Interference from future mobile network services in frequency band MHz to digital TV in frequencies below 790 MHz.

Transcription

1 Interference from future mobile network services in frequency band MHz to digital TV in frequencies below 790 MHz. TABLE OF CONTENTS 1 SUMMARY BACKGROUND /INTRODUCTION APPROACH, ASSUMPTIONS AND LIMITATIONS APPROACH ASSUMPTIONS LIMITATIONS TYPICAL DVB-T HOME INSTALLATIONS AND RECEPTION DVB-T RECEIVERS CHARACTERISTICS OF THE DVB-T ROOF TOP ANTENNA FEEDER CHARACTERISTICS PORTABLE OUTDOOR/INDOOR AND MOBILE RECEPTION DVB-T PLANNING PARAMETERS AND DEFINITIONS Coverage definitions Orthogonal polarization discrimination CHARACTERISTICS FOR THE MOBILE NETWORK SITE AND TRANSMITTER DATA BASE STATION ANTENNA PATTERN INTERFERENCE SCENARIO WITH HIGHER BASE STATION ERP SELECTION OF SITES FOR THE MOBILE NETWORKS PROTECTION RATIOS AND DVB-T RECEIVER OVERLOAD THRESHOLD PROTECTION RATIOS CHANNEL ARRANGEMENT AND CHOICE OF PROTECTION RATIOS DVB-T RECEIVER OVERLOAD THRESHOLD DESCRIPTION OF DIFFERENT INTERFERENCE CASES OVERVIEW/GENERAL HOLES IN DVB-T COVERAGE AREA DVB-T ROOFTOP RECEPTION DVB-T PORTABLE RECEPTION RECEIVER OVERLOAD INDOOR SITES INTERFERENCE FROM THE UPLINK DESCRIPTION OF WORKING METHODOLOGY SOFTWARE AND DATABASES USED THE WORKING PROCESS: DVB-T CALCULATIONS...25 Page 1 (65)

2 8.3 THE WORKING PROCESS: THE MOBILE NETWORK CALCULATIONS THE WORKING PROCESS: SUMMATIONS AND ANALYSIS RESULTS OF CALCULATIONS OVERVIEW OF THE CALCULATIONS DVB-T POPULATION AND COVERAGE AREA WITHOUT INTERFERENCE INTERFERENCE CALCULATION FOR DVB-T ON CHANNEL 60 FOR THE WHOLE COUNTRY FIXED RECEPTION IMPACT OF POLARIZATION DISCRIMINATION IMPACT OF HIGH POWER MOBILE NETWORK IMPACT FOR PORTABLE INDOOR RECEPTION IMPACT FOR PORTABLE OUTDOOR RECEPTION INTERFERENCE CALCULATION FOR DVB-T CH 58 AND 59 FOR TWO AREAS ANALYSIS OF RESULTS GENERAL INTERFERENCE INTO AREAS WHERE DVB-T CHANNEL 60 IS USED CALCULATIONS IN SMALLER GEOGRAPHICAL AREAS- CHANNEL 60 IN NORRKÖPING AND SKÖVDE INTERFERENCE INTO DVB-T CHANNEL 58 AND CONCLUSIONS POSSIBLE TECHNIQUES TO REDUCE INTERFERENCE RECOMMENDED FUTURE STUDIES REFERENCES...44 ANNEX 1: DVB-T SYSTEM PARAMETERS...46 ANNEX 2: SITE CHARACTERISTICS AND DISTRIBUTION OF CHANNEL 60 DVB-T SITES...47 ANNEX 3: SITE CHARACTERISTICS AND DISTRIBUTION OF CHANNEL 59 DVB-T SITES...49 ANNEX 4: SITE CHARACTERISTICS AND DISTRIBUTION OF CHANNEL 58 DVB-T SITES...52 ANNEX 5: SITES AND DISTRIBUTION OF THE MOBILE NETWORK...54 ANNEX 6: RESULTS OF THE CALCULATIONS...56 ANNEX 7: GENERAL ON LOGNORMAL ADDITION...62 ANNEX 8: EXTRACT FROM THE REPORT RADIATION PATTERN CHARACTERISTICS OF UHF BAND V TELEVISION RECEIVING ANTENNAS...63 ANNEX 9: RESULTS OF MEASUREMENTS OF DVB-T RECEIVER OVERLOAD THRESHOLDS FROM TDF [11], DEC Page 2 (65)

3 1 Summary This report presents the results of a study of potential interference from a future mobile service in the frequency range MHz into digital TV broadcasting below 790 MHz in Sweden. The focus of the study is to assess potential interference from the mobile downlink into DVB-T channel 60 assuming fixed rooftop antennas. Calculations of the existing DVB-T networks coverage on channels have been used as a basis, while the mobile network is based upon an existing GSM 900 network in Sweden. With the assumptions made, the result of the calculations shows that the potentially interfered population in the channel 60 areas is between 2.5% and 17%. The results are very sensitive to the assumed protection ratio between the mobile network service and the DVB-T service. The results presented above are valid for protection ratios -30 db and -15 db, respectively. The area where interference may occur is relatively small compared to the population affected. This is due to the large number of mobile base stations present in densely populated areas where there is need for high capacity in the mobile networks. In order to minimize the interference from the mobile network it seems necessary to insert a frequency separation (guard band) of 2 MHz above 790 MHz. This will reduce the impact of interference to less than 8% of the population where DVB-T channel 60 is received. Interference into DVB-T channels 58 and 59 from the mobile downlink seems to be significantly less compared to the case of DVB-T channels 60, due to the increase of the protection ratios at larger frequency separations. Interference into portable DVB-T reception seems to affect a larger relative area compared to the rooftop reception cases, when receiving antenna discrimination is applicable. However the relative population affected remains in the same order in the two cases. Further to the insertion of a 2 MHz frequency separation between the broadcast and the mobile downlink (guard band), other mitigations techniques may be needed to reduce the impact of the interference. These techniques could involve: Use of vertical polarisation at the mobile base stations, will reduce the impact of interference significantly Power limitations for the mobile base stations in areas where channel 60 is used Use of a more stringent spectrum mask at the mobile base stations (BS) Insertion of a passive filter before the RF front end in the DVB-T receiver Further studies are recommended to access the impact of these measures and the possible impact of DVB-T receiver overload. Further analysis when protection ratios between LTE and DVB-T become available is also suggested. 2 Background /Introduction ITU-R WRC 2007 decided to allocate 72 MHz, in the frequency range MHz (TVchannels 61-69) in region 1 for mobile services on a co-primary basis. Following this the Swedish government decided in December 2007 that this sub band should no longer be allocated to Page 3 (65)

4 broadcasting. Broadcasting services in band IV and V would be limited to the frequency band MHz, i.e. to broadcasting channels When deploying mobile services in the frequency range MHz there is a potential risk of adjacent channel interference into existing digital TV services using channels below channel 61 (below 790 MHz). The purpose of this study is to investigate the potential impact of the interference from a future mobile network into existing broadcast services and to suggest possible solutions to reduce the interference into the broadcast services. 3 Approach, Assumptions and limitations 3.1 Approach The final channel allocation for the mobile service is not yet defined. However it seems probable that FDD (Frequency Division Duplex) with reversed duplex will be used. Reversed duplex means that the downlink, from mobile base station (BS) to the mobile terminal (MS) will use the lower part of the sub band. The uplink from the mobile terminal to the base station will be allocated in the upper part of the sub band. This arrangement is used in order to minimise interference from the mobile terminal (UE) into broadcasting services below 790 MHz. A likely channelization for the mobile service would be 6 channels of 5 MHz each for the downlink, a duplex gap between uplink and downlink of 10 or 12 MHz and 6 channels of 5 MHz each for the uplink see figure 1. Mobile Service D1 D2 D3 U1 U2 U Figure 1: Possible channel arrangements in the sub band MHz, without guard band The focus of the study is potential interference from the downlink of mobile service into existing broadcast services on the channels Possible interference from the mobile terminals (UE) into broadcast is not analysed in this report. The limitations and assumptions listed below are mainly related to the actual interference calculations and the procedures. 3.2 Assumptions In order to complete this study it has been necessary make a number of assumptions. The main assumptions are: Page 4 (65)

5 o The coverage criterion for DVB-T is 95% of locations, with a lognormal field strength standard deviation of 5.5 db. This coverage criterion is in accordance to the GE06 agreement [9] o A limited number of antenna patterns were available for DVB-T. In case the antenna pattern was not available for a DVB-T site, omni-directional antennas has been assumed. This approach is however believed to have limited impact upon the results o DVB-T transmissions are based upon using the system variant 8k, 64-QAM, code rate 2/3 o Calculations are made for fixed rooftop reception for the broadcasting service. For one geographical area calculation has also been done assuming portable reception o The same antenna pattern is used for all of the mobile networks transmitters, with 1-3 antenna directions (sectors) o Interfering signals are also subject to lognormal fading with a standard deviation of 5.5 db 3.3 Limitations When evaluating the impact of interference the following limitations exist: o Protection ratios for broadcasting interfered by UMTS (WCDMA) and WiMAX should be regarded as preliminary. In the report protection ratios are used that can represent both WiMAX and W-CDMA. Protection ratios for DVB-T interfered by LTE (Long Term Evolution) are currently not available o For DVB-T channel 58 and 59, evaluation is only made for two geographical areas (Kalix/Överkalix and Skellefteå). The distribution of channel 58 and 59 in Sweden is describe in annex 3 and 4 o For portable DVB-T reception on channel 60, the investigation is only made for two areas (Norrköping and Skövde). The areas where channel 60 is used is shown in annex 2 o Calculations assuming orthogonal polarization discrimination are only made for one area (Norrköping) o Calculation assuming high output EIRP (Effective Isotropic Radiated Power) at the base stations is only made for one area (Norrköping) o For the field strength predictions, a height and clutter database with a resolution of 50 metres has been used. This resolution is not always sufficient to detect small coverage holes in urban areas when base stations with low EIRP are used 4 Typical DVB-T home installations and reception 4.1 DVB-T receivers Most of the DVB-T receivers today are so called set-top boxes connected, using a SCART connector or the antenna input, to an ordinary analogue TV receiver. New flat screen TVreceivers often has an integrated DVB-T receiver, IRD (Integrated Receiver Decoders). With the large sale of flat screen TV s it can be assumed that integrated receivers may dominate the market within a few years. Technical characteristics of DVB-T receiver and its ability to operate under non-interference and interference conditions are described in terms of a number of parameters. Important DVB-T receiver parameters are: o Receiver noise figure Page 5 (65)

6 o Required C/N value o Protection ratios for both co-channel and adjacent channels for different services o The overload threshold of the DVB-T receiver (the input level of any interfering signal when the receiver looses the ability to decode any wanted input signal) These parameters are described in a number of documents, for example ITU-R BT [6], the GE06 agreement [9] and the DVB-T implementation guideline [10]. The two first bullet points describes the DVB-T receiver sensitivity and C/N values under different receiving conditions. They are well defined and established since several years. Protection ratios between broadcast services are also well defined and internationally agreed, for example in the GE06 agreement. Protection criteria for broadcast services interfered by UMTS (WCDMA) services are still not fully defined. However it is expected that these will be agreed in the near future, within the CEPT and ITU-R. In the case of, E-UTRA (LTE), there is currently a lack of measurement results. Since LTE is expected to be the main mobile standard implemented in the sub band MHz, this represents a difficulty. This is discussed in more detail in section 6, protection ratios. 4.2 Characteristics of the DVB-T roof top antenna Normally DVB-T transmissions in Sweden are horizontally polarized. This is mainly explained by the wish to reuse existing (earlier analogue) transmitting antennas and rooftop receiving antennas. In some countries like the Netherlands and Germany vertical polarization is commonly used. Vertical polarization provides higher field strength at low receiver antenna height and is therefore considered to improve portable (indoor and outdoor) reception. There is wide range of consumer roof top antennas with different characteristics. The most common antennas are grid- and yagi antennas. The characteristics of these antennas differ slightly. Figure 2 and 3 shows examples of grid and yagi antennas. Figure 2: Grid antenna Figure 3: Yagi antenna Grid antennas have a rather wide radiation pattern in the horizontal plane (assuming horizontal polarization) and have a relatively high gain across the whole UHF frequency band. The yagi antennas however, have a narrower beam width but offers higher gain compared to the grid antenna. Page 6 (65)

7 The grid antennas are more suitable in areas with good coverage and were several DVB-T transmitters are combined into an SFN. At the edge of the DVB-T coverage area, where the field strength is lower, the yagi antenna with high gain and directivity is appropriate. Table 1 summarizes some important characteristics for the yagi and grid antenna from two main manufactures. Table 1: Antenna characteristics Antenna # Manufacture Type # of elements Gain F/B ratio (db) Beam width º (H) (db) 1 A Grid A Grid B Grid A Yagi A Yagi B Yagi B Yagi These characteristics are also confirmed by a measurement of radiation pattern characteristics for a number of UHF, TV receiving antennas, summarized in a draft ITU-R report [1]. In the summary of measurements [1] from ITU-R, a comparison is made between the antenna mask in the planning recommendation ITU-R BT.419 [2], and the measured antenna patterns. The report shows that a number of antennas are well within the mask, but others are not due to their wider horizontal pattern and lower gain. An interview, [7], with a representative from company installing DVB-T receiving antennas, confirms that the most common choice for new antennas for TV reception today is the grid antenna. It is used in areas with good DVB-T coverage. The wide horizontal antenna pattern also provides a benefit when used in Single Frequency Networks, SFN s. Yagi antennas are used where the field strength is lower and there is a need for additional antenna gain to achieve acceptable coverage. In this study we have assumed receiving antenna characteristics according to ITU-R BT.419 since most of the measured antenna diagrams are within this antenna mask. The antenna gain for channel 60 (786 MHz) used in the study is 11.9 dbd. The beam width of the antenna is about 25 degrees (-3dB). The maximum directivity discrimination is reached at 60 degrees (-16dB). A polarization discrimination of 16 db is applied in case of use of orthogonal polarization for the wanted broadcast and the interfering mobile service. In houses with a rooftop antenna, there is often no need for signal amplifiers, if less than four antenna sockets are used. In larger installation with more than three antenna sockets and longer feeder cables, there may be a need for antenna amplifiers. Antenna amplifiers often have a broadband input, which may result in interference from a future mobile service into broadcast. The impact of use of antenna amplifiers has not been considered in this study. Page 7 (65)

8 4.3 Feeder characteristics The antenna feeder used today is often triple-shielded feeders. A triple-shielded feeder has good HF -isolation with values over 85 db. This means that theoretically there is no impact of high field strength with respect to the feeder. In some practical cases the output plates in the rooms/houses may be a limitation. In such cases the main problem will be uplink interference from a mobile terminal operating in the vicinity of the TV-receiver, operating in the same room or house. Interference from indoor mobile sites (downlink) may also be a problem in such cases. The feeder loss depends on both the frequency and the dimension of the cable. The feeder cable loss is between 10 db/100 metres and 25 db/100 metres at 800 MHz depending on feeder dimension. An average home installation has a cable length of metres. Common feeder dimensions are (Ø for the inner conductor) 1.6 mm, 1.0 mm and 0.8 mm. The choice of feeder dimension depends on both the cable length and possible restrictions present at each installation. Table 2 shows the feeder loss for different feeder dimensions and cable length. Table 2: Feeder loss at different cable length Cable length Feeder loss 13.4 MHz for 1.6 mm. Feeder loss 20.0 MHz for 1.0 mm. Feeder loss 25.5 MHz for 0.8 mm In the study a feeder loss of 4.9 db is used as an assumed value for a medium quality feeder (1.0 mm) of 25 m and at channel 60 (786 MHz), which is close to the 5 db value highlighted in table Portable outdoor/indoor and mobile reception Portable and mobile DVB-T reception differs from fixed rooftop reception in several ways: Receiving antenna height is lower compared to fixed reception, calculation of portable and mobile DVB-T coverage normally assume an antenna height of 1.5 metres For portable and mobile reception, the antenna gain is much lower. In the GE06 frequency plan the antenna gain is set to 0 dbd for portable indoor antennas For portable reception no antenna directivity is assumed Page 8 (65)

9 No polarization discrimination can normally be applied in case of orthogonal polarization between the wanted and the interfering service Active antennas with built-in amplifiers are often used for portable reception in caravans and boats The coverage area for portable and mobile DVB-T reception is much smaller compared to fixed reception. The reason is that the required field strength needs to be higher compared to the fixed reception case. In this study we also will analyse the impact of interference into portable reception. However the main focus in upon fixed rooftop reception, as pointed out earlier. 4.5 DVB-T Planning parameters and definitions Coverage definitions The DVB-T coverage definition is based upon GE06 agreement for both rooftop and portable reception, which are internationally agreed. Unless otherwise stated the GE06 parameters are used as a basis in this study. Both for DVB-T rooftop and portable reception a coverage level of 95% of locations is used. It is graded as Good reception and regarded as sufficient by network operators for providing coverage. An area (one pixel, of say 50 x 50 metres) is considered as covered when the location probability for reception exceeds 95%. This value should not be mixed up with the total coverage area, which consists of all the pixels which fulfil the coverage of 95%. When calculating the coverage area for DVB-T so-called lognormal summation is used, using the Schwartz and Yea summation method. This applies to both the interfering (DVB-T and mobile base stations) as well as the wanted DVB-T signals. A lognormal standard deviation of 5.5 db has been used for both wanted and interfering signals. Since the distance between the mobile base station and DVB-T receiver normally is quite short and free space propagation may apply it can be discussed whether a 5.5 db standard deviation should always be applied for signals from the interfering mobile base station. However this is not believed to have a significant influence on the results presented here. Annex 1 gives a summary of the most important DVB-T system and planning parameters used in the study Orthogonal polarization discrimination Most of the mobile networks today use cross-polarized transmitting antennas, so called X-pol. For this reason we have not applied any polarization when DVB-T is interfered by a mobile base station. In the case of interference from DVB-T into DVB-T, polarization discrimination has been applied when applicable. However a large majority of the DVB-T transmissions use horizontal polarization, which means that polarization discrimination is normally not applied, even for the DVB-T vs. DVB-T case. Page 9 (65)

10 When applied the polarization discrimination factor is set to 16 db, according to [2]. This means a further interference reduction of 16 db in the log-normal summation compared to if polarization discrimination would not be applied. When polarization discrimination is applied, no antenna discrimination is applied. It should also be noted that the polarization discrimination is not a fixed value, it shows a considerable spread. Measurements have shown that values between 5 db and 25 db can be expected. 5 Characteristics for the mobile network 5.1 Site and transmitter data In order to make a realistic estimation of possible interference from a future mobile network in the MHz sub band into broadcasting, it is important to make a realistic assumption about the structure and characteristics of the mobile network. In order to achieve a high population and area coverage it is believed that a future mobile network in the MHz frequency band will be rolled out across the whole country. To simulate such a mobile network in our calculations, we have used a mobile network based upon an existing GSM900 network. The reason for this choice is that existing GSM900 MHz networks has a very high population and area coverage percentage. The existing GSM900 mobile network is distributed across the whole country using all types of base stations, from large macro cell base stations in rural areas, to small micro cell base stations in urban areas. Using data from existing network as a basis for our simulations makes it possible to create a new mobile network in the sub band MHz which can be assumed to be close to what a future LTE or WCDMA network would look like. The antenna pattern given in section 5.2 has been applied to each of existing GSM900 base stations. The assumed mobile network has the following characteristics: Site positions are based upon an existing GSM900 network Antenna heights are the same as an existing GSM900 network Number of sectors on each site is the same as an existing GSM900 network In case a higher number of sectors than 3 are indicated in the input data, the numbers of sectors have been limited to 3 Antenna directions are the same as an existing GSM900 network Each site in the mobile network has the following characteristics: Antenna data: o Gain: 16.5 dbi o Half power (-3dB) beam width (H-pol): 65 deg o Half power (-3dB) beam width (V-pol): 10 deg o Electrical down tilt: 3 deg Same output Base station transmitter power as in the existing GSM900 network Page 10 (65)

11 Resulting in a Base station ERP in the range from: 1.5 W 1 kw (32-60dBm); while typical ERP values are between 150 and 300 W (52-54 dbm), this is about db below the maximum ERP of 62 dbm indicated by Telia Sonera in [14] Polarization: slant polarization (X-Pol) (+45º / -45º polarization) Feeder: 7/8 (commonly used in mobile networks like GSM900) Frequency reuse factor: 1 Channel bandwidth: 5 MHz 5.2 Base station antenna pattern The antenna patterns for the base stations is a Kathrein antenna. Antennas with this type of antenna pattern are commonly used in the existing mobile networks like GSM900. The BS antenna pattern is given in figure 4 and 5. Figure 4: Horizontal polarization Figure 5: Vertical polarization 5.3 Interference scenario with higher base station ERP For one DVB-T area (Norrköping) interference calculations has also been made assuming higher base station output power, relative to those powers previously given by the GSM900 input data. The background to this calculation is that mobile operators have indicated the interest to use substantially higher ERP at some of the base stations. In [14] TeliaSonera indicates a maximum EIRP of 64 dbm (representing an ERP of approximately 62 dbm). This is value is higher than the original values available in the GSM900 input data, described earlier. The base station ERP is calculated using the following parameters: 46 dbm (40 W) base station transmitter output power 16.5 dbi antenna gain 3.5 db/100m feeder loss. The feeder length was set to antenna height and the base station ERP was calculated. Table 3 shows the feeder loss for a 7/8 feeder with 3.5 db/100m feeder loss at different antenna heights. Page 11 (65)

12 Table 3: Feeder loss at different antenna heights Antenna height (m) Feeder Loss (db) Table 4 shows the base station ERP for different antenna heights with the parameters above. Table 4: ERP for different antenna heights Antenna height (m) ERP (dbm) ERP (W) In this scenario the base station ERP is between W (54-60 dbm), with typical ERP s in the range W (58-59 dbm). 5.4 Selection of sites for the mobile networks To limit the calculation time in the interference simulations, we decided to reduce the number of mobile base station sites. Only mobile network sites located inside the DVB-T coverage areas (95% coverage probability), without taking the interference from the mobile network into account, were included in the calculations. The interference limited coverage area was estimated for each DVB-T transmitter included in the study. In the summation, each DVB-T transmitter was used as a wanted transmitter, while all other co-channel DVB-T transmitters in Sweden and neighbouring countries were used as interfering sites. Two different methods were used when deciding which mobile base station sites to include depending on the size of the coverage area of the DVB-T transmitter. Method 1(small DVB-T sites, fill in sites) Mobile base stations, which are within the 95% coverage area of the DVB-T fill site, are included in the interference calculation. The 95% coverage area is calculated using (1% time) interference from neighbouring DVB-T sites. Method 2(large DVB-T sites) For the DVB-T main sites with large coverage areas, a slightly different method has been applied. The reason for this is that there is often a large coverage overlap between neighbouring DVB-T main stations. We have then assumed that the viewers have their antenna turned in the Page 12 (65)

13 direction of the strongest DVB-T station. Only mobile base stations located within the DVB-T best server area are included in the calculations. Figure 6 8 describes the two different selection methods. DVB-T coverage area based upon 95% probability. Mobile network sites inside the coverage area are included in the calculation. Figure 6: Method 1. Mobile sites within the interference limited coverage (95%) for a small fill-in site are taken into account Page 13 (65)

14 DVB-T coverage area for 95% probability. DVB-T best server area for the same site. Mobile network sites inside the best server area that is included in the calculation. Figure 7: Interference limited coverage for a DVB-T site with a large coverage area. Figure 8: Best server coverage for the same site. 6 Protection ratios and DVB-T receiver overload threshold 6.1 Protection ratios In order to assess the impact of possible interference from mobile services above 790 MHz into DVB-T below 790 MHz there are two main important parameters. These are the RF protection ratios and the DVB-T receiver overload threshold. The protection ratio is defined as a ratio between the wanted and the unwanted signal at a certain quality criteria, which will ensure error-free reception. In order to reduce the impact of noise, protection ratios are normally measured at the wanted input signal level, which is 3 db above the noise floor: PR (db) = FS (wanted) [DVB-T] / FS (unwanted) [UMTS/WIMAX/LTE] More details on protection ratios are given in [11] an input to the CEPT ECC TG4. Within a framework of CEPT ECC TG4 (Digital Dividend) a number of measurements has been carried out in order to verify the protection ratios. Some of these studies has been summarised in an input document to ITU-R JTG 5-6 [8] and CEPT ECC TG4, from October It should be pointed out the values are preliminary and they may be changed slightly as more results become available. In [8] the following adjacent channel protection ratios are given for DVB-T interfered with by UMTS and WiMAX. Page 14 (65)

15 Table 5: Measured protection ratio in db for different wireless systems Network type N+6.5 MHz N+1 N+2 UMTS BS without TPC UMTS BS with TPC UMTS UE with TPC WiMAX BS (10 MHz) Mobile WiMAX UE with TPC (5 MHz) UMTS and WiMAX networks use so-called Transmit Power Control (TPC), both for the uplink and the downlink. TPC is used in order to reduce interference in the mobile networks and will reduce the power consumption of the mobile terminal (UE). In table 5 protection ratios are given with and without TPC. It should be noted that the measurements indicates a difference of 21 db in the PR between the TPC and the non TPC case, at a frequency difference of +6.5 MHz. The reason for this large difference is not clear. Further investigation may be needed. However, one possible explanation is that the DVB-T receivers seem to be sensitive to rapid changes in interfering levels caused by the BS TPC. When calculating the impact of the interference from the UMTS BS into DVB-T there is in principle two alternatives: Use of PR valid for the non TPC case and use the maximum allowed ERP of the UMTS base station. This alternative has the advantage that it is easy to know what the find possible limitations in terms of BS ERP Use of PR valid for the TPC case combined with the use of average ERP of the mobile base station. Since all UMTS base stations use TPC this would relate more closely to reality. In this case there would be a need to specify the average ERP of the base station relative to the maximum ERP. A figure of -6 (-12 db) is often used In this study we have decided to use the latter alternative. The reason is: All UMTS base stations will use TPC. This is also true for LTE networks. It is a more critical case since the difference between average and the maximum power is only 8 db while the increase of PR is in the order of 20 db at a frequency offset of +6.5 MHz In the measurement report [8} the Protection Ratio for UMTS BS and UE with TPC it is stated that the PR is a worst case and better values will be found in a live network with a large number of mobile stations active. This comment applies in particular to the UMTS BS (downlink). However, since it cannot be assumed that the mobile network is always fully loaded we have decided to use the PR s applicable for the TPC case. Use of these values will also mean that the protection ratio values given refer to the average power of the unwanted UMTS or WiMAX signals. Page 15 (65)

16 For this reason, the ERP of the mobile base stations used in this study should be regarded as an average ERP, not a maximum ERP. This is valid for all types of area, rural as well as urban areas. A closer look at the site data (section 5) shows that the ERP values given are 10 to 30 db below the maximum ERP of 62 dbm suggested by Telia Sonera. This verifies that this assumption is well founded. Some comments to table 5: The measurements for WiMAX BS are made with a 10 MHz channel bandwidth and are therefore not possible to use since the new mobile services are likely to have a 5 MHz channel raster. A WiMAX service using 10 MHz bandwidth will overlap with the DVB-T service, using 8 MHz bandwidth. New mobile services above 790 MHz are believed to be OFDM based using the LTE standard. Currently there are almost no measurements of protection ratios for DVB-T interfered by LTE. For this reason the PR:s used in this study is based on the measurements on Mobile WiMAX UE for 5 MHz, high lighted in grey in table 5. This protection ratio will be applied for the downlink, PR in uplink and downlink are quite symmetrical. Comparison between PR for UMTS with TPC in downlink and uplink confirms this. In the absence of real LTE measurements, there is some uncertainty here. In order to handle this we have evaluated the impact of several PR values. The foundation for these values is described in the following section. 6.2 Channel arrangement and choice of Protection Ratios The measurements [8] of PR are made using an 8 MHz channel raster. As previously pointed out, the new mobile services above 790 MHz are likely to use 5 MHz channel raster. Therefore we have made an interpolation in order to obtain the appropriate PR for the correct difference in frequency, to be used in the calculation. The interpolation is based on the frequency difference between the DVB-T channel and the mobile base station (BS) channels 1-3. Since it is still discussed, whether a frequency separation distance of 2 MHz (guard band) should be used between DVB-T and the downlink of the mobile network, the interpolation is made both with and without a 2 MHz guard band. Figure 9 shows the channel arrangement with centre frequencies for DVB-T and the mobile BS without a guard band of 2 MHz. BROADCAST 790 MHz MOBILE 8 MHz 5 MHz DVB-T ch58 DVB-T ch59 DVB-T ch60 BS ch1 BS ch2 BS ch3 BS ch.. fc=786 MHz Figure 9: Channel arrangement without guard band. fc=792.5 MHz Page 16 (65)

17 Interpolation of PR without a guard band gives the following PR values (tables 6 8): Table 6: PR interpolation for DVB-T channel 60 without guard band. Mobile BS channel fc for Mobile BS ch fc for DVB-T Δf = DVB-T and BS MHz 786 MHz 6.5 MHz MHz 786 MHz 11.5 MHz MHz 786 MHz 16.5 MHz Interpolated PR (db) Table 7: PR interpolation for DVB-T channel 59 without guard band. Mobile BS channel fc for Mobile BS ch fc for DVB-T Δf = DVB-T and BS MHz 778 MHz 14.5 MHz MHz 778MHz 19.5 MHz MHz 778MHz 24.5 MHz Interpolated PR (db) Table 8: PR interpolation for DVB-T channel 58 without guard band. Mobile BS Δf = DVB-T Interpolation PR fc for Mobile BS ch fc for DVB-T channel and BS (db) MHz 770 MHz 22.5 MHz MHz 770 MHz 27.5 MHz MHz 770 MHz 32.5 MHz Figure 10 shows the channel arrangement with centre frequencies for DVB-T and mobile downlink with a guard band of 2 MHz. BROADCAST 790 MHZ MOBILE 8 MHz 2 MHz 5 MHz DVB-T ch58 DVB-T ch59 DVB-T ch60 BS ch1 BS ch2 BS ch3 BS ch.. fc=786 MHz fc=794.5 MHz Figure 10: Channel arrangement with a guard band of 2 MHz. Interpolation of PR with a guard band gives the following PR values (tables 9 11):. Table 9: PR interpolation for DVB-T channel 60 with a 2 MHz guard band. Mobile BS channel fc for Mobile BS ch fc for DVB-T Δf = DVB-T and BS MHz 786 MHz 8.5 MHz MHz 786 MHz 13.5 MHz MHz 786 MHz 18.5 MHz Interpolation PR (db) Page 17 (65)

18 Table 10: PR interpolation for DVB-T channel 59 with guard band. Mobile BS channel fc for Mobile BS ch fc for DVB-T Δf = DVB-T and BS MHz 778 MHz 16.5 MHz MHz 778MHz 21.5 MHz MHz 778MHz 26.5 MHz Interpolation PR (db) Table 11: PR interpolation for DVB-T channel 58 with guard band. Mobile BS Δf = DVB-T Interpolation PR fc for Mobile BS ch fc for DVB-T channel and BS (db) MHz 770 MHz 24.5 MHz MHz 770 MHz 29.5 MHz MHz 770 MHz 34.5 MHz From the tables above the following PR are used in the calculations: For interference calculation between DVB-T channel 60 and mobile services the PR values -15 db, db and db is used. By using these PR values most of the scenarios with and without a guard band are covered. For interference calculation between DVB-T channel 59 and mobile services the PR value db is used. For channel 59 an evaluation is only made for one DVB-T area without a guard band. Using a guard band gives better result since the PR is even better. For interference calculation between DVB-T channel 58 and mobile services the PR value db is used. For channel 59 an evaluation is only made for one DVB-T area without a guard band. Using a guard band gives better result since the PR is even better. Since there is a large difference when using a PR of -15 db and db, we have also made calculations using a PR of db. The choice of the value db is because this value is between -15 db and 29.5 db. The reason being the uncertainty regarding the protection ratio for DVB-T interfered with by LTE. 6.3 DVB-T receiver overload threshold The second important DVB-T receiver parameter is the so-called receiver overload threshold, Oth. Results of measurements of the Oth are reported, in [11]. In this document the overload threshold is defined as: "Overloading threshold (O th ) is the maximum interfering signal level expressed in dbm, where above that level the receiver loses its ability to discriminate against interfering signals at frequencies differing from that of the wanted signal. Page 18 (65)

19 This means that (DVB-T) reception will fail regardless of the wanted input signal level when it is exposed to a signal, which exceeds the overload threshold. The overload threshold varies with the difference in frequency between the wanted (DVB-T) and the unwanted (mobile service) signal. In [11] the Oth has been measured both using BS TPC as well as without TPC. The results of the measurements are presented in Annex 9. The Oth values are between 9 dbm and -1 dbm for the non TPC case. f I - f W (MHz) O th (dbm) Table 12: DVB-T receiver overload thresholds as measured in [11] In section 7.5, we will analyse the consequences of these measurement results in terms of interference radius around a base station. 7 Description of different interference cases 7.1 Overview/General There are a number of different interference cases between DVB-T broadcast and the new mobile sub band above 790 MHz. The mobile service uplink as well as the downlink may interfere with DVB-T. As stated before the focus of this report is to study interference from a mobile service downlink into DVB-T, below 790 MHz. In the case of interference from the downlink of the mobile service there are a number of different interference cases. Interference into DVB-T fixed rooftop reception Interference into DVB-T portable (indoor and outdoor) reception In both these cases interference can be caused either by DVB-T receiver overload that is by interfering signal levels exceeding the overload threshold or when the protection criteria is not fulfilled. We will treat the latter case first. 7.2 Holes in DVB-T coverage area The network structure for DVB-T consists of a relatively small number of sites often using high output power in combination with high antenna heights. The mobile network has a denser network structure, with a higher number of base stations and lower power levels. DVB-T receivers that are located in areas close to mobile network base stations are likely to receive relatively high field strength from mobile network base stations, compared to the field strength from DVB-T transmitter. The coverage holes appears when: Page 19 (65)

20 FS DVB-T < Fs mobile basestation +PR + Protection margin (95% locations) + DVB-T Antenna discrimination (orthogonal or direction) The coverage hole in the DVB-T coverage is schematically described in figure 11. DVB-T coverage Coverage holes caused by mobile network base stations. Figure 11: Holes in DVB-T coverage area caused by high interfering field strength from several mobile base stations The size and the shape of the coverage holes created will differ between fixed rooftop reception and portable DVB-T reception 7.3 DVB-T rooftop reception In the case of DVB-T rooftop reception the receiving antenna is normally directed towards the DVB-T transmitter. This will mean that antenna directivity needs to be considered in our calculations. In a case where the mobile base station is using X-pol it can be assumed that the horizontally and the vertically polarized field components have the same level. There are then two cases: No polarization discrimination is applied for the BS horizontal field component if the DVB-T transmissions are horizontally. Only antenna directivity discrimination is then applied For the BS vertical component no directivity discrimination is applied. Polarization discrimination is applied. The first cases will dominate, and be of interest since we do not add both polarization and directivity discrimination in accordance with the GE06 DVB-T planning parameters. However it can be discussed if a -3dB polarization discrimination should be applied due to the 45 degree slant polarization of the mobile BS. The size of coverage holes created by the mobile base station varies with the field strength for the wanted DVB-T transmissions, the ERP of the mobile base station and of course the protection ratio. Because of the DVB-T antenna directivity the size of the coverage hole will be larger at locations where there is a mobile base station in the direction of the DVB-T transmitter. An example is given in figure 12 below: Page 20 (65)

21 DVB-T transmitter Mobile base stations Figure 12: Coverage holes in DVB-T coverage area caused by high interfering field strength from several mobile base stations It can also be seen that the coverage holes get somewhat larger as you move away from the DVB-T transmitter. The reason is that the field strength of the wanted transmitter is lower. Very close to the edge of the coverage area the coverage holes can be quite large. Simulations carried out by EBU show that to size of a coverage hole for fixed reception could be between meters depending on power levels of the BS, using the Okumura Hata prediction model. This is in line with the results found here, although the prediction models differ. The Swedish DVB-T network is primarily planned for rooftop reception. For this reason it is the most important interference case to evaluate. 7.4 DVB-T portable reception For portable reception an omnidirectional antenna pattern is normally assumed. This also means that neither directivity nor polarizations discrimination can be applied. Page 21 (65)

22 This means that a coverage hole will be more or less round in shape, neglecting the fact the DVB-T field strength varies within the coverage area. DVB-T transmitter Mobile base stations Figure 13: Round coverage holes in DVB-T portable coverage area caused by high interfering field strength from several mobile base stations 7.5 Receiver overload In areas close to the mobile network base stations where the field strength is high there is a potential risk for DVB-T receiver front-end overload. Receiver front-end overload occurs when a strong unwanted signal makes it impossible for the receiver to detect the wanted signal. If the interfering signal is so strong that it overloads the receiver front-end the receiver becomes blind and is unable to receive anything at all. An overload receiver is unable to detect the wanted signal whatever the level of the wanted signal is. The overload threshold is the power level of the interfering signal at the receiver antenna connector when the receiver becomes overloaded. Page 22 (65)

23 Measurements of DVB-T receiver overload [11] on 10 different DVB-T receivers shows that the average overload threshold for N+6.5 MHz is about -9 dbm (table 12). One example of a overload distance assuming free space propagation Base station EiRP DVB-T Oth = 57.5 dbm = -9 dbm Needed path loss (PL) to avoid receiver overload can be calculated using the following formula: PL= BS EIRP (Oth) + antenna gain (dbi) - feeder loss (db) This gives: Portable reception PL = 57.5 (-9) = 66.5 db Rooftop reception: PL = 57.5 (-9) = 75.5 db (worst case) Applying the free space propagation formula results in Portable outdoor case 66.5 db 65 m Rooftop reception case 75.5 db 170 m (worst case- in antenna direction) A path loss of 66.5 db is equal to 65 m and 75.5 db is equal to 170 m calculated with free space at 800 MHz. This means that a base station with an EIRP of 57.5 dbm can overload DVB-T receivers at distances up to 170 m. This situation can occur if a base station transmitter antenna and TV receiver antenna is located near each other on a large roof and the antennas is pointed directly to each other. Also for portable reception the problem can occur for a TV viewer with a window facing towards a base station and using a small indoor portable antenna. It should be pointed out that the example presented above is a worst case. In many cases there is not line of sight between the two antennas. Additionally there may be some reduction of the interference radius due to the VRP of the base station as well as due to the DVB-T receiving antenna in the case of fixed reception. If the base station EIRP is reduced by say 20 db it will result in interference distances of a few meters. 7.6 Indoor sites Mobile network indoor base stations are built for both capacity and coverage. The indoor base stations are often found in office and commercial buildings. With an indoor base station the downlink field strength level is adjusted to lowest possible level with maintained coverage and capacity. When the indoor base station is transmitting on a low level and with DVB-T rooftop antennas for reception the risk of adjacent channel interference is limited. Page 23 (65)

24 7.7 Interference from the uplink Uplink interference is mainly a problem with portable (outdoor and indoor) DVB-T reception. The reason for this is that the wanted receiving antenna and the interfering mobile station (MS) need to be close to each other for interference to occur. The main interference in this case is DVB-T receiver front-end overload. Measurements by EBU/Technical Department [3] shows that standard commercial DVB-T set top boxes with an indoor antenna is interfered at distances of 2-6 m between the MS and the DVB-T receiver. To avoid interference the transmit power of the MS had to be reduced with 10 to 12 db. For a separation distance of 1 m, the power reduction required to avoid overloading would be more than 20 db. The measurement also shows problems with N+9 channel interference. To avoid N+9 channel interference the MS transmit power also required 10 to 12 db reduction. The measurement referred above [3] was made with an UMTS terminal simulation with WCDMA modulation. Measurements with other modulations such as OFDMA are unknown but one can assume that a MS with OFDMA modulation also will cause uplink interference problems to DVB-T set-top boxes. 8 Description of working methodology 8.1 Software and databases used All field strength calculations have been made using Progira s network planning software GiraPlan. The following databases have been used in GiraPlan for the coverage predictions: Digital terrain (height) data with 50 metres resolution for Sweden Clutter data (terrain description data) with 50 metres resolution Site data for existing DVB-T sites provided by PTS for field strength calculations Site data for a mobile network provided by PTS for field strength calculations Population data (raster data) with 100 metres in urban areas and 250 metres resolution in rural areas The relatively low resolution (of 50 metres) of booth the clutter and height data means that it will not be possible to detect and identify coverage holes in the DVB-T coverage caused by interference from the mobile base stations which has a smaller radius than about 50 metres. This will mainly be a problem in urban areas as the EIRP of the base stations is quite low there, resulting in smaller coverage holes. Field strength prediction model The field strength prediction model used is CRC-Predict version It is a terrain based model using height and clutter data. It is developed by Communications Research Centre (CRC) in Canada. The prediction model is used with good experience in DVB-T/H and DAB planning in several countries such as for example the Netherlands, Sweden, Finland and Switzerland. The used version has been tuned to provide low error based upon measurement data from the digital networks (DVB-H and DVB-T) in several countries. Page 24 (65)

25 Single Frequency Networks (SFN) DVB-T Transmitters in the same area, using the same frequency, form Single Frequency Networks (SFNs). When estimating the total coverage of the SFN the field strengths from each of the involved transmitters are combined using so called lognormal summation, where the (location) coverage probability is estimated. The lognormal summation takes into account both external interference (co-channel and adjacent channel) as well as SFN self-interference from wanted network itself, due to signal delays outside the guard interval. A certain pixel (for example of size 100 x 100 metres, 10 x 10metres or 2x 2 metres) having a coverage probability exceeding 95% is normally considered covered. 8.2 The working process: DVB-T calculations The working process can be summarized in the following steps: Import and verification of DVB-T site data provided by PTS Import of ITU site data for calculations of DVB-T interferers Calculation of DVB-T field strength on channels 58, 59 and 60 Lognormal summation of each DVB-T site using external DVB-T interferers to determine which mobile bases stations should be included in the calculations When a DVB-T transmitter has a large coverage area, a best server calculation is also made in order to decide which mobile bases stations are of interest in the calculation 8.3 The working process: the mobile network calculations The working process can be summarized in the following steps: Import of site data for a mobile network for each DVB-T transmitter s coverage area. The site data is provided by PTS For each base station a horizontal and vertical antenna pattern is applied. Same antenna pattern for all transmitters The number of transmitters for each base station is limited to 3 in the case that the site data provided by PTS contained more than 3 transmitters for a site The antenna direction for each transmitter is applied as given by the PTS site data. Calculation of the coverage for each mobile network base station 8.4 The working process: summations and analysis The working process can be summarized in the following steps: Lognormal summation of each DVB-T site with external DVB-T interferers Lognormal summation of each DVB-T site with external DVB-T interferers and also with the mobile network base station in the DVB-T coverage area Calculation of coverage area and population without interference from a mobile network Calculation of coverage area and population with interference from a mobile network Page 25 (65)