Low Multipath Fixed Reception It can be shown that the two groups of locations fit different curves (using potential interpolation). This is caused by the dependence of the Pc-BER with the environment characteristics. Mean Pc-BER 1.00E+00 1.00E-01 1.00E-02 1.00E-03 1.00E-04 1.00E-05 1.00E-06 0.00 10.00 20.00 30.00 40.00 50.00 Mean C/N (db) Figure 3. Pc-BER vs. C/N for Strong Multipath Fixed Reception The measurements carried out detected a minimum value for error free reception of Pc-BER equal to 2.1E-02. Minimum values which correspond to this error free reception condition are shown in Table 3. The Pc-BER value used is quite conservative and ensures the absence of SF-CRC errors. Table 3. Minimum values for Fixed Reception Low Multipath Strong Multipath Pc-BER 2.1 E-02 2.1 E-02 C/N 12.5 15.5 Field Strength (MS-1) 50.5 53.5 Field Strength (MS-2) 41.6 44.6 The results agree with the theoretical 14 db value. Variations from this value depend on the multipath conditions at each location. According to the classification formerly made in this paper, a value of 12.5 db should be considered for low multipath environments, whereas high multipath environments will need a 15.5 db value. An important remark is that there is no obvious correlation between the low multipath and strong multipath locations and the previously planned urban and dense urban areas. This may be due to the fact that the previous classification made at the measurement planning stages was based on building height and degree of density around the measurement van, rather than the clearance or shadowing of the transmitter s direct ray.
B. Mobile Reception Two different analyses have been made with the data collected from mobile measurements in 19 routes. First, each route has been studied independently. It has to be considered that each route passes through different urban density areas and also has variable speeds depending upon the traffic circumstances. Therefore, the value assigned to each route corresponds to the most critical location in that route. The result from this analysis is then the minimum C/N that ensures correct reception along the whole route. Table 4 shows the values obtained. Once again, it can be seen that some routes provide minimum C/N values which are below the 14 db threshold, corresponding to mobile reception in low multipath environment at low vehicle speeds. The highest values for the minimum C/N are mainly obtained in the high-speed urban roads or in a very strong multipath environment. Table 4. Minimum C/N values for Mobile Reception Lowest Value Mean Value Highest value 11.6 14.3 16.7 The second analysis tried to evaluate the influence of speed on the minimum C/N. Each route has been divided into sections with constant speed. The minimum C/N value has been calculated for each section. All the sections resulting from the 19 routes have been grouped by speed and shown on Fig. 4. Minimun C/N (db) 18 17 16 15 14 13 12 11 10 9 8 0 10 20 30 40 50 60 70 80 90 Vehicle speed (Km/h) Figure 4. Minimum C/N values recorded at different speeds Fig. 4 shows that C/N is influenced by both multipath and speed. Logically, the minimum C/N takes higher values as speed increases. However, this increase is also affected by the multipath intensity at speeds below 50 Km/h. Most sections with speeds below 50 Km/h have a minimum C/N value below 14.3 db. It should be also noticed that some sections with low speeds (< 20 Km/h) present a high value of minimum C/N (up to 16.7 db). Such sections correspond to dense urban areas where multipath is very strong. C. Signal Strength Local-area Variations The results for the 358 areas measured showed an average standard deviation of 4.2 db for the slow fading. This result is close to the values obtained in Germany [11] and Great Britain [12] for Band-III. VI. CONCLUSIONS Results obtained in this study confirm that the minimum C/N 14 db threshold is adequate for planning purposes. Most of the tested locations in this measurement campaign presented a SF-CRC error free reception condition associated to a minimum C/N value lower than 14.3 db. However, in order to achieve coverage inside areas where multipath is very strong and also to ensure urban reception at vehicle speeds up to 80 Km/h a more restrictive value of 16.7 db is recommended. Different C/N values have been measured depending upon the multipath intensity. Such values could be useful when planning for small urban service areas where multipath can be estimated. It should be noticed that the minimum field strength needed for good coverage depends upon the receiver system being used. If network planning has the minimum field strength value as the reference parameter, the receiver system characteristics should be carefully specified. Considering a minimum C/N value of 14 db, the minimum field strength associated with our measurement systems is 52 dbµv/m for MS-1 and 43.1 dbµv/m for MS-2. According to coverage measurements and as described in [8], it s highly recommended that L- Band receivers include an antenna amplifier. Further work needs to be done in order to test the behavior of the L-Band Single Frequency Networks in urban environments.
Detailed L-Band network planning will also need further information about the fast and the combined fading in urban environments. ACKNOWLEDGMENTS The authors would like to thank Telefónica Sistemas Audiovisuales TSA, which has made all this work possible. REFERENCES 1. ETSI (European Telecommunications Standards Institute), "Digital Audio Broadcasting (DAB) to mobile, portable and fixed receivers, ETSI EN 300 401 v1.3.3, May 2001. 2. ITU Recommendation, System for Terrestrial Digital Sound Broadcasting to Vehicular, Portable and Fixed Receivers in the Frequency Range 30-3000 MHz, ITU-R BS. 1114-1, 1995. 3. ITU Report, Digital Sound Broadcasting to Vehicular, Portable and Fixed Receivers using Terrestrial Transmitters in the UHF/VHF Bands, ITU-R BS. 1203-1, 1994. 4. ITU-R Special Publication, Terrestrial and Satellite Digital Sound Broadcasting to Vehicular, Portable and Fixed Receivers in the VHF/UHF Bands, Geneva 1995. 5. R. Paiement, Experimental DRB System in Ottawa, CRC Technical Memo, Ottawa, 1997 6. ITU-Radiocommunication Study Groups, Coverage Measurements in Tunnels with Digital System A (DAB), Document 10B/xxx, 20 August 1999. 7. OFTA (Office of the Telecommunications Authority Hong Kong), Digital Audio Broadcasting Technical Trials, Final Report, November 1999. 8. ETSI (European Telecommunications Standards Institute), "Digital Audio Broadcasting (DAB); Signal strengths and receiver parameters; Targets for typical operation, ETSI TS 101 758 v2.1.1, November 2000. 9. R. Schramm, Pseudo Channel BER an objective quantity for assessing DAB coverage, EBU Technical Review, Winter 1997. 10. ERC (European Radiocommunications Committee) Field Strength Measurements along a Route with Geographical Coordinate Registrations, ERC Recommendation (00)08, November 2000 11. A. Lau, M. Pausch, W. Wütschner, First results of field tests with the DAB single frequency network in Bavaria, EBU Technical Review, Autumn 1994. 12. M.C.D. Maddocks, I.R. Pullen, J.A. Green, Field trials with a high-power VHF single frequency network for DAB. Measurement techniques and network performance, EBU Technical Review, Autumn 1994. BIOGRAPHIES The five authors of the paper belong to the TSR (Radiocommunications and Signal Processing) research group and work at the University of the Basque Country. Their research interests include digital TV and radio planning and Electronics for Radio and Television Applications. Manuel Mª Vélez received the M.S. the Basque Country, Spain, in 1993. In 1995 he joined the Department of Electronics and Telecommunications of the University of the Basque Country, where he is currently an assistant professor. Currently he is working toward his Ph. D. Degree in digital audio broadcasting (DAB). Pablo Angueira received the M.S. the Basque Country, Spain, in 1997. In 1998 he joined the Department of Electronics and Telecommunications of the University of the Basque Country, where he is currently an assistant professor. Recently, in May 2002 he has received the Ph. D. degree for a study of the urban portable outdoor reception of DVB-T signals in
single frequency networks. His research interests include several aspects of network planning for digital broadcasting services. David de la Vega received the M.S. the Basque Country, Spain, in 1996. In 1998 he joined the TSR (Radiocommunications and Signal Processing) research group at the Department of Electronics and Telecommunications of the University of the Basque Country. He is currently an assistant professor at the University of the Basque Country teaching and researching on circuit theory and radar and remote sensing systems. His research interests focuses on new digital TV services and applications. Juan Luis Ordiales received the M.S. Engineering from the Polytechnic University of Madrid in 1983. He received the Ph. D. Degree in Telecommunication Engineering from the University of the Basque Country, Spain in 1996. Since 1997 he is Associate Professor at the University of the Basque Country teaching and researching on Broadcasting. Since 1993 he is Vice-Dean of the Telecommunication Engineering School in Bilbao. Amaia Arrinda received the M.S. the Basque Country, Spain, in 1993. Before receiving the degree, she studied for a year at ENST Bretagne, and worked as a researcher at CNET, both in France. In September 1993 she joined the Department of Electronics and Telecommunications of the University of the Basque Country, where she is currently an assistant professor. In February 2001 she presented her doctoral thesis dealing with interferences between terrestrial analog and digital TV transmissions and she received the Ph. D. Degree. Her current research interests include digital TV and radio signal propagation, measurement and simulation inside the TSR research group.