FM RECEIVER STUDY

FINAL REPORT
Nozema N.V. May 2004 Version 1.0

FM receiver study

Abbreviations
AF STIR IPC MAD MOS OBU PWC WTISR Audio Frequency Signal-To-Interference Ratio Industrial Personal Computer Mean Absolute Deviation Mean Opion Score Output Balancing Unit PairWise Comparison (Radio-Frequency) Wanted To Interfering Signal Ratio

Page 1

Table of content
Introduction... 10 Measuring protection ratio curves... 11
2.1 Receiver selection.... 11 2.2 Method.... 11 2.3 Measurement arrangement.... 12 2.4 Results..... 14 2.4.1 Radio-frequency protection ratio curves... 14 2.4.2 Sensitivity.... 15 2.5 Selection of the reference receiver... 16 2.6 Selection of the good and the bad receiver.. 18
Recording of the sound samples... 20
3.1 3.2 3.3 3.4 Test setup.... 20 Source sound samples.... 22 Sound sampled recorded with the reference receiver.. 22 Sound samples recorded with the good and the bad receiver.. 23
Subjective assessment of the sound samples.. 24
4.1 Participants.... 24 4.2 Procedure.... 24 4.2.1 Pairwise comparison.... 24 4.2.2 Mean Opinion Score.... 24 4.3 Analysis and results of the pairwise comparison... 25 4.4 Analysis and results of the mean opinion score.. 26 4.5 Protection ratios.... 26
Consequences for frequency planning.. 32
5.1 General.... 32 5.1.1 ITU.... 32 5.1.2 This study.... 32 5.2 Considerations on protection ratios for conventional networks.. 33 5.2.1 Comparison... 33 5.2.2 Conclusions regarding conventional networks.. 33 5.3 Considerations on protection ratios for same programme networks. 34 5.3.1 Comparison... 34 5.3.2 Conclusions regarding same programme networks.. 34 5.4 Considerations on protection ratios for HF-synchro networks.. 35 5.4.1 Comparison... 35 5.4.2 Conclusions regarding HF-synchro networks.. 35 5.5 Planning considerations.... 36 5.5.1 Usable field strength calculations... 36 5.5.2 Optimised networks... 36 5.5.3 International frequency coordination... 36 5.5.4 General observations... 37

6.1 6.2 6.3 6.4

Comparison with zero-base results in The Netherlands.. 38
General.... 38 Conventional network.... 38 Same programme network... 38 Synchronised network.... 39 Page 2

7.1 7.2

High signal performance... 40
Approach.... 40 Results..... 41

RDS Switching... 44

8.1 Approach.... 44 8.1.1 RDS switching behavior due to differences in radio frequency levels. 44 8.1.2 RDS switching behavior due to multipath.. 44 8.2 Results..... 45 8.2.1 RDS switching behavior due to differences in radio frequency levels. 45 8.2.2 RDS switching behavior due to multipath.. 46
Appendices A Detailed results of the protection ratio measurements B Example of a same programme network C Reception problems due to inter-modulation D Photographs of the tested receivers

Page 3

Management summary
The goal of this study is twofold. First, the technical characteristics of a representative group of present day FM receivers should be assessed. Second, the protection ratios for the network configurations conventional, same programme and HF-synchro should be determined using a reference receiver. From a set of thirty present day FM receivers, consisting of ten car radio, ten portables and ten handhelds, the radio-frequency protection ratio curve according to Recommendation ITU BS 641 was used to characterize a receiver. Initial tested showed that some of the car radios and most of the portables and handhelds were not able to meet the requirements of this Recommendation. With an adjusted value for the audio-frequency signal-to-interference ratio the protection curves of most receivers were determined. The receivers that were still not able to meet the revised audio-frequency signal-to-interference ratio were discarded. Based on the measured protection ratios the Sanyo DC-DA1000 was selected as reference receiver. The reference receiver was used to record an extensive set of sound samples for the network configurations: conventional, same programme and HF-synchro. The quality of the recorded sound samples was assessed by a representative panel. This assessment resulted in protections ratio curves for the network configurations conventional, same programme and HF-synchro were determined. In general the results of these subjective tests are in fair agreement with the Recommendation ITU-R BS.412-9. For planning of conventional network it is advised to use the radiofrequency protection ratio values of this recommendation. The Zero-Base study resulted in a 3 to 8 dB reduction for the protection ratios for conventional networks. These results, however, could not be confirmed by this study. For frequency planning of transmitters carrying the same programme it is advised to use the protection ratios based on a grade of 3,5 in case of steady interference. In case of tropospheric interference it is advised to use the protection ratios based on a grade of 3,0. The values for same programme and steady interference do not differ much from the ZeroBase values for MPX synchronisation, except for 100 kHz frequency difference. It should be noted that the Zero-Base value is optimistic. The Zero-Base study gave no directions for tropospheric interference. Because the results for synchronized networks are not significantly better than those of same programme and former are more complicated to operate than the latter it is advised to use same programme in stead of HF-synchro between transmitters carrying the same programme. Compared to the Zero-Base study the protection ratios for synchronized transmitters are higher for frequency difference of 0 and 100 kHz. The reason for this could be the different synchronisation used. High signal performance tests have shown that interference from third order inter-modulation products may occur around FM transmission sites where many frequencies are used. There the reception of signals from either a station with much lower power than the others, or from other more distant sites could lead to problems. In these situations frequencies should be Page 4

FM receiver study chosen such that no third order inter-modulation products occur in the pass-band of the receiver tuned to the low level signal. If this is not possible either the power of the low power transmission should be increased or a gap-filler should be installed to achieve the required protection ratio.

Page 5

Summary
Introduction The Expertengruppe UKW 2001 has recommended to optimize FM-networks because FM will remain the main modulation technique for radio for the coming fifteen to twenty years. FM networks could be optimized by revising the protection ratios used for frequency planning. The goal of this study is twofold. First, the technical characteristics of a representative group of present day FM receivers should be assessed. Second, the protection ratios for the network configurations conventional, same programme and HF-synchro should be determined using a reference receiver. Approach From a set of thirty present day FM receivers, consisting of ten car radio, ten portables and ten handhelds, the technical characteristics were determined as follows. First, the characteristics of the individual receivers were measured according to Recommendation ITU-R BS.641. Based on these measurements a good, a reference and a bad receiver was selected from the total group of receivers. Next, the protection ratios for frequency planning for three different network configurations based on subjective assessment of sound samples recorded with the good, the reference and the bad receiver were determined. Finally, the high signal performance and the RDS switching behaviour of a select group of receivers was investigated. This summary will present the results and the conclusions for each of these steps. Results and conclusions For this study the radio-frequency protection ratio curve was used to characterize a receiver. This protection curve is determined according to Recommendation ITU BS 641. Initial measurements showed however that some of the car radios and most of the portables and handhelds were not able to meet the minimal audio-frequency signal-to-interference ratio of 56 dB. With a audio-frequency signal-to-interference ratio of 46 dB it was possible to measure about eighty percent of the receivers. The rest was discarded. The results of these measurements are presented in the figure below.

KENWOOD KDC-3024A [C] P NASONIC CQ-RDP162N [C] A P NASONIC CQ-RDP003N [C] A BECKE MEXICO PRO CD 4627 [C] R BLAUPUNKT WOODSTOCK DAB 52 [C] SUP ERTECH AR-921 CD [C] JVC KD- SX997R [C] JVC KS- FX480REX [C] VDO DA YTON CD 2200 [C]

wtisr[dB]

0 -400 -300 -200 -100 -300 400
SANYO DTA-300M [P] P HILIP AZ3012 [P] S SANYO DC- DA1000 [P ] SONY CFD-S550L/SC [P] THOMSON RR 600CD [P] DIGITALWAY FD100 [H] P HILIP AZT9500 [H] S SONY ICF-C1200 [H] UNITE DM2595-2 [H] D AIWA HS-RM539 [H] SONY WM-FX491[H]

f [kHz]

Figure S1: Radio-frequency wanted-to-interfering signal ratio (wtisr) for twenty receivers recorded according to ITU Recommendation BS.641. The category to which the receivers belongs is indicated between brackets. The letters C, P and H are used for respectively the category car radios, portables and handhelds.

Page 6

FM receiver study From all the radio-frequency protection ratio curves an average and a median radio-frequency protection ratio curve was calculated. The Sanyo DC-DA1000 was chosen as reference receiver since it had the lowest mean absolute deviation from the mean and median radiofrequency protection ratio curve. Based on their protection ratio curves it was decided that the Blaupunkt Woodstock DAB52 represents a good receiver and the Sony WM-FX491 represents a bad one. The radio-frequency protection ratio curves of the good, the reference and the bad receiver as well as the Zero-Base reference receiver are shown in the figure below.
-400 -300 -200 -100 -10 -20 -30 -300 400
AVERAGE RECEIVER MEDIAN RECEIVER SANYO DC-DA1000 [P] ZEROBASE REFERENCE RECEIV E R
Figure S2: Comparison between the Zero-Base reference receiver, the Sanyo DC-DA1000, the mean and the median receiver of this study.
The reference receiver was used to record an extensive set of sound samples for the network configurations: conventional, same programme and HF-synchro. This extensive set covered a predetermined range of different receiving conditions such as frequency differences, delays and signal-to-noise ratios for wanted signals speech and classical music. To limit the total number of receiving conditions a limited set of sound samples per network configuration were recorded with the good and the bad receiver. The recorded sound samples were used in subjective tests. The goal of these tests was to assess the quality of the recorded sound samples using pairwise comparison and the mean opion score. From these results the protections ratio curves for the network configurations conventional, same programme and HF-synchro were determined. The protection ratio curve based on a grade of 3,5 for speech samples recorded with the reference receiver is presented in the figure below.

FM receiver study higher. Also an increase in protection ratio with delay times was expected. The results show, in particular for the 0 s case, a relatively high value. Compared to the Zero-Base study the protection ratios for synchronized transmitters are higher for frequency difference of 0 and 100 kHz. The reason for this could be the different synchronisation used. Since a synchronized network is more complicated to operate an the results are not significantly better than those of same programme it is advised to use same programme in stead of HF-synchro between transmitters carrying the same programme. High signal performance tests have shown that interference from third order inter-modulation products may occur around FM transmission sites where many frequencies are used. There the reception of signals from either a station with much lower power than the others, or from other more distant sites could lead to problems. In these situations frequencies should be chosen such that no third order inter-modulation products occur in the pass-band of the receiver tuned to the low level signal. If this is not possible either the power of the low power transmission should be increased or a gap-filler should be installed to achieve the required protection ratio.

Page 9

1 Introduction
This report describes the results of the FM receiver study according to the request for tender from SRG/SSR/ide Suisse from 25 June 2002 and the conclusions reached at the meeting between representatives of the Swiss Companies and Nozema on 26 May 2003 in Lopik.

Page 10

2 Measuring protection ratio curves
The first step in this receiver study is to determine the radio-frequency protection ratio curves of thirty receiver. This should be done according to Recommendation ITU-R BS.641. Based upon the radio-frequency protection ratio curves a reference receiver will be selected. The following paragraphs describe the selection procedure. 2.1 Receiver selection Thirty receivers were used in this study: ten car radios, ten portables and ten handhelds. SRG/SSR provided a list with the make and type of the radio. Per category a high and a low end type, based on price, was selected. The tables 2.1 to 2.3 give an overview of the selected receivers. Also the Zero-Base reference receiver, the NAD 1600, was included in the tests. 2.2 Method Recommendation ITU-R BS.641 is used to the determine the radio-frequency protection ratio curves of the receivers. This recommendation indicates that the audio-frequency signal-tointerference ratio should be at least 56 dB. Initial tests with only a few receivers indicated that the 56 dB audio-frequency signal-to-interference ratio could not always be reached. To be able to objectively compare receivers it is important that all receivers use the same minimal audio-frequency signal-to-interference ratio as starting point for the determination of the radio-frequency protection ratio curve. Receivers which are not able to reach the minimal can not be taken into account. Therefore, it was decided to determine the maximum attainable audio-frequency signal-to-interference ratio per receiver first. During this test the audio distortion per receiver is also measured. The audio distortion also gives an indication of the quality of the receiver. The results of this test are given in tables below. #

2.4.1 Radio-frequency protection ratio curves Most receivers were not able to meet the required start value of 56dB for the audio-frequency signal-to-interference ratio dictated by Recommendation ITU-R BS.461. Therefore, this start value has been lowered to 46 dB. Since this start value is lower than the original stop value for the audio-frequency signal-to-interference ratio this value was lowered to 40 dB. With this adaptation it was possible to measure the radio-frequency protection ratio curve of twenty receivers. The results are depicted in figure 2.2. The frequency difference is defined as the frequency of the unwanted transmitter minus the frequency of the wanted transmitter.

Page 14

KENWOOD KDC-3024A [C] PANASONIC CQ-RDP162N [C] PANASONIC CQ-RDP003N [C] BECKER MEXICO PRO CD 4627 [C] BLAUPUNKT WOODSTOCK DAB 52 [C] SUPE RTE AR-921 CD [C] CH JVC KD-SX997R [C] JVC KS-FX480REX [C] VDO DAYTON CD 2200 [C]
SANYO DTA-300M [P] PHILIP AZ3012 [P] S SANYO DC-DA1000 [P] SONY CFD-S550L/SC [P] THOM SON RR 600CD [P ] DIGITALWAY FD100 [H] PHILIP AZT9500 [H] S SONY ICF-C1200 [H] UNITE DM2595-2 [H] D AIWA HS-RM 539 [H] SONY WM-FX491[H]
Figure 2.1: Radio-frequency wanted-to-interfering signal ratio (wtisr) for twenty receivers recorded according to ITU Recommendation BS.641. The category to which the receivers belongs is indicated between brackets. The letters C, P and H are used for respectively the category car radios, portables and handhelds.
Details of the measurements can be found in Appendix A. About half of the receiver did not have an antenna input. This meant the signal from the multi-source generator had to be transmitted via the fixed or wire antenna of the receiver under test. Disadvantage of this method is that the input level at the input of the receiver is not known. This, however, is not necessary for the determination of the radio-frequency protection ratio if it is assumed that ratio between the output level of the transmitter and the input level of the receiver is the same for both the wanted signal and the unwanted signal. 2.4.2 Sensitivity Besides the radio-frequency protection ratio curve, the sensitivity of each receiver was also measured. The sensitivity is the signal level, in dBV, that is needed for an audio-frequency signal-to-interference ratio of 20dB. The 20 dB audio-frequency signal-to-interference ratio was chosen so that the sensitivity of all receiver could be measured. Not all receivers can be compared based on sensitivity since sensitivity depends on the way the signal is fed to the receiver. For this study three different feeds are used. The first one is a direct feed. This type of feed can be used for receivers that are equipped with an RF-input connector. All receivers from the category car radios have such a feed. Advantage of this type of feed is that the signal level at the input of the receiver is equal to the signal level at the output of the transmitter. The second type of feed uses a alligator clip to transmit the signal onto the fixed antenna. For the third type of feed the receive antenna, a wire, is wrapped around the transmit antenna. In both cases it is not possible to determine the exact input level. This means that only the sensitivity of receivers that use the same type of feed can be compared. Page 15

FM receiver study display using a keyboard. The presentation of the sample was followed by a warning signal to warn the participants that they should enter their ratings (there was a 2-sec interval between samples). Each sample was presented only once. The participants were tested in groups of 3 or 4, in separate testing booths with individual headphones. For each group of participants, the order of the speech samples (MOS run 1) and of the classical music samples (MOS run 2) was different. About twenty (half) of the participants started with the speech samples and the other participants with the classical music samples. Prior to each MOS run, the participants were presented with eight practice trials to familiarize them with the experimental procedure and with the quality range for both the speech and classical music samples to be expected in the experiment. In total, the MOS test (2 runs) took approximately 25 minutes, including a short break. 4.3 Analysis and results of the pairwise comparison For each participant, the preference matrix of the samples was determined. Table 4.1 illustrates an artificial comparison between five samples. A cell value of 1 indicates that the column variable is preferred over the row variable, a cell value of 0 indicates that the row variable is preferred. The sum of the column reflects the number of times the column variable is preferred over all row variables. Sample A B C D E Column total A 3 B 0 Preference C 4
Table 4.1: Quality preference matrix based on an artificial pairwise comparison of five samples.
A cumulative preference matrix of the preference matrices of all participants was calculated for each network configuration, and for speech and classical music separately. In order to determine the rank order of the samples and to normalize the distances between rankings, we constructed a new matrix from the cumulative preference matrix in which the proportion of times a sample was preferred above another is depicted. Then we converted the proportions into Z-scores using the cumulative normal distribution. In order to get an idea of the spread of the Z-scores caused by the individuals participants and to be able to perform statistical analyses on these data, we applied the Quenouille-Tukey-jackknife method. According to this simulation procedure, we calculated the Z-scores 40 times (40 listeners participated in the experiment) by leaving out the data from one participant each time. With the obtained Zscores it is now possible to rank order the sample conditions within each network configuration. In order to obtain one rank order that includes the results of the three network configurations, we re-scaled the data. For this purpose we used the Z-scores for the best (ref1) and worst references (ref4) to derive the translation and scaling factors, because these anchor points were included in each of the PWC runs

5 4,3,2,1,-30 -25 -20 -15 -10 -35 SIR [dB]
F0_D0 F0_D10 F0_D20 F0_D50 F100_D50 F200_D50 F300_D50
Figure 4.3: Mean opinion score as a function of the signal-to-interference ratio for speech recorded in a HF synchro network with the Sanyo receiver.
Figures 4.4, 4.5 and 4.6 present the results for classical music and the Sanyo receiver, for the conventional, the same program and the HF-synchro networks, respectively. Page 27
Sanyo-Classical music-Conventional
Figure 4.4: Mean opinion score as a function of the signal-to-interference ratio for classical music recorded in a conventional network with the Sanyo receiver.
Sanyo-Classical music-Same programme

5 4,3,2,1,-5 0

MOS -30 -25 -20 -15 -10
Figure 4.5: Mean opinion score as a function of the signal-to-interference ratio for classical music recorded in a same programme network with the Sanyo receiver.
Sanyo-Classical music-HF synchro
5 4,3,2,1,-20 -15 -10 -35 SIR [dB]
Figure 4.6: Mean opinion score as a function of the signal-to-interference ratio for classical music recorded in a HF synchro network with the Sanyo receiver.
For the Blaupunkt and Sony receivers a MOS test was carried out for speech only. Figures 4.7 and 4.8 show MOS scores as a function of SIR (dB) for Blaupunkt and Sony, respectively.

Blaupunkt-Speech

5,4,3,5

Page 28

F0_CONV F200_CONV F0 SAME
Figure 4.7: Mean opinion score as a function of the signal-to-interference ratio for speech recorded in a conventional, same programme and HF-synchro network with the Blaupunkt receiver.

Sony-Speech

6 5,4,3,2,1,80 SIR [dB]
F0_CONV F200_CONV F0_SAME F200_SAME F0_D0_SYNC F0_D50_SYNC F200_D50_SYNC
Figure 4.8: Mean opinion score as a function of the signal-to-interference ratio for speech recorded in a conventional, same programme and HF-synchro network with the Sanyo receiver.
If a quality criterion of MOS = 3.51 is regarded as acceptable, then the following protection ratios can be obtained.
The criterion of MOS = 3.5 is similar to the criterion applied in the ZeroBase study.

Page 29

Figure 4.9: The protection ratio as function of the frequency distance based on a mean opion score of 3,5 for speech recorded with the Sanyo receiver.

Sanyo-Classical music

Figure 4.10: The protection ratio as function of the frequency distance based on a mean opion score of 3,5 for classical music recorded with the Sanyo receiver.

0 --20 -30

CONV SAME SYNC_D50 SYNC_D100 f [kHz] 150 200
Figure 4.11: The protection ratio as function of the frequency distance based on a mean opion score of 3,5 for speech recorded with the Blaupunkt receiver.

Page 30

f [kHz] 150 200
CONV SAME SYNC_D50 SYNC_D0
Figure 4.12: The protection ratio as function of the frequency distance based on a mean opion score of 3,5 for speech recorded with the Blaupunkt receiver.

It should be noted that the protection ratios are obtained by means of interpolation and extrapolation and may differ from the values obtained through real experimental measurement. The protection ratios show some surprising results. First, the network configuration same programme resulted in a lower protection ratio than HF-synchro. The opposite was expected. Second, a higher delay didnt always result in a higher protection ratio for the network configuration HF-synchro. Third, the Sony required a higher protection ratio for f=200 kHz than for f=0 kHz. This result is in contradiction with the results of the objective measurements. Fourth, the bad receiver required a lower protection ratio for f=0 kHz for both the network configuration HF-synchro and same programme than the good receiver. The consequences of these protection ratios for frequency planning will be discussed in the next chapter.

Page 31

5 Consequences for frequency planning

5.1 General

5.1.1 ITU Recommendation ITU-R BS.412-9 indicates that protection ratios for steady interference provide approximately 50 dB signal-to-noise ratio. The protection ratios for tropospheric interference correspond closely to a slightly annoying impairment condition. According to Recommendation ITU-R BS.562.3, table 1, slightly annoying relates to grade 3 of the fivegrade impairment scale. It is not clear why the curves for steady and tropospheric interference of Recommendation ITU-R BS.412-9 merge from 200 kHz onwards. Possibly the effect that stations with higher frequency separations may be closer and therefore only steady interference is interest, has been brought in the curves. Recommendation ITU-R BS.412-9 states that in case of identical programmes an improvement of the protection ratio is expected at least for monophonic signals. In case of the same frequency and modulation, with synchronised signals, the protection ratios for monophonic signals are much lower than the one for different programmes. For stereophonic signals the protection ratios depend on the propagation delay and stereophonic content. Recommendation ITU-R BS.641 recommends to use an objective method using a sinusoidal tone of 500 Hz as wanted signal and a standard coloured noise signal as interfering source. 5.1.2 This study In objective tests according to Recommendation ITU-R BS.641, the test conditions for achieving an audio signal-to-noise ratio of approximately 50 dB could not be reached with most receivers. Therefore the audio signal-to-noise ratio for the tests has been lowered to 40 dB. The results should therefore not be compared to the ITU-R protection ratio values for steady interference, which provide approximately 50 dB signal-to-noise ratio. It raises however the question if the quality standard for steady interference is not too high for present day FM reception. The objective results show a great variety in performance of the tested receivers. A receiver with an average protection ratio curve has been selected as reference receiver. Car radios are in general considered as better than the reference receiver (higher achievable audio signal-tonoise-ratio and better selectivity), and walkman-radios are in general considered as worse receivers. If frequency planning is based on the average receiver, it should be born in mind that a large number of receivers may be subject to more than slightly annoying interference, in particular resulting from transmitters with frequency differences of 200 kHz and 300 kHz. On the other hand in particular car radios may still give acceptable reception outside the calculated coverage area. Three network configurations have been subjectively tested: Conventional. Same programme. Page 32

From this table it can be seen that the inter-modulation effects are caused by third order intermodulation products. The effect of these third order inter-modulation products is most noticeable when they are located in the pass-band of the receiver.

Page 42

FM receiver study Experience in The Netherlands has shown that interference from third order inter-modulation products may take place if reception is near an FM transmission site where many frequencies are used and: a) at least one with much lower power (for instance a local station) than the others b) reception of signals from another, more distant, site is required In these situations frequencies should as far as possible be chosen in such a way that no third order inter-modulation products occur in the pass-band of the receiver tuned to the low level signal. If it is not possible to avoid third order inter-modulation products in case a) the power of the low power transmission may need to be increased in order to achieve the required protection ratio for this situation. In case b) a fill-in transmitter may be required that fulfils the conditions indicated for case a). An example of case b) in the Netherlands is shown in Appendix C

Page 43

8 RDS Switching
This receiver study focuses on two aspects of RDS switching behavior. The first one investigates the RDS switching behavior due to differences in radio frequency levels of two different sources transmitting the same program on different frequencies. The second one investigates the RDS switching behavior due to multipath. The approach followed in both cases will be explained in the next paragraph. 8.1 8.1.1 Approach RDS switching behavior due to differences in radio frequency levels
This test represents the situation when traveling from the coverage area of one transmitter to the coverage area of another transmitter belonging to the same program chain and thus transmitting the same program on a different frequency. This test investigates the behavior of a receiver under such conditions and can be split up in the following two steps: 1 - Setting up the transmitters. Sources A and B of the multi source generator are frequency modulated with a 500 Hz sinusoidal tone and RDS. For both sources the frequency deviation due to RDS is set to 2 kHz. The combined output of sources A and B is fed to the receiver under test. 2 - Measuring the radio frequency switching level The following procedure is repeated for the following combinations of carrier frequencies: Combination Source A, f (MHz) Source B, f (MHz) 1 90,0 105,90,0 90,105,0 90,90,0 89,9
Table 8.1: Frequency combinations for RDS switching behavior due to differences in radio frequency levels.
The carrier frequency of sources A and B are set according to table 8.1. The radio frequency levels of both sources are set to the same level. The receiver under test is tuned to source A. Next the radio frequency level of source A is lowered, in steps of 1 dBV, until the receiver under test switches to source B. The radio frequency level of source A at the time of the switch over is the required radio frequency switching level. 8.1.2 RDS switching behavior due to multipath

This test represents the situation where the receiver receives a number of reflected signals. Normally a channel simulator is used for multi path testing. Due to the fact that such a device was not available an alternative approach for measuring multi path effects is explained below. This alternative test can be split up into the following two step: 1 - Setting up the transmitters. Page 44
FM receiver study Sources A and B of the multi source generator are frequency modulated with a 500 Hz sinusoidal tone and RDS. For both sources the frequency deviation due to RDS is set to 2 kHz. Station C of the multi source generator is frequency modulated with pop music. The outputs of sources A, B and C are combined and fed to the receiver under test. 2 - Measuring the radio frequency switching level The following procedure is repeated for the following combinations of carrier frequencies: Combination Source A, f (MHz) Source B, f (MHz) 1 90,0 105,90,0 90,105,0 90,90,0 89,7
Table 8.1: Frequency combinations for RDS switching behavior due to multipath.
The carrier frequency of sources A and B are set according to Table 8.1. The carrier frequency of source C is made equal to that of source A. The radio frequency levels of sources A and B are set to the same level. The receiver under test is tuned to source A. The radio frequency level of source C is chosen such that it does not lead to any interference on channel A. Next the radio frequency level of source C is raised, in steps of 1 dBV, until the receiver under test switches to source B. The radio frequency level of source C at the time of the switch over is the required radio frequency switching level. 8.2 Results
8.2.1 RDS switching behavior due to differences in radio frequency levels The level of sources A and B is set to 60 dBV. The radio-frequency level of source A at the time of switch over is given in table 8.3. Receiver
Blaupunkt Woodstock DAB52 Jvc KD-SX997R Vdo Dayton CD2200 Sony CDX-M850MP Becker Mexico Pro CD4627
f, source A (MHz) & f, source B (MHz)
90,0&105,0 90,0&90,1 105,0&90,0 90,0&89,9
Jvc KS-FX480REX Panasonic CQ-RDP003N Panasonic CQ-RDP162N
Table 8.1: Results for RDS switching behaviour due to differences in radio-frequency levels. The value in the cell is the radio-frequency level, in dBV, of source A when the receiver switches from source A to source B.

Based on the results he following conclusions can be drawn: Page 45
FM receiver study When the frequency difference is small, a small reduction of the radio-frequency level makes all receivers switch over. When the frequency difference is large (f>300 kHz), some receivers accept a low radio-frequency level before switching over while other receivers already switch over after a small reduction of the radio-frequency level. The Blaupunkt Woodstock DAB52 and the JVC KD-SX997 show an asymmetric switching behaviour for large frequency differences. The reason for this behaviour is not clear. The Sony and the Becker switch over after a small reduction in radio-frequency level for both small and large frequency differences. For the other car radios the switching behaviour depends on the frequency differences.
8.2.2 RDS switching behavior due to multipath The level of sources A and B is set to 60 dBV. The radio-frequency level of source C at the time of switch over is given in table 8.4. Receiver
90,0&105,0 90,0&90,3 105,0&90,0 90,0&89,7
Table 8.1: Results for RDS switching behaviour due to multipath. The value in the cell is the radio-frequency level, in dBV, of source C when the receiver switches from source A to source B.
In general all receivers, except the Jvc KS-FX480REX, show a comparable multipath behaviour.

Page 46

Appendix A: Detailed results of the protection ratio measurements
Tof Rf-w2isr max Wt-rfl: rfw2isr max M/s Ut-afl: dev=32 kHz Wt-afl: afs2ir=46 dB Ut-dev Wt-dev Wt-rfl: SNR=20dB Wt-rfl: SNR=20dB corrected Wt-rfl: afs2ir=46 dB Rf-w2isr: df= 400 kHz Type of feed. The following tof were used:1-Direct feed (D), 2-Indirect feed using an alligator clip (I-AC) and 3-Indirect feed using a wire antenna (I-W). The maximum radio-frequency wanted-to- interfering signal ratio. The radio-frequency level of the wanted transmitter to obtain the maximum radio- frequency wanted-to-interfering signal ratio. Mono/stereo. The audio frequency level of the unwanted transmitter to obtain a frequency deviation of 32 kHz. The audio frequency level of the wanted transmitter to obtain a audio- frequency signal-to- interference ratio of 46 dB. The frequency deviation of the unwanted transmitter. The frequency deviation of the wanted transmitter. The radio frequency level of the wanted transmitter to obtain a signal-to-noise ratio of 20 dB. The corrected, due to using an indirect feed, radio frequency level of the wanted transmitter to obtain a signal-to-noise ratio of 20 dB. The radio frequency level of the wanted transmitter to obtain a audio- frequency signal- to-interference ratio of 46 dB. The radio-frequency wanted-to- interfering signal ratio for a frequency difference between the wanted and unwanted transmitter of 400 kHz.

Table A.1: Detailed results of the radio-frequency protection ratio measurements.
Appendix B: Example of a same programme network
Table B.1: Example of a same programme network in The Netherlands
Appendix C: Reception problems due to inter-modulation
An FM site in Rotterdam transmits the frequencies indicated in the table 1 below. Station Radio TV West Business News Radio RTV Rijnmond Noordzee FM Sky Radio Radio 538 Veronica RTL FM
Table C.1: FM transmissions fromRotterdam
f (MHz) 89.3 91.3 93.4 100.4 101.5 102.7 103.2 104.6
The ERP on 102.7 MHz is highest (100 kW) and within about 6.5 km from the transmitter the field strength is 95 dBV/m or more. FM networks Radio 1, 2 and 3 are not transmitted from the site in Rotterdam, but from Lopik about 45 km away. The field strength of the Lopik transmissions near the site in Rotterdam is about 61 dBV/m. Reception of Radio 1, 2 and 3 is interfered by third order inter-modulation products within 6.5 km from the Rotterdam site as indicated in table 2. Station Radio 1 Radio 2 Radio 3 Frequency (MHz) 98.9 92.6 96.8 Frequency of 3rd order inter-modulation-products (MHz) 98.6 98.7 99.0 99.1 99.2 92.3 92.4 92.5 92.7 92.9 96.5 97.0 97.1
Table C.2: Frequency of the third order inter-modulation products near Rotterdam.
Frequencies could not be changed and the solution was to install at the Rotterdam site three fill-in transmitters to improve coverage of Radio 1, 2 and 3. It was not possible to find intermodulation free frequencies for the fill-in transmitters. Consequently the ERP of the fill-in transmissions should be high enough to respect the required protection ratios in case of high signal performance.
Appendix D: Photographs of tested receivers

KENWOOD KDC-3024A

PANASONIC CQ-RDP162N
BECKER MEXICO PRO CD 4627
BLAUPUNKT WOODSTOCK DAB 52

SUPERTECH AR-921 CD

Car radios continued

JVC KD-SX997R

JVC KS-FX480REX

SONY CDX-M850MP

Table D. 1: Photographs of the tested car radios.

VDO DAYTON CD 2200

SANYO DTA-300M

GRUNDIG LUNA RP 9200 PPL

Portables continued

GRUNDIG OCEAN BOY 350

PANASONIC RX-EX1

PHILIPS AZ3012

SANYO DC-DA1000

SONY CFD-S550L/SC

SONY ICF-C743L

THOMSON AM1180

Table D. 2: Photographs of the tested portables.

THOMSON RR 600CD

SONY ICF-M33RDS

GRUNDIG CITY BOY 52