hash stringlengths 32 32 | doc_id stringlengths 7 13 | section stringlengths 3 121 | content stringlengths 0 2.2M |
|---|---|---|---|
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 9.6.2.2 Experimental results | |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 9.6.2.2.0 Overview | This clause presents some experimental results regarding forward link alignment accuracy from [i.63]. The test configuration is shown in Table 9.9. Link Margin vs Pointing Error -15.00 -12.50 -10.00 -7.50 -5.00 -2.50 0.00 2.50 5.00 7.50 10.00 12.50 15.00 0.5 0.7 1 φ [deg] Link Margin [dB] 0.45m Antenna 0.6m Antenna 1.2m Antenna ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 173 Table 9.9: Test Configuration for experimental result of forward link alignment accuracy |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 9.6.2.2.1 Pointing losses | The test motor is driven at 0,1-degree steps measuring the received signal strength at each step. Two sets of graphs are generated - one obtained from horizontal (azimuth) movements and the other from vertical (elevation) ones, starting from the position of optimum pointing. Figure 9.17: Pointing loss as azimuth changes at 0,1-degree steps ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 174 Figure 9.18: Pointing loss as elevation changes at 0,1-degree steps The results shown in Figures 9.15 and 9.16 are combined into the 2-D plot in Figure 9.19. As shown, the point of minimum signal loss coincides with the centre of the "excellent zone". Figure 9.19: Pointing loss (Horizontal: along azimuth) (Vertical: along elevation) |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 9.6.2.2.2 Bit Error Ratio Results | The main parameter to quantify the quality of the link is the BER: the number of received bits that have been altered due to noise, interference and distortion, divided by the total number of transferred bits during a studied time interval. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 175 The test motor is driven at 0,1-degree steps measuring the received signal strength at each step. Two sets of graphs are generated - one obtained from horizontal (azimuth) movements and the other from vertical (elevation) ones, starting from the position of optimum pointing. BER measurements are collected over a 3-seconds observation interval for each step. Figure 9.20: BER as azimuth changes at 0,1-degree steps - Range: 33,8 - to - 38,3 degrees Figure 9.21: BER as azimuth changes at 0,1-degree steps - Range: 35,1 - to - 37,1 degrees ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 176 Figure 9.22: BER as elevation changes at 0,1-degree steps - Range: 33,7 - to - 38,1 degrees Figure 9.23: BER as elevation changes at 0,1-degree steps - Range: 35,7 - to - 37,1 degrees ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 177 Figure 9.24: BER variation along azimuth (horizontal) and elevation (vertical) Figure 9.24 shows that the point of minimum BER level coincides with the centre of the "excellent zone". Figure 9.25 shows the impact of polarization angle - with linear polarization - on the BER. After the forward link alignment is achieved, the polarizer is rotated at 1-degree steps. The BER is measured over 3-second intervals at each step. Figure 9.25: BER variation along skew [-5, +5] degrees ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 178 |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10 System Implementation Guidelines | |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10.0 Introduction | Examples of transmit/receive performance characteristics are provided to be used as guidelines in system performance evaluation, dimensioning and network capacity assessment. |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10.1 Typical RCST RF Signal Characteristics | |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10.1.0 Overview | This clause describes a set of typical RCST performance characteristics to be used as guidelines for evaluating the overall system performance. |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10.1.1 Phase Noise | Phase noise characteristics are typically dominated by the phase noise produced by the IDU/ODU. Table 10.1 presents typical phase noise characteristics of the RCST. In order to evaluate the overall link performance, it is assumed that the combined effect of all other phase noise sources in the link is at least 10 dB lower than the typical values presented here. Table 10.1: RCST Transmit Phase Noise Item Description Overall RCST Remarks A SSB Phase Noise (for Symbol rate ≥ 128 kBaud) 10 Hz 100 Hz 1 kHz 10 kHz 100 kHz > 1 MHz ≤ -16 dBc/Hz ≤ -54 dBc/Hz ≤ -64 dBc/Hz ≤ -74 dBc/Hz ≤ -89 dBc/Hz ≤ -106 dBc/Hz It is assumed that the combined effect of other phase noise sources in the transmission path (hub and satellite included) is at least 10 dB better. The split of SSB phase noise between IDU and ODU is left to the manufacturers' decision. B SSB Phase Noise (for 128 kBaud > Symbol rate ≥ 8 kBaud) 10 Hz 100 Hz 1 kHz 10 kHz 100 kHz > 1 MHz ≤ -30 dBc/Hz ≤ -60 dBc/Hz ≤ -70 dBc/Hz ≤ -74 dBc/Hz ≤ -89 dBc/Hz ≤ -106 dBc/Hz |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10.1.2 Carrier Frequency Accuracy | Clause 7.3.9.3 of ETSI EN 301 545-2 [i.1] defines the requirement on the carrier frequency accuracy of the RCST as 10-8 (root mean square) relative to the nominal carrier frequency. The corresponding maximum error value should be 6 × 10-8 relative to the nominal carrier frequency. The frequency accuracy of the terminal burst is the result of a number of contributors. Typical fixed contribution - i.e. the frequency accuracy typical of a classical system - is provided in Table 10.2. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 179 Table 10.2: Carrier Frequency Accuracy Contributor Value Application Source Ku-band (note 1) (Hz) Ka-band (note 2) (Hz) Terminal Frequency Accuracy 6 × 10-8 U/L Worst case 870 Hz 1 800 Hz Gateway Frequency Accuracy 10-8 D/L Typical 128 Hz 202 Hz Satellite Frequency Accuracy 10-7 delta (D/L, U/L) Typical 175 Hz 980 Hz Satellite motion (Doppler effect) 10-8 (note 3) U/L + D/L Typical 418 Hz 802 Hz Total contribution 1 590 Hz 3 784 Hz NOTE 1: Ku-band: 14,5 GHz uplink, 12,75 GHz downlink. NOTE 2: Ka-band: 30,0 GHz uplink, 20,2 GHz downlink. NOTE 3: Dependent on station keeping strategy, could be improved to typically 5 × 10-9. Terminal frequency accuracy: this value corresponds to the maximum error value of the RCST normalized frequency accuracy (considering to 6 times the RMS value). This value excludes Doppler shift and should be considered as normalized with respect to master synchronization reference at the NCC. Gateway frequency accuracy: typical value for the gateway receiver (ODU+IDU) frequency accuracy. Satellite Frequency Accuracy: typical value for the satellite uplink and downlink frequency accuracy. For transparent satellite, the resulting effect of frequency accuracy is computed on the difference between uplink and downlink frequency. Doppler Effect due to satellite motion: typical value for the satellite Doppler shift. This value depends on the station-keeping strategy. This effect results in two contributions on the return path from the terminal to the gateway: 1) offset in RCST transmit frequency due to Doppler shift on the forward path, and 2) Frequency offset due to Doppler shift on the return path. The first contribution is a consequence of locking the terminal local frequency to the transmit PCR reference which has induced Doppler (NCR time drift). This would cause a frequency offset on the terminal transmit frequency. The second contribution is the classical Doppler frequency offset due to satellite motion, which has to be accounted for on the uplink (from terminal to satellite) but also on downlink (from satellite to gateway). |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10.1.3 Amplitude Variation | Table 10.3 provides peak-to-peak limits of the carrier output amplitude variations, assuming a continuous wave transmit signal, for different frequency ranges. Table 10.3: Amplitude Variation Amplitude Variation Overall RCST Remarks In any 3 MHz band In any 20 MHz band In any 40 MHz band < 0,5 dB p-p < 1,5 dB p-p < 2,0 dB p-p The split of amplitude variation between IDU and ODU is left to the manufacturers' decision |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10.1.4 I/Q Imbalance | The I/Q amplitude imbalance is typically less than ±0,5 dB p-p. I/Q quadrature imbalance is typically less than ±1 degree p-p. The I/Q amplitude offset is typically lower than 0,5 dB. The maximum misalignment between I and Q symbols is typically less than 5 % of the symbol period. |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10.1.5 Spurious Levels | Within the transmit band, the RCST should meet the spurious radiation such that for each spurious signal that it transmits outside the nominated bandwidth, the total EIRP of each spurious signal should not exceed a level of 60 dB below the total EIRP of the transmitted carrier (modulated or unmodulated). Within the transmit band, the transmission enabled RCST should not generate a noise EIRP density (dBW/Hz) exceeding: Nominal EIRP (dBW) - 122 dBHz outside the nominated bandwidth. In the transmission disabled state the limits is 30 dB more stringent. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 180 For outside the transmit band, the requirements governing the RCST operation are specified in ETSI EN 301 459 [i.6] and ETSI EN 301 428 [i.7]. |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10.1.6 Non-linearity | |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10.1.6.0 Overview | The non-linear distortion impacting the RCST transmit RF signal is mainly caused by the high power amplifier at the ODU. Non-linear characteristics of the amplifier depends on several factors, including the technology (i.e. TWTA vs. SSPA), frequency band (e.g. C, Ku or Ka), transmit waveform (modulation scheme and pulse shaping filter), amplifier operating point (input signal level), environment conditions (e.g. temperature) as well as the device ageing. The amplitude and phase characteristics of a non-linear device is typically provided as the power transfer (AM/AM) and phase transfer (AM/PM) functions that relate the output signal instantaneous power and phase to the input power. Typically, such characteristics are provided based on single-tone measurements of unmodulated carriers. |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10.1.6.1 Power Transfer Characteristics | Figure 10.1 presents examples of power transfer of an SSPA. As shown in this figure, depending on the input signal power, the amplifier operates linearly at low input power (1 dB/dB gain), saturates at a high input power. In a mid-power range (depending on the device), the input signal power could be compressed or expanded. Simulations involving nonlinearities usually require computation of the transfer characteristic at arbitrary points. Although measured date along with the interpolation between points could be used for such computation, the slope discontinuities introduced by nearly all interpolation techniques such as splines or polynomial techniques introduce artificial signal distortion, which could appear for example as increased side lobe levels or signal distortion. Figure 10.1: Example AM/AM non-linearity characteristics (measured data) A mathematical expression that is commonly used to model SSPA's AM/AM characteristics has been described in [i.38]. Similar expression is presented below: s s sat A A g A g A G 2 1 2 1 ) ( + = where s is the smoothness factor, Asat is the saturation amplitude and g defines the gain of the amplifier in the linear region. This model can be applied to modulated signals as a memoryless function to compute the instantaneous signal amplitude values. An example is presented in Figure 10.2, with s = 6, g = 1,122 and Asat = 1,0351. Asymptotically, the model can present an ideal clipping law (for large values of s). -10 -5 0 5 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 Input Power [dB] Output Power [dB] ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 181 Figure 10.2: An example of AM/AM non-linearity according to Rapp Model [i.38] (with s = 6, g = 1,122 and Asat = 1,0351) The impact of the non-linear distortion on the overall system performance can be categorized as: • Power Saturation: The output power does not proportionally increase with respect to the input power. This is reflected in the AM/AM characteristics of the amplifier. • Modulation Dependent Power Loss: Depending on the modulation scheme, the output power loss can vary (compared to continuous wave, there is a higher output power loss for linearly modulated signals). • Out-of-Band power (spectral regrowth): varies depending on the operating point and the input waveform characteristics. • In-band distortion: The non-linearity will impact the signal characteristics. The impact can be seen as in- band interfering components, transmitted together with the original signal (inter-modulation interference). |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10.1.6.2 Phase Transfer Characteristics | The phase transfer (AM/PM) characteristic of SSPA's is usually more benign than that found in TWT amplifiers. In the open literature, the impact of AM/PM conversion of SSPA on the RF signal is often considered to be negligible. There are however examples of measured data that indicate otherwise [i.44]. Examples of phase transfer models have also been provided in [i.44]. For DVB-RCS2 systems that can utilize higher order modulations particularly 16 QAM, it is recommended that the phase transfer characteristics of the non-linear device to be examined and to be considered in system performance evaluation. It should be noted that there are pre-compensation techniques to be deployed at the transmitter to reduce the impact of such phase (and amplitude) distortions. |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10.1.6.3 Power Amplifier Operating Point Stability | In order to maximum the efficiency of the power amplifier, it is desirable to minimize the output back-off. However, the actual operating point of the amplifier and the level OBO can vary due to several reasons, such as: • Power Detector Instabilities (assuming a power detector used in a closed feedback loop, as shown in Figure 10.3): - Temperature stability - Detector frequency response: production test limit - Detector response to different modulations and symbol rates -14 -12 -10 -8 -6 -4 -2 0 2 4 -14 -12 -10 -8 -6 -4 -2 0 2 Input Amplitude (dB) Output Amplitude (dB) ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 182 • IBO instabilities: - Cable amplitude variation over frequency (e.g. 500 MHz frequency range) - IBO granularity (due to the driver settings) • SSPA-related instabilities (ageing, temperature, etc.): - Saturation power - AM/AM and AM/PM curve shape For linear modulation schemes, a closed loop control could be utilized in order to reduce the operating point instability when using. An example of such feedback control loop is shown schematically in Figure 10.3. As shown, in this figure, the transmit power level at the out of the power amplifier (SSPA) is measures (power detector) and the measurement results are used (in this example, power measurement values are communicated to the IDU via DISEqC™ protocol [i.20]). The RCST should indicate to the NCC the relative OBO used between QPSK and 8PSK, and between QPSK and 16QAM. The RCST will enforce this OBO autonomously in order to avoid violation of the PSD mask. The relative OBO indicated and enforced by the RCST is dependent on the local minimum OBO configured for the specific installation and the characteristics of the HPA. Figure 10.3: Feedback Loop for IDU/ODU Operating Point Power Stability |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10.2 Receiver Performance | |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10.2.1 Linear Modulation | Turbo Decoder performance results in AWGN channel are presented in tables below. Results are based on software simulations with the following parameters: • Number of Turbo Decoder Iterations: up to 8 iterations; • Decoder Architecture: BCJR, Normalized Max-log-map; • Synchronization: Ideal. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 183 Table 10.4: AWGN Performance results for control bursts Waveform ID Table A-1 of [i.1] Burst Length Modulation Code rate Es/N0 (dB) @ PER = 10-1 10-2 10-3 10-4 10-5 2 262 QPSK 1/3 -0,53 0,06 0,50 0,87 1,29 41 3 236 BPSK 1/3 -4,00 -3,65 -3,38 -3,17 -2,98 Table 10.5: AWGN Performance results for Short Bursts (536 symbols) Waveform ID Table A-1 of [i.1] Burst Length Modulation Code rate Es/N0 (dB) @ PER = 10-3 Es/N0 (dB) @ PER = 10-5 3 536 QPSK 1/3 -0,27 0,22 4 536 QPSK 1/2 1,92 2,34 5 536 QPSK 2/3 3,90 4,29 6 536 QPSK 3/4 4,93 5,36 7 536 QPSK 5/6 6,11 6,68 8 536 8PSK 2/3 7,71 8,08 9 536 8PSK 3/4 8,90 9,31 10 536 8PSK 5/6 10,43 10,85 11 536 16QAM 3/4 10,83 11,17 12 536 16QAM 5/6 12,16 12,56 Table 10.6: AWGN Performance results for Long Bursts (1 616 symbols) Waveform ID Table A-1 of [i.1] Burst Length Modulation Code rate Es/N0 (dB) @ PER = 10-3 Es/N0 (dB) @ PER = 10-5 13 1 616 QPSK 1/3 -0,80 -0,51 14 1 616 QPSK 1/2 1,49 1,71 15 1 616 QPSK 2/3 3,46 3,69 16 1 616 QPSK 3/4 4,50 4,73 17 1 616 QPSK 5/6 5,64 5,94 18 1 616 8PSK 2/3 7,29 7,49 19 1 616 8PSK 3/4 8,56 8,77 20 1 616 8PSK 5/6 10,02 10,23 21 1 616 16QAM 3/4 10,55 10,72 22 1 616 16QAM 5/6 11,86 12,04 Table 10.7: AWGN Performance results for Very Short Bursts (266 symbols) Waveform ID Table A-1 of [i.1] Burst Length Modulation Code rate Es/N0 (dB) @ PER = 10-3 Es/N0 (dB) @ PER = 10-5 44 266 QPSK 5/6 6,52 7,3 45 266 8-PSK 2/3 8,2 8,71 46 266 8-PSK 3/4 9,41 10,04 47 266 8-PSK 5/6 10,83 11,59 48 266 16-QAM 3/4 11,24 11,73 49 266 16-QAM 5/6 12,56 13,18 Table 10.8: AWGN Performance results for BPSK Modulated Bursts (very long bursts) Waveform ID Table A-1 of [i.1] Burst Length Modulation Code rate Es/N0 (dB) @ PER = 10-3 Es/N0 (dB) @ PER = 10-5 42 3 236 BPSK 1/3 -3,81 -3,52 43 3 236 BPSK 1/2 -1,53 -1,30 ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 184 A summary of Spectral Efficiency for different waveforms are shown in Table 10.9 where the impact of the unique words, pilot symbols and guard symbols are included in the computation of the efficiency for each waveform. Reference burst waveforms, as listed in Table 10.9, are designed to fit all into a unified timeslot grid comprising 270 Physical layer symbols. The use of constant burst size can simplify RRM, ACM and payload size adaptation. The RCST HPA saturation effects may require an RCST to reduce its emitted RF power (i.e. increasing the OBO) when utilizing higher order modulation schemes, thus requiring a higher link margin. The RCST may indicate the level of such RF power reduction to the NCC so the most appropriate waveform can be selected (as described in clause 7.1.1). The NCC may assume a worst case RF power reduction scenario if this information is not provided by the RCST. Figure 10.4 illustrates the bit/symbol efficiency as a function of Es/N0 for 28 reference bursts (as per Table 10.9). A target packet error ratio of PER = 10-5 has been considered. Figure 10.4: Efficiency versus sensitivity thresholds for 28 reference burst waveforms of 4 distinct sizes, at ideal synchronization in an AWGN channel ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 185 Table 10.9: Linear Modulation Waveforms Efficiency in AWGN Channel Waveform ID (Table A-1 of [i.1]) Burst Size (symbols) Guard (symbols) Payload (bits) Efficiency (Bits/Symbol) Es/N0 @ PER = 10-5 44 266 4 408 1,51 7,3 45 266 4 440 1,63 8,71 46 266 4 496 1,84 10,04 47 266 4 552 2,04 11,59 48 266 4 672 2,49 11,73 49 266 4 744 2,76 13,18 3 536 4 304 0,56 0,22 4 536 4 472 0,87 2,34 5 536 4 680 1,26 4,29 6 536 4 768 1,42 5,36 7 536 4 864 1,60 6,68 8 536 4 920 1,70 8,08 9 536 4 1 040 1,93 9,31 10 536 4 1 152 2,13 10,85 11 536 4 1 400 2,59 11,17 12 536 4 1 552 2,87 12,56 13 1 616 4 984 0,61 -0,51 14 1 616 4 1 504 0,93 1,71 15 1 616 4 2 112 1,30 3,69 16 1 616 4 2 384 1,47 4,73 17 1 616 4 2 664 1,64 5,94 18 1 616 4 2 840 1,75 7,49 19 1 616 4 3 200 1,98 8,77 20 1 616 4 3 552 2,19 10,23 21 1 616 4 4 312 2,66 10,72 22 1 616 4 4 792 2,96 12,04 42 3 236 4 984 0,30 -3,52 43 3 236 4 1 504 0,46 -1,3 |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10.2.2 CPM | |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10.2.2.1 Simulation Model | The CC-CPM simulation block diagram used for the performance assessment is shown in Figure 10.5. The signal from the modulator is transmitted through the channel such that the signal at its output is given by: ( ) ( ) ( ) ( ) r t s t n t i t = + + , where: • n (t) is AWGN, with single-sided power spectral density of N0 (Watt/Hz). • i(t) denotes interference (ACI) from four equally spaced carriers (i.e. two carriers on either side of the desired carrier s(t)). If if f and jf f are the centre frequencies for the ith and jth carriers, then the frequency separation between two immediate neighbours is, | | f i j f f Δ = − Δ = |f −f | , for all, | | 1 i j − = |i −j| = 1 . The spectral efficiency is measured as, 2 log c f s R M T η = Δ and sT is the symbol durationη = × . The carriers are homogenous, in that they all use the same CPM modulation parameters. The interfering carriers are assumed to be 3 dB stronger than the desired carrier. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 186 Figure 10.5: Block diagram of CPM transmitter, channel and the receiver The block diagram shown above depicts the receiver blocks. A perfect timing recovery and carrier phase synchronization is assumed at the receiver. Additionally, no information is exchanged with the adjacent carriers' receivers. A low-pass filter can be applied at the receiver front-end to mitigate the ACI at higher spectral efficiencies. The CPM demodulator consists of a front-end filter-bank followed by the Soft-In, Soft-Out (SISO) CPM detector. The CPM detector generates the Maximum A Posteriori (MAP) probabilities for the transmitted codebits using a trellis [i.39] or factor graph [i.40] describing the modulation. The convolutional decoder also performs SISO detection by executing the BCJR algorithm [i.41] on the trellis describing the convolutional code. Noting that the constraint length 3 convolutional code decodes on a 4 state trellis, whereas the constraint length 4 code requires an 8 state trellis. A maximum of 30 iterations are performed between the CPM SISO detector and the SISO convolutional decoder, during which extrinsic information is exchanged between them. In summary: • Five carriers with identical channel code and CPM parameters are simulated • Two interfering carriers on either side of desired carrier, each is 3 dB stronger than desired carrier • No multi-user detection/ ACI cancellation is applied • An ideal synchronization is assumed (no carrier frequency offset or phase noise is present) • 30 iterations between the CPM detector and the convolutional decoder |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10.2.2.2 Performance Results for CPM Waveforms | Performance results for three information block sizes and different spectral efficiencies are summarized in the following tables. Results are obtained using software simulation models described in clause 10.2.2.1 Waveform parameters as summarized tables below are in line with those identified in ETSI EN 301 545-2 [i.1] (see Table A-2). It should be noted that the impact of known symbols, trellis termination symbols and the guard time are not taken into account in the computation of the spectral efficiency and the performance threshold values. Convolutional Encoder Bit-Source Bit Interleaver Bit-to-Sym Mapping CPM Channel Convolutional Decoder Bit Bit Interleaver(2 bytes) CPM Detector Filter-Bank ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 187 Table 10.10: Simulation Results for bursts containing 400 information bits Waveform ID, Table A-2 of [i.1] Spectral efficiency b/s/Hz Modulation Index h Pulse shape AV Code rate FEC Conv. Code Normalized Carrier Spacing (fT) Eb/No(dB) @ PER = 10-3 Eb/No(dB)@ PER = 10-5 3 0,5 2/5 αRC = 0,98 0,5 (5,7)octal 2,000 2,2 2,75 4 0,75 1/3 αRC = 0,75 0,5 (5,7)octal 1,333 2,75 3,25 5 1,10 2/7 αRC = 0,75 2/3 (5,7)octal 1,210 3,7 4,4 6 1,25 2/7 αRC = 0,75 2/3 (5,7)octal 1,067 4,4 5,2 7 1,50 1/4 αRC = 0,75 4/5 (15,17)octal 1,067 6,1 7,2 8 1,80 1/5 αRC = 0,625 6/7 (15,17)octal 0,974 9,2 11,1 Table 10.11: Simulation Results for bursts containing 1 024 information bits Waveform ID, Table A-2 of [i.1] Spectral efficiency b/s/Hz Modulation Index h Pulse shape AV Code rate FEC Conv. Code Normalized Carrier Spacing (fT) Eb/No(dB)@ PER = 10-3 Eb/No(dB)@ PER = 10-5 9 0,5 2/5 αRC = 0,98 0,5 (5,7)octal 2,000 1,75 2,0 10 0,75 1/3 αRC = 0,75 0,5 (5,7)octal 1,333 2,3 2,6 11 1,10 2/7 αRC = 0,75 2/3 (5,7)octal 1,210 3,2 3,5 12 1,25 2/7 αRC = 0,75 2/3 (5,7)octal 1,067 3,6 4,2 13 1,50 1/4 αRC = 0,75 4/5 (15,17)octal 1,067 5,4 6,0 Table 10.12: Simulation Results for bursts containing 1 504 information bits Waveform ID, Table A-2 of [i.1] Spectral efficiency b/s/Hz Modulation Index h Pulse shape AV Code rate FEC Conv. Code Normalized Carrier Spacing (fT) Eb/No(dB) @ PER = 103 Eb/No(dB)@ PER = 10-5 15 0,5 2/5 αRC = 0,98 0,5 (5,7)octal 2,000 1,6 1,8 16 0,75 1/3 αRC = 0,75 0,5 (5,7)octal 1,333 2,2 2,4 17 1,10 2/7 αRC = 0,75 2/3 (5,7)octal 1,210 3,0 3,4 18 1,25 2/7 αRC = 0,75 2/3 (5,7)octal 1,067 3,5 3,9 19 1,50 1/4 αRC = 0,75 4/5 (15,17)octal 1,067 5,2 5,6 20 1,80 1/5 αRC = 0,625 6/7 (15,17)octal 0,974 8,1 9 Examples of performance curves measured as a function of Packet Error Ratio (PER) are illustrated in Figures 10.6 and 10.7. Figure 10.6: Performance results for CC-CPM and spectral efficiency = 1,25 bits/s/Hz 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 10 -5 10 -4 10 -3 10 -2 10 -1 10 0 1.25 b/s/Hz Eb/N0 (dB) Packet Error Rate for Center Carrier AWGN + ACI: 400 information bits AWGN + ACI: 1024 information bits AWGN + ACI: 1504 information bits ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 188 Figure 10.7: Performance results for CC-CPM and spectral efficiency = 1,8 bits/s/Hz |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10.3 RLE Efficiency | |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10.3.0 Overview | The RCS2 Return Link Encapsulation protocol (RLE) has been designed based on a number of assumptions where one of the most critical one is efficiency. In particular, the protocol is designed such that the overhead, in terms of additional bits compared to the payload, be as low as possible for different desired operation scenario. It is worth noting that the relatively small burst sizes in the return link are especially challenging compared to that on the DVB-S2 forward link with a nominally longer frame size. Based on the additional assumption that RLE runs at the link level and that the link is tightly controlled by the hub or NCC, preference has been given to out-of-band signalling for protocol operation when useful. This means that protocol options that are not expected to change during a single session are not signalled over the air interface because this would cost additional bits which would always have the same value. Nevertheless, several options are supported by the RCS2 LLS that can influence efficiency. This clause will analyse the costs associated with these options and will also make some general statements about the RLE efficiency. The encapsulation efficiency will be measured as the ratio between the number of useful bits (the bits of the IP packet, for example) and the total number of bits in the payload of the physical layer burst (before spreading and CRC). Ideally, this ratio should be as high as 1, meaning that each bit on the air interface is a useful bit. In practice however, there are a number of sources for bits that are not considered useful in this context, as presented below: • Bit padding. The RLE protocol is byte-oriented meaning that it can use the burst payload only in multiple of 8 bits. If the burst payload length is not a multiple of 8 bits, the remaining bits cannot be used. By careful design of the coding this does not occur with the standard RCS2 burst sizes. • Padding. Padding can occur at the end of a burst for two reasons: - There is no data available to fill the burst with another RLE packet. - The available space at the end of the burst is too short to start another RLE packet. • Per-burst overhead. This is overhead that occurs once in a burst. For RLE this consists of: - Signalling byte. This is the optional first byte of the burst payload which signals the length of the burst label and the length of the frame label. Because in all standard RCS2 RLE profiles these labels have the same length (see Table 7-10 in ETSI EN 301 545-2 [i.1]) this byte is completely optional and is used only if the use_explicit_payload_header_map field of the Frame Payload Format Descriptor is set to 1 (see clause 6.4.17.14 of ETSI EN 301 545-2 [i.1]). 3 4 5 6 7 8 9 10 11 12 13 14 10 -5 10 -4 10 -3 10 -2 10 -1 10 0 1.8 b/s/Hz Eb/N0 (dB) Packet Error Rate for Center Carrier AWGN + ACI: 400 information bits AWGN + ACI: 1504 information bits ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 189 - Burst label. For transparent star configurations this label is present only for the random access methods in which case its size is fixed for each method and defined in Table 7-10 of ETSI EN 301 545-2 [i.1]. - Burst CRC. No CRC is used in RCS2 because there is already a CRC present for error detection below the spreading sub-layer. • Per-fragment overhead. This is overhead that occurs for each RLE packet (which carry higher layer packets or packet fragments). This overhead consists of the following parts: - RLE packet header. This is the two byte header which starts each RLE packet. - Fragment label. This is the label that optionally maybe attached to each RLE packet. For all standard RLE profiles this label has the length 0. - Start packet header. For RLE packets that contain the start of a higher layer packet (but do not contain the complete higher layer packet) there is an additional two byte header. - Reassembly check. RLE packets that contain the end of a higher layer packet (but not the start of it) contain the reassembly check bytes. This is either a 1-byte sequence number or a four byte CRC. Whether to use the sequence number or the CRC is signalled by the hub in the allow_alpdu_crc and allow_alpdu_sequence_number fields of the Frame Payload Format Descriptor (see clause 6.4.17.14 of ETSI EN 301 545-2 [i.1]). If both are set the terminal can decide on which mechanism to use. For each fragmented higher layer packet the decision is signalled in-line in the start packet header. • Per packet overhead. This is overhead that occurs once for each higher layer packet. It is produced by the encapsulation sub-layer and consists of the following parts: - Protocol type field. This field may be 0, 1, 2 or 3 bytes long depending on the actual protocol type, the allow_ptype_omission, the use_compressed_ptype and the implicit_protocol_type fields of the Frame Payload Format Descriptor (see clause 6.4.17.14 of ETSI EN 301 545-2 [i.1]). - Packet label. This field maybe 0, 1 3 or 6 bytes long if the standard RLE profiles are used. For a transparent star only the lengths 0 and 1 apply. - Extension headers. This is a variable length field the use of which is not specified by the standard. • In order to analyse the performance of RLE a simulator has been used (see Figure 10.8). Figure 10.8: RLE performance simulation environment In this simulator a packet source creates higher layer packets which are sent to the encapsulator and scheduler. The encapsulator and scheduler additionally receive burst descriptions from the burst generator. It creates burst payloads according to these descriptions and forwards them to the channel. The channel may introduce drops and bit errors. The resulting bursts are sent to the decapsulator which extracts the higher layer packets and forwards them to the packet sink. Two different packet source algorithms have been used for the following analysis: • A source producing packets of equal size with a fixed rate, a fixed label and a fixed protocol type (CPR source). Simulations have been done for the following packet sizes: 20, 40, 60, 90, 120, 180, 270, 400, 600, 800, 1 000, 1 200, 1 400 and 1 500 byte. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 190 • A source producing multiple streams of packets. Each of the streams can be configured independently and produces packets with a fixed rate, normal distributed packet sizes and fixed label and protocol type (NRM source). For the burst generator a single algorithm is used: • A generator producing configurable combinations of ModCod and burst length for fixed amounts of time (LIN generator). The time for a single combination has been set to 600 seconds for all simulations. Because measurements are required only at the encapsulator, the channel is ideal for all simulations and the decapsulator just drops the packets. The simulations have been done with fully loaded links so that padding because of no available data does not occur. This kind of padding is related to resource management not to protocol efficiency. |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10.3.1 Optimum efficiency case | Optimum performance can be reached with the following configuration: • No burst signalling byte used. • User traffic and management traffic carried in a single SVN (single SVN operation). • Use of sequence number for reassembly check. • Only IP and signalling traffic. • Default protocol type is 0x30 (IPv4/IPv6). • Protocol type compression allowed. • No extension headers used. Under these conditions the maximum efficiency of RLE is reached which is shown in Figure 10.9. This figure shows the efficiency for DAMA traffic where no burst label is used. Figure 10.9: RLE optimum performance case ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 191 The figure shows the encapsulation efficiency for each combination of the selected set of packet sizes, the standard combinations of the modulation, coding and burst sizes corresponding to two groups of Linear Modulation waveforms (536 and 1 616 symbols) as defined in Table A-1 of ETSI EN 301 545-2 [i.1]. The most critical combination is small packets and small bursts. The lowest efficiency reached is around 0,74 which means the maximum overhead is 26 %. For large bursts and large packets the efficiency reaches 0,994 meaning an overhead of 0,6 %. Small packet sizes (resulting from header-compressed TCP ACKs or VoIP packets) are generally more critical than small burst sizes. The spikes in the figure are resonance effects when the RLE packet sizes and the burst sizes have a ratio denoted by small numbers. For the CRDSA encapsulation efficiency is smaller than that for DAMA because of the burst label. Figure 10.10 compares the efficiency of DAMA bursts vs. CRDSA bursts for the lowest Modcod and two burst sizes. Figure 10.10: DAMA vs. CRDSA encapsulation efficiency For the minimum bursts the maximum overhead is 30 % (for very small packets) and never goes below 19 % even for very large packets. For the large bursts the numbers are 16 % and 6 % resp. When configuring a system for CRDSA larger bursts should be considered instead of the minimum bursts. |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10.3.2 Default protocol and protocol type compression | The protocol type field that normally occupies two bytes for each higher layer packet can be compressed in RLE by two mechanisms: default protocol types and compressed protocol types. First the hub can provide the terminals with a default protocol type in the implicit_protocol_type field of the Frame Payload Format Descriptor and set the allow_ptype_omission flag (see clause 6.4.17.14 of ETSI EN 301 545-2 [i.1]). When the terminal transmits a packet which has this protocol type, it can omit it completely and set the protocol_type_suppressed flag in the start packet header thus saving two bytes. This is especially helpful for very short packets like header-compressed VoIP or TCP acknowledgements. The implicit_protocol_type defines the default protocol type for label types 0 to 2. The default protocol type for label type 3 is fixed to L2 signalling and cannot be changed. 20 40 60 90 120 180 270 400 600 800 1000 1200 1400 1500 0.7 0.75 0.8 0.85 0.9 0.95 1 Encapsulation efficiency per packet size packet size [bytes] efficiency [bit/bit] DAMA encapsulation efficieny QPSK1/3; 536 DAMA encapsulation efficieny QPSK1/3; 1616 CRDSA encapsulation efficieny QPSK1/3; 536 CRDSA encapsulation efficieny QPSK1/3; 1616 ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 192 Second the hub can allow the terminal to compress the protocol type to one byte instead of two bytes by setting the use_compressed_ptype field of the Frame Payload Format Descriptor. If this bit is set the terminal looks up the protocol type in the table of compressed protocol types (Table 7-3 in ETSI EN 301 545-2 [i.1]). If the protocol type is found the one byte value is inserted into the encapsulated packet. If it is not found, the escape value 0xff is used instead and the real two byte protocol type is inserted after the label. In this case the protocol type is actually expanded instead of compressed. In case a system makes heavy use of a protocol type not contained in Table 7-3 of ETSI EN 301 545-2 [i.1], a new compressed protocol type may be used for this system in the user defined area (0x80 to 0xfe). In order to estimate the savings of using default and compressed protocol types simulations are done with two configurations: without default protocol type, but compressed protocol types and without default and compressed protocol types. The results are compared to the optimum case when both mechanisms are allowed. Figure 10.11 shows the loss in encapsulation efficiency when not using default protocol types, but only compressed protocol types. In this case the default type has been set to 0x30 (IP) and all packets are IP packets. The change in efficiency is independent from the Modcod, but heavily depends on the packet sizes. It is largest for the small packet sizes where can reach 5 % for packets of 20 bytes. Figure 10.11: Loss of encapsulation efficiency without default protocol type The case when always the full protocol type is transmitted is show in Figure 10.12. In this case the efficiency maybe up to 8,5 % worse than in the optimum case. The most critical cases are very short higher layer packets. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 193 Figure 10.12: Loss of encapsulation efficiency with 2-byte protocol type For maximum efficiency, especially in systems with a large number of small packets it is strongly recommended to use an appropriate default protocol type and protocol type compression. |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10.3.3 Sequence number vs. CRC reassembly control | For error checking of reassembled packets two mechanisms are for seen: one uses a one-byte sequence number and the other a 32-bit CRC. The sequence number is appropriate for normal RCS2 systems with a reasonable packet loss ratio. It works well for packet loss ratios up to 95 %. Use of CRC is recommended for regenerative systems requiring on-the-fly translation of RLE packets to GSE packets. The CRC has been specified to correspond to the CRC of the translated GSE packet. The CRC may also be appropriate for mobile systems with large bursts of packet loss. The use of the CRC results in a less efficient encapsulation. This effect is shown in Figure 10.13. The effect is most visible for small packet sizes and small to medium burst sizes. In the worst case efficiency drops by approximately 5,3 %. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 194 Figure 10.13: Loss of encapsulation efficiency with CRC instead of sequence number |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10.3.4 GSE compatibility mode | For on-the-fly translation of RLE into GSE the configuration should not use default or compressed protocol types and use CRC for reassembly check. Additionally fragmentation of the protocol type and the label should be switched off and a packet label will need to be used for destination addressing (assumed to be 3 byte). The resulting efficiency is shown in Figure 10.14. The corresponding difference to the optimal configuration is in Figure 10.15. Figure 10.14: Encapsulation efficiency for GSE compatibility mode ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 195 Figure 10.15: Loss of encapsulation efficiency for GSE compatibility mode If can be see that there is a noticeable loss of efficiency for packets smaller than 500 byte. For very small packets the efficiency reaches just slightly more than 0,6 bit/bit. The encapsulation efficiency for very small packets and small bursts in this case is just 0,66 bit/bit. For small packets in large bursts (1 608 symbols and 16QAM 5/6) the efficiency is 0,82. For very large packets the efficiencies are 0,91 and 0,99 resp. The difference to the optimum configuration is shown in Figure 10.16. These numbers do not take into account additional fragment or burst labels that may be required for regenerative cases. Figure 10.16: Loss of encapsulation efficiency for GSE compatibility mode ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 196 |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10.3.5 Traffic mix | In order to estimate the performance for a mix of packets a simulation with three packet sources and the following parameters has been done: • Source 1: 400 packets/s with normal distributed packet size (µ = 60 bytes; σ = 40 bytes; minimum size 10 bytes; maximum size 120 bytes). This represents VoIP packets and TCP acknowledgments. • Source 2: 40 packets/s with normal distributed packet size (µ = 512 bytes; σ = 200 bytes; minimum size 120 bytes; maximum size 1 200 bytes). This represents HTTP requests, DNS packets and other signalling packets. • Source 3: 40 packets/s with normal distributed packet size (µ = 1 400 bytes; σ = 400 bytes; minimum size 1 000 bytes; maximum size 1 500 bytes). This represents TCP data packets. Two simulations have been carried out: with the optimum RLE configuration and in GSE compatibility mode. The results for the different ModCods and burst sizes are shown in Figure 10.17. For small bursts configuration 1 reaches 0,88 bits/bit while configuration 2 gets 0,82 bits/bit. For large bursts the numbers are 0,98 and 0,99 bits/bit resp. Figure 10.17: Encapsulation efficiency for traffic mix |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10.4 Examples of System Performance Evaluation | |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10.4.0 Introduction | An example of system application scenario for DVB-RCS2 utilization is described below. Based on a set of system parameters, several examples of link budget analysis as well as system capacity assessment are presented. |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10.4.1 System Scenario Definition | As a reference system scenario, a multi-spot beam satellite is considered. The system coverage is Europe using a single satellite located at 33° East using Ka-band frequency on the feeder and user links. The user link coverage consists of 98 beams. The user beams are served by 7 gateways (as shown in Figure 10.19). The baseline frequency plan is illustrated in Figure 10.18. A single polarization frequency re-use based on 4 colouring scheme has been considered. The frequency plan is based on a 4 colour reuse scheme with 125 MHz per spot beam. The forward and return user links are assigned to different polarizations as shown in Figure 10.18. 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 QPSK1/3; 536 QPSK1/2; 536 QPSK2/3; 536 QPSK3/4; 536 QPSK5/6; 536 8PSK2/3; 536 8PSK3/4; 536 8PSK5/6; 536 16QAM3/4; 536 16QAM5/6; 536 QPSK1/3; 1616 QPSK1/2; 1616 QPSK2/3; 1616 QPSK3/4; 1616 QPSK5/6; 1616 8PSK2/3; 1616 8PSK3/4; 1616 8PSK5/6; 1616 16QAM3/4; 1616 16QAM5/6; 1616 Encapsulation efficiency efficiency [bit/bit] GSE compatibility mode Optimum RLE configuration ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 197 Figure 10.18: Four-Colour, Single Polarization Frequency Plan The receiver antenna gain (on board or the satellite) is presented in Figure 10.19. Figure 10.19: On-board Rx antenna gain on the user link |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10.4.2 Link Budget Examples | For the system scenario defined in clause 10.4.1, link budget analysis is presented. Three RCSTs located at different geographical locations are considered. Figure 10.20 illustrates RCST 1, 2, 3 which are served by gateways A, B and C respectively. RHCP LHCP 27.5GHz 29.5GHz 30.0GHz RHCP LHCP 17.7GHz 19.7GHz 20.2GHz User User Forward Return Sub-channel granularity 10MHz, 4 colour re-use at sub-channel level GW GW GW GW GW GW GW GW Normal Up Down ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 198 Figure 10.20: On-board Rx antenna gain on the user link Table 10.13 provides a summary of EIRP realization for RCSTs at three different geographical locations. For these RCSTs a nominal SSPA saturated power of 2 Watts is considered. An additional power variation per RCST has been considered to take into account saturated power variations over the terminal population. The nominal OBO setting is according to the modulation scheme used by the RCST (as noted in the table). Table 10.13: Examples of EIRP Values Examples of RCST Tx characteristics RCST #1 RCST #2 RCST #3 RCST geographical Location Latitude: (deg) 46 51 37,25 Longitude: (deg) 5 15 43,25 Altitude: (m) 292 341 1 312 Elevation angle: (deg) 30,33 29,18 45,48 Up-Link Tx frequency: (GHz) 29,75 29,75 29,75 Nominal SSPA Saturated Output Power: (see note 1) (W) 2 2 2 Terminal Power Fluctuation: (W) 0,05 0,17 0,1 Terminal OBO (see note 2) (dB) 0,97 0,56 2,29 Terminal TX Losses (see note 3): (dB) 1,5 1,5 1,5 Terminal Antenna Diameter: (m) 0,6 0,6 0,6 Terminal Antenna Efficiency: 0,6 0,6 0,6 Antenna Peak Gain: (dBi) 43,22 43,22 43,22 EIRP: (dBW) 43,87 44,52 42,64 NOTE 1: The maximum output power is assumed to fluctuate from unit to unit and in time with 3σ ≤ 0,5 dB. NOTE 2: The OBO setting is a function of the modulation: 0,5 dB for QPSK, 0,7 dB for 8PSK and 2,5 dB for 16QAM. NOTE 3: Tx Losses include both coupling losses (typically 0,5 dB) as well pointing losses (considered below 1 dB). A clear sky link budget has been evaluated at three different geographical locations taking into account system and payload parameters. Results are presented for three different baud rates corresponding also to three different selection of modulation and coding per each RCST. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 199 Table 10.14: User Up-Link DVB-RCS2 Link Budget for Linear Modulation Schemes RCST#1 RCST#1 RCST#1 RCST Tx (See note 1) Carrier Symbol Rate: (Mbaud) 7,69 3,85 1,93 Modulation 8-PSK 8-PSK 16QAM Code Rate 5/6 3/4 5/6 Occupied Bandwidth (20 % Roll-off) (MHz) 9,22 4,62 2,32 Up-Link Tx frequency: (GHz) 29,75 29,75 29,75 EIRP per carrier (see note 1) (dBW) 43,87 44,52 42,64 Atmospheric Attenuation (w/o rain): (dB) 0,53 0,5 0,3 Polarization Losses: (dB) 0 0 0 Rain Attenuation (dB) 0 0 0 Slant Range: (Km) 38 580,39 38 683,54 37 377,87 Path Loss: (dB) 213,64 213,66 213,36 SAT RX Sat RX Gain: (dBi) 53,56 50,13 47,92 Sat RX NF: (dB) 2 2 2 Sat RX Total T: (K) 490,94 490,94 490,94 Sat RX G/T: (dB/K) 26,65 23,21 21,01 Up-link C: (dBW) -116,74 -119,52 -123,1 Up-link C/N0: (dBHz) 84,95 82,17 78,59 Up-link C/N: (dB) 16,1 16,31 15,74 Up-link C/I co-channel and adj-channel (see note 2) (dB) 14,72 12,52 18,65 Up-link C/IM (see note 3) (dB) 22,89 21,22 22,71 NOTE 1: RCST locations, EIRP values and other characteristics are outlined in Table 10.13. NOTE 2: C/I is computed both for co-channel interference from co-colour beams as well adjacent channel interference taking into account the OBO of the adjacent carriers. Carrier spacing of 1,2 Rs is assumed. NOTE 3: C/IM reflects the impact of the non-linear distortion taking into account the spectral regrowth and out-of-band power computed per each modulation and coding. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 200 Table 10.15: Feeder Down-link DVB-RCS2 Return Link budget RCST#1 RCST#2 RCST#3 SAT TX Down-Link frequency: (GHz) 18,7 18,7 18,7 Satellite Saturated Output Power: (W) 130 130 130 Sat TX OBO: (dB) 4 4 4 Satellite resulting fixed Gain: (dB) 108,87 108,88 108,87 Sat TX Losses: (dB) 1,55 1,55 1,55 Sat TX Gain: (dBi) 50,32 53,25 51,63 EIRP per carrier: (dBW) 40,9 41,06 35,85 Atmospheric Attenuation (w/o rain): (dB) 0,34 0,38 0,21 Down link polarization Losses: (dB) 0,2 0,2 0,2 Path Loss: (dB) 209,61 209,63 209,33 GATEWAY RX GW Latitude: (deg) 50,33 56 40,04 GW Longitude: (deg) 10,17 25,93 33,77 GW Elevation: (deg) 28,5 25,9 43,72 GW Altitude: (m) 333 140 943 GW Slant Range: (Km) 38 744,81 38 984,51 37 502,18 GW Antenna Diameter: (m) 8 8 8 GW Antenna Efficiency: 0,67 0,67 0,67 GW RX Gain: (dB) 62,17 62,17 62,17 GW RX Noise Figure: (dB) 2 2 2 GW RX Antenna Temperature: (K) 44,01 52,15 31,31 GW RX Total T: (K) 314,71 322,85 302,01 GW RX G/T: (dB/K) 37,19 37,08 37,37 GW RX Losses: (dB) 1 1 1 RX Signal Power C: (dBW) -108,08 -107,98 -112,72 Down Link C/N0: (dBHz) 95,51 95,46 91,05 Down Link C/N: (dB) 26,65 29,6 28,2 Down Link C/I: (dB) 33,49 27,09 43,18 Total C/(N+I): (dB) 11,8 10,46 13,25 Required threshold per Selected Modulation and Coding (see note) (dB) 9,3 11,00 13,00 NOTE: Reported threshold values correspond to PER = 10-5 for long bursts (1 616 symbols) Compared to AWGN threshold results reported in Table 10.6, implementation losses and performance degradation due to phase noise has also considered. |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10.4.3 System Capacity Evaluation | |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10.4.3.0 Introduction | Examples of system capacity evaluation for the return channel are presented. The baseline system is as per the system definition presented in clause 10.4.1 corresponding to a multi-beam satellite system. The focus of this analysis is on the return link using DVB-RCS2 waveforms according to Linear Modulation. |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10.4.3.1 System Assumptions | In order to perform system analysis certain system assumptions have been considered: a) within each beam bandwidth, carrier segregation is applied (i.e. co-channel carriers have the same bandwidth); b) the organization of carriers within a beam is assumed to be repeated for all the user beams; c) the network is fully loaded and a perfect packet scheduler is considered, implying that all the bursts are filled; d) both co-channel and adjacent channel interferers are always considered in clear sky, which is a worst case assumption although quite pertinent for both availability and capacity computations; e) uniform traffic request is assumed; f) no external interference (e.g. inter satellite interference) was considered in the simulation; ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 201 g) no adaptive uplink power control for satellite terminals is considered. Performance analyses have been carried out for two different classes of RCST terminals (non-coexisting in the same network), one with 2 W nominal saturation power and one with 0,5 W saturation power. System performance results are reported for both classes. System performance evaluation has taken into account a minimum peak data rate requirement of 10 Mbits/s over 90 % of the coverage area. This requirement has been considered for both set of simulations based on 0,5 W and 2 W nominal saturation power. It should be noted that the reported results here represent merely an indication of the overall system performance and are meant to serve as examples. The selection of the system parameters, waveforms and system assumptions will have significant impact on the actual performance results. |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10.4.3.2 Selected waveforms for system simulations | Table 10.16 shows a set of selected waveforms used for system performance assessment The required thresholds values for each waveform includes RX implementation losses based on a bench system implementation prototype. Table 10.16: Selected Linear Modulation Waveforms and their Performance Waveform ID Table A-1 of [i.1] Burst Length Modulation Code rate Es/N0 (dB) @ PER = 10-5 AWGN Ideal Synch. Required Es/N0 (dB) @ PER = 10-5 Including RX Impl. Loss (See note) Bits/Symbol 13 1 616 QPSK 1/3 -0,51 0,0 0,61 14 1 616 QPSK 1/2 1,71 2,3 0,93 15 1 616 QPSK 2/3 3,69 3,9 1,30 16 1 616 QPSK 3/4 4,73 5,0 1,47 17 1 616 QPSK 5/6 5,94 6,1 1,64 18 1 616 8PSK 2/3 7,49 8,2 1,75 19 1 616 8PSK 3/4 8,77 9,3 1,98 20 1 616 8PSK 5/6 10,23 11,0 2,19 21 1 616 16QAM 3/4 10,72 11,6 2,66 22 1 616 16QAM 5/6 12,04 13,0 2,96 NOTE: Thresholds include the receiver implementation loss and performance degradation due to realistic receiver algorithms. |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10.4.3.3 Interference and Fading Mitigation Techniques | The selected DVB-RCS2 waveforms, as listed in Table 10.16, provide a range of operating points (in terms of SNIR requirements) while maintaining a quasi-constant burst sizes. From a system perspective, this allows a flexible use of link adaptation techniques such as Adaptive Coding and Modulation (ACM) or Dynamic symbol Rate Adaptation (DRA) as well as combination of both. By changing the physical layer configuration based on the channel conditions, ACM optimizes the system efficiency and consequently its throughput. The use of low symbol rate carriers (in addition to ACM) can enhance the system availability by switching the satellite terminals that experience high fading onto a low symbol rate carrier. In such conditions, the link is closed by increasing the instantaneous SNIR due to the increased signal power spectral density, although with a penalized data rate. However, the use of DRA is not only to maintain the link availability in the presence of atmospheric fading, but also to cope with the variable co-channel and adjacent channel interference in multi-beam satellite systems using MF-TDMA as return link access scheme. In particular, the use of combined DRA and ACM can considerably reduce the effect of adjacent channel interference while maintaining a higher peak rate distribution. The use of combined DRA and ACM techniques in DVB-RCS2 systems has been recently investigated and reported in [i.45]. Similar methodology has been used here for system capacity analysis. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 202 |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 10.4.3.4 System Performance Results | Systems performance results are summarized in Table 10.17 and Figures 10.21 to 10.23. Results are obtained using a computer system simulation tool taking into account the system assumptions described in previous sections. Further details regarding the simulation tool specifications and system models can be found in [i.45]. Comparative system performance analyses indicate that the use of combined ACM and DRA will allow a higher distribution of peak data rate among the satellite terminals when compared to a system using ACM only, without compromising other performance measures such as the averaged total offered capacity and service availability, as summarized in Table 10.17. In this context, the availability figures indicate percentage of coverage area where the return link is available at least 99,7 % of the time. As indicated in Table 10.17 for systems scenario with 2 W transmit power per ST, there is even some gains in terms of the average capacity and service availability when combined DRA and ACM is deployed. Figure 10.22 and Figure 10.23 illustrates the utilization of 10 Linear modulation waveforms (shown in terms of modulation and coding distribution) for two different adaptive return link techniques (namely (1) ACM and (2) combined ACM and DRA). The results correspond to the system with 0,5 W and 2 W transmitting power RCSTs respectively. Table 10.17: Capacity Evaluation Results Linear Modulation, two values of TX power RCST Class Tx Power = 0,5 W Tx Power = 2,0 W Link Adaptation Technique ACM ACM+DRA ACM ACM+DRA Capacity (Gbits/s) 16,1 15,6 20,6 21,2 Spatial Availability (see note 1) 98,32 % 98,54 % 98,4 % 100 % average spectral efficiency (bit/s/Hz) (see note 2) 1,32 1,28 1,69 1,74 NOTE 1: The term Availability refers to Spatial availability of service for 99,7 % of the time availability. NOTE 2: The term "average" refers to an averaging both in space, i.e. over the coverage, and time, i.e. over the various fading events with the related probabilities. Figure 10.21: Complementary Cumulative Distribution of available bit rate in Clear Sky over coverage area 0 5 10 15 20 25 30 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Datarate CCDF in Clear Sky over the coverage Rb Complementary CDF Rb over the coverage [Mbps] ACM, 2 W ACM + DRA 2 W ACM, 0.5 W ACM + DRA 0.5 W ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 203 Figure 10.22: The utilization of different modulation/coding schemes for ACM and combine ACM and DRA schemes with a nominal Tx power of 0,5 W Figure 10.23: The utilization of different modulation/coding schemes for ACM and combine ACM and DRA schemes with a nominal Tx power of 2 W 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Traffic weighted modcod PDF vs Modcod index , 0.5 W PDF(Modcod) QPSK 1/3 QPSK 1/2 QPSK 2/3 QPSK 3/4 QPSK 5/6 8PSK 2/3 8PSK 3/4 ACM ACM + DRA 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Traffic weighted modcod PDF vs Modcod index, 2W PDF(Modcod) QPSK 1/3 QPSK 1/2 QPSK 2/3 QPSK 3/4 QPSK 5/6 8PSK 2/3 8PSK 3/4 8PSK 5/6 16QAM 3/4 16QAM 5/6 ACM ACM + DRA ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 204 |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11 Non-Geostationary Orbit Satellites | |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.0 Introduction | Following the commercial requirements detailed in [i.76], new modifications and enhancements were introduced to specifications, aiming at supporting the operation of a DVB-RCS2 based equipment within NGSO systems. [i.76] included requirements for facilitating operation within a NGSO system, requirements for enhancing the capabilities of the system, and for analysing its performance and limitations. Some of the work described herein has also been described in [i.91] and [i.98]. |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.1 Doppler Shift and Time Drift for NGSO Satellites | |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.1.0 The Doppler Shift Levels | CR100 in [i.76] referred to the operation with high Doppler shifts and high Doppler variations present when operating with NGSO satellites. Previous sections discussed the Doppler shift and time drift experienced by mobile terminals of geostationary satellite systems. For NGSO systems, the shift is much higher and diverse in view of the wide range of satellite orbits considered. In this clause the extent of Doppler shift and time drift, which are essentially two facets of the same Doppler effect, are given, with the purpose of demonstrating the order of magnitude of the expected shift and its rate of change. It should be noted that the goal of this clause is only to provide representative values and do not cover the entire range of possible trajectories and use cases. The Doppler shift depends on the relative velocity between the transmitter and receiver along the line-of-sight connecting them. Thus, it depends on the satellite velocity in orbit and its component in the direction of the terminal. While the velocity is a function of the orbit height and eccentricity, the line-of-sight component depends on the orbit position in space and the viewing angle of the terminal, which is a function of its location. Figure 11.1 shows the Doppler shift expected for circular orbits at heights of 200, 1 000, 2 000 and 10 000 km, ranging from VLEO to MEO. The orbit inclination is also included, as earth rotation is not negligible. The calculations were made for a terminal located on the equator, with the satellite passing overhead, to present the largest possible shift. The shift is given as a function of the angle from the zenith, which is common to all orbits, ranging between -70° and 70°, and is presented for a 1 GHz centre frequency, as a common baseline. The actual Doppler shift should be scaled according to the relevant frequency. Figure 11.2 depicts the variation of the Doppler shift along the orbit, again for a 1 GHz reference, presented as a function of the angle from zenith. Values for relevant frequencies in the Ku and Ka band are given in Table 11.1 and Table 11.2, respectively, to be compared with Tables 6.1 and 6.2. As it can be observed, the Doppler Rates (defined here as ⁄ , the ratio between the relative radial velocity and the speed of light) are much higher than those presented n Tables 6.1 and 6.2. The Doppler shift affects the time drift as well. As, for any periodic signal the frequency is inversely proportional to the period, a Doppler shift of relative to, in our example, 1 GHz, is equivalent to time drift in (namely, a Doppler shift of 1 Hz at fc = 1 GHz is equivalent to a time drift of 1 ). The time drift is presented as well in Tables 11.1 and 11.2. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 205 Figure 11.1: Doppler Shift at 1 GHz for Different Orbit Heights and Inclinations Figure 11.2: Doppler Shift Variation at 1 GHz for Different Orbit Heights and Inclinations Table 11.1: Doppler shift in Ku-band for different types of satellite orbits Orbit Type Orbit Height (km) Maximal Doppler rate Uplink Doppler frequency shift (kHz) (note 1) Downlink Doppler frequency shift (kHz) (note 2) Time drift (µs/s) Uplink frequency drift (kHz/s) (note 1) Downlink frequency drift (kHz/s) (note 2) VLEO 200 2,7E-05 387 341 26,7 16 14,1 LEO 1 000 2,3E-05 330 290 22,7 2,6 2,29 LEO 2 000 1,9E-05 276 243 19,1 1,04 0,919 MEO 10 000 7,5E-06 108 95 7,48 0,07 0,06 NOTE 1: Uplink frequency: 14,5 GHz NOTE 2: Downlink frequency: 12,75 GHz ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 206 Table 11.2: Doppler shift in Ka-band for different types of satellite orbits Orbit Type Orbit Height (km) Maximal Doppler rate Uplink Doppler frequency shift (kHz) (note 1) Downlink Doppler frequency shift (kHz) (note 2) Time drift (µs/s) Uplink frequency drift (kHz/s) (note 1) Downlink frequency drift (kHz/s) (note 2) VLEO 200 2,7E-05 802 540 26,7 33 22 LEO 1 000 2,3E-05 682 459 22,7 5,38 3,6 LEO 2 000 1,9E-05 572 385 19,1 2,15 1,45 MEO 10 000 7,5E-06 224 151 7,48 0,146 0,098 NOTE 1: Uplink frequency: 30,0 GHz NOTE 2: Downlink frequency: 20,2 GHz |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.1.1 Doppler Compensation Errors | |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.1.1.0 General | To mitigate the effects caused by the Doppler shift, the terminal would require Doppler compensation circuitry. This circuitry can be driven by a tracking mechanism that would track the shift in a closed-loop manner. Alternatively, an open-loop mechanism which uses the nominal trajectory data of the satellites can be used, and as satellites trajectories are typically predictable, the implementation of this technique might be simpler. However, errors in the terminal location, timing or satellite location may cause residual frequency and timing errors, which may result in residual compensation errors. In practice, a combination of techniques would be used, where a closed loop circuitry is used to reduce the residual errors following open-loop compensation and other frequency and timing errors in the system. Open-loop prediction is often needed for initial acquisition, accurate enough to get the signal into the demodulator's capture range. Subsequent tracking may or may not be doable without further compensation, depending on the tracking bandwidth of the demodulator. This is a system trade-off; one can open up the tracking loops to get better tracking range, but this results in them being noisier; i.e. less accurate. In turn, this costs demodulation performance. The errors estimated in this clause are the errors caused by inaccurate information about the terminal location and the satellite location. These errors include: 1) Timing errors resulting from a synchronization error of the terminal. 2) Location errors, including satellite height location error and error along the satellite trajectory. |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.1.1.1 Timing Errors | Figure 11.3 shows the estimation error as it develops along the trajectory, as a function of time (t=0 refers to the satellite at the zenith), for a satellite at height of 200 km, and inclination 180°, with timing errors ranging between -8 seconds to 8 seconds. Figure 11.3 (a) shows the Doppler estimation error, while Figure 11.3 (b) shows the errors of the estimation of the Doppler variation. As above, all the values refer to a centre frequency of 1 GHz. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 207 (a) (b) Figure 11.3: The error in Doppler Estimation (a) and Doppler Variation Estimation (b), for different timing errors- along the trajectory, orbit height - 200 km, centre frequency 1 GHz Figure 11.4 shows the maximal errors for different trajectories, again at heights of 200, 1 000, 2 000 and 1 000 km. (a) (b) Figure 11.4: The maximal error in Doppler Estimation (a) and Doppler Variation Estimation (b), as a function of timing errors- for different orbit heights, (1 GHz centre frequency) |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.1.1.2 Location Errors | In case the satellite location is not accurate, or equivalently, the location of the RCST is not accurate. The analysis was made for the case that the location is along the satellite trajectory, which is the worst case. Figure 11.5, shows, similarly to Figure 11.3 above, the estimation error as a function of time along the trajectory for location errors ranging from -500 km to 500 km. The maximal estimation errors of the Doppler shift and its variation are presented in Figure 11.6. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 208 (a) (b) Figure 11.5: The error in Doppler Estimation (a) and Doppler Variation Estimation (b), for different location errors- along the trajectory. Orbit height- 200 km, centre frequency 1 GHz (a) (b) Figure 11.6: The maximal error in Doppler Estimation (a) and Doppler Variation Estimation (b), as a function of location errors - for different orbit heights, (1 GHz centre frequency) |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.1.1.3 Height Errors | The maximal errors resulting from inaccuracy of the satellite height, are presented in Figure 11.7, for heights between -10 km to +10 km around the nominal value. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 209 (a) (b) Figure 11.7: The maximal error in Doppler Estimation (a) and Doppler Variation Estimation (b), as a function of location errors- for different orbit heights, (1 GHz centre frequency) |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.1.1.4 Discussion | The residual compensation error for the case of timing and height errors is linear with those errors, while it can be approximated as proportional to the square of the location error, for location errors up to a few kilometres. The range of timing, height and location errors plotted in those figures is quite large. Table 11.3 below presents the level of errors that can cause a 10 kHz error in the frequency compensation. Table 11.3: Timing, Height and Location Errors Resulting in 10 kHz residual Doppler Compensation Error Orbit Type Orbit Height (km) Band Frequency (GHz) Timing Error (s) Orbit Height Error (km) Location Error (km) VLEO 200 Ku 12,75 0,71 13,5 65,4 14,5 0,62 12,0 57,5 Ka 20,2 0,45 9,0 41,3 30,0 0,30 5,9 27,8 LEO 1 000 Ku 12,75 4,37 62,0 389,5 14,5 3,85 57,6 361,8 Ka 20,2 2,76 39,5 304,9 30,0 1,86 27,4 244,2 LEO 2 000 Ku 12,75 10,96 119,2 751,4 14,5 9,64 105,1 702,1 Ka 20,2 6,92 79,0 588,1 30,0 4,66 51,5 477,5 MEO 10 000 Ku 12,75 161 875 3 750 14,5 142 769,7 3 473 Ka 20,2 102 552,7 2 862 30,0 68 372,4 2 294 The residual error in the Doppler variation (or, equivalently the drift), is highly dependent on the orbit height, For MEO the drift can be as low as 1 - 3 Hz/s in the Ku and Ka bands, for LEO it would be in the range of 5 - 40 Hz/s while for VLEO it could reach tens of kHz/s in the error ranges presented above. Compare, for the sake of example, a terminal without GNSS, or at logon stage, which is not synchronized to the network and one that is synchronized and well located. For the first assume a timing error of 3 s, location error of 1 km and height error of 1 km, while for the second a 1 s timing error, 100 m location error and 100 m height error. Table 11.4 presents the expected residual Doppler error and residual Doppler variation for the two terminals. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 210 Table 11.4: Compensation Error Non GNSS Terminal Synchronized Terminal Orbit Type Orbit Height (km) Band Frequency (GHz) Doppler Error (Hz) Doppler Variation Error (Hz/s) Doppler Error (Hz) Doppler Variation Error (Hz/s) VLEO 200 Ku 12,75 42 274 1 500 14 077 494 14,5 48 077 1 706 16 010 562 Ka 20,2 66 975 2 377 22 303 783 30,0 99 469 3 530 33 123 1 162 LEO 1000 Ku 12,75 6 868 44,53 2 287 14,6 14,5 7 810 50,64 2 601 16,6 Ka 20,2 10 881 70,54 3 623 23,13 30,0 16 160 104,77 5 381 34,35 LEO 2000 Ku 12,75 2 740 8,13 912 2,65 14,5 3 116 9,25 1 037 3,02 Ka 20,2 4 341 12,88 1 445 4,21 30,0 6 447 19,13 2 147 6,25 MEO 10000 Ku 12,75 187 0,066 62 0,02 14,5 212,7 0,075 71 0,02 Ka 20,2 296,3 0,10 98 0,03 30,0 440 0,15 146 0,05 It should be noted that the values are given for very specific cases, with the goal of providing an order of magnitude estimation. |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.1.2 The Impact of the Doppler Shift | To evaluate the impact of the Doppler shift, two aspects were referred to in clause 6.2.2, and another one was presented in the context of beam hopping acquisition in clause C.4.4.2 of [i.77]. 1) The uncompensated error has to be lower than some given percentage of the symbol rate, depending on the implementation. The minimal symbol rates are given there for various mobile terminals. 2) The uncompensated frequency drift and timing drift limit the burst duration, as shown in clause 6.2.2.1. 3) The effect of frequency and symbol rate error on the acquisition performance of the burst forward signal. For the minimal symbol rate, it can be observed that in terms of the order of magnitudes, comparing Table 11.4 to Tables 6.3 and 6.4, the residual frequency shift for NGSO is similar to that of mobile terminals. For VLEO satellites (200 km orbit) the residual Doppler error, can be comparable to that of an aeronautical terminal (tens of kHz, up to 100 kHz), for low LEO satellites (1 000 km height) that error can be comparable to that of high-speed trains in the Ku band (around 10 kHz) or vehicular terminals in the Ka band, for high LEO satellites (2 000 km) the error is comparable to that of vehicular or maritime terminals, while for MEO, the values are less than that of a pedestrian terminal. Thus, the results presented in Tables 6.3 and 6.4 can be modified to take into account the NGSO Doppler shift. Of course, for mobile terminals of an NGSO system, the components of uncertainty presented in Tables 6.3 and 6.4 should be added to those of Table 11.4. Considering the limit on the burst duration, as described in clause 6.2.2.1, assuming that the demodulators compensate for both Doppler shift and Doppler drift. Table 11.5 details the limit on the burst duration, under the criterion of 4° maximal phase rotation at the last symbol of the burst, as described in clause 6.2.2.1. The limit is given for two cases: • With compensations but for the Doppler variation, for a "Non-GNSS" terminal described above. • With drift compensation for the "GNSS synchronized" terminal. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 211 Table 11.5: Limit on Burst Duration Orbit Type Orbit Height (km) Band Frequency (GHz) Compensated Non GNSS Terminal (ms) Compensated Synchronized Terminal (ms) VLEO 200 Ku 12,75 3,88 6,71 14,5 3,64 6,29 Ka 20,2 3,09 5,33 30,0 2,53 4,37 LEO 1 000 Ku 12,75 22,57 39,01 14,5 21,17 36,58 Ka 20,2 17,93 31,00 30,0 14,72 25,43 LEO 2 000 Ku 12,75 52,97 91,49 14,5 49,67 85,80 Ka 20,2 42,09 72,69 30,0 34,53 59,65 MEO 10 000 Ku 12,75 621,01 1 058,44 14,5 582,33 992,52 Ka 20,2 493,37 840,90 30,0 404,85 690,02 Comparing the results to those in Tables 6.5 and 6.6. The maximal burst duration for mobile terminals ranges between 3,6 ms (for an aeronautical terminal in Ka band) to 21,4 ms for a pedestrian terminal at Ku. For all the LEO cases, without frequency drift compensation the maximal burst size is lower than that, however with compensation, even for a non-GNSS terminal, the minimal burst duration for LEO satellites is in the order of magnitude and even longer than those specified in clause 6.2.2.1. Slightly smaller burst durations are needed for VLEO. Another criterion limiting the burst duration, addressed in clause 6.2.2.1 is that the timing drift resulting from both symbol timing inaccuracy and Doppler Effect should not induce a timing error of any symbol within the burst higher than of 0,1 symbol duration. Table 11.6 specifies the maximal number of symbols in a burst, assuming no symbol rate compensation, in which the timing drift is still under 0,1 of the symbol duration. Table 11.6: Limit on Number of Symbols in a Burst Orbit Type Orbit Height (km) Maximal Doppler rate Time drift (µs/s) Doppler Drift rate (1/s) Maximal Number of symbols VLEO 200 2,7E-05 26,7 1,10E-06 3 704 LEO 1 000 2,3E-05 22,7 1,79E-07 4 348 LEO 2 000 1,9E-05 19,1 7,17E-08 5 263 MEO 10 000 7,5E-06 7,48 4,83E-09 13 333 These burst lengths are compatible with the burst sizes as specified in Annex A of [i.1] for the return channel but are too small for the DVB-S2/X frames either on the forward channel or over the return channel. A symbol rate compensation or symbol synchronization is required, to take care of the drift. In [i.77] simulations were performed with carrier offset shift of 340 kHz and drift of ±30 kHz/s, symbol clock offset of 15 ppm and clock drift of 1 ppm. The performance, in terms of probability of false alarm and probability of detection are detailed in that clause. Referring to Tables 11.1 and 11.2, it can be observed that this carrier offset shift is higher than the downlink Doppler shift in the Ku band, but lower than that of the shift observed for Ka band (except for the MEO case). The downlink frequency drift of 30 kHz is higher than any of the cases specified in the tables. The symbol clock offset of 15 ppm should be compared to the maximal Doppler rate. Whereas for MEO the Doppler rate is lower than 15 ppm, it is not the case for LEO and VLEO. The clock drift, on the other hand, is higher than the actual Doppler drift for all cases, as indicated in Table 11.6. To comply with the conditions as specified in the above simulations, Doppler compensation for the centre frequency will need to be performed for LEO satellites in the Ka band, and symbol rate adjustment will be needed for any band. Further analysis should be made if those conditions cannot be fulfilled. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 212 |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.1.3 Doppler Compensation on the forward and return links | Within and RCS system there are some differences in the application of the Doppler compensation described in clause 11.1.2: • The return link channel is a many-to-one channel, and the RCST should transmit in the timeslot allocated to it by the system. Consequently, Doppler compensation for that link is made on transmit. The compensation should be made based on the satellite nominal trajectory. • The RCS2 bursts on the return channel are typically short, so frequency drift compensation or symbol rate compensation are not needed. However, for continuous transmission, DVB-S2/X bursts or long spread-spectrum bursts, such measures might be needed as well. • The forward link channel is one-to-many, so it is again the RCST which should perform the compensation. In that case both Doppler, Doppler rate and symbol rate compensation should be made. • On receive the compensation can be based on the nominal trajectory but tracking can be used as well. • For acquisition of beam-hopping bursts, the RCST receiver is required to shift its centre frequency and adjust the symbol rate according to the satellite trajectory. • In case of a transparent system, the hub should correct for the Doppler shift in its transmission to the satellite, in order to avoid the extra complication of computing the effects of both links at the RCST. |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.2 Dynamic Range | |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.2.0 Overview | Clause 10.4.2 in the present document presents an example of link budget calculation for a GEO satellite. Similar calculations can be made for the NGSO satellites, however, one has also to consider the fact that the signal power received by an RCST in a NGSO system varies as it tracks the satellite along its orbit. In this clause the main components of this variability are detailed. One distinguishes between instantaneous dynamic range, which refers to the difference between the strongest and weakest signals received simultaneously by a receiver, and the long-term dynamic range which is to the difference between the strongest Received Signal Strength (RSS) and weakest RSS during the operation time, or in the NGSO case, the satellite pass time. This clause focuses on the later. The expression for the RSS is given by: = − − − + [ ] Where all the variables are in dB. This relation applies to both the downlink and the uplink, however, the discussion below will refer only to the downlink. The uplink direction can be inferred by reciprocity. • EIRP depends on the transmission power, and transmission antenna gain and its offset from the beam centre. The satellite beam might be fixed on the ground cell or sweep the ground cell as the satellite passes by. While in the first case variability might be caused by terminal mobility which is rather slow compared to the satellite speeds, in the latter case the RSS at the RCST would vary accordingly. See above clauses 7.1 and 7.2. • LFS is the free-space attenuation given by the expression: = 10 ⋅ With is the slant range between the RCST and the satellite. Assuming the satellite altitude is given by H and the elevation angle of the antenna varies from θl to θh, the range varies between to , so the RSS varies by: 10 ⋅ ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 213 Typically, the satellite tracking starts at elevation angles of 200 - 300, so the variation of the free space attenuation can reach 9,3 dB (for θl = 200, θh = 900), within the time interval of the visible trajectory. • Latm is the atmospheric attenuation, which is a function of the meteorological conditions. The values taken for the link budget in clause 10.4.2 are derived by the link availability statistics according to Recommendation ITU-R P.618 [i.96] and related recommendations. The fade rate has been discussed in the literature, e.g. [i.99]. The atmospheric attenuation is directly related to the propagation path length in the atmosphere, which depends on the RCST antenna elevation angle. Thus, it is expected to vary, in the first order, according to the expression: 10 ⋅ A detailed evaluation of the losses, in specific cases, is presented in the subsequent sub-sections. • is the pointing error loss given by = 12 ⋅ , which holds approximately for small offset values of the aperture angle (angle from boresight as shown in Figure 11.8). The angle !" represents the 3 dB beam width, where the power flux density has reduced by 3 dB w.r.t. the maximum power flux density Θ#. Figure 11.8: Beam width $%& and aperture angle of a dish antenna with reflector diameter D All of the above factors, except for the rain fading, change at a rate proportional to the observable speed of the satellite, and can be handled by traditional AGC. • is the receive antenna gain. For the beam tracking the satellite, the offset from the beam centre depends on the tracking error and is random in nature. If the antenna is an Electronically Steerable Antenna Array (ESAA), there is an additional variability which is proportional to '( , where is the angle between the satellite direction and the antenna boresight. The factor "1" in the exponent stems from the reduced effective aperture of the ESAA, while the factor "r" is the result of the deviation of the array element from an ideal isotropic element. |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.2.1 Evaluation of Slant Range and Free-Space Loss | Numerical evaluations of a LEO satellite fly-by were conducted for: • Three different RCST locations on Earth: - Central Europe (Erlangen, 11.0°E, 49.5°N) - Dry Region (Egypt, 29°E, 26°N) - Wet Region (Manaus, 60°W, 3°S) • Elevation angles from 20° to 90° ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 214 • Four different orbit altitudes: 200 km, 500 km, 700 km, 1 200 km (= border to inner Van-Allen belt) • Six different carrier frequencies: 5 GHz, 12 GHz, 14 GHz; 19 GHz, 29 GHz, 55 GHz • Service availability of 99,9 % • Perfect antenna pointing, i.e. = 0 In order to distinguish between the effects of geometry related free-space loss and statistics-related atmospheric losses, they are analysed separately in the following. For all three locations, the free-space loss has been calculated using [i.97]. The same curves and dynamic range of LFS values for slant range and free-space loss resulted for the same satellite altitudes. For low elevation angles, the distance or slant range to the satellite is the largest. Accordingly, the free-space loss is maximum and as well as the signal propagation through the attenuating atmosphere. Consequently, minimum losses of the U-shaped curves will be observed at 90° elevation when the Nadir point meets the RCST location. Table 11.7 summarizes the free space loss at elevation angles of 20° and 90° and its dynamic range at frequencies 5 GHz and 55 GHz. Table 11.7: Free Space Loss Dynamic Range Slant Range Dependent Dynamic Range @ 5 GHz Frequency Dependent Dynamic Range @ θh = 90° Satellite altitude (km) LFS @ θh = 90° (dB) LFS @ θl = 20° (dB) Dynamic Range (dB) LFS @ 5 GHz (dB) LFS @ 55 GHz (dB) Dynamic Range(dB) 200 152 161 9 152 173 21 500 161 168 7 161 182 21 700 163 171 8 163 171 21 1 200 168 174 6 168 189 21 Note that also the time of visibility increases significantly with higher satellite altitudes. |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.2.2 Evaluation of Atmospheric Loss (Rain, Clouds, Gasses, Scintillation) | The evaluation of the atmospheric losses was carried out using [i.97], which implements the relevant set of Recommendation ITU-R P in the Python library ITU-Rpy. In general, these models are recommended for at least the frequencies between 4 and 55 GHz (exception: scintillation, only up to 20 GHz). Central Europe atmospheric loss (rain, clouds, gasses, scintillation), 99,9 % service availability for different frequencies is presented in Table 11.8: Table 11.8: Atmospheric Loss Dynamic Range - Central Europe Slant Range Dependent Dynamic Range Frequency (GHz) Latm @ θh = 90° (dB) Latm @ θl = 20° (dB) Dynamic Range (dB) 5 0,15 0,5 0,35 12 1,1 2,6 1,5 14 1,9 4,2 2,3 19 3,6 7,5 3,9 29 7,9 14,9 7,0 55 45 109 64 The slant range dependent dynamic range of Latm loss values are the same per carrier frequency for different satellite altitudes, as it is the loss along the path within the atmosphere, and is a function of the elevation angle. While Latm loss values in the few dB range are observed for carrier frequencies up to 19 GHz, significantly increased loss values can be observed for carrier frequencies of 29 and 55 GHz. Similar observations hold for the following dry region evaluation results with reduced attenuation values. Table 11.9 presents the results for a dry region atmospheric loss (rain, clouds, gasses, scintillation), 99,9 % service availability, for different frequencies: ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 215 Table 11.9: Atmospheric Loss Dynamic Range - Dry Region Slant Range Dependent Dynamic Range Frequency (GHz) Latm @ θh = 90° (dB) Latm @ θl = 20° (dB) Dynamic Range (dB) 5 0,13 0,46 0,33 12 0,24 0,78 0,54 14 0,29 0,94 0,65 19 0,7 1,6 0,9 29 0,62 2,0 1,38 55 28 80 52 The results for the wet region atmospheric loss (rain, clouds, gasses, scintillation), 99,9 % service availability is presented in Table 11.10: Table 11.10: Atmospheric Loss Dynamic Range - Wet Region Slant Range Dependent Dynamic Range Frequency (GHz) Latm @ θh = 90° (dB) Latm @ θl = 20° (dB) Dynamic Range (dB) 5 0,5 1,7 0,5 12 6,8 15,3 8,5 14 11 22,6 11,6 19 19,9 37,5 17,6 29 42 70 28 55 130 230 100 For the wet region evaluation results, significantly increased loss values are evident compared to the central European values and dry region values. |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.3 NGSO System Architecture and Implementation | |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.3.1 Introduction | The Lower Layers (DVB-S2/S2X and DVB-RCS2) were optimized for efficiency, flexibility, scalability and cost of use. For GSO applications, the terminal was the focus of the optimization. For NGSO payload implementation is also an important factor in the overall consideration, and it is affected as well by the NGSO system architecture. NGSO satellite system architectures have evolved significantly in recent years, offering various configurations to meet diverse communication needs. This clause will cover the main types of NGSO satellite system architectures, focusing on mesh and hub-based systems, transparent and regenerative payloads, and the inclusion or exclusion of Inter-Satellite Links (ISLs). [i.92] provides a survey of NGSO satellite systems at date. [i.93] compares the performance of four NGSO systems, with various architectures, based on their FCC filing, and compares their expected performance. Mesh architectures in NGSO satellite systems allow direct communication between terminals without the need for ground-based intermediaries. It offers lower latency by reducing the number of hops required for data transmission, increased independently of the gateway location, resilience thanks to its distributed nature. Hub-based architectures, also known as star topologies, offer easier network management and control from ground stations, reduce the complexity of routing functions at the payload. The payload implementation, which can be used in both architectures, can be either transparent or regenerative. While transparent payloads are less complex due to minimal on-board processing, are more flexible and can handle various signal types and in general less expensive to implement, regenerative payloads enjoy improved signal quality, use efficiently the gateway links and perform air interface and protocol translations if needed. NGSO constellations equipped with ISLs offer reduced ground infrastructure, providing seamless global coverage without relying on ground stations and decrease the need for ground hops. However, payload complexity with ISL is obviously higher. As stated in ETSI TS 101 545-1 [i.3], from the point of view of the terminal the satellite constellation can be viewed as a virtual satellite server. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 216 |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.3.2 NGSO System aspects considerations | The satellite in a NGSO fundamentally serves a varying set of terminals along its path. Typically, terminals located in a closed geographical vicinity, covered by the satellite antenna beam are grouped in cells. A satellite may serve one or several cells, depending on the numbers of antennas it carries or, more accurately, the number of beams available at the payload. For each of the cells the system has to allocate and schedule its resources- bandwidth, polarization, power and time, according to the demand, the terminal capabilities, service level agreements and reception conditions. More specifically the system allocates for each cell: • Carrier channels (centre frequency, bandwidth, polarization) on the forward and return links. • Beam Hopping Time Plans (BHTP), for the case of beam hopping on the forward link. • Burst Timing Plans (BTP) for the MFTDMA superframes, frames and time slots on the return channel. Being on different frequency bands, the forward and return links can be planned separately. For example, in case of a beam-hopping system, with a payload that has separate antennas or independent beams for the down and up links, the BHTP and BTP do not have to be identical. Return-channel hopping is transparent to the user terminal: it is the responsibility NCC to ensure that capacity assignments coincide with cell illumination. However, there might be constraints that link the two: • Some processes, e.g. the correction of the delay, frequency and power performed via the CMT require short return trip time. • For some systems, it might be beneficial to support half-duplex terminal wherein transmission and reception are not simultaneous, leading to a lower cost and lower power consumption of the terminals. In this case, the differential delay, the difference in return trip time between all the terminals within the entire area of all cells served by the satellite becomes a relevant design constraint. At the lower layers, the satellite system (gateways and satellites) has the support the following functions: 1) Demodulation and decoding of the forward link signals 2) Demodulation and decoding of the return link signals 3) Demodulation and decoding of the ISL signals 4) Extracting and framing the SDU's received from each of the above channels (decapsulation, defragmentation) 5) Routing the SDU's to the destination 6) Manage the queues for SDU transmissions 7) Fragment, encapsulate the frame the physical layers PDU's on the forward link, return link and ISL's 8) Encode and modulate the baseband frames 9) Schedule the transmissions, according to the BHTP on the forward link, and the gateway and other satellite availability on the return link and ISL. 10) Control the RCST in all the cells the satellite is serving: adjust transmission frequency, timing and power of each terminal, measure and determine forward link modcods and return link waveforms. 11) Manage the RCST's in terms of fault, configuration, accounting, performance and security. The main network-level architectural challenge in NGSO system is the rapidly changing connection topology between ground endpoints and therefore the need for link re-routing (handover). Transparent satellites are simpler but can only serve fixed one-to-one connections between ground endpoints (or, in the case of star network like DVB-RCS2, between beams). • Link handover still needs to be addressed, presumably out-of-band. This means that, unlike in GSO systems, the satellite network is not 100 % transparent. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 217 • Beam-hopping can in principle be used but requires synchronization and a burst format for the signal (e.g. DVB-S2X Annex E and/or DVB-RCS2), making the system 'even less transparent'. An example of a synchronization mechanism between a gateway transmissions and payload switching is given in [i.95] • DVB-RCS2 assumes a common forward channel carrier. The scope of a DVB-RCS2 network is therefore limited to one user cell. A multi-user-beam satellite needs to support multiple DVB-RCS2 networks, each with its own ground NCC. • Hopping requires a DVB-S2X Annex E forward link. No change to the DVB-RCS2 return link is needed other than ensuring that the frequency-burst plan and NCC capacity allocation are aligned with hop coverage. • Multiple (fixed) connections can be created for a single ground endpoint through frequency multiplexing. • ISLs can in principle be used if they provide transparent RF channel transport (through channel sampling or RF over optics). A regenerative satellite is more complex, as it implements the functionality listed above, but has the benefits of efficient gateway usage, improved signal quality and flexible air interface protocol allocation, and, in addition it can support any- to-any connectivity. There are two cases: 1) The simpler one has signal regeneration but still fixed routing: - Connections are fixed and one-to-one (as in transparent satellites). - DVB-RCS2 NCCs are on the ground. The payload only has to perform trans-modulation, frame translation and burst-framing synchronization. 2) The most general one is signal regeneration with any-to-any packet routing: - The system should employ a network-level (layer 3) packet routing scheme, in which the ground terminals are the edge points and the satellites are the network routers. - DVB-RCS2, being a layer 2 protocol, operates under the layer-3 protocol and is limited to the direct link between the satellite and a user-cell. - The DVB-RCS2 NCC needs to be placed at the payload. At the same time, there are ways to offload some of the NCC processing the ground. This set of functionalities, in addition to the tasks of managing the satellite bus itself, is complex and requires computational resources and power consumption. Hence the effort to reduce the complexity by off-loading the functionalities to the ground. Transparent payloads, also known as bent pipes, avoid the computational complexity by off-loading the handling of the forward and return links (items 1, 2, 4, 6, 7, 8, 9 in the list of functions above) to the gateway on the ground, as well as the control and management of the RCST. For this kind of payload, the signals handled by the payload itself are essentially analogue RF signals. In the forward link the gateway signals are transmitted with the same modcod as the one destined for the terminal and similarly for the return link waveforms. The payload typically shifts the received signals in frequency, amplifies them and retransmits them back to the ground. However, in modern satellite systems, especially NGSO satellite systems, the signals should be routed from their source to the destination. In conventional satellites the routing is done by a channelizer in the frequency domain, of which the implementation typically involves highly complex digital signal processing. Time domain routing is applicable as well, for example in beam- hopping systems. In this case the transmissions from the gateway should be synchronized with the beam switching on board, an example of such a synchronization mechanism is given in [i.95]. For NGSO satellite systems with transparent payloads that include ISLs, assuming the inter-satellite links themselves carry the same analogue RF signals, the signal routing issue, be it in the frequency or the time domain becomes even more complex consider the variable latencies experienced in the ISL as well as the need to manage the various frequency channels within the ISL's. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 218 |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.3.3 Operation with ISL | Inter-Satellite Links (ISL) are used by many NGSO satellite constellation to enable connection between any point on earth to any other point without using a terrestrial gateway, thus reducing the cost of the ground segment considerably. However, using ISL introduces delay to the system. [i.94] analyses the delay performance of satellite systems with ISL, comparing different topologies and ISL connection methodology. The results demonstrate that satellite networks improve delay performance compared to terrestrial optical fibre connections as the transmitter-receiver distance increases, System level simulations were performed to estimate the effect of ISL. A satellite constellation with satellite altitude of 600 km was used for the simulations. Consider a satellite system consisting of a satellite serving terminals arranged in cells in the satellite coverage area. The simulations use Monte Carlo methods, collecting the required statistics over a set of "drops" (the statistical case). In each drop the simulation: 1) Randomly selects the terminals in each of the observed cells. 2) Randomly selects the traffic to be carried by each terminal according to a chosen traffic model. 3) Based on those calculations, the simulation evaluates, for each of the terminals the packets delay, The results of each drop are collected, and their statistics are processed. The simulations were made for a zero-hop, single hop and 2 ISL hop scenarios. It was also performed for a system with a Fractional Re-use Factor of 3 (FRF-3) and for a static beam hopping scenario, which also introduces delay to the transmitted packets. Figure 11.9 shows the statistical Cumulative Distribution Function (CDF) of the 6 scenarios described above for two traffic models: Constant Bit Rate (CBR) wherein packets arrive at a constant rate, and FTP3 where packets arrive randomly according to Poisson statistics. In both scenarios the average bit rate of the packets was 1 MB/s, with packets arriving with 10 ms intervals in average. Figure 11.9: Packet delay probability distribution for ISL and beam-hopping In this simulation it can be seen that compared to a system with no inherent delay (FRF-3, no beam hopping and zero ISL hops), a single hop ISL contributes relatively less to the total delay, while the effect of the second hop is significant. Beam hopping, in which packets need to queue for transmission to their destination cell affects the CBR scenario less than the FTP3 scenario. A detailed description of the simulation is available in [i.100]. |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.4 Return Link Capability and Capacity Enhancements | |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.4.0 Introdcution | Historically, the satellite return channel was designed for asymmetric type of communication wherein the return channel bandwidth and transmission time is allocated to many terminals, in short and relatively narrow bandwidth bursts. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 219 CRs 109,110, 111 and 113 of [i.77] aim at enabling higher capacity and higher efficiency on the return link, according to the tendency of modern telecommunication systems to provide more symmetric and higher capacity communication. |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.4.1 Lower roll-off waveforms | Originally the RCS2 specifications specified a value of 20 % roll- off for the linear modulation waveforms. To enable a denser grid in the frequency domain, roll-off values of 10 % and 5 % were introduced. Obviously lower roll-off values entail longer impulse response for the shaping filter. An extended burst amplitude mask was introduced as well. The pulse response of the shaping filter is shown in Figure 11.10, for 5 %, 10 % and 20 % roll offs. The masks for each roll-off are presented as well. Figure 11.10: Pulse Responses and Burst Amplitude Masks for 5 % , 10 % and 20 % Roll offs The burst duration is indeed extended by 2 symbols (for 10 % roll-off) and by 8 (for 5 %), relative to the original mask, which is an extension of 0,7 % or 3 % for the shortest burst defined for those waveforms (262 symbols, annex A of [i.1]), compared to an increase of 12 % or 14 % in the number of available channels for 10 % and 5 % roll-off respectively. |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.4.2 Extending symbol rate | Unlike other standards, the DVB satellite communication specifications do not specify any particular numerology and does not limit the implementor in selecting the symbol rate. However, the variables used in various tables defining the basic MF-TDMA grid as well as other implied parameters need to be changed to accommodate the higher symbol rate of 500 Msps, required by the CRs. Thus, extending the symbol rate involved the following modifications: • Creating a new FCT3 table within which the btu_duration, btu_carrier_bw and btu_symbol_rate within the FCT table are extended from 24 to 32 bits. • Enabling a bigger number of frames in a superframe within the SCT table. • Introducing an "enhanced capacity" mode in the Lower layer Service Link descriptor, which enables the terminal to request capacity by larger chunks. • Enable finer resolution for time correction in both Correction Message Table and Correction Message Descriptor • Introduce symbol level time-offset for all of the linear-modulation waveforms, including LM-SS waveform, where it was missing. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 220 • Introduce a new version of the Control Assign Descriptor, Control Assign Descriptor Version 2, increases the width of the control_frame field, provides finer resolution for the time threshold, allows a time-limited assignment and allows transmission of tolerances only without assigning actual control slots. It also allows specification of the repeat period without actually assigning a control a slot sequence. However, some technical issues need to be addressed: • A wideband channel is more susceptible to frequency response flatness impairments. These result from the non-ideal frequency response of the RF chain, multipath effects resulting from reflections from various objects along the path and, in case of NGSO satellites, high antenna squints. • The resulting degradation in performance can be solved by equalizing the channel. Equalizing can be performed at the transmitter, if the channel is known, or at the receiver, which is, if the channel is not known, capable of measuring it and compensating for the impairments. However, in the case of RCS2, that might be problematic for a receiver which needs to receive a large number of terminals. Looking at the potential causes and nature of the impairments: 1) RF Chain impairments The RF chains at both ends of the link are composed of components which are not ideal, and impedance mismatches in the connection of those components cause frequency sensitive responses. This kind of impairment is generally static, namely it does not vary in time, although it may depend on temperatures and voltages that may vary slowly. Some of those impairments can be corrected during manufacturing and installation. Calibration may also be used to compensate for them, or to measure them such that they can be equalized, when relevant. 2) Multipath Multipath might be a major source of impairments in terrestrial channels, however for satellite channels, where the line-of-sight path of propagation is dominant, and antennas are very directional the multipath component is typically negligible. Indeed, for mobile terminals with a small antenna, multipath might need to be taken into consideration. Furthermore, for mobile terminals, or for NGSO satellites the multipath might change as the terminal or the satellite move. However, since the path difference between the direct path and reflected path are small, the net effect is that of flat fading over a rather wideband. 3) Antenna squint This is a problem unique to phased array antennas. Phased array antennas are used to track the satellite electronically (mobile terminals and/or NGSO satellites). They operate by coherently summing the signals received by the array elements, after compensating for the carrier phase at each element for a given direction. Since this phase is dependent on the frequency, a wideband signal would experience distortion, because each frequency component would "squint" to another direction. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 221 As an example, consider Figure 11.11 below: Figure 11.11: Antenna patterns Figure 11.11 shows the radiation patterns of a circular phased array antenna of 2 m diameter, for three signals spaced at centre frequencies of 29 750 MHz, 30 000 MHz and 30 250 MHz. The antenna is scanning to 70° off boresight, which is the case for an antenna pointing to the zenith and covering the satellite at 20° elevation angle. The squint is about 1,3° around the centre frequency. If, instead of 3 CW signals, a wideband signal would have been transmitted, the resulting pattern, in terms of the average power received per direction would be given by the blue curve above. Another way to observe the phenomenon is to look at the frequency response of a wideband signal. This frequency response is given in Figure 11.12 below: Figure 11.12: Frequency Response This frequency response would result in an ISI level of 12,1 dB without equalization. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 222 As a rule of thumb, the limit on the bandwidth is given by the equation: < ) * Where L is the antenna size, + is the scanning angle and c is the speed of light. Figure 11.13 below shows the bandwidth limit, as a function of antenna size and scanning angle. Figure 11.13: Bandwidth Limit as a function of antenna size and scanning angle The actual bandwidth limitation depends on the deployment scenario, the antenna size and scanning angle required for the particular application. Specifically for 500 MHz, a 1 m phased array antenna would be limited to 35° scanning angle and a 2 m antenna would be limited to 15° scanning. In case such a problem exists, it should be noted that for NGSO, the resulting impairment is dynamic, and the frequency response varies along the satellite trajectory. Possible solutions could be: • Use true time delay arrays, which are already available in the market. Since the frequency response of the signal is known, as determined by the scanning angle of the phase array antenna, the signal can be dynamically equalized, either on transmit and on receive. |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.4.3 Supporting Jumbo frames on the return link and GSE | The protocol defined by the DVB-RCS2 standard for carrying traffic over the physical layer burst is the Return Link Encapsulation (RLE) protocol [i.78]. The Protocol Data Unit (PDU) in the RLE protocol is limited to ~4 kbytes, while the CR called for Jumbo frames, commonly understood to mean 9 kbytes per PDU. Accordingly, the use of Generic Stream Encapsulation (GSE) [i.79] in the return link, as an alternative to RLE was introduced. GSE supports PDU sizes of up to 64 kbytes. This is far in excess of what is required; Using an existing encapsulation protocol facilitates the introduction of this functionality with limited effort. It should also be noted that the use of GSE in the return link is well aligned with the use of DVB-S2X in burst and/or continuous mode in the return link, which is required as well. The use of GSE encapsulation is introduced on a per-timeslot basis. Hence, the use of the current RLE is not impeded in any way, including applications for random-access and mesh transmission. GSE requires a slightly larger overhead and does not support all features of RLE; notably, it is lacking some of the label features and other facilities required to support certain ways of operating random-access traffic, including contention-based control bursts. This is not considered a major drawback for the envisaged use, which does not involve random access other than for logon transmissions. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 223 |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.4.4 Support of DVB-S2X on the Return Link | The DVB-RCS2 standard supports a continuous carrier mode. But the waveform specified for it in [i.1] was not adopted by industry. On the other hand, the DVB specified the DVB-S2/S2X standard, which is well accepted in industry, highly efficient and versatile and can be readily used as an adequate waveform for the continuous carrier mode. Accordingly, it was decided to deprecate the existing continuous carrier waveform, which had been defined prior to the finalization of the DVB-S2X specification and replace it with DVB-S2/S2X. Moreover, following the introduction of beam-hopping waveforms in DVB-S2X, the same beam-hopping waveform can now be used as a bursty version over the return channel, which opens the possibility to use it in any mode on the return link. Any DVB-S2X waveform, specified in [i.80], can be used either over a continuous carrier or within timeslots allocated within the MF-TDMA grid. Within a grid, following a logon process including synchronization, regular DVB-S2/S2X frames can be used. In other cases, implementors might choose the beam-hopping DVB-S2X superframe, to make use of its long Start-of-Superframe preamble for burst acquisition. The continuous carrier allocation can be persistent in which the RCST is semi-permanently operating using the continuous return link, or "non-persistent", in which the expected mode of operation is that the RCST logs on using regular MF-TDMA but has the option of switching to continuous DVB-S2X mode and back as the system decides, dictated e.g. by traffic volumes at different times. There are four basic operating modes for the RCS terminal: a) Burst mode using DVB-RCS2 waveforms (linear modulation, spread-spectrum and CPM). b) Burst mode using DVB-S2X waveforms. c) Non-persistent continuous mode using DVB-S2X waveforms. d) Persistent continuous mode using DVB-S2X waveforms. Due to implementation consideration, a terminal is not required to support all modes simultaneously. The specifications allow transitions between modes that may require re-start of the modulator or may take relatively long time (2 s). In a DVB-RCS2 system, the selection of the waveform, namely control of the modulation and coding (modcod), is done by the hub. However, in the new revision, the option of letting the terminal determine the S2X modulation and coding (modcod) is enabled as well. The terminal can select the modcod, if allowed to, from a list of modcods provided by the hub over the newly defined RTMST. |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.5 Operation without GNSS | |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.5.1 Introduction | A satellite terminal operating within a NGSO System, needs to determine accurately its own location, the satellite location and the satellite velocity at the time of transmission. This information is needed for the terminal to be able to point its antenna to the satellite, compensate for the expected Doppler, synchronize to the satellite transmissions and schedule its own transmissions taking into account the propagation delay. GNSS receivers, which are widely available and affordable for most applications, can be used for accurate geo-location and timing. However, GNSS systems are vulnerable to jamming and interruption of service. Alternative means can be used for synchronization and Geo-location, which may not be as accurate as GNSS navigation, but still enable the operation of the terminal, using the signals and information provided by the DVB-S2X/ RCS2 specification. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 224 Using NGSO constellations, mainly LEO ones, as an alternative to GNSS systems has been a subject of research in the navigation community. [i.85] and [i.86] present surveys of such research. [i.85] compares the accuracy expected from a LEO-based Position Navigation and Timing (PNT) based on various existing LEO constellations. [i.86] covers the whole set of components of a LEO system and discusses the requirements from a PNT point of view. The above references focus on solutions for an alternative global navigation system, while [i.87] and [i.88] present a method of localization based on signals from one or two satellites of the Globalstar system for the purpose of quick location registration, to enable paging the terminal. The accuracy required in this case is not comparable to GNSS. The position determination is based on two-way communications between the terrestrial gateway and the terminal, and the capabilities to perform delay and frequency measurements in both. This is not always the case for initial acquisition for example. In other cases, such as the one described in [i.89] and [i.90], without some feedback it might be difficult to acquire the downlink signal, due to the large sensitivity to Doppler of that system. Since the DVB specifications can be used by a large variety of NGSO systems, this clause does not provide any specific implementation or performance estimation but focuses in presenting the means provided within the specifications would enable operation without GNSS, with enough precision to achieve: • Antenna pointing to the satellite within its beamwidth. • Doppler compensation within the acquisition and tracking error. • Synchronization accurate enough to enable transmission within the allocated time slot. These means include: • The downlink signal, based on DVB-S2X, has a set of preambles and pilots, which make it possible for the receiver to perform accurate timing and frequency measurements. • NCR packets, which can be used for ranging and synchronization to the satellite, and the Time Association Message which associates the NCR time stamp to the time-of-day. • Accurate knowledge of the satellite trajectories provided by the satellite position information in an accurate OEM format of [i.81]. • A feedback loop, using the Correction Message Descriptor, correcting the time and frequency offset of the terminal. With the current enhancement the feedback of time correction can be of a resolution of 0,58 ns (1/64 of an NCR tick). |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.5.2 Procedures for operation | There are three main stages in the operation of a terminal in an NGSO system: • Satellite search, acquisition, and operation initialization • Operation- communication with the satellite. • Satellite handover. The most demanding stage is, undoubtedly, the initial one. One can assume that the following information is available to the terminal at this stage: • The satellite trajectory is known, at least, according to TLE data [i.82]. • Time of day is known within a few second accuracy. • The terminal own location, one can assume that the accuracy of a tenth of a longitude/latitude degree (about 100 km) can be readily obtained. • Parameters (polarization, frequency symbol rate, etc.) of the initial channel that can be received by the terminal. This channel is referred to as the "Startup TDM" in ETSI EN 301 545-2 [i.1]. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 225 Except for terminals using omni-directional antennas (usually at relatively low frequency RF bands), the first step would to be to point the antenna toward the satellite. As the terminal location is not known well enough, a search for the signal will be required. With the time of day and location information available, it is possible to point the antenna at the general direction where the satellite is expected to rise, at some given elevation (typically 200 - 300 above ground obstacles) and then scan in azimuth until the signal is detected. Other search patterns can also be used. For illustration it should be noted that for a satellite at height of 1 000 km, with a 1° beamwidth antenna, an accuracy of less than 50 km in location is sufficient for antenna pointing. Followingly, or simultaneously with the previous stage, a measurement or scan in frequency needs be performed in order to estimate and compensate for the signal Doppler shift. The error in the Doppler compensation is discussed in clause 11.1.1 above. The next step will be to acquire the signal, namely, detect the start of frame (or start of superframe, in case of beam-hopping), and decode the received signal. A DVB-RCS2 system would convey the essential information as part of its Layer 2 Signalling (L2S): • SAT table containing satellite position information, which is now available in an accurate OEM format [i.81], for the transmitting satellites and other satellites. • NCR packets indicating a specific transmission time of a start-of frame or start-of superframe. • Time association message, associating the NCR clock to the time of day. With that information, the terminal can now synchronize to the satellite Network Clock Reference and derive the time-of day. At this point the terminal has the information of the satellite location extracted from the ephemeris data at the time of transmission and it can perform one or all the following measurements: • Antenna pointing direction. • Doppler shift • The reception time of the NCR reference point (the SOSF for example) at the terminal. Using this information and measurements, the terminal can now form equations expressing those measurements as functions of the terminal position, and biases inherent in the system. Ephemeris determination, namely the accuracy if the ephemerides provided over the satellite position tables is out of scope of the present document. Some constellations may use on-board GNSS receivers for that. Repeating the process at other points in time along the satellite trajectory, or with other visible satellites in the constellation, would make it possible to determine the terminal location, or more importantly, get a good estimate of the propagation delay that will allow it to transmit a logon message to the hub, following which it can tune in to the correct location. Note that achieving the necessary accuracy, based on a single satellite might take longer than when using a GNSS for geo-location, but remains entirely feasible. During this time, while collecting the information and performing the measurements, the terminal will need to track the satellite in position and in frequency (due to the changing Doppler) and estimate the reception time of the NCR reference point. The problem is more complicated for mobile terminals, unless they carry their own non-GNSS means of navigation (e.g. airborne or maritime). Simultaneous measurements from several satellites might be needed for that. |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.5.3 Measurement Accuracy | |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.5.3.0 Introduction | In this clause, the basic expressions and estimation of the expected accuracy and time needed to join the network and handover, without GNSS are presented. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 226 |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.5.3.1 Required Accuracy | The following may be considered for determining the required accuracies for operation: Position accuracy required for pointing the antenna is given by: ∆ = ℎ∗ , where Beamwidth is the antenna beamwidth and Rsat is the range to the satellite, ∆P is the error in position, which may be the horizonal position error, in case the satellite is in zenith, or a combination of the horizontal and vertical errors in case the satellite is in other elevation angle. This requirement is quite loose (tens of kilometres). The position accuracy also affects Doppler compensation ability, as discussed in clause 11.1.1 above. The required accuracy would then be determined by the receiver tolerance to frequency shift and frequency variation. Note that some location methods may cause an ambiguity that should be removed. Timing accuracy is a function of the size of the timeslot allocated for logon and control, which should be determined by the system operator according to the specific system requirements. |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.5.3.2 Measurements Accuracy | During operation, the terminal will need to perform time, Doppler, and direction measurements. Estimates of the standard deviation error of those measurements (derived from the Cramer Rao Lower Bound for high SNR case) are given by: ≥ ' "√- , . ≥ ' /√- , ≥ 0 √- In which , ., are the standard deviations of the timing, frequency, and direction respectively, B is the signal bandwidth, T is the observation time allocated for the frequency measurement, and Θ the antenna beamwidth. The timing measurement performed by a DVB-S2X receiver over the downlink, relies on timing measurement of a specific marker in time- the first symbol of the start of a frame (SOF) or the first symbol of the start of a superframe (SOSF). The measurement relies on correlation of the known SOF or SOSF with the received signal, thus the effective signal to noise ratio is = 1 2 -, where 1 is the number of symbols used for correlation and 2 - is the energy per symbol, as received after matched filtering. 1 can take the value of 26, in case of regular DVB-S2/X frame, 270, for a superframe with unknown format, 720 for an annex E superframe with known format and 1224 for superframe format 6, using the Extended Header Field. |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.5.3.3 Location Estimation | Using the NGSO constellation as an alternative GNSS requires a much deeper and detailed analysis, however for the question of enabling operation, based on the DVB-S2X/RCS2 specifications the most challenging is the timing requirement. Let the ! "indicate the transmission time (over ToD clock) of the first symbol (marker) of frame/superframe i with an NCR packet referring to it. Let ! " indicate the true reception time (over ToD clock). Let b indicate the bias between the receiver ToD clock and the true (satellite) ToD clock. The propagation delay as estimated by the receiver is: !! " = # ! " + $% − ! " = ' ) |&345# ! "% −&'|+ (11.5.1) Where: • &345# ! "% is the satellite location vector at the time of transmission, extracted from the available satellite ephemeris. The ToD value is derived from the received NCR value, and the association message received in the SAT message. • &' is the unknown location vector of the terminal. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 227 • is the timing measurement error. At least 4 of these measurements need to be collected, to extract the unknown terminal coordinates and the bias. Additional measurements would provide a better estimation of the parameters. This set of equations are very similar to the set of equations used in a GNSS system. The solution can be made directly, or as more commonly implemented, by iterative techniques based on linearization, or using Extended Kalman filtering. Further analysis is beyond the scope of the present document, however, the location estimation for the purpose of predicting the propagation delay, is less sensitive to the horizontal Depletion of Precision (DOP) resulting from the fact that all the measurements are taken along the satellite tracks. To illustrate the effect of the vertical DOP, or rather the DOP in the direction of the satellites, consider an example for a satellite of altitude 1 000 km. Figure 11.14 depicts the satellite trajectory, as seen by a terminal located at the origin in a coordinate system at the plane of the orbit and the terminal. The satellite locations every 10 s are indicated by a blue dot. Figure 11.14: Location Determination by TDOA Using the Time Difference of Arrival method (TDOA) method, by taking the differences between the propagation delay measurements of Equation (clause 11.5.1) at two points, the equi-difference locus takes the form of an hyperbola depicted (partially) in Figure 11.14. Two such hyperballs are plotted in Figure 11.14. One is the difference between measurements taken at t = 0 and t = 100 s. The other is the difference between the measurements taken at t = 0 and t = 200 s. The intersection of the two hyperbolas is the estimated location of the terminal. Figure 11.15 is a zoom view of the above around the terminal location (x=0 and y=0), where for each set of measurements two hyperbolas are plotted one with +10 ns measurement error, and the other with -10 ns measurement error. The uncertainty area, shaded in grey is indeed elongated and represents a large DOP. It should be noted that in our case, the DOP at the direction of the latest position of the satellite should be considered, as this is the error introduced in the timing transmission. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 228 Figure 11.15: Location Determination by TDOA- zoom around the terminal location For this specific example, the timing errors, as a function of time, is given in Table 11.11: Table 11.11: Timing Errors for TDOA, 1 000 km satellite Measurement Time (s) Timing Error 40 ± 28,5 µs 100 ±9 µs 200 ±240 ns Compared to the measurement 10 ns error, these values represent a DOP at the order of magnitude of 14 to 2 850, which are quite high, however, if the guard time allocated for logon is larger than that, the RCST would be capable to transmit a logon burst and use the CMT mechanism quite early in the process. |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.5.4 Operation without GNSS- Summary | The DVB-RCS2 specifications provides the means of operating without relying on GNSS. The terminal should have the capability of scanning and acquiring the satellite signal, make measurements of timing, frequency offsets and direction, tracking the satellite and perform location calculations. Acquisition time depends on the guard times allocated for the logon channel, and for systems designed for it, it may be achieved quite quickly using a single satellite orbit. |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.6 System Enhancements | |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.6.0 Introduction | CRs 105-115 of [i.76] aim at facilitating, or rather improving, the performance and ease the implementation of the system, and its components: terminals and payloads, within a NGSO system. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 229 |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.6.1 Satellite Positioning Format | While previous versions of the DVB set of specifications referred to satellites in geostationary orbits, for which a single longitude designation suffices, operation with a NGSO satellites require the terminal to know the entire orbit. DVB first introduced a new table in Version 1.18.1 of the DVB-Service Information specification (DVB-SI) [i.83], the SAT table, containing the satellite position information in the well-known TLE format [i.82]. In Version 1.19.1 of [i.83], a new satellite position information format, derived from the Orbit Ephemeris Message (OEM) format introduced in [i.81], was added. While the TLE format is light weight, and orbit parameters of all the satellites are publicly available [i.82], its accuracy is limited, it cannot account for variations, planned manoeuvres and collision avoidance activities. The new OEM format contains time-series state vectors (time stamp, position, velocity, accelerations vectors, and covariance data), and provides much higher accuracy of the satellites' ephemerides. It also includes "metadata" information, such as start time, stop time, interpolation method between points, etc. As this format may require several sections of the SAT table, when the data set for one satellite needs to be split between several sections, the values of the fields in the "Header" and "Metadata" groups should be identical in all relevant sections. From the terminal point of view, using this format means that, essentially, no measurement or tracking of the satellite is needed. If the terminal position is known, antenna pointing direction and Doppler compensation can be accurately extracted from the trajectory data. |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.6.2 Support of handover | A NGSO satellite is visible from a given point on earth for only a limited time. Hence uninterrupted communication requires frequent handovers between satellites, even for fixed terminals. The original version of the DVB-RCS2 document [i.1] has support for mobility and for handovers of terminals between cells belonging to the same satellite. Both distributed approach, with the detection and recommendation taking place in the terminal, and the centralized approach, in which the detection and recommendation for handover are made by the network controller, are supported. However, there is no explicit support for make before break handovers. Make before break handovers require an auxiliary channel: an additional antenna (or an additional beam of a multi-beam antenna), and a separate receive and transmit chains, to manage the transition. It should be noted, though, that when the NGSO system uses beam hopping one RCST receive/transmit chain can be used. That requires the antenna system to enable fast switching between satellites, e.g. using an Electronically Steerable Antenna. In the new revision of the DVB-RCS2 specification [i.1], in order to enable full control over the terminal operation (mainly if the centralized approach is preferred by the operator) new commands to initiate the handover of the auxiliary channel were added to the mobility command descriptor. As for the mobility management, additional mobility management tables were added to enable signalling of a specific channel in a specific satellite as a target of handover. |
cb3dadab4f22493b63142dd0ebb828dd | 101 545-4 | 11.6.3 Support of Beam Hopping | The importance of beam hopping for any multi-beam satellite system was recognized by DVB and it initiated the amendment of the DVB-S2X specification [i.80] to support it. The specification was approved in 2019. Signalling to support it in the DVB-RCS2 [i.1]. DVB-SI [i.83] and DVB- GSE-LLC [i.84] were introduced as well. Beam hopping waveforms, based on Annex E superframes in [i.80] were added. A descriptor for those waveforms was added to the relevant signalling specifications, as well as beam hopping time plans, NGSO satellite positions and cell-fragment information were added to the newly defined SAT table. To simplify the implementation of the payload, a small, but significant change was made to the reference point of the Network Clock Reference (NCR). The NCR packet contains a time stamp referring to an easily distinguishable point of time within the transmitted signal. The new reference point is now the start of superframe thus making it possible to synchronize the terminal synchronization mechanism with the beam hopping time plan. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 230 Figure 11.16 shows the options that exist for the NCR packets: Figure 11.16: Association of NCR Value to SOSF When the forward link uses a DVB-S2X superframe format (described by the S2Xv2 descriptor, specified in [i.83]) the NCR_reference flag indicates if the NCR packets contains a time stamp associated with the first symbol in the Start-of Superframe (SOSF) field or, it refers to the first symbol of a SOF field of a DVB-S2X frame. The NCR_version field of that descriptor indicates if the NCR packet resides within the first XFECFRAME, of the third one. To simplify the implementation of the new NCR reference point, a constraint avoiding the insertion of frame fragments before the first XFECFRAME was introduced. This constraint applies only to Superframe Format 5, as defined in [i.80], and enables the receiver to estimate the time of the first symbol of the SOSF by estimation of other known time-instances, e.g. by estimating the time of arrival of the first symbol of the first frame's SOF, the first symbol of the SOSF can be estimated by deducting the time of 1 476 symbols. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 231 Annex A: Generalized CPM Waveform Definition A.1 Introduction A.1.0 General ETSI EN 301 545-2 [i.1] defines a combination of convolutional coding and Continuous Phase Modulation (CPM) as one of the two physical layer schemes adopted for use in DVB-RCS2 (see clause 7.1 of [i.1]). There are more general approaches to define CPM waveforms for the return satellite interactive channel. Such generalized approach provides a framework that can incorporate future improvements by introducing new features in terms of performance, spectral efficiency and implementation. This clause provides a unified approach in defining CPM waveforms including the CC-CPM normative scheme of [i.1] as well as alternative combinations of coding, interleaver and CPM modulation. The definition of Generalized CPM waveform has impact on several elements of ETSI EN 301 545-2 [i.1] for potential future use (as a user defined waveform). Figure A.1 illustrates the general framework for defining generalized CPM waveforms. Figure A.2 shows two examples to use the general framework to define the normative CPM waveform (CC-CPM) and an alternative waveform using a different FEC, interleaver and modulation parameters. Figure A.1: Generalized CPM Waveform Definition ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 232 Figure A.2: CC-CPM and eBCH-CPM schemes in Generic CPM mode A.1.1 CC-CPM scheme A Concatenated Coding with CPM (CC-CPM) modulation scheme is defined in clause 7.3.5.2 of ETSI EN 301 545-2 [i.1]. ETSI EN 301 545-2 [i.1] also defined binary, non-systematic, non-recursive convolutional codes as part of the CC-CPM scheme. The constraint length K is either 3 or 4. The generator polynomials for the rate 1/2 constraint length K=3 code are: • GNS1 = 1 + x2 (5 in octal) • GNS2 = 1+ x + x2 (7 in octal) The generator polynomials for the rate 1/2 constraint length 4 code are: • GNS1 = 1 + x + x3 (15 in octal) • GNS2 = 1+ x + x2 + x3 (17 in octal) Code rates >1/2 are obtained by puncturing the rate 1/2 code. The puncturing patterns are given in ETSI EN 301 545-2 [i.1]. A.1.2 eBCH-CPM scheme The functional blocks in the eBCH-CPM encoder include the eBCH encoder, bit interleaver, bit-to-symbol mapping, and CPM modulator. All possible scheme parameters are reported in clause A.4. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 233 Two e-BCH codes are specified: (k: information bits of sub-block, n: coded bits of sub-block, dmin: minimum distance): • (k=51, n=64, dmin=6) adopted for low spectral efficiencies ( 1 ≤ η ); • (k=113, n=128, dmin=6) adopted for high spectral efficiencies ( 1 η > ). where k, n and dmin represent the information bits of sub-block, the coded bits of sub-block, and the minimum distance respectively. The bit interleaver is based on the following permutation rule: ( ) 1 ,..., 2 ,1, 0 mod 2 1 − = + × = N i N P P i j where P1 and P2 are two prime numbers and N is the interleaver length, i.e. the codeword length. The interleaver is a classical S-random interleaver. The algorithm needed to generate the interleaver pattern is summarized in the following steps: 1) Random permutations are created by generating random integers i, 1 ≤ ≤ , without replacement. 2) The permutation rule is defined as follows: - each randomly selected integer is compared to a spread value S previously selected integers; - if the current selection is equal to any S previous selections within a distance of ± S, then the current selection is rejected; - this process is repeated until all N integers are selected. The CPM modulation parameters are specified in clause A.3.2. A.2 Forward Link L2S A.2.0 Overview This clause describes necessary changes to Lower Layer Signaling in order to support the Generic CPM scheme. A.2.1 Description of FL L2S Components A.2.1.1 The FCT2 content Table 6-15 of ETSI EN 301 545-2 [i.1] defines all transmission types used in the frame type. In order to support the generalized CPM waveform, a new entry has been added to this table as shown in Table A.1: • tx_format_class: This field indicates the transmission format class of all transmission types used in the frame type. The values are assigned in Table A.1. Table A.1: Coding of Transmission Format Classes Value tx_format_class 0 Reserved 1 Linear Modulation Burst Transmission 2 Continuous Phase Modulation Burst Transmission 3 Continuous Transmission 4 Spread-Spectrum Linear Modulation Burst Transmission 5 to 127 Reserved 128 Generic Continuous Phase Modulation Burst Transmission 129 to 255 User defined ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 234 A.2.2 Syntax and Coding of FL Signals for L2S A.2.2.1 The ICT content A.2.2.1.0 Interleaver Configuration Table 6-16 of ETSI EN 301 545-2 [i.1] specifies the different transmission types. In order to support a generalized CPM waveform, an Interleaver configuration table ICT is added as shown in Table A.2. The ICT Table_id is set to 0xC0 (see Table 6-1 of ETSI EN 301 545-2 [i.1]). Table A.2: Syntax of the Interleaver Configuration Table Content Syntax No. of bits Mnemonic Reserved (see note 1) Information interleaver_configuration_table_content() { Interleaver_loop_count 8 uimsbf for (j=1;j<=interleaver_loop_count;j++) { parameterized_interleaver 7 1 uimsbf if(parameterized_interleaver==1){ interleaver_seed_P1 4 12 uimsbf interleaver_seed_P2 4 12 uimsbf spread_value 4 12 uimsbf } else{ interleaver_id 8 uimsbf pi_data_size 16 uimsbf for (k=0;k<pi_data_size;k++) { pi_data_bit 1 uimsbf } while (!bytealigned) { stuffing_bit 1 uimsbf } } } NOTE 1: Reserved bits are of type bslbf, and should precede the Information bits on the same line. NOTE 2: The interleaver loop may contain zero, one or more of all the interleavers. • interleaver_loop_count: The amount of interleavers present in the next loop. • parameterized_interleaver: This is a 1 bit field. When set to 1, it stipulates that the CPM bit interleaver permutations be computed using the parameters interleaver_seed_P1, interleaver_seed_P2, spread_value. When set to 0, the interleaver is specified by its interleaver id. • interleaver_seed_P1: This 12 bit field is an positive prime integer number used in generation the interleaver permutations. • interleaver_seed_P2: This 12 bit field is an positive prime integer number used in generation the interleaver permutations. • spread_value: This 12 bit field is an positive integer number used in generation the interleaver permutations. • interleaver_id: This 8 bit field identifies the interleaver that is defined here. • pi_data_size: This 16 bit field specifies the data size in bits for interleaver PI, the value should always be a multiple of 12. • pi_data_bit: This field contains a data bit for the interleaver PI. • stuffing_bit: Since the UW description, interleaver PI are byte aligned, stuffing bits are present until the next byte boundary. The stuffing bits may take any value and should be discarded by the terminal. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 235 A.2.2.1.1 Data Block for Generic CPM The data block format providing the configuration for the generic CPM mode is specified in Table A.3. Table A.3: Data Block format for the Generic CPM Transmission Syntax No. of bits Mnemonic Reserved (see note) Information gen_cpm_data_block { tx_block_size 8 uimsbf threshold_es_n0 8 uimsbf tx_start_offset 12 20 uismbf modulation_mh 1 3 uimsbf modulation_ph 1 3 uimsbf encoder_type 1 1 uimsbf cpm_encoder_memory_length 2 uimsbf modulation_type 4 uimsbf for (i=0;i<cpm_encoder_memory_length x 16;i++) { phase_pulse_sample 16 uimsbf } for (i=0;i<modulation_ph;i++) { 8 uimsbf normalization_sequence } number_of_subblocks 8 uimsbf subblock_parity 1 uimsbf encoding_polynomial1 23 uimsbf encoding_polynomial2 8 uimsbf subblock_input_length 16 uimsbf subblock_output_length 16 uimsbf one_bit_longer_subblocks 8 uimsbf interleaver_id 8 uimsbf uw_length 16 uimsbf for (i=0;i<uw_length;i++) { uw_bit; 1 uimsbf } while (!bytealigned) { stuffing_bit 1 uimsbf } normalization_sequence_length 3 nbr_uw_sequences 1 4 for (i=0;i<nbr_uw_sequences;i++) { uw_sequence_start; 15 uimsbf normalization_sequence_flag 1 uimsbf uw_segment_length 8 uimsbf } } NOTE: Reserved bits are of type bslbf, and should precede the Information bits on the same line. Semantics for the gen_cpm_data_block: • tx_block_size: The number of consecutive BTUs required for transmission of the physical layer block used by the specific tx_type. This indicates the size of the timeslot required for the burst. • threshold_es_n0: This is the nominal sensitivity for the transmission type encoded as (5 × threshold) + 120 with the threshold given in dB, and serves as a reference for ACI control as specified in clause 7.3.8 of ETSI EN 301 545-2 [i.1]. • tx_start_offset: A 20-bit field that gives the nominal offset for burst start from the start of the timeslot in units of NCR ticks. • modulation_mh: This 3 bit field specifies the numerator in a fraction representing the modulation index. The numerator mh equals the value of this field +1. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 236 • modulation_ph: This 3 bit field specifies the denominator in a fraction representing the modulation index. The denominator p equals the value of this field +1. • encoder_type: This 1 bit field specifies the encoder type configuration. A value '0' indicates FEC variant 1 (systematic). A value '1' indicates FEC variant 2 (non-systematic). • modulation_type: This 4 bit field specifies the modulation type and the symbol mapping option, as defined in Table A.4. Table A.4: Modulation_type value Modulation_type value Modulation Symbol mapping 000 binary Linear 001 quaternary Gray mapping 010 quaternary Linear mapping 011 octal Gray mapping 100 octal Linear mapping 101 reserved Reserved 110 reserved Reserved 111 user-defined User-defined Table A.5: Bit to Symbol mapping for binary linear mapping MSB LSB b0 Symbol value 0 -1 1 1 Table A.6: Bit to Symbol mapping for quaternary linear mapping MSB b1 LSB b0 Symbol value 0 0 -3 0 1 -1 1 0 1 1 1 3 Table A.7: Bit to Symbol mapping for quaternary gray mapping MSB b1 LSB b0 Symbol value 0 0 -3 0 1 -1 1 1 1 1 0 3 Table A.8: Bit to Symbol mapping for octal linear mapping MSB b2 b1 LSB b0 Symbol value 0 0 0 -7 0 0 1 -5 0 1 0 -3 0 1 1 -1 1 0 0 1 1 0 1 3 1 1 0 5 1 1 1 7 ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 237 Table A.9: Bit to Symbol mapping for octal gray mapping MSB b2 b1 LSB b0 Symbol value 0 0 0 -7 0 0 1 -5 0 1 1 -3 0 1 0 -1 1 1 0 1 1 1 1 3 1 0 1 5 1 0 0 7 • cpm_encoder_memory length: This 2 bits specifies the length of CPM encoder. Table A.10: CPM encoder memory length MSB b1 LSB b0 Memory length(L) 0 0 1 0 1 2 1 0 3 1 1 Reserved • phase_pulse_sample: This 16 bit field specifies a sample of the applicable phase pulse. • number_of_subblocks: This 8 bit field specifies the number of subblocks in the encoder. • subblock_parity: This 1 bit when set (value = 1) means that there is an additional parity bit for the systematic code. When not set (value = 0) there is no parity bit. • encoding_polynomial1: This 23 bit field specifies encoding polynomial 1. See clause A.3.1.1 for the definition. • encoding_polynomial2: This 8 bit field specifies encoding polynomial 2. See clause A.3.1.1 for the definition. • subblock_input_length: This 16 bit field specified the encoding input length for the subblock. • subblock_output_length: This 16 bit field specified the encoding output length for the subblock. By definition, this value is larger or equal than the value of the subblock_input_length. • interleaver_id: This 8 bit field identifies the interleaver defined in ICT (interleaver configuration table) to be used. • uw_length: This is an 8 bit field specifying the UW length in bits. The loop which follows is aligned however on byte boundaries. This means that for example if the UW length is 14 bit, the loop over the UW bytes will consist of 2 Byte with the last 2 bit being merely stuffing bits. • uw_bit: This is a 1 bit field specifying a bit of the UW. As the UW is not scrambled, a proper sequence should be selected in order to comply with ETSI requirements concerning off-axis EIRP. The bits are listed in transmission order (first bit listed = first bit send on air interface). • stuffing_bit: Since the UW description, interleaver PI are byte aligned, stuffing bits are present until the next byte boundary. The stuffing bits may take any value and should be discarded by the terminal. • normalization_sequence_length: This 3 bit field specifies the number of normalization symbols. • nbr_UW_sequences: This 8 bit field specifies the number of UW sequences to be inserted in the bursts. • uw_sequence_start: This 15 bit field provides the position (expressed in number of bits) of the first bit of the UW sequence within the burst. A value of zero means the first bit of the burst. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 238 • normalization_sequence_flag: This 1 bit flag when set (value = 1) means that the UW sequence consists of a UW segment and a preceding normalization sequence. When not set (value = 0) this means there is no preceding normalization sequence and the UW sequence only consists of the corresponding UW segment. • uw_segment_length: This 8 bit field specifies the number of bits in the UW segment. • normalization_sequence: This 8 bit field contains the entries of the normalization sequence lookup table. If the normalization_sequence_length in symbols multiplied by the log2(alphabet size M) is less than 8 bit the normalization sequence bits will start at the left most bit in the field and will be followed by a sufficient number of stuffing bits set equal to zero. The lookup table will always be signalled even in case no normalization is required. • tx_block_size: The number of consecutive BTU's required for transmission of the physical layer block used by the specific tx type. For TDMA this indicates the size of the timeslot required for the burst, given in BTU's. For TDM this parameter is irrelevant and it is set to zero. • tx_start_offset: For TDMA this is the nominal offset from the start of the timeslot in NCR ticks. For TDM this parameter indicates the time in the future the transmission may start. A.3 Return Link -Transmission Bursts A.3.0 Introduction This clause defines the burst generation according to the generic CPM burst structure. A.3.1 Generic Coding and Interleaving A.3.1.1 The Generalized Sub-Block Polynomial Encoder The Generalized Sub-block Polynomial Code (GSPC) is defined by FEC parameters signalled to the terminal in the BCT Table. Main principles of the coding scheme are defined below. The rate K/N Generalized Sub-block Polynomial encoder segments the uncoded K-bit data sequence delivered by the CRC encoder, [ ] 0 2 1 , , , x x x X K K L − − = in Nb blocks sub-blocks with length k1, k2,…kNb. K k b N j j = =1 [ ] ) ( 0 ) ( 2 ) ( 1 ) ( , , , j j k j k j x x x X j j L − − = ) , ( ) ( j m m j m x x ′ = + = ′ + = Nb j j j k m j m m 1 ' ) , ( = ≥ + = < shortened j shortened j k k q j k k q j 1 1 1 ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 239 = ≥ + = < shortened j shortened j n n q j n n q j 1 1 1 Each sub-block is encoded using a binary systematic or non-systematic linear polynomial or convolutional code (signalled by the FEC parameters). FEC variant 1 In case of FEC variant 1, the encoder appends a number of parity check bits equal to the degree of the encoding polynomial for the jth block. It also has the option to add an overall parity bit. Denote the output bits of the jth sub-block as: [ ] [ ] ) ( 0 ) ( 1 ) ( 0 ) ( 2 ) ( 1 ) ( 0 ) ( 2 ) ( 1 ) ( , , , , , , , , , j j d j j k j k j j n j n j p p x x x y y y Y j j j j j L L L − − − − − = = where: j j j d k n = − N n b N j j = =1 The mathematical definition of the sub-block polynomial encoder is: )1( ) ( ) ( ) ( ) ( ) ( ) ( ) ( j reg j e j reg j Y e D D Y D Y j ⋅ + ⋅ = ) ( ) ( ) ( ) ( ) ( ) ( D P D D X D Y j d j j reg j + ⋅ = where the binary polynomials X, Yreg, Y and P are defined as: − = = 1 0 ) ( ) ( ) ( j k t t j t j D x D X − − = = ) ( 1 0 ) ( ) ( ) ( j j e n t t j t j reg D y D Y − = = 1 0 ) ( ) ( ) ( j n t t j t j D y D Y − − = = ) ( 1 0 ) ( ) ( ) ( j j e d t t j t j D p D P And where the generator polynomial is defined as: − = = ) ( 0 ) ( j j e d t j t S D g D G The 'parity' polynomial is obtained as the remainder of the division as defined in: ) ( mod ) ( ) ( ) ( ) ( ) ( D G D D X D P S e d j j j j − ⋅ = Note that the code reduces to a regular polynomial code if the e(j) bit is set to zero. If the e(j) bit is set to one, an overall parity bit is added to the polynomial code. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 240 FEC variant 2 FEC variant 2 is a non-systematic binary convolutional code. The encoder generates the following codeword: D D X D G D X D G D Y j NS j NS j ⋅ ⋅ + ⋅ = ) ( ) ( ) ( ) ( ) ( 2 ) ( 2 2 ) ( 1 ) ( where the generator polynomials are given by the expression below for i=1,2: = − = j j d t t i t d NSi D g D G 0 ) ( Note that dj has the meaning of polynomial degree, here (and not the amount of parity bits). Practical implementation Note that both encoders (systematic and non-systematic) can be implemented using the same circuit. A possible implementation is shown below. The canonical implementation of both schemes is a shift register with feedback (that can be enabled or disabled) starting in the all-0 state. In case of the systematic code, feedback is enabled and the feedback taps are equal to the generator polynomial of the polynomial code. In case of a non-systematic code, feedback is disabled and the filter taps are equal to the generator polynomial of the convolutional code. Note that for the non-systematic encoder, this circuit should be run twice (once for each polynomial). Figure A.3: Practical implementation Signalling The number of sub-blocks Nb, the generic channel interleaver (see below) and the code type (systematic / non-systematic) are signalled by Table BCT. In addition, for each sub-block and for each transmission mode, the encoding polynomials G and the parameters jk , jd and ej are indicated by Table BCT. The range of all these parameters and the mapping to the signalling table from section is defined in the signalling section. g0 t>k and FEC var1 t>k g1 g22 synchronous rst at start of each sub-block e last parity bit data bits di ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 241 Table A.11: Mapping parameters to Signal Name Parameter Signal name FEC variant encoder_type Nb number_of_subblocks ej subblock_parity GS or GNS1 encoding_polynomial 1 GNS2 encoding_polynomial 2 kshortened subblock_input_length nshortened subblock_output_length q1 one_bit_longer_subblocks The two polynomials are defined by 23 and 8 bits respectively. The first bit g0 is always 1. Mapping on the signalled bits is defined in the table below. The first bit in the signalled sequence is denoted by x1. Polynomial 1 g0=x1,...,g22=x23 Polynomial 2 g0=x1,...,g7=x8 In case of FEC variant 1, there is only one polynomial and the second polynomial definition is ignored. A.3.1.2 Generic Interleaver and Puncturer The interleaver should be totally programmable and is therefore implemented as lookup tables, which are signalled in the BCT Table. The look-up table of the interleaver defines the mapping between the position of a bit in the coded data block entering this interleaver and the position of this bit in the data block leaving the interleaver. The look-up table defines the function n(m), with index m = 0,1 …, M-1 and values n = 0, 1, ... , N-1, whereas N is the size (number of bits) of the coded data block and M is the size of the coded data block after (optional) puncturing. Value n = 0 corresponds to the first bit entering the interleaver, index m = 0, corresponds to the first bit leaving the interleaver. The look-up table will be signalled as M words of 12 bit, whereas each word represents a value n for the corresponding index m. The word for index m = 0 is sent first. A.3.2 Generic Continuous Phase Modulation The CPM modulation is defined through the following parameters (programmable and signalled in the Table BCT): • alphabet size, M: - values M = 2, 4, 8 • modulation index, h: - the modulation index h is rational fraction h = m/p <=1 with m and p relative prime positive integers in the range [1,2,…,8]. • phase pulse samples, q(t): - generic with fixed duration L=2 symbols: q(t) = 0 0, 0 ( ) ,0 0.5, t s s t g d t LT t LT τ τ < ≤≤ > , where L is the memory length of the CPM modulation(L=2) ( ) g t is the frequency response, ( ) RC g t and ( ) REC g t represent the raised-cosine(RC) and rectangular(REC) pulse. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 242 1 (1 cos ),0 4 ( ) 0, s s s RC t t LT T T g t otherwise π − ≤ ≤ = g t = 1 1 cos ,0 2 4 0,otherwise s s s t t T T T π − ≤≤ and 1 ,0 4 ( ) 0, s s RC t LT T g t otherwise ≤ ≤ = g t = 1 ,0 2 4 0,otherwise s s t T T ≤≤ - generic with fixed duration L=3 symbols: Note this includes effective duration Leff<=3 symbols - symmetrical representation q(t)=1/2 - q(L-t) • symbol mapping method The CPM waveform is expressed by the following formula: 0 ( , ) exp 2 ( ) n n s t j h q t nT α π α +∞ = = ⋅ − % The real transmitted signal (at carrier frequency f0) is: { } t f je t s t s 0 2 ) , ( ~ Re ) , ( π α α = where: • T is the symbol period. • αn is the sequence of uncorrelated data information symbols to be transmitted and each taking one of the values {±1, ±3, ±(M-1)}. • q(t) is the phase pulse. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 243 A.4 Reference waveforms for Generic CPM (eBCH-CPM scheme) Table A.12 lists the reference waveforms for generic continuous phase modulation format class bursts considering eBCH-CPM scheme. The parameters follow the syntax specified in clause A.2.2.1.1. Table A.12: Reference Waveforms for Generic Continuous Phase Modulation bursts (eBCH-CPM scheme) Waveform id Content Type FEC input bit length (K) FEC output bit length (N) Preamble (uw) length Data #1 bit length Normalization _seq. #1 Midamble (uw) length Data #2 length Normalization _seq. #2 Burst symbol length Alphabet size (M) Modulation index (h) Code rate eBCH (subblock_output_length, subblock_input_length) Number of subblock (number_of_subblocks) Carrier Spacing Spectral Efficiency Memory length (L) UW bits (uw_bits) (Preamble+Midamble) Phase Response 1 Logon 456 768 64 64 6 64 704 6 454 4 4/7 0,591 (64,51) 24 2,3259 0,5 2 7CD593ADF7818AC8 Raised Cosine 2 Control 168 272 64 64 6 64 208 6 206 4 4/7 0,591 (64,51) 8 2,3259 0,5 2 7CD593ADF7818AC8 Raised Cosine 3 Traffic /Control 400 673 64 64 6 64 609 6 407 4 4/7 0,591 (64,51) 21 2,3259 0,5 2 7CD593ADF7818AC8 Raised Cosine 4 Traffic /Control 400 556 64 64 6 64 492 6 348 4 3/7 0,7053 (64,51) 12 1,8773 0,75 2 7CD593ADF7818AC8 Raised Cosine 5 Traffic /Control 400 556 64 64 6 64 492 6 348 4 2/7 0,7162 (64,51) 12 1,3766 1 2 7CD593ADF7818AC8 Raised Cosine 6 Traffic /Control 400 475 64 64 4 64 411 4 306 4 1/4 0,847 (128,113) 5 1,15 1,5 2 7CD593ADF7818AC8 Raised Cosine 7 Traffic /Control 400 460 64 64 6 64 396 6 298 4 1/5 0,871 (128,113) 4 0,99 1,8 2 7CD593ADF7818AC8 Raised Cosine 8 Traffic /Control 1 024 1 726 64 64 6 64 1 662 6 933 4 4/7 0,591 (64,51) 54 2,3259 0,5 2 7CD593ADF7818AC8 Raised Cosine 9 Traffic /Control 1 024 1 440 64 64 6 64 1 376 6 790 4 3/7 0,7053 (64,51) 32 1,8773 0,75 2 7CD593ADF7818AC8 Raised Cosine 10 Traffic /Control 1 024 1 427 64 64 6 64 1 363 6 784 4 2/7 0,7162 (64,51) 31 1,3766 1 2 7CD593ADF7818AC8 Raised Cosine 11 Traffic /Control 1 024 1 204 64 64 4 64 1 140 4 670 4 1/4 0,847 (128,113) 12 1,15 1,5 2 7CD593ADF7818AC8 Raised Cosine 12 Traffic /Control 1 024 1 174 64 64 6 64 1 110 6 657 4 1/5 0,871 (128,113) 10 0,99 1,8 2 7CD593ADF7818AC8 Raised Cosine 13 Traffic /Control 1 504 2 544 64 64 6 64 2 480 6 1 342 4 4/7 0,591 (64,51) 80 2,3259 0,5 2 7CD593ADF7818AC8 Raised Cosine 14 Traffic /Control 1 504 2 128 64 64 6 64 2 064 6 1 134 4 3/7 0,7053 (64,51) 48 1,8773 0,75 2 7CD593ADF7818AC8 Raised Cosine 15 Traffic /Control 1 504 2 089 64 64 6 64 2 025 6 1 115 4 2/7 0,7162 (64,51) 45 1,3766 1 2 7CD593ADF7818AC8 Raised Cosine 16 Traffic /Control 1 504 1 774 64 64 4 64 1 710 4 955 4 1/4 0,847 (128,113) 18 1,15 1,5 2 7CD593ADF7818AC8 Raised Cosine 17 Traffic /Control 1 504 1 714 64 64 6 64 1 650 6 927 4 1/5 0,871 (128,113) 14 0,99 1,8 2 7CD593ADF7818AC8 Raised Cosine ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 244 A.5 CPM Complexity A.5.0 Introduction This clause provides an overview of algorithmic complexity of implementing CPM scheme at the transmitter and the receiver. A.5.1 eBCH-CPM Modulator A.5.1.1 eBCH Encoder Figures A.4 and A.5 depict the functional operation flow in the process of eBCH encoding. First, the 'Message' data is fed into eBCH Encoder controller block. The BCH parity can be generated by the waveform ID about K_short data input. The tail bit generator makes one tail bit and eBCH output controller constructs the entire eBCH_output data. The specific operation procedure is as follows: 1) According to the waveform ID, 'BCH Encoder Controller' selects BCH 'Generator polynomial Coefficients'. 2) According to the waveform ID, 'BCH Encoder Controller' sends to 'Parity_calulation' block by appending 'zero' bits, the number of N_zero to K_short message information. 3) While 'Message' data including the number of K(N_zero+K_short) is fed, it is open in gate1 and off in gate2 in 'Parity calculation block'. 4) The gate1 is off when the number of K, all 'Message' data is fed. As soon as gate2 becomes "on" mode, the parity bits outputs through shift register operation. 5) While calculating the parity in 'BCH Encoder Controller', 'Message' data outputs and 'Party bits' outputs, subsequently. 6) The tail bit should be generated through 'N_short-1' in eBCH output controller. 7) Finally, 'eBCH_Out' (N_short) data bits are the outputs of the encoder. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 245 Figure A.4: eBCH encoder functional architecture Figure A.5: Timing diagram of eBCH Encoder process ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 246 Table A.13 represents the generator polynomial coefficient according to the different waveform ID. Table A.13: BCH Code Generator Polynomial G0 G1 G2 G3 G4 G5 G6 G7 G8 G9 G10 G11 G12 G13 G14 eBCH (51, 64) 1 0 0 1 1 1 0 0 1 0 1 0 1 - - eBCH (113, 128) 1 1 1 0 1 1 1 0 1 1 0 0 0 0 1 A.5.1.2 eBCH+CPM Modulator Based on the proposed eBCH+CPM design, it has been implemented using an FPGA Xilinx XC5VLX110 model chip. The only core module such as eBCH encoder, S random interleaver, CPM modulator and digital up-convertor have been considered in the target FPGA. The other interface blocks such as PCI communication, CPU module, RLE, Digital Analogue Convertor (DAC) and forward link signalling parsing have not been considered to compute complexity. In the modulator aspects, LUT amount used for S random interleaver design has been occupied when it has been implemented as tabularized approach. Table A.14: FPGA utilization Slice logic Utilization Utilization (Used/Total) Number of Slice Register 2 % (1 420/69 120) Number of Slice LUTs 1 % (506/69 120) Number of occupied Slices 2 % (504/69 120) Number of fully used LUT-FF pairs 16 % (279/1 647) A.6 CPM mobile performance This clause illustrates the applicability of generic CPM modulation to mobile environment. In particular a channel model based on Rice-distributed fading is assumed. Software simulations have been carried out under the assumption of highly correlated Doppler spectrum by considering directive antenna deployment. Two cases are envisaged, depending on the directivity of the antenna, which can be linked to both Rice factor and normalized Doppler frequency. Specifically, higher antenna directivity values correspond to higher Rice factors and to lower normalized Doppler frequencies. To investigate the effect of mobile channel on the CPM waveform, an ideal channel estimation at the receiver. Table A.16 presents the simulation results, comparing AWGN and Mobile channel. As shown, for mobile LOS (K=17 dB, Rice Factor), more than 2 dB loss should be considered compared the results obtained with the ideal AWGN channel. Table A.15: Simulation Parameters Parameter Value(s) Spectral Efficiency [bit/s/Hz] 0,5 0,75 1 Modulation Index 4/7 3/7 2/7 Modulation Order (log2 M) 2 Target Code Rate 0,591 0,7053 0,7162 Code Block size [bit] 1 024 Baud Rate [kbaud] 512 Rice Factor K [dB] 17 Speed [km/h] 150 Normalized Doppler Frequency 8 × 10-4 (Ku:14 GHz, Ka:30 GHz) ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 247 Table A.16: eBCH+CPM Thresholds (PER = 10-3) Spectral efficiency AWGN Mobile(K=17) Gap(degradation) 0,5 1,8 dB 4,1 dB 2,3 dB 0,75 2,1 dB 4,5 dB 2,4 dB 1 3,15 dB 5,4 dB 2,25 dB ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 248 Annex B: Examples of CRDSA Implementation and Performance B.1 Introduction This annex provides some examples of performance and implementation results for the CRDSA technique [i.57], [i.58] applied to random access return channels based on software simulations and a hardware implementation which has served as a testbed to validate this technology [i.60]. Slotted Aloha (SA) protocol [i.55] or enhanced version of the scheme, such as Diversity Slotted Aloha (DSA) [i.56], are used in TDMA systems with low efficiency and reliability typically for logging into the network. Recently an enhanced version of DSA dubbed Contention Resolution Diversity Slotted Aloha (CRDSA) has been introduced [i.57]. The CRDSA key idea is to transmit two or more (total) replicas of the same packet at random locations within the same frame as in DSA but with a slightly higher signalling to point to the location of other packet replicas. In case of successful packet reception at the gateway this extra signalling information allows to locate the twin packet within the frame and to accurately cancel it in addition to the one successfully decoded. To be remarked that using a powerful Forward Error Correcting (FEC) code and an adequate signal to noise ratio, there is probable to correctly detect the packet even in the presence of other colliding bursts. By iterating the process a number of times some of the initially lost packets can be recovered, improving the DSA performance. Results reported in [i.58] and [i.59] indicate that the CRDSA throughput is significantly higher than that of conventional Slotted Aloha and Diversity Slotted Aloha. The throughput gain of the CRDSA compared to conventional random access scheme is mainly due to successive interference cancellation. The CRDSA performance improves noticeably with burst power unbalance (received from different user terminals within the same time slot). Such power unbalance facilitates the successive cancellations performed by the burst demodulator. The high throughput/high reliability CRDSA technique in a random access scenario makes this scheme suitable to support a low-duty cycle bursty traffic such as VSAT and broadband access return link signalling packets as well as SCADA applications. B.2 CRDSA Parameter Settings Example B.2.0 Introduction In this clause some examples on how to get the best performance of the CRDSA TDMA random access scheme are provided. Examples are based on relevant literature or unpublished results from which some key results have been extracted. Simulation results reported in this clause are based on the 3GPP FEC coding scheme [i.64] and using a burst containing100 information bits. Each frame is composed of 100 time slots and the traffic is assumed to have a Poisson distribution. The number of slots per frame should be kept possibly above 60 since lowering the number of slots per frame may degrade the CRDSA performance. In an MF-TDMA system, CRDSA time slots can be distributed in a two- dimensional time/frequency plane, for example using 10 time slots combined with 6 frequency bins (a total of 60 time slots). It should be noted that a different selection of the coding scheme may impact the performance results, however the same trend as a function of other system parameters are expected. In the following simulated scenarios, the performance is measured in terms of the maximum normalized throughput and the Packet Loss Ratio (PLR) as a function of the system traffic loading. The performance results are obtained in the presence of AWGN with a noise density corresponding to a link quality of Es/N0 = 10 dB (excluding the signal power fluctuation). In addition, signal power fluctuations (among the bursts received from different RCSTs) are lognormally distributed with zero mean and parameterized standard deviation σ (measured in dB). B.2.1 Performance dependence on the number of replicas It is important to assess the impact of the number of burst replicas on the CRDSA performance. Results shown in Figure B.1, indicate that at a target PLR of 10-3, the best performance is obtained with four replicas. The steepness of the PLR characteristic versus the MAC load is increasing with the number of replicas but using more than 4 replicas reduces the throughput (measured at the target PLR = 10-3). ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 249 Figure B.1: CRDSA performance versus the number of replicas, 3GPP FEC coding rate ½, Es/N0 = 10 dB, equipowered burst B.2.2 Performance dependence on the burst power unbalance Figure B.2 shows the CRDSA performance assuming that bursts' power is randomly distributed according to a lognormal distribution with zero mean and parameterized standard deviation σ. It can be observed that the MAC throughput improves reaching a value above 1 for σ = 3 dB (Throughput values above 1 is possible thanks to the interference cancellation). At the same time a PLR floor appears for σ = 3 dB due to the non-negligible probability that the received packet power is too low to avoid packet errors in the presence of AWGN. This effect can be mitigated by increasing the operating Es/N0, reducing the FEC code rate (see clause B.2.3) or limiting the value of σ. It should be noted that the assumption of lognormal burst power distribution may be pessimistic as in practice system power fluctuations can be better represented by a truncated lognormal distribution that mitigates the PLR floor effects. Figure B.2: CRDSA performance versus the burst power unbalance, 3GPP FEC coding rate ½, Es/N0 = 10 dB, lognormal burst power distribution B.2.3 FEC Coding rate impact A final aspect to be assessed is the impact of the FEC coding rate on the CRDSA performance. Also in this case the results are counterintuitive as being TDMA an orthogonal access scheme one can expect that the highest throughput is obtained with the highest FEC coding rate. Instead being interested in the aggregate MAC random access throughput the results reported in Figure B.3 indicate that the best performance is obtained for r = 1/3. The throughput is similar to that of r = 1/2 but the floor for σ = 3 dB is now reduced because of the better FEC performance. Note that normalized load refer to the information bits and not to the coded symbols to allow making a fair comparison. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 250 Figure B.3: CRDSA performance versus the burst power unbalance, 3GPP FEC coding rates ½ and 1/3, Es/N0 = 10 dB, lognormal burst power distribution B.3 Reference CRDSA Implementation B.3.1 CRDSA Frame Structure This clause describes the hardware implementation, configuration and the lab performance results of a reference design of CRDSA scheme [i.60]. The main configuration parameters used to validate the CRDSA access scheme in [i.60] are reported in Table B.1. Table B.1: Hardware Configuration Parameters for CRDSA Implementation Parameter Value Modulation ♦ Coding Rate QPSK ♦ 1/2 Payload (bits) 488 Pre-amble ♦ Post-amble ♦ Pilots ♦ Guard (symbols) 40 ♦ 12 ♦ 54 ♦ 6 Symbol Rate (k symbols/s) 128 Burst Length (symbols) ♦ Burst Duration (msec) 600 ♦ 4,69 Bit Rate (kbits/s) 104,11 Bursts per Frame ♦ Frame Duration (msec) 66 ♦ 309,38 Number of Replicas (including original) per frame 4 Max Number of overlapping Bursts 8 Each CRDSA frame consists of 66 time-slots located on a single carrier where the original bursts and their replicas are allocated. The number of replicas (including the original burst) is four in each CRDSA frame. Each burst is 600 symbols long, including preamble and pilot symbols and the adopted modulation and coding rate are respectively QPSK and 1/2. The MODEM works with a baud rate of 128 ksymbols/s and the demodulator is able to perform up to seven iterations to erase the replicas within a given time-slot. The maximum number of overlapping bursts within a given time-slot is limited to eight bursts. The underlying coding used in the testbed is according to the DVB-RCS2 specifications (16-states duo-binary Turbo-ϕ Coding) [i.1]. The position of the replicas inside a given frame is obtained through running a Pseudo-Random Generator (PRG) for a sufficient number of times. The 8-bits seed of this PRG, sent to the receiver together to the 16-bits Terminal ID, is exploited by the receiver to re-generate the position of the replicas. Once a burst has been correctly demodulated (i.e. the CRC is found to be correct) it can be regenerated and subtracted from the time-slots where its replica(s) are located (see clause B.3.2 for more details). ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 251 Successive interference cancellation iterations over the whole frame are necessary to eliminate most of the burst replicas present. In a simplified scenario used for better illustrating the iterative cancellation approach the process starts with the detection of a "clean" burst (no overlapping bursts such as B1 as shown in Figure B.4), continue cancelling the demodulated bursts up to it is possible, end when no more cancellation can be done (e.g. because a burst collision, as in the case of bursts B7 and B8 in Figure B.4). In reality, because of the powerful FEC adopted or possible received bursts power unbalance, packets can also be detected in the presence of one or even more colliding bursts. So in general any detected burst (successful CRC check) is then removed together with its replica(s) from the frame being processed. At the receiver, the replica to be subtracted is re-generated by the demodulator using the whole burst decoded and re-encoded data to perform maximum likelihood channel parameter estimation. In particular, since the replicas of a given burst are transmitted by the same modulator and being the frame duration limited in time, it can be reasonably assumed that they have the same amplitude, timing and frequency inside a frame. Therefore the re-generation can be done through using the timing and the frequency parameters estimated on the slot where the first clean burst of a given set of replicas was detected. This is not true for the replica's carrier phase which need to be re-estimated being rapidly time variant because of carrier phase noise or Doppler effects. This operation is carried out on the whole burst belonging to the slot where the replica should be subtracted through cross-correlating it with its already demodulated replica burst. Exceptionally this process may be required also for the burst amplitude variation if it is rapidly fluctuating due to fading effects. More details on channel estimation for CRDSA can be found in [i.61] and [i.62]. Figure B.4: Successive Interference Cancellation Process B.3.2 MODEM Architecture This clause describes the implementation of the modulator and demodulator in the reference testbed supporting CRDSA capability [i.60]. The transmitter baseband physical layer functionality is shown in Figure B.5. Compared to a conventional modulator, the CRDSA modulator uses frame memory to allow bursts' replication before transmitting. It can be tailored according to the maximum number of slots per frame (S < 66). The factor S determines the memory overhead required for supporting the CRDSA mechanism. As shown in Table B.1, S=66 is assumed. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 252 Figure B.5: CRDSA Modulator Since the frame memory is typically based on a ping-pong (PP) implementation, exploiting 1 bit/symbol representation, this overhead can be estimated as: bits S H SYMBS PP IQ MOD CRDSA 400 158 600 2 2 = ⋅ ⋅ ⋅ = − Where 600 symbols are considered in each transmitted burst as per Table B.1. Each burst is labelled with the SEED used to initialize the PRG which provides the addresses of the bursts' replicas. When the frame memory has been filled with the original bursts, it is read back accumulating those allocated to the same slot, up to a maximum of eight bursts. It should be noted such accumulation step is not necessary in a realistic RCST and is merely required as part of the testbed functionality (single transmitter device). The typical resources available in hardware (such as FPGA) are memories (usually organized in blocks of RAM), flip-flops, embedded multipliers and combinatorial logic implemented through look-up tables. The hardware implementation of the transmitter in the testbed indicates a maximum of 100 % increase in the memory usage in RCST, considering the frame size of 66 time slots. Figure B.3 illustrates a functional diagram of the CRDSA receiver (Physical Layer base-band elements). The demodulating process is arranged into seven, double step iterations, correlating and cancelling interference within each frame. The following proceeding steps are carried out in the demodulator: • During the first step of the first iteration all the bursts of a frame are processes and those successfully demodulated are stored into the "Symbol Frame B Memory" (as shown in Figure B.6). At the same time a complete image of the frame is stored into the "Samples Frame Memory" and into the "Symbol Frame A Memory". • In the second step of the first iteration the demodulator carries out the cross correlation between the A and B Symbol Frame Memories filling the "Synch Data Frame Memory" which is used to update the "Samples Frame Memory" cancelling from it the bursts demodulated during the first step. • In the second iteration this double step procedure starts again and the cleaned content of the "Samples Frame Memory" is demodulated as far as it is possible storing the results in the A and B Symbol Frame Memories and repeating the same process carried out during the previous iteration. • This procedure is repeated up to a pre-defined maximum number of iterations (set to 7 in the testbed configuration). ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 253 Figure B.6: CRDSA Demodulator The complexity overhead of a CRDSA capable baseband demodulator, with respect to a standard DVB-RCS demodulator, can be investigated considering the two main architecture extensions required to support the cancellation process: • the insertion of frame memory buffers, • the execution of a given number of iterations. The first item increases the quantity of memory (i.e. the employed silicon area), the latter increases the processing speed requirement (in order to achieve a given throughput). A higher processing speed may require parallelization of running processes in order to cope with technological limitation of hardware. The implemented operating mode (see Table B.1) allows improving the speed of the structure to support up to four times the real-time iterations, hence D=2 demodulators are necessary to support seven iterations. Actually the D demodulators are employed in frame division fashion: each one executes seven iterations working on one frame each. In order to assess the complexity, consider bursts of 600 symbols and frames of S=66 slots and taking into account the memories shown in Figure B.3 together with the following assumptions: i) Samples Frame Memory, complex 12 bits wide 4 samples per symbol; ii) Symbols Frame A Memory, complex 12 bits wide; iii) Symbol Frame B Memory, complex 1 bit wide; and iv) Synch Data Frame Memory having a negligible complexity. Overall complexity increase of the CRDSA demodulator (in terms of memory usage), compared to a conventional demodulator is around 4 times in this specific implementation. This estimate does not include the parts that are in common with a normal MF-TDMA burst demodulator like the digital front-end and the frequency demultiplexer. B.3.3 Performance Results The behaviour of the implemented CRDSA demonstrator has been assessed based on the configuration reported in Table B.1 and under different assumptions concerning the users' power distribution. The traffic is assumed to be Poisson distributed but the maximum number of burst simultaneously transmitted is limited to 8 because of the modulator implemented limitations. Figures B.7 to B.12 show the throughput achieved, in terms of correctly demodulated bursts per frame and the associated burst Packet Error Rate (PER). Performance results obtained from the prototype hardware closely follow the software simulation results, as illustrated in each figure. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 254 As it is known from literature results [i.59] and the experimental results shown in Figures B.7, B.9 and B.11, the power unbalance between the transmitted burst replicas can considerably improve the system throughput. It is worth noting that overall performance of the CRDSA in terms of the achievable throughput, among other factors, depends on the error correction capability of the underlying FEC scheme, particularly in the presence of interference. This can be observed by comparing the performance results shown in figures below for Turbo-phi code compared to those reported in Figure B.3 for 3GPP code. Overall, it is expected that an optimization of the FEC code may further improve the CRDSA throughput which even with the existing FEC designs already offers a significant improvement compared to conventional random access scheme such as SA or DSA. Figure B.7: Physical Layer Throughput (Bursts with Power Imbalance among the received bursts) 0,00 0,10 0,20 0,30 0,40 0,50 0,60 0,70 0,00 0,20 0,40 0,60 0,80 1,00 1,20 1,40 1,60 Normalized Throughput (T) Normalized Traffic Load Throughput (SNR= 10 dB, Sigma= 0 dB) Test Bed Floating Point Simulations ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 255 Figure B.8: Physical Layer Packet Error Ratio Figure B.9: Physical Layer Throughput 1,E-06 1,E-05 1,E-04 1,E-03 1,E-02 1,E-01 1,E+00 0,00 0,20 0,40 0,60 0,80 1,00 1,20 1,40 1,60 PER Traffic Load (bursts/frame) PER (SNR= 10 dB, Sigma= 0 dB) Test Bed Floating Point Simulations 0,00 0,10 0,20 0,30 0,40 0,50 0,60 0,70 0,80 0,00 0,20 0,40 0,60 0,80 1,00 1,20 1,40 1,60 Throughput (bursts/frame) Traffic Load (bursts/frame) Throughput (SNR= 10 dB, Sigma= 1 dB) Test Bed Lab Measurements Floating Point Simulations ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 256 Figure B.10: Physical Layer Packet Error Rate Figure B.11: Physical Layer Throughput 1,E-06 1,E-05 1,E-04 1,E-03 1,E-02 1,E-01 1,E+00 0,00 0,20 0,40 0,60 0,80 1,00 1,20 1,40 1,60 PER Traffic Load (bursts/frame) PER (SNR= 10 dB, Sigma= 1 dB) Test Bed Lab Measurements Floating Point Simulations 0,00 0,10 0,20 0,30 0,40 0,50 0,60 0,70 0,80 0,90 1,00 0,00 0,20 0,40 0,60 0,80 1,00 1,20 1,40 1,60 Throughput (bursts/frame) Traffic Load (bursts/frame) Throughput (SNR= 10 dB, Sigma= 2 dB) Test Bed Lab Measurement Floating Point Simulations ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 257 Figure B.12: Physical Layer Packet Error Rate 1,E-06 1,E-05 1,E-04 1,E-03 1,E-02 1,E-01 1,E+00 0,00 0,20 0,40 0,60 0,80 1,00 1,20 1,40 1,60 PER Traffic Load (bursts/frame) PER (SNR= 10 dB, Sigma= 2 dB) Test Bed Lab Measurement Floating Point Simulations ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 258 Annex C: Evolution towards SC-FDMA C.1 Introduction This annex describes the possible merits of using the Single Carrier Frequency Division Multiple Access (SC-FDMA) scheme in a future evolution of the DVB-RCS2 standard. The SC-FDMA scheme has been adopted for use in the up-link of LTE [i.64] and, more recently, in the satellite component of DVB-NGH [i.65]. Potential advantages of the SC-FDMA scheme in the DVB-RCS2 context are the possible increase in spectral efficiency due to the possible use of zero roll-off waveforms as well as simplification in the implementation of Multi Carrier Demodulators (MCD) at the GW or at the terminal side in mesh applications. It is also worth to note that an SC-FDMA receiver chain would have significant similarity with classic OFDM receiver chains which are commonly used in terrestrial applications. The SC-FDMA scheme has also other important advantages. In particular, the most relevant one is the powerful equalization capability which allows to effectively counteracting frequency selective fading. This property is not relevant in the DVB-RCS2 context and will not be further considered here. Conversely, the main disadvantage of SC-FDMA is the requirement for synchronization between users and the associated need of a Cyclic Prefix (CP) to absorb residual timing errors between users. The CP was introduced in OFDM/OFDMA like schemes to counteract the channel impulse response dispersion thus allowing ISI free operation. In the present context the channel is assumed non-dispersive. However, CP can be used to absorb asynchronicity between transmitting users. If timing error between users is within the CP, a single FFT can be used at the receiver for demultiplexing all different users thus significantly simplifying the MCD hardware implementation. Regarding user timing synchronization accuracy requirements, this is not necessarily more stringent than in current RCS2 systems. In fact, actual acceptable timing errors are depending on the selected system parameters, i.e. FFT size, frequency resolution and CP overhead as will be discussed later in this annex. The basic access scheme proposed here for SC-FDMA is still MF-TDMA with basically the same organization as in conventional DVB-RCS(2) systems. In particular, the MF-TDMA frame can be organized in time frequency slots as in current RCS system, with most of the slots used for carrying traffic and some others for control purposes (i.e. for signalling and synchronization acquisition/ maintenance). C.2 SC-FDMA Principles SC-FDMA differs from straight OFDMA for its use of DFT precoding for reducing envelope fluctuation of the generated signal (Figure C.1). Different strategies for sub-carrier mapping can be devised [i.66]. Two of the possible mapping strategies are indicated in Figure C.2: Localized FDMA (LFDMA) and Interleaved FDMA (IFDMA). In the former strategy each user is allocated a set of contiguous tones. With IFDMA tones allocated to a user are interleaved with those of the other users. For higher flexibility in resource management, the tone mapping strategy could be deferred at the time of actual resource allocation without an a-priori mapping strategy being defined. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 259 Figure C.1: SC-FDMA Schematic Diagram. It is assumed that M < N Figure C.2: Two common sub-carrier mappings in SC-FDMA It should be noted that the LFDMA approach is equivalent to perform a sin(x)/x interpolation of the original time domain symbols. The obtained signal is actually the same as that obtainable by a time domain generation with ideal Nyquist pulse shaping filter (zero roll-off shaping filter). The use of zero roll-off implies some higher signal envelope fluctuation of the generated signal with respect to that of signals with higher roll-off. Such envelope fluctuation are however, still much lower than that experienced in true multicarrier (OFDM) systems. It should also be noted that, with SC-FDMA a non-zero roll-off can also be implemented still operating in the frequency domain. With the same roll-off, TDMA and SC-FDMA have the same PAPR. However, a zero roll-off is most often used in SC-FDMA, for example in [i.64] and [i.65], given its bandwidth advantage. The IFDMA subcarrier mapping produces a repetition of the corresponding time domain symbols apart for a phase rotation applied between successive copies of each input symbol. No envelope signal fluctuations are thus generated (at least in the ideal case). In practice, PAPR advantages of IFDMA mapping over LFDMA are minimal when the need for signal oversampling is considered and LFDMA is actually preferred to IFDMA for the lower sensitivity to synchronization errors. Figure C.3 shows the Total Degradation (TD) experienced by a terminal transmitting multiple carriers (8 carriers per terminal) assuming to use straight OFDMA transmission (i.e. with no DFT precoding) or SC-FDMA (either LFDMA or IFDMA). The advantages of SC-FDMA with respect to OFDMA are evident from the figure. Also the losses with respect to DVB-RCS2 are quite modest (about 0,4 dB). For completeness, Figure C.4 also shows the resulting TD assuming 16QAM modulation. It should be stressed that in case a single tone is assigned to a user, the DFT precoding becomes a no-op and the transmitted signal is actually equivalent to a signal with rectangular pulses. In such case no envelope fluctuation is present and no degradation from non-linearity would be experienced as with CPM modulation. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 260 NOTE: A linearized TWTA was used at the terminal (see annex H of DVB-S2 specifications [i.2]). Figure C.3: QPSK ½ total degradation (TD) curves evaluated at BER ≈ 10-3 (15 users each transmitting 8 tones) NOTE: A linearized TWTA was used at the terminal. Figure C.4: 16QAM ½ total degradation (TD) curves evaluated at BER ≈ 10-3 (15 users each transmitting 8 tones) C.2a Demodulator Processing Assuming that user terminals have already achieved a coarse timing (and frequency) synchronization before transmitting traffic bursts. As for conventional RCS terminals it is assumed that receiver synchronization is first achieved before transmitting the first burst on a random access slot. Before going into steady state, terminals have to achieve a time synchronization accuracy such that the residual error is within the adopted CP (as briefly discussed in clause C.5). Assuming that terminals are already in steady-state synchronization and transmit traffic burst and control burst for synchronization /signalling purposes (which refered to here as SYNC slot in analogy with first generation RCS terminology) on the NCC assigned resources. The demodulator processing, briefly illustrated below, can be used for both Traffic and Control bursts. 0.00 0.50 1.00 1.50 2.00 2.50 3.00 0 0.5 1 1.5 2 2.5 3 TD (dB) OBO (dB) OFDMA TD LFDMA TD IFDMA TD DVBRCS2 TD 0.00E+00 1.00E+00 2.00E+00 3.00E+00 4.00E+00 5.00E+00 6.00E+00 7.00E+00 8.00E+00 9.00E+00 1.00E+01 0 1 2 3 4 5 6 7 TD (dB) OBO (dB) OFDMA TD LFDMA TD IFDMA TD ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 261 The following definitions are used for "modulation symbols", "OFDM symbols" and "tones": • modulation symbols, the sequence of QPSK or 8PSK or 16QAM symbols as per a conventional TDMA system. • OFDM symbols: indicate the signal obtained after DFT precoding of the "modulation symbols. Hence, an OFDM symbol will include from a minimum of 1 "modulation symbol" to a number of "modulation symbols which is only limited by the system bandwidth (and frequency granularity of SC-FDMA). • Tones: In order to avoid confusion between "OFDM symbols" and "modulation symbols", the "modulation symbols" will be referred to as tones given the fact that with DFT precoding a number M of "modulation symbols" is transformed into an equal number of tones in the frequency domain (although there is no one-to one correspondence between a modulation symbol and a tone). A burst is assumed composed of a sequence of encoded payload OFDM symbols intermixed with pilot OFDM symbols. Pilot OFDM symbols are added with a certain constant periodicity. The burst is assumed starting with a pilot OFDM symbol. The final OFDM symbol of a burst is also assumed to be a pilot OFDM symbol. This last pilot OFDM symbol may be at a lower distance from the previous pilot OFDM symbol, due to the fact that the encoded payload length in OFDM symbols may not be a multiple of the chosen pilot OFDM symbol repetition period. The pilot OFDM symbol repetition period, referes to the number of payload OFDM symbols comprised between two pilot OFDM symbols (the actual number may be lower when considering the last pilot OFDM symbol of a burst). Burst organization is pictorially shown in Figure C.5. NOTE: An OFDM symbol is here composed of either all pilot tones (green one) or payload symbols. Figure C.5: Organization of a SC-FDMA burst after DFT precoding. Green Box indicates pilot tones The reason to have a given OFDM symbol composed of either all pilot tones or all information symbols is due to the fact it is desirable to have constant amplitude pilot tones in the frequency domain (hence, not only in the time domain) for reasons connected to channel estimation. Channel estimation is, in fact, easier to perform in the frequency domain. At this regard time error estimation requires multiple pilot tones to be present in a single OFDM symbol. The overall sequence of operation of the Demodulator is shown in Figure C.6. The demodulator, in addition to demodulated data, also provides measurement which may be fed-back to transmitter for adjusting their parameters (time, frequency, power level, possible MODCOD). Frequency ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 262 Figure C.6: Schematic Diagram of the demodulator It is generally convenient to have constant tone amplitude in both the frequency domain and time domain. This has led to the widespread usage of the Zadoff-Chu sequences (also known as a CAZAC sequence - constant amplitude zero autocorrelation waveform) as pilot symbols in SC-FDMA. The use of such sequences is assumed here for pilot symbols although in applications where the channel is not frequency selective, having constant amplitude of pilot tones in the frequency domain is not strictly required. Concentrating pilot energy at the edge of the band may be, in fact, a better strategy as it would allow higher accuracy in timing error measurements. Also, to reduce the pilot overhead, mixed pilot and traffic symbols are often used. This leads to higher envelope fluctuation on those mixed symbols. The associated performance penalty is however acceptable given that such higher envelope fluctuation is concentrated in a small percentage of the symbols. Moreover, some solutions for controlling the PAPR of the mixed symbol when employing SC-FDMA already exist (e.g. pilot design of the DVB-NGH satellite component [i.65]). Performance results presented here do not assume the use of mixed pilot symbols and make use of conventional Zadoff-Chu sequences for such pilot symbols. Looking at the scheme in Figure C.6 it appears that after CP elimination and transformation of the signal in the frequency domain (through the N-point FFT), tone demux can be performed in order to separate tones according to their users. Once tones are separated, narrowband processing per user can take place. All subsequent demodulation algorithms can be performed in the frequency domain, coming back to the time domain (with the M-point IFFT) only for final decoding of the data. The symbol timing recovery can be carried out in the frequency domain thanks to the duality of time and frequency with respect to Fourier Transform (a time shift in the time domain is equivalent to a linear phase ramp in the frequency domain). Hence, estimation of timing error in the frequency domain can actually be performed with conventional algorithm for time-domain frequency estimation. For the actual timing error estimator the Mengali-Morelli frequency estimator [i.67] has been used to get the performance results which will be presented in clause C.3. Once the timing error has been estimated, timing correction is performed by rotating the phase of each tone according to the corresponding frequency-domain caused linear phase drift per tone within each OFDM symbol. Frequency Offset Recovery also operates on the pilot tones. However, processing is now done between pilot tones of the same frequency index but on different pilot OFDM symbols. After frequency error estimation, frequency correction takes place by rotating the phase of tones in subsequent OFDM symbols. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 263 It should be noted that an implicit assumption here is that the frequency error is always less than a frequency bin. In this case there is no need of coarse frequency estimation as the frequency error is always less than a frequency bin. This is certainly a reasonable assumption for terminals already in steady-state synchronization status. C.3 Performance Results Some simulation performance results are reported here for both QPSK r = ½ and 8PSK r = 2/3 with a Turbo Coding scheme in line with the RCS2 normative document [i.1]. Regarding SC-FDMA simulation parameters, a tone granularity of 128 KHz is assumed. Two adjacent users were simulated each with 8 tones (for a total baud rate per user of 1,024 MBaud) out of a total number of tones (size of the wideband FFT) equal to 128. A 128 point FFT was thus used at the receiver side to demultiplex the various users which are then subject to independent demodulation. Figure C.7 shows the performance in linear channel of QPSK 1/2. Payload size of the bursts was 472 info bits according to the waveform ID=4 of annex A of ETSI EN 301 545-2 [i.1]. One of the SC-FDMA users had pilot periodicity equal to 12 OFDM symbols (corresponding to a total of 7 OFDM symbols used for pilots within the burst) and the other had pilot repetition equal to 15 OFDM symbols (corresponding to a total of 5 OFDM symbols used for pilots within the burst). Frequency error was 2,56 KHz for both users. Phase noise according to DVB-RCS guideline [i.47] was added. Timing error for both users was different from zero but lower than the Cyclic Prefix (CP). The corresponding reference performances for the DVB-RCS2 waveform are shown in Figure C.8. The performance loss of the SC-FDMA case (with pilot periodicity equal to 12 OFDM symbols) is about 0,3 dB. Performance of the SC-FDMA system in presence of a non-linearity (modelled according to the Rapp model [i.38]) is shown in Figures C.9 and C.10 for pilot periodicity of 15 OFDM symbols and 12 OFDM symbols, respectively. The AM/AM of the non-linearity is shown in Figure C.11. The loss with respect to the standard ID=4 waveform of DVB-RCS2 is only 0,35 dB for the case of pilot periodicity 12 at IBO = 0 dB. To such value about 0,1 dB should be also added for the difference in OBO due to the use of a zero roll-off factor in SC-FDMA. NOTE: Payload size 472 info bits. Figure C.7: Performance in linear channel of SC-FDMA QPSK 1/2 1.E-06 1.E-05 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 0 0.5 1 1.5 2 2.5 3 3.5 4 Es/No (dB) FER Pilot Period 12 Pilot Period 15 QPSK 1/2 k=472 8 tones per user Tone Spacing 128 Kbaud Linear Channel Frequency Error: 2.5 KHz Phase Noise ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 264 NOTE: Payload size 472 info bits. Figure C.8: Reference performance of DVB-RCS2 (Waveform ID=4) in linear channel with and without phase noise NOTE: Payload size 472 info bits. Figure C.9: Performance of SC-FDMA with QPSK ½ (pilot period 15) in a nonlinear channel with different IBO 1.E-05 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 0 0.5 1 1.5 2 2.5 3 3.5 Es/N0 (dB) FER 0 dB / 0.78 dB 1 dB / 1.05 dB 2 dB / 1.64 dB 3 dB / 2.36 dB IBO / OBO QPSK 1/2 K=472 Norm. Freq. Error = 0.02 Phase Noise Pilot Period 15 8 tones per user 128 KHz bin-width granularity ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 265 NOTE: A RAPP model HPA was considered. Payload size 472 info bits. Figure C.10: Performance of SC-FDMA with QPSK ½ (pilot period 12) in a non-linear channel with different IBO As a second study case, a payload of 920 info bit with 8 PSK modulation and r = 2/3 was considered (Waveform ID=8 of annex A in ETSI EN 301 545-2 [i.1]). Performances with different tone periodicity are shown in Figures C.11 and C.12 respectively for a linear and non-linear channel. Two adjacent users were simulated each with 8 tones (tone granularity 128 KHz). Frequency error was about 2,56 KHz for both users. Phase noise according to DVB-RCS guideline, [i.47], was added. Using the same simulation models, the performance of conventional waveform ID=8 of [i.1] has been obtained as shown in Figures C.13 and C.14 respectively for the linear and non-linear case. For the non-linear channel (with IBO = 0 dB) SC-FDMA (with pilot periodicity equal to 9 OFDM symbols) has a loss of about 0,3 dB (0,4 dB when the OBO difference is also accounted) with respect to the reference DVB-RCS2 case. The small difference in performance between SC-FDMA and DVB-RCS2 is mostly due to the frequency error impact. For smaller frequency error the performance difference is smaller approaching the difference in OBO. The impact of the frequency offset can be reduced by using mixed data and pilot symbols, which would allow having more frequent pilots in the time domain without increasing the overall pilot overhead. The small performance loss of SC-FDMA is anyway compensated by the higher spectral efficiency achievable by SC-FDMA. 1.E-06 1.E-05 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 0 0.5 1 1.5 2 2.5 3 3.5 Es/N0 (dB) FER 0 dB / 0.78 dB 1 dB / 1.05 dB 2 dB / 1.64 dB 3 dB / 2.36 dB QPSK 1/2 K=472 Norm. Freq. Error = 0.02 Phase Noise Pilot Period 12 8 tones per user 128 KHz bin-width granularity IBO / OBO ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 266 NOTE: Payload size 920 info bits. Figure C.11: Performance in linear channel of SC-FDMA 8PSK 2/3 NOTE: Payload size 920 info bits. Figure C.12: Performance in non-linear channel of SC-FDMA 8PSK 2/3 with IBO = 0 dB (OBO = 0,78 dB) 1.E-06 1.E-05 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 6 6.5 7 7.5 8 8.5 9 9.5 Es/No (dB) FER Pilot Period 12 Pilot Period 15 Pilot Period 9 Pilot Period 10 8 PSK 2/3k=920 Norm. Freq. Error = 0.02 Phase Noise Linear Channel 1.E-05 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 7 7.5 8 8.5 9 9.5 10 Es/No (dB) FER Pilot Period 9 PilotPeriod 10 Pilot Period 12 Pilot Period 15 IBO =0 dB 8 PSK 2/3 k=920 Norm. Freq. Error = 0.02 Phase Noise ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 267 NOTE: Payload size 920 info bits. Waveform ID=8. Figure C.13: Performance in linear channel of standard DVB-RCS2 8PSK 2/3 (with and without phase noise) NOTE: Payload size 920 info bits. Waveform ID=8. Figure C.14: Performance in non-linear channel of DVB-RCS2 8PSK 2/3 at IBO = 0 dB Tables C.1 and C.2 show spectral efficiencies respectively achieved in DVB-RCS2 with waveform ID=4 and 8 and SC-FDMA with QPSK ½ (pilot periodicity 12 OFDM symbols) and 8PSK 2/3 (pilot periodicity 9 OFDM symbols) for two different CP lengths. 1.E-06 1.E-05 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 7 7.5 8 8.5 9 Es/No (dB) FER IBO=0 dB OBO=0.71 dB 8PSK 2/3, k=920 Freq. Error =1.5 KHz Phase Noise ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 268 It can be observed that for both waveform examples, SC-FDMA provides a spectral efficiency improvement with respect to RCS2 with a small performance loss (less than 0,4 dB at FER = 10-4). It should however be noted that the current simulation results are based on pilot distribution which was not optimal as the last OFDM pilot symbols was not spaced as the other symbols. The spectral efficiency of SC-FDMA could be likely further increased by considering mixed data and pilot symbols and better payload length. In DVB-NGH for example, for the SC-FDMA satellite component, mixed data and pilot symbols with controlled low PAPR are used [i.65]. Table C.1: Overall Spectral efficiencies of two DVB-RCS2 standard waveforms DVB-RCS2 Waveform Useful Bits Payload Symbols Preamble / pilots Guard symbols Roll-Off Spectral Efficiency bit/s/Hz ID=4 (QPSK ½) 472 472 64 6 0,2 0,738 ID=8 (8PSK 2/3 920 460 76 6 0,2 1,447 Table C.2: Overall Spectral efficiencies of two SC-FDMA mode roughly having the same power efficiencies as the corresponding DVB-RCS2 ones of previous table. SC-FDMA Waveform Useful Bits OFDM Payload Symbols OFDM Pilot Symbols Spectral Efficiency (bit/s/Hz) CP=1/8 CP=1/16 QPSK ½ 472 59 7 0,795 0,841 8PSK 2/3 920 57,5 9 1,57 1,67 C.4 Implementation Aspects The SC-FDMA scheme has other attractive properties as mentioned in the introduction. With regard to Multi-Carrier Demodulator implementation, SC-FDMA is advantageous compared to conventional techniques when implementing MCD capability at the terminal as demultiplexing of different user signals is straightforward via FFT processing. As an example, a single 512-points FFT processor at the mesh receiver demodulator would be able to extract any signal within a 36 MHz transponder with a frequency tone resolution of around 80 KHz allowing compatibility with any allocated user bandwidth which is an integer multiple of 80 kHz. The 512-point FFT would require 3 204 real multiplication and 12 420 real additions should a radix-8 algorithm be selected. This is equivalent to about 6 real multiplications and 25 real additions per input complex sample. Should a real sampling be used, a 1 024 points FFT operating on a signal sampled at about 80 MHz is required. However, such 1 024 FFT can still be computed via a complex 512-points FFT provided that an additional stage performing the real-to-complex conversion is added to the FFT processor (Figure C.15 - see dashed block in the figure). The last stage for real to complex conversion will require 1 018 real multiplications and 2 560 real additions. Even with this addition, there are about 8 multiplication and 30 additions per complex sample. At 40 MHz complex sampling rate these would amount to 320 Million Multiplies/s and 1,2 Billion additions/s. Figure C.15: A 512-point complex FFT with radix-8 decomposition Should the same capability be implemented in a conventional system, it would have required to implement for example a fast convolution approach (Figure C.16) with a fixed size FFT (in the order of at least 8K points to minimize guard band between channels) followed by spectral modification and a number of smaller size IFFT (one per actual carrier to be separated). Overlapped processing of the windows is also required. The Direct FFT is common for all signals to be demultiplexed. A separate IFFT (and inverse windowing if windowing is used) should be done for each carrier. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 269 Figure C.16: Demultiplexer implemented via Fast_Convolution The direct FFT, if implemented with a radix 2 algorithm, requires about 200 000 real multiplications or 24 real multipliers per input sample. Assuming an overlapping factor between FFTs of 25 %, the direct FFT would become equivalent to about 30 real multipliers /sample. To that number, the complexity of the IFFTs has to be added. The actual impact of IFFTs would depend on the actual usage of the channel. Assuming IFFTs complexity is equal to that of the FFT the amount of calculations would end up with 60 real multiplications/sample just for the FFT/IFFTs, i.e. 10 times more than with a 512 point FFT. The complexity ratio of the two demultiplexing approaches is at least of a factor 10 in terms of number of multiplications in favour of the SC-FDMA approach. Hence, SC-FDMA would actually provide a cost saving in the modem at least when multi-carrier reception capability is desired. Another potential advantage of SC-FDMA is its flexibility in resource assignment. In conventional MF-TDMA, carrier bandwidths have to be defined in advance according to the foreseen needs of the system and may only be reconfigured occasionally. So a few carrier bandwidths are typically available and this would make interference and fading mitigation techniques like Data Rate Adaptation (DRA) not very effective. In SC-FDMA, instead, resources can be dynamically assigned in quantum of both time and frequency. DRA would become a quite natural way to overcome fading in such a case. Recalling at this regard that DRA, when feasible, is more efficient than ACM as fading countermeasure. C.5 Synchronization Requirements In order to exploit the SC-FDMA technology it is needed that user terminals are mutually time synchronized with a maximum error less than the CP width. The absolute time accuracy which is required with SC-FDMA is thus dependent on the CP duration. For a given CP relative overhead, the absolute time duration of the CP can be made larger by increasing the duration of the OFDM symbol. For a given bandwidth, the duration of the OFDM symbols is increased by increasing the FFT size (hence reducing the frequency bin width). As examples, for a 1/8 CP size and about 41 MHz total gross bandwidth spanned by the FFT, Table C.3 shows the duration of the CP as function of the FFT size. Table C.3: Relationship between CP duration and FFT parameters FFT Size Total bandwidth (MHz) Frequency Bin Width (kHz) OFDM Symbol duration (μs) CP rel. Overhead CP Duration (μs) 128 40,96 320 3,125 1/8 0,39063 256 40,96 160 6,25 1/8 0,78125 512 40,96 80 12,5 1/8 1,5625 1 024 40,96 40 25 1/8 3,125 NOTE: Duration of the OFDM symbol is without CP. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 270 It appears that timing accuracy requirements are similar or slightly more stringent than those of required in DVB-RCS2 system. It should be noted that longer FFTs might also require a corresponding increase of the frequency accuracy to reduce degradation due to user frequency errors. Synchronization requirements of SC-FDMA are thus not necessarily more stringent than those of conventional MF-TDMA systems. The same frame organization principles as in current DVB-RCS may be applied to SC-FDMA. Regarding random access bursts like logon burst which is typically transmitted before having the opportunity of achieving strict TX synchronization, several alternatives can be considered. In the preferred alternative, such slots are segregated in a reserved temporal section of the frame in order that their lacks of synchronization do not produce interference on frequency adjacent users. Random access bursts transmitted in such slots may then use a different waveform than SC-FDMA. These bursts can be used for initial logon and for acquiring steady-state synchronization. C.6 Compatibility with normative specifications The Current RCS2 Lower Layer normative document does not foresee the use of SC-FDMA. Transmission roll-off is fixed to 20 % for RCS2 Linear Modulation against the 0 % roll-off which should be preferred for SC-FDMA. Also, CP extent needs to be defined for SC-FDMA. Current signalling for the Data Block for the TC-LM Transmission Format Class (see Table 6.18 of ETSI EN 301 545-2 [i.1]) is however sufficient for describing SC-FDMA bursts. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 271 Annex D: Time Slot Sharing D.1 Introduction Time Slot Sharing (TSS) is an effective solution for increasing the bandwidth efficiency of mesh connections over transparent satellites. The technique relies on the fact that two peer terminals, if served by the same beam, may reuse the same time slot as each peer is able to cancel its own signal from the received signal thus allowing it to demodulate the unknown signals coming from the other terminal. D.2 Applicability The following advantages of TSS can be highlighted: • TSS is compatible with different transmit waveforms (independent of the modulation and coding). Both linear modulations and CPM are compatible with TSS. Different MODCODs could also be used by the terminals sharing the same time/frequency slot. Similarly ACM operations can be also supported. • TSS can be used in both bandwidth limited and power limited situations. Whilst its advantages are obvious in a bandwidth limited situation, it may have also advantage in a power limited situation as the saved bandwidth provided by this approach may be exploited for reducing the spectral efficiency of transmitted carriers which implies an increase of the achievable power efficiency. As a matter of fact, if the available bandwidth doubles with such a technique one may, for example, use QPSK ½ instead of 8PSK 2/3. Doing so the same spectral efficiency as in the original system is achieved but there is a gain in power as QPSK 1/2 requires an Eb/N0 about 2,5 dB lower than 8PSK 2/3 for a codeword size of about 500 bits. The following constraint should be considered for the technique applicability: • TSS can only be used when the transmitting earth station is also able to receive its own transmitted signal. That generally precludes its general adoption in multi-beam systems apart for the cases where the two peer entities happen to be in the same beam. • Although TSS can also be employed in star configurations, it is best suited for symmetric links such as transparent mesh profile where peer-to-peer links can be established using balanced carriers (in terms of power and bandwidth). • The use of TSS could induce constraints in the resource assignment strategy. Since it operates on peer-to-peer connections, it is necessary that the same resources (in terms of MF-TDMA slots) be assigned to both parties involved in a peer-to-peer connection. In addition to more complex (and perhaps less efficient) resource management, the traffic should be balanced. Such constraint will somewhat reduce the actual gain of the technique with respect to the theoretical maximum. • TSS is sensitive to channel non-linearity. For linear modulations, terminal non-linearity effects could be mitigated by using the SSPA output as the reference signal. The use of constant-envelope CPM signals is effective since the CPM waveform is insensitive to terminal non-linearity. Satellite non-linearity, given the multicarrier satellite operation, cannot be however fully compensated through a linear canceller. Maximum theoretical gain is a factor of two in case of bandwidth limited systems, assuming no limiting constraint due to the increase of power spectral density at the satellite input. It should be noted that TSS is an analogue form of a network coding scheme similar to those proposed in literature to reduce bandwidth requirements in regenerative satellite systems. TSS can be extended to multicast multi-party conference (see Figure D.1), although the theoretical gain rapidly decreases with number of parties in the conference as in general (N-1) slots are required for an N-party conference. The resulting theoretical gain is thus N/(N-1) which is maximum with just two users. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 272 Figure D.1: TSS in Multi-Party Conference D.3 TSS Echo Cancellation The TSS technique relies on a peer terminal to cancel its own (round trip delayed) signal from the received signal. As shown in Figure D.2, an "echo" canceller is required at the receiver side to remove the self-interfering signal from the received signal. After Echo Cancellation a standard RCS2 demodulator can be used. Figure D.2: TSS cancellation technique The filter in the echo canceller may be a Digital Anti-Aliasing Filter (DAAF) to reduce noise outside the signal bandwidth. For linear modulation the DAAF may actually implement the shaping filter moving this function outside the standard DVB-RCS2 receiver. The filtering, echo parameter estimation and echo reconstruction can be actually performed via a sequence of functions as identified in Figure D.3. These functions are independent of the modulation of the echo as the echo signal is fully known apart for parameters like frequency, phase and timing. The only assumption here is that the known samples are provided with the same sampling rate as the actual sampling rate of the incoming signal to be processed. Such sampling rate it is assumed to be at least three times larger than actual signal bandwidth. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 273 The minimum sampling frequency needs to satisfy Nyquist sampling theorem. However, a somewhat larger sampling frequency may be desirable with Farrow interpolation and to simplify fine timing recovery. Figure D.3: Echo parameter estimation and reconstruction The first active process for the echo estimation and reconstruction is the burst synchronizer. This block performs simultaneously both the estimation of the arrival time of the echo and the coarse estimation of the echo frequency. Given the unknown frequency of the echo, the matched filter needs to be split in smaller sections which are coherently combined according to different frequency error hypothesis (Figure D.4). The largest result for correlation in a time window corresponding to the time uncertainty range of the echo arrival is then selected as the best coarse timing and frequency hypotheses. A fractional delay can also be estimated with respect to the echo Start of Burst sample by parabolic interpolation between the best correlation value achieved by the burst Synchronizer and the two adjacent correlation values. A Farrow Interpolator is then used to compensate for such fractional delay. The input sequence of samples is then multiplied by the complex conjugate of the known echo interpolated samples to remove phase modulation thus obtaining a signal usable for fine frequency, phase and amplitude estimation. The obtained frequency correction values and amplitude / phase values can then be also applied to the reference signal in order to have the echo replica ready for cancellation from the original input signal stream. Figure D.4: Burst Synchronizer for coarse timing and frequency estimations In Burst Synchronizer Fine Frequency Estimator (Mengali-Morelli) Fine frequency correction Phase / Amplitude Recovery ffine Phase In Input Signal Known samples Shaping Filter Farrow Interpolator Coarse frequency correction tcoarse fcoarse Modulation Removal Known Samples Frequency, Phase, Amplitude Correction Signal to be used for cancellation ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 274 D.4 Simulation results Some simulation results with echo cancellation are reported below for two possible DVB-RCS2 LM waveforms. In particular, RCS2 Waveform ID=4 (QPSK ½) and waveform ID=8 (8PSK 2/3) were considered respectively for the QPSK and 8 PSK cases. Such waveforms are representative of short bursts in DVB-RCS2 and should be thus more critical as far as echo cancellation is concerned (with respect to longer bursts). Figure D.5 shows results for QPSK1/2 (ID=4) with and without TSS both in linear and non-linear channel. For the non-linear channel a Rapp-Model with parameter s=6 was used for representing the terminal SSPA (satellite HPA is assumed to operate linearly). The Rapp model only considers the AM/AM non-linearity as it assumes that the SSPA AM/PM is negligible. The amplifier gain is modelled as: The assumed signal baud rate was 128 kBaud and phase noise (if included) was compliant with guidelines of DVB-RCS. A frequency error of 2 % of the baud rate was also considered. As apparent from curves in Figure D.5, the loss caused by TSS in case of phase noise is limited to 0,2 dB only. The optimal SSPA IBO was 0 dB (at least for the selected Rapp model). For that IBO, the loss with respect to the linear case was 0,15 dB with TSS (against 0,1 dB in absence of TSS). Overall, the loss due to TSS in the worst case as far as phase noise, non-linearity and frequency error is only 0,2 dB (FER = 10-4) with respect to the conventional case without TSS. Figure D.5: Performance of DVB-RCS-2 QPSK ½ (Waveform ID=4) with and without TSS and different channel conditions Figure D.6 shows the effect of carrier unbalance in TSS. This is justified by the fact that lower power echo, although cancelled less accurately, is clearly less disturbing. Conversely, a higher power echo is easier to cancel more accurately. Figure D.7 shows results with 8PSK 2/3 (Waveform ID=8) under different channel conditions. In presence of non-linearity (with IBO = 0 dB) TSS has about 0,65 dB penalization with respect to the conventional case. Even with perfect echo estimation, a residual interference would be present due to the non-compensated HPA distortion. This effect amounts to 0,2 dB (as can be inferred by comparison with and without TSS in a linear channel as shown in Figure D.8 and in non-linear channel as shown in Figure D.9). That 0,2 dB loss could be recovered if information about the non-linearity is available and exploited to reconstruct the echo. 1.E-05 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 1 1.5 2 2.5 3 FER ES/N0 (dB) FER FER_PN FER_IBO(0) FER_IBO(0)_PN FER_REF_TSS FER_PN_TSS FER_IBO(0)_TSS FER_IBO(0)_PN_TSS WF = "4" QPSK 1/2 TurboPHI (472 bits, 8 it, MAX) ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 275 Figure D.6: Performance of Echo cancellation under carrier unbalance Figure D.7: Performance of waveform ID=8 (8PSK 2/3, short burst) under different channel conditions with and without TSS 1.E-05 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 7 7.5 8 8.5 9 9.5 FER ES/N0 (dB) FER FER_PN FER_IBO(0) FER_IBO(0)_PN FER_REF_TSS FER_PN_TSS FER_IBO(0)_TSS FER_IBO(0)_PN_TSS WF = "8" 8PSK 2/3 TurboPHI (920 bits, 8 it, MAX) ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 276 NOTE: Losses due to TSS is about 0,4 dB. Figure D.8: Comparison of performance in linear channel with and without TSS Waveform ID=8 (8PSK 2/3, short burst), 128 kBaud, Phase Noise as in DVB-RCS guidelines NOTE: Waveform ID=8 (8PSK 2/3, short burst), 128 kBaud, Phase Noise as in DVB-RCS guidelines [i.47]. Losses due to TSS is about 0,65 dB. Figure D.9: Comparison of performance in non-linear channel (IBO = 0 dB) with and without TSS D.5 TSS Implementation Support for Time Slot Sharing (TSS) can be included as an option within the signalling for dynamic mesh links control. Assuming that mesh signalling is supported by DCP, terminals can notify of TSS support with the DCP logon request message sent after logon. It should be noted that the name of "RCST capability request" (RCSTCapReq) is taken from current ETSI C2P specifications (ETSI TS 102 602 [i.75], "Connection Control Protocol for DVB-RCS"), although the message is clearly a notification. Anyway terminology in ETSI C2P is to name as "request" the originating messages and as "response" the related answer. Hence the corresponding message send by NCC in response to the RCST capability request" is "the RCST capability response". In this message, new IEs should be included for communicating various capabilities relevant to mesh networking (e.g. number of mesh receivers available in the RCST) as well as the support for TSS. The new IEs can be added in the range of User Defined IEs. The DCP messages allow the introduction of additional IEs. 1.E-05 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 7 7.5 8 8.5 9 9.5 FER ES/N0 (dB) FER_PN FER_PN_TSS WF = "8" 8PSK 2/3 TurboPHI (920 bits, 8 it, MAX) 1.E-05 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 7 7.5 8 8.5 9 9.5 FER ES/N0 (dB) FER_IBO(0)_PN FER_IBO(0)_PN_TSS WF = "8" 8PSK 2/3 TurboPHI (920 bits, 8 it, MAX) ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 277 After the NCC has been informed about "Time Slot Sharing" support, the usage of it may be managed either statically or dynamically: • for static management, it is sufficient the NCC knows if an RCST is "Time Slot Sharing capable": for every connection requested, if the peer RCST is also "Time Slot Sharing capable", the connection is assigned by NCC as "Time Slot Sharing type"; • for dynamic management, an RCST "Time Slot Sharing capable", after signalling to NCC, is required to request for each connection if it wants to use "Time Slot Sharing"; In the last case, at mesh link service establishment, an RCST should explicitly signals if it wants to use "Time Slot Sharing" for the new connection request, adding the new "Time Slot Sharing" IE to the Link Service Establishment Request message; the peer RCST should be informed accordingly by the NCC. Also for this case, the NCC may assign burst by burst resources in "Time Slot Sharing" mode in an implicit way through the TBTP2 (as the same slot may be assigned to more than one terminal). This clearly requires that each terminal detect this condition by comparing the list of slots to receive (i.e. those assigned to its peer for transmission) with those in which he has transmitted. The "Time Slot Sharing" IE should also include additional fields to allow one of different policies (and their parameter, if any) to be specified. An example of "Time Slot Sharing" connection establishment for dynamic management case is described below: • after the logon, an RCST sends DCP logon request message with "Time Slot Sharing" IE (for specific details on "Time Slot Sharing" usage); • an initiating RCST sends Link Service Establishment Request to NCC with "Time Slot Sharing" IE (if it wants to use now "Time Slot Sharing"); • the NCC verify if the peer is "Time Slot Sharing" capable: if so, sends Link Service Establishment Request to the peer with "Time Slot Sharing" IE; • after that, the NCC sends Link Service Establishment Response to the originating RCST with "Time Slot Sharing" IE; • the broadcasted TBTP2s should assign the same slot to both parties. Slots to be used in TSS should be associated to a specific resource identifier (which may be different for each the terminal and is assigned by the NCC at the mesh link service establishment request as an assignment_id similar to those assigned at logon by the NCC at every terminal) which has been exclusively allocated for that peer-to-peer connection. The "Time Slot Sharing" technique may be extended to allow a partial channel sharing also in multicast multipart video-conferences with 3 or 4 participants max (only for multicast contributions) if all participants are "Multicast Time Slot Sharing capable". The bandwidth gain will reduce with increasing participant number, so for more than 4 participants (25 % bandwidth gain) it is no more practical continue to use this technique. A multiparty video-conference is typically established by other higher level (usually at application level) signalling which is also in charge of managing the communication of all relevant parameters (video/audio format, speed, etc.) and the multicast addresses of all participants. In case of "Multicast Time Slot Sharing" with more than 2 participants, the wanted signal is recovered with iterative multiple cancellations at physical layer. Finally TSS also requires that signals transmitted in the same time slot are uncorrelated for correct cancellation. Different preambles and scrambling sequences should be used by each of the two peer terminals to uncorrelate the peer transmitted bursts. At present, this is not supported by the RCS2 lower layer specs as preamble are defined per time slot; also scrambling is fixed for all users and is not applied on the preamble. In order to not change current specs, one may assume that the "Time Slot Sharing" IE sent by the NCC to terminals also contain a flag which may change the semantic of some physical layer spec. In particular, such flag, if active, could change the order with which preamble symbols are transmitted (i.e. one of the peer would transmit a time mirrored preamble instead of the normal one) and the initialization word of the burst scrambler. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 278 Annex E: Forward Link Spectrum Spreading E.1 Introduction ETSI EN 301 545-2 [i.1] (see clause 5.4.5) introduces the optional use of spread spectrum techniques for the forward link. The use of spread spectrum is particularly useful for applications such as interactive services for mobile RCSTs with small antenna aperture as well as extending the operating ranges in the presence of severe atmospheric fading conditions. Two alternative solutions for the forward link spectrum spreading are introduced, namely Direct Sequence Spectrum Spreading (DSSS) and Frame Repetition Spectrum Spreading (FRSS). These techniques have their strengths and weaknesses that make them more suitable for different applications as summarized in Table E.1. Table E.1: Two Forward Link Spread Spectrum Techniques Spectrum Spreading Technique Strong Points Weak points Direct Sequence Spectrum Spreading • Support of very lower SNR • Robust against carrier instability • Not compatible with DVB-S2 waveform. Therefore, cannot share the same carrier with a conventional DVB-S2 service. Frame Repetition Spectrum Spreading • Coexistence with non-spread DVB-S2 waveform [i.2] (backward compatibility) • Impact of carrier instability E.2 Technical Description E.2.1 Direct Sequence Spectrum Spreading A DVB-S2 forward link transmission based on symbol repetition can be spread in bandwidth using the provisions in this clause. Such spreading is applied in two stages: spreading and scrambling. The first operation, spreading, multiplies every (I+jQ) PL symbol by a sequence of chips to enlarge the bandwidth of the signal (the PL frame duration in absolute time remains the same). The number of chips per symbol is called the Spreading Factor (SF). The waveform bandwidth expansion is proportional for SF. When SF=1, the transmission is equivalent to a conventional DVB-S2. The second operation, scrambling, applies a scrambling code to the spread signal. The processing is illustrated in Figure E.1. Figure E.1: Direct Sequence Spreading Technique PLFRAME SPREAD PLFRAME SCRAMBLING SEQUENCE Scrambling RESET Scrambled PLFRAME NPLFRAME symbols NPLFRAME *SF spread symbols NPLFRAME *SF scrambled sequence NPLFRAME *SF scrambled symbols ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 279 The spectrum spreading should be applied on a PLFRAME basis, following the conventional DVB-S2 physical layer scrambling process. Each symbol in a PLFRAME, including the PLHEADER and pilot symbols, if used, should be spread by a repetition of a real-valued spreading code C (i). The output of the spreading for each symbol on the I and Q branches should thus be a sequence of SF chips corresponding to the spreading code chip sequence, multiplied by the corresponding, real-valued symbol component value. The spreading code sequence should be time-aligned with the symbol boundary. If {d[k]}, k = 0, 1,…, NPLFRAME-1, represents the (I+jQ) symbols of the PLFRAME, where NPLFRAME is the number of symbols in one PLFRAME, then the spreading operation yields the spread sequence s(i): ( ) ( )) , mod( / ) ( SF i C SF i d i s = for i = 0, 1,..., (NPLFRAME × SF)-1 Spreading codes C(i) are defined for spreading factors of 1, 2, 3 and 4 and are signalled in the extended Satellite Forward Link Descriptor in clause 6.4.17.6 (Table 6-50 of ETSI EN 301 545-2 [i.1]). The extended signalling description is given in Table E.2. In terms of the reference modulator signal flow defined in DVB-S2 [i.2], the spreading should be performed immediately prior to the physical layer scrambling. The second operation, scrambling, is achieved through the use of the same method as that defined for physical layer scrambling in clause 5.5.4 of [i.2], except that: • the length of the scrambling sequence is here equal to NPLFRAME × SF, rather than NPLFRAME; and • the scrambling sequence is applied to the entire spread PLFRAME, including the PLHEADER and pilots if used. The scrambling sequence should be aligned with the PLFRAME epoch, and it should be re-initialized at the beginning of each PLFRAME. The sequence of complex valued chips is scrambled (complex chip-wise multiplication) by the complex-valued scrambling code, w(i), defined in clause 5.5.4 of [i.2], when SF is greater than 1. The Spread PLFRAME duration depends on the selected modulation and the adopted spreading factor. The scrambled symbols, z(i), is obtained by directly multiplying the spread symbols, s(i), by the scrambling sequence, w(i), as follows: z(i) = s(i) × w(i modulo 66420), i=0,1,2,…,(NPLFRAME × SF)-1 After scrambling, the signal {z(i)} is filtered using square root raised cosine filter with a pre-selected roll-off factor as described in clause 5.6 of [i.2]. It is necessary to define an explicit scrambling sequence in the corresponding satellite forward link descriptor (clause 6.4.17.6 of ETSI EN 301 545-2 [i.1]) when SF is greater than 1. Table E.2: Syntax of the Satellite Forward Link descriptor for Direct Sequence Spread Syntax No. of bits Information Mnemonic Reserved Information Satellite_forward_link_descriptor() { descriptor_tag 8 uimsbf descriptor_length 8 uimsbf satellite_ID 8 uimsbf beam_ID 16 uimsbf NCC_ID 8 uimsbf multiplex_usage 3 bslbf local_multiplex_ID 5 uimsbf frequency 32 uimsbf orbital_position 16 bslbf west_east_flag 1 bslbf Polarization 2 bslbf transmission_standard 2 uimsbf if (transmission_standard == 0) { "001" 3 bslbf } else if ((transmission_standard == 1) or (transmission_standard == 2)) { ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 280 Syntax No. of bits Information Mnemonic Reserved Information scrambling sequence selector 1 bslbf roll_off 2 uimsbf } symbol_rate 24 uimsbf if (transmission_standard == 0){ FEC_inner 4 bslbf Reserved 4 bslbf } else if ((transmission_standard == 1) or (transmission_standard == 2)) { Input_Stream_Identifier 8 uimsbf if (scrambling_sequence_selector == 0) spreading_code_selector 3 scrambling_sequence_index 3 18 uimsbf } for (i=0; i<N; i++) { private_data_byte 8 bslbf } } Semantics for the Satellite_forward_link_descriptor: • spreading_code_selector: This 3-bit field defines the chip sequence used to achieve spectrum spreading, in accordance with Table E.3. Table E.3: Forward link spreading sequences for Direct Sequence Value Spreading factor Chip sequence 000 2 1, 1 001 2 1, -1 010 3 1, 1, 1 011 4 1, 1, 1, 1 100 4 1, 1, -1, -1 101 4 1, -1, 1, -1 110 4 1, -1, -1, 1 111 1 (no spreading) 1 E.2.2 Frame Repetition Spectrum Spreading A DVB-S2 forward link transmission can be spread in bandwidth based on PL-frame repetition using the provisions described in this clause. The frame repetition spectrum spreading is applied in four stages (see example in Figure E.2 for SF=2): 1) π/2 BPSK symbol mapping from short LDPC codeword as the same symbol mapping method of PLHEADER block defined in clause 5.5.2 of DVB-S2 [i.2], PLFRAME is constructed with addition of PLHEADER block. 2) Xspread Frame is constructed through PLFRAME repetition according to the number of SF(Spreading Factor) that is called as the number of identical PLFRAME to enlarge the bandwidth of the signal. When SF=1, the transmission is a conventional DVB-S2 signal that can transmit π/2 BPSK modulation. 3) Scrambler applies a scrambling sequence to the Xspread Frame. Spreading is applied on π/2 BPSK Xspread Frame basis, following the DVB-S2 physical layer scrambling process defined in clause 5.5.4 of [i.2]. The scrambling process is applied to the Xspread Frame. 4) After scrambling process, Spread Frame is constructed through recombination of a pair of PLHEADER and Xspread Frame fragmented from Xspread Frame in Figure E.2. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 281 The value of SF, i.e. 1, 2 and 3, is signalled in the extended Satellite Forward Link Descriptor in clause 6.4.17.6 of ETSI EN 301 545-2 [i.1] and in the modified MODCOD table as shown in Table E.4. The extended signalling description is given in Table E.5. The Spread Frame is transmitted using a π/2 BPSK modulation. Code rates 1/4 and 1/3 with a short LDPC codeword and no pilot mode are considered. When spread signal is transmitted, the MODCOD table (Table 12 of DVB-S2 [i.2]) should be replaced with Table E.4 as shown below. Specifically, in case of the specified MODCOD of 16APSK corresponding to 18D, 19D, 20D, 21D, 22D and 23D, non-pilot and short frame size mode, the relevant MODCOD and TYPE field is used for spread signal transmission in coexistence network with spread and non-spread signal. After Spread Frame construction, the signal is filtered using a squared root raised cosine filtered with a pre-selected roll-off factor as described in clause 5.6 of [i.2]. π π π π π π Figure E.2: Forward link spectrum spreading (in case of SF 2) Table E.4: MODCOD coding for spreading signal Mode MODCOD The MSB of the TYPE field(16K/64K) The MSB of the TYPE field(Pilot/Nonpilot) π/2BPSK 1/4 and Spreading Factor(SF) 1 18D 1 0 π/2BPSK 1/3 and Spreading Factor(SF) 1 19D 1 0 π/2BPSK 1/4 and Spreading Factor(SF) 2 20D 1 0 π/2BPSK 1/3 and Spreading Factor(SF) 2 21D 1 0 π/2BPSK 1/4 and Spreading Factor(SF) 3 22D 1 0 π/2BPSK 1/3 and Spreading Factor(SF) 3 23D 1 0 ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 282 Table E.5: Syntax of the Satellite Forward Link descriptor for Frame Repetition Syntax No. of bits Information Mnemonic Reserved Information Satellite_forward_link_descriptor() { descriptor_tag 8 uimsbf descriptor_length 8 uimsbf satellite_ID 8 uimsbf beam_ID 16 uimsbf NCC_ID 8 uimsbf multiplex_usage 3 bslbf local_multiplex_ID 5 uimsbf frequency 32 uimsbf orbital_position 16 bslbf west_east_flag 1 bslbf Polarization 2 bslbf transmission_standard 2 uimsbf if (transmission_standard == 0) { "001" 3 bslbf } else if ((transmission_standard == 1) or (transmission_standard == 2)) { scrambling sequence selector 1 bslbf roll_off 2 uimsbf } symbol_rate 24 uimsbf if (transmission_standard == 0){ FEC_inner 4 bslbf Reserved 4 bslbf } else if ((transmission_standard == 1) or (transmission_standard == 2)) { Input_Stream_Identifier 8 uimsbf if (scrambling_sequence_selector == 0) spreading_code_selector 3 scrambling_sequence_index 3 18 uimsbf } for (i=0; i<N; i++) { private_data_byte 8 bslbf } } Table E.6: Forward link spreading sequences for Frame repetition Value Spreading factor MODCOD 000 1 π/2BPSK + code rate 1/4 001 1 π/2BPSK + code rate 1/3 010 2 π/2BPSK + code rate 1/4 011 2 π/2BPSK + code rate 1/3 100 3 π/2BPSK + code rate 1/4 101 3 π/2BPSK + code rate 1/3 110 reserved 1, reserved 111 1 (no spreading) Conventional DVB-S2 waveform (except π/2BPSK) ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 283 E.3 Performance Evaluation E.3.1 Direct Sequence Spectrum Spreading The impact of Direct Sequence Spread Spectrum on the Frame Error Ratio has been investigated. In particular, the following system parameters have been considered: • MODCOD: 1/4-QPSK • Chip rate = 27,5 Mchip/s • Symbol rate = Chip rate / Spreading factor • Train speed = 300 km/h • Propagation channel: AWGN and correlated Ricean channel with Rice factor = 17 dB • Satellite HPA IBO = 0,5 dB • No interference from adjacent satellites The robustness of the DVB-S2 spread signal with respect to non-linear distortion by different IBO values of satellite HPA is presented in Figure E.3 in an AWGN channel without mobility effect. The comparison between the spread and non-spread signals indicates that the signal spreading slightly improves the physical layer performance in the presence of a non-linear distortion. Figure E.3: Comparison between spread and not spread signal in AWGN channel in the presence of non-linear distortion The impact of terminal mobility and nonlinear HPA effect on the performance has been analysed. Results are presented in Figure E.4. The same chip-rate (27,5 Mchip/s) has been considered, thus the transmission symbol rate is 13,75 MBaud and 6,85 MBaud for SF=2 and SF=4, respectively. As noted before, spectrum spreading introduces a slight gain with respect to that of a conventional signal. The real benefit of the spreading factor is in the link margin, showing a gain of 3 dB and 6 dB for spreading factor 2 and 4, respectively. Spreading vs No Spreading with non linear HPA 1.00E-06 1.00E-05 1.00E-04 1.00E-03 1.00E-02 1.00E-01 1.00E+00 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Eb/N0 [dB] BER QPSK 1/4 IBO=2 - SPREADING FACTOR=2 QPSK 1/4 IBO=2 - NOSPREADING QPSK 1/4 IBO=0.5 - SPREADING FACTOR=2 QPSK 1/4 IBO=0.5 - NOSPREADING ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 284 Figure E.4: Comparison between spread and not spread signal in LOS channel with the presence of non-linear HPA E.3.2 Frame Repetition Spectrum Spreading The impact of Direct Sequence Spread Spectrum on the Frame Error Ratio has been investigated. In particular, the following system parameters have been considered: • MODCOD: 1/4-π/2BPSK • Symbol rate = 1 Msymbol/s and 10 Msymbol/s • Train speed = 300 km/h • Propagation channel: AWGN and correlated Ricean channel with Rice factor = 17 dB • Satellite HPA IBO = -1 dB • Phase noise (DVB-S2 typical) • No interference from adjacent satellites The performance of the DVB-S2 spread signal based on frame repetition scheme is presented in Figure E.5 with all channel impairment except adjacent channel interference. The comparison between the spread and non-spread signals highlights that the signal spreading slightly deteriorates as spread factor increases and symbol rate and SNIR decreases. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 285 -10.5 -10.0 -9.5 -9.0 -8.5 -8.0 -7.5 -7.0 -6.5 -6.0 -5.5 10 -5 10 -4 10 -3 10 -2 10 -1 10 0 Frame Error Rate Es/No[dB] SF1 1Mbaud SF1 10Mbaud SF2 1Mbaud SF2 10Mbaud SF3 1Mbaud SF3 10Mbaud Ideal πι/2 BPSK Figure E.5: Frame Error Ratio performance for Frame Repetition Spectrum Spreading ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 286 Annex F: Return Link Spectrum Spreading By Burst Repetition F.0 Introduction In Mobile LOS scenarios, return link carriers may require the use of spectrum spreading in order to reduce the spectral density and in particular off-axis EIRP emission density. ETSI EN 301 545-2 [i.1] specifies as an optional feature the use of direct-sequence spread spectrum link to achieve spectrum spreading in the return link. The first generation of DVB-RCS supports an alternative solution for spectrum spreading based on burst repetition. This technique has a minimum impact on the RCST hardware implementation. Burst repetition can be employed for certain special applications, but is not recommended for general spread-spectrum transmission purposes due to possible performance degradation (compared to direct-sequence spread spectrum techniques). This clause describes the burst repetition technique and implementation aspects as a user defined feature. F.1 Spreading description The burst repetition consists of increasing the symbol rate of the signal by a factor N without increasing the power. This modification reduces the Es/N0 at the receiver side. In order to recover the required Es/N0, each burst is repeated N times at the transmitter side. The signal after spreading is depicted in Figure F.1. Figure F.1: Spectrum spreading by burst repetition In order to avoid discrete lines in the output spectrum, a random phase shift should be applied to each replica before transmission. This has no impact on the de-spreading since the relative phases between replicas are estimated in the receiver. F.2 De-spreading description De-spreading is achieved by summing together corresponding signal samples of successive replicas. Prerequisite of the approach is that the timing error and Doppler of the transmitter with respect to the gateway is not such that a significant drift of the sampling time at the gateway happens during the reception of the replicas to be combined. This is actually the case in typical operation scenarios. In fact, even considering the worst case Doppler experienced in the aeronautical domain (1 100 ns/s), more than of 45 000 symbols would be required for timing errors to accumulate up to value larger than 10 % of the symbol period. Obviously combining should be such that homologous samples are added in phase; therefore a phase alignment is required before combining. The principle for the recombination is shown in Figure F.2. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 287 ...... ...... ...... ...... N copies Phase estimator Phase compensation Phase compensation Phase compensation Figure F.2: Coherent recombination The de-spreading solution by recombination needs to memorize at least 2 replicas in order to perform the relative phase estimation. This estimation can be done via a block correlator because exp(jΔφ) is needed and not Δφ itself. An iterative recombining scheme will minimize the needed memory. In order to have a good phase alignment, the differential frequency shift between replicas should be as low as possible. The corresponding frequency shift tolerance will depend on the burst size. The frequency tolerance and the phase shift tolerance are related by the following formula: S s N T f ⋅ Δ = ⋅ Δ π ϕ where Ns is the number of symbols inside a replica, Ts is the symbol duration after spreading, Δϕ is the maximal phase tolerance and Δf is the frequency tolerance between replicas. The longer is the burst, the lower is the frequency tolerance. Typically, in order to have negligible degradations, a 15° error phase between replicas have been considered. Considering a 512 kBaud carrier and a 15° tolerance, the maximal frequency tolerance between 2 consecutive replicas is 45 Hz for a burst of 1 000 symbols length. This value is compatible with a 1 kHz/s frequency drift in Ku-Band. After the de-spreading, the estimations of the absolute delay, frequency and phase are done by conventional timing and carrier recovery algorithms, so a classical DVB-RCS demodulator can be used as shown in Figure F.3. Figure F.3: Receiver architecture including dispreading When spectrum spreading by burst repetition is applied, the transmission instants of individual replicas of the same burst should not vary relative to the start instants of their respective sub-timeslots by more than ±5 % of the symbol period. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 288 F.3 Burst Repetition Implementation F.3.1 Description of FL L2S Components The necessary signalling is supported in the BCT in a user defined field. Table 6-15 of ETSI EN 301 545-2 [i.1] defines all transmission types used in the frame type. In order to support the burst repetition, a new entry needs to be added to this table as shown in Table F.1: • tx_format_class: This field indicates the transmission format class of all transmission types used in the frame type. The values are assigned in Table F.1. Table F.1: Coding of Transmission Format Classes Value tx_format_class 0 Reserved 1 Linear Modulation Burst Transmission 2 Continuous Phase Modulation Burst Transmission 3 Continuous Transmission 4 Spread-Spectrum Linear Modulation Burst Transmission 5 to 127 Reserved 128 User Defined (Generic Continuous Phase Modulation Burst Transmission) 129 Burst Repetition 130 to 255 User Defined Table 6-26 of ETSI EN 301 545-2 [i.1] defines the data block for the return link spreading. Similar description for the burst repetition data block can be defined: • modulation_scheme: This is an 8-bit field which serves as an identifier of the modulation scheme as defined in Table F.2. When spread-spectrum transmission is employed, the three LSB of the field indicate the spreading factor as defined in Table F.3. Table F.2: Modulation Scheme Code Values Modulation Scheme Value Reserved 0x00 to 0x04 π/2-BPSK (No Spreading) 0x05 Reserved for future use 0x06 to 0x0f π/2-BPSK with burst repetition spreading 0x10 to 0x17 QPSK with burst repetition spreading 0x18 to 0x1f π/2-BPSK with direct-sequence spreading 0x20 to 0x27 Reserved 0x28 to 0x7f User defined 0x80 to 0xff Table F.3: Return link spreading factors Spreading factor Modulation Scheme LSB's 2 000 3 001 4 010 6 011 8 100 10 101 13 110 16 111 ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 289 Annex G: RCST IDU/ODU Cable IFL Protocol Description G.0 Introduction This annex provides a description of a cable protocol between the IDU and ODU where the IDU acts as a Master and the ODU acts as a Slave. The protocol is based on an extension of the Eutelsat DiSEqC™ bus specification Version 4.2 [i.20]. This protocol was also reported in annex B of the implementation guidelines for the first generation of the DVB-RCS systems. The control and management of a DiSEqC™ ODU is not completely defined by this specification and system dependent issues should be expected. There are as well optional protocol elements. It should be expected that an IDU need some special adaptation to partly and fully exploit a specific type of DiSEqC™ ODU. G.1 Command and request processing Only one command or status request can be processed at a time. Once the IDU has issued a command or status request to the ODU, a new command cannot be issued until the IDU has received a valid response (ACK or NACK) or the command has timed out. In the case of either a NACK or time-out, the IDU may issue a given command up to three times before declaring a fault on the interface. G.2 Alarms When a hardware alarm occurs within the ODU, the ODU should: 1) disable the SSPA to inhibit transmission by removing power to the Tx circuit; 2) disable the frequency reference signal provided to the IDU (if implemented); 3) buffer the fault indication until read or cleared by the IDU. After detecting the ODU fault (via loss of reference, hub report of abnormal logoff or time-out of command request), the IDU should send a status request message to the ODU to identify the type of alarm. G.3 Dynamic behaviour Unless otherwise specified in the subsequent clauses, the ODU should respond to a command or request received from the IDU within the allowable timeout period (TOODU) of 150 ms. In general, to force the ODU to respond immediately to a command/request from the IDU, the DiSEqC™ command 0x01 may be sent after any power-on or (re-) initialization procedure. NOTE: DiSEqC™ devices will by default apply a random response time between 15 ms and 115 ms, and even 135 ms in case of need for collision avoidance. The TOODU may be disabled during installation. G.4 Error recovery mechanism When the IDU does not receive its expected answer (no answer or NACK), it may re-send the message twice, after which an alarm should be raised. From the ODUs point of view, there will be no limitations on the number of times that the IDU can attempt to send a message. If the ODU keeps receiving messages in error, it will continually respond with the error code. If there is an invalid password, and the ODU requires command authorization, it will "lock-up" after the sixth attempt (see clause G.5.4). ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 290 G.5 Message level description G.5.0 Introduction The Monitoring and Control Protocol message is depicted in Figure G.1. A dedicated protocol will be used for extended messages longer than 8 bytes (e.g. software downloads). It is described in detail in clause G.5.5. Bytes are transmitted MSB first, and each byte is followed by an odd parity bit. Framing P Address P Command P Optional data P 1 byte 1 byte 1 byte odd parity bit up to several bytes Framing P Optional data P Monitoring and Control Message Reply Message (ACK or NACK) Figure G.1: Message format "Framing", "Address" and "Command" fields are detailed in clauses G.5.1, G.5.2 and G.5.3 accordingly. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 291 G.5.1 Framing field description The Framing Field is described in Table G.1. Table G.1: Framing definitions Hex Byte Binary Framing byte Function 0xE2 1110 0010 Command from Master, Reply required, First transmission 0xE4 1110 0100 Reply from Slave, "OK", no errors detected 0xE5 1110 0101 Reply from Slave, Command not supported by slave 0xE6 1110 0110 Reply from Slave, Parity Error detected - Request repeat 0xE7 1110 0111 Reply from Slave, message format not recognized - Request repeat. 0xE8 1110 1000 Extended Command from Master, Reply required only after last message block, First transmission 0xE9 1110 1001 Extended Command from Master, Reply required only after last message block, Repeated transmission 0xEA 1110 1010 Extended Command from Master, Reply required after each message block, First transmission 0xEB 1110 1011 Extended Command from Master, Reply required after each message block, Repeated transmission 0xEC 00 1110 1100 0000 0000 Reply from Slave, command understood, task not yet completed, unknown time to execute 0xEC nn 1110 1100 nn Reply from Slave, command understood, check if task completed after nn seconds (1 to 127 binary) 0xED nn 0xED E1 0xED Fp 0xED F0 0xED Fn 0xED FE 0xED FF 1110 1101 1111 nnnn 1110 1101 1110 0001 1110 1101 1111 pppp 1110 1101 1111 0000 1110 1101 1111 nnnn 1110 1101 1111 1110 1110 1101 1111 1111 Reply from Slave, repeat block nn (where nn is between 01 and 2C) Reply from Slave, EUI-64 of IDU not valid Reply from Slave, relating to password commands, where p indicates: Reply from Slave, password in the incoming string not valid (attempt 1 to n - unidentified number of non-critical attempts) Reply from Slave, password in the incoming string not valid (n identifies the sequence number of a non-critical failing attempt) Reply from slave, password in the incoming string not valid, pen-ultimate attempt (e.g. attempt 5) Reply from slave, ODU Locked (installer required - 6 or more attempts made in a row applying wrong password) 0xEE 00 1110 1110 Reply from Slave, CRC not valid (no additional information) 0xEF 1110 1111 Reply from Slave, additional blocks to follow 0xF0 1111 0000 Request from Slave 0xF1 1111 0001 Reply from Master, OK 0xF2 1111 0010 Reply from Master, Error NOTE: The framing commands are grouped in pairs, where the value of the 2nd LSB of the first bytes gives an indication whether a further response is expected ("1") or not ("0"), although this is not a "hard" rule this should assist with low level detection software. A positive acknowledgement (0xE4) should be used to reply that the message from the Master has been successfully received. Any data requested by the IDU (defined by the original command from the IDU) will be sent directly after the positive acknowledgement byte. If the reply is more than 8 bytes in total then it should use the extended message structure (see clause G.5.5.3). Negative acknowledgements should use values 0xE5, 0xE6 and 0xE7 (no additional data is permitted) as defined in Table G.1. 0xE5 should additionally be used when a command is not supported or cannot be implemented due to a functional problem in the ODU. In this case the ODU should also flag an alarm to the IDU using the mechanism described in clause G.2. 0xE6 Parity error detected (this will, in practice, occur as the result of transmission error), parity check is performed on each byte of each command. Notice that in case of CRC error, the reply is 0xEE. 0xE7 should be used to flag an incompatibility between a command and any other field rendering execution of that command impossible for example incorrect message structure, wrong number of bits or bytes. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 292 In cases where the password is used in the command, the ODU may reply with 0xEB, 0xEC, 0xED or 0xEE. G.5.2 Address field description This field (encoded in one byte) specifies the destination subsystems for each message according to definitions in Table G.2. Table G.2: Address definitions Hex Byte Binary Address byte Function 0x80 1000 0000 Any RCST (and VSAT) 0x81 1000 0001 IDU of RCST 0x82 1000 0010 ODU of RCST 0x83 1000 0011 Other (Future extension) G.5.3 Command field description (IDU→ODU) This field (encoded in one byte) specifies the required action for the addressed subsystem according to definitions in Table G.3. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 293 Table G.3: Command definitions Hex. Byte Use password Command Status Action 00 No Reset O Reset all the ODU functions (same as power down reset) 0A No Soft Reset O ODU Software Reset 12 No Monitoring M Request the general status of the ODU 5C No Manufacturer's ID M Request Manufacturer's Identification 5D No Product ID M Request the Product's Identification C1 No Download start O Allow the ODU to enter into download mode C2 No Download data O Download software data C3 No Download abort O Abort the download process C4 No Download valid O Grab the "new" software into a non volatile memory C5 No Download toggle O Toggle between current software version and previous one C6 Optional SSPA ON M Enable the ODU amplification output C7 No SSPA OFF M Disable the ODU amplification output C8 Optional Set power level O Set SSPA power level C9 No Mod ON O Normal operation CA Optional Mod OFF O Modulation Off, transmit Continuous Wave (CW) CB Yes Change password O Enable the ODU password modification CC Yes Validate password O Activate the new password CD Yes Reset ODU locked O Reset the "Faulty password counter" and back to default password CE No Transmitter Disable M ODU to power down transmitter CF Optional Transmitter Enable M ODU to power up transmitter (SSPA remains off) D0 No Get calibration table O Get the ODU calibration data for temperature and frequency variations D1 No Get Temperature O Report temperature of the ODU D2 No Get power output value O Get the ODU measured power output D3 No Get Location O Get the ODU geographical location (latitude, longitude, altitude) D4 Optional Set Location O Set the ODU location (when stored in ODU) D5 No Serial Number M Request the ODU serial number D6 No Firmware version M Request the ODU firmware version D7 No Set Rx_Freq. O Set Rx Carrier Frequency to ODU D8 No Set Beacon_Freq O Set Beacon Frequency to ODU D9 No Set Tx_Freq O Set Tx Carrier Frequency to ODU DA No Set Satellite_ID O Set Satellite ID to ODU DB No Track OFF O Report Tracking status(OFF) to IDU DC No Track ON O Report Tracking status(ON) to IDU DD to DE - - - Reserved for future standard commands DF - - - Reserved for ODU manufacturer dependent commands NOTE: M = Mandatory, O = Optional (in case function is not supported). G.5.4 Password (optional) A password may be required by the ODU for some commands in order to avoid inadvertent transmission or unapproved use of the ODU. The password will consist of 4 bytes and may be required together with the designated commands shown in Table G.3. In these cases the password will immediately follow the relevant command byte, if there is any associated data used by these commands, it will be sent in a following block (see clause G.5.5). The ODU should refuse commands if the password does not correspond to its current valid password. After 5 consecutive erroneous passwords, the ODU should warn that only 1 try remains. After the 6th faulty password, ODU should refuse all commands except the status request, transmitter disable command and the download commands. These last commands allow an expert to reset the "Faulty password counter" and reset to the default password. The default password is system dependent. When the ODU becomes locked due to 6 consecutive incorrect passwords, the SSPA should be disabled and the transmitter powered down. In clause G.5.5, the password is noted "PWD" in the commands description. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 294 G.5.5 Extended message format G.5.5.0 Introduction In order to maintain backward compatibility with existing DiSEqC™ processors (typically 8 bit microprocessors) it is not possible to have more than 8 bytes of continuous code without the risk of potentially crashing existing devices. Therefore, to allow for the transmission of much longer messages, these will be subdivided into blocks of 8 bytes. Between each block there should be a short pause (Tb) of between 5 ms and 10 ms to allow existing microprocessors, and systems with small hardware buffers, to process each block without a data overflow. G.5.5.1 Extended messages for commands (IDU → ODU) The structure of the first block will always be a standard DiSEqC™ message which has a framing, address and command byte, and does NOT contain any of the subsequent data which is to be error protected (e.g. CRC verified). This block will identify that the subsequent blocks are mostly data and will have a different structure, namely the first byte will be a block identifier which increments in each block, and the last byte will again be reserved for error protection. The framing byte (0x/E2/E8/EA) of the first block defines whether a reply is required to THIS initial block (before the data is transmitted), only after the last block or to all blocks. Also within the first block it will be possible to define how many blocks there are in total. An advantage of the optional reply here is that the slave can be given some time to "prepare" itself for the main data processing task (e.g. clearing a block of memory), and could delay the reply for (say) up to 100 ms, if it needed to (assuming the master has asked for a reply). If not all the subsequent blocks are to be replied to, then the LAST block could then have a reply of the form "E4" (OK), or "ED nn [nn]" (repeat block number[s] nn). All subsequent (continuation) blocks would be of the form: "Ax dd dd dd dd dd dd [pp]" where A is A, B or C indicating the high nibble of the block count, x is the low nibble of the block count, d are data nibbles and pp is a simple (optional) checksum of the 6 bytes in the block. "A0" will be reserved as a "wildcard" block number for applications where it is unnecessary to update the block identifier byte for each block. The last block contains data (or "stuffed" bytes if appropriate) AND the 16-bit CRC. The reason for this is mainly that the CRC is processed in exactly the same way as the data bits, and then if the result is 0000 the data is valid. In this way, with 6 bytes per data block, this fits 256 data bytes (+ 2 CRC bytes) neatly into exactly 43 blocks (plus the initial "header" block) which would be carried in the range of 0xA1 to 0xCB. The extended message structure is shown below: Tb Tb F A C L R R R S P P P P P P P P B1 D D D D D D S P P P P P P P P Bn D D D D V V S P P P P P P P P Header block Continuation block Last block F = Framing byte, P = Parity bit, A = Address byte, C = Command byte, L = Length of message; R = Reserved byte (for reply strategy etc.), Bn = Block identifier, D = Data byte, S = checkSum (optional); V = Verification (CRC as described in clause G.5.6), 5 ms < Tb < 10 ms. Figure G.2 ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 295 G.5.5.2 Simplified structure for short fixed length extended messages (IDU → ODU) To simplify structure for short fixed messages of two or three blocks, for example password protected commands, it is possible to drop the data verification (CRC) since the likelihood of errors is much lower. As the message length (number of 8 byte blocks) is fixed and is defined by the command itself, byte "L is not required, in this case the subsequent block identifier(s) is set to "A0". To give an example for the case of a password protected command the structure could be as follows: Tb Tb F A C PWD S P P P P P P P P Bn= A0 D D D D D D S P P P P P P P P A0 D D D D D D S P P P P P P P P Header block Continuation block Last block PWD PWD PWD F = Framing byte, P = Parity bit, A = Address byte, C = Command byte, PWD = Password byte, Bn = Block identifier set to A0, D = Data byte, S = checkSum (optional), 5 ms < TD < 10 ms. Figure G.3 G.5.5.3 Extended messages for replies (ODU → IDU) For certain commands, the replies have additional data attached. If the total number of data bytes expected in the reply (as defined by the originating command) is more than 6 bytes then it is necessary to use the extended message structure shown below. The framing byte will usually be "E4" and the last byte is reserved for a checksum (whether it is used or not). This gives a "payload" of 6 bytes per block. Tb Tb F D P P P P P P P P 1st block 2nd block nth block (as defined by the command) D D D D D S F D P P P P P P P P D D D D D S F D P P P P P P P P D D D D D S F = Framing byte, P = Parity bit, D = Data byte, S = checkSum (optional), 5 ms < Tb = 10 ms. Figure G.4 G.5.6 CRC definition Some commands require a CRC (see Figure G.5) at the end of the payload in order to secure the communication. The framing, destination address, command and other bytes of the first block are not included within the calculation. Only the data bytes in the subsequent blocks are processed to calculate the CRC). The CRC used is: CRC = x16 + x12 + x9 + x5 + x + 1 Figure G.5: CRC calculation ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 296 G.5.7 General implementation of functions G.5.7.0 Overview In this clause the breakdown of each function into message exchanges between IDU and ODU is shown. These commands are not used during transmissions to avoid generating any spurious noise in the ODU. G.5.7.1 Reset status and parameter request G.5.7.1.1 ODU reset (0x0A) (optional) Reset of all software ODU functions (reload PLL divider, reset register status, alarm, etc.). Note that the "Faulty password counter" will not be reset. This command (see Table G.4) should not be sent if the SSPA is On. Table G.4: ODU reset Direction Message Comment IDU ODU E2 82 0A IDU sends reset command to the ODU. ODU IDU E5 Command rejected, not supported by ODU (optional function) ODU IDU E6 Command rejected, parity error during transmission. ODU IDU E7 Command rejected, message format not recognized. ODU IDU ED FF Command rejected, ODU locked (due to use of faulty password in at a previous stage). ODU IDU E4 Command accepted. ODU will perform a complete reset (software reset). Faulty_passwd_counter will not be reset. Note that "reset command" will be rejected if the ODU is locked. In fact the only way to download new software is to perform an ODU hard reset (cycling power). The IDU should wait at least 10 s after an "ODU reset" to send any command (ODU loader boot time). G.5.7.1.2 ODU Status (0x12) G.5.7.1.2.0 General This command requests the ODU status (see Table G.5). The ODU returns the general status information to the IDU. The ODU should reply to this command even if it is in locked state. Means that the 0xED answer is not possible to this command. Alarms are buffered until the IDU reads the status register or until the IDU performs an ODU reset (0x0A). Table G.5: ODU status Direction Message Comment IDU ODU E2 82 12 IDU sends status command to the ODU. ODU IDU E5 Command rejected, not supported by ODU (should never occur). ODU IDU E6 Command rejected, parity error during transmission. ODU IDU E7 Command rejected, message format not recognized. ODU IDU E4 aa bb cc Command accepted. ODU will give its status with 3 bytes (aa bb cc). When the Status request command is launched while the ODU is in download mode, only "Software Download alarm" and "ODU main" fields are relevant. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 297 G.5.7.1.2.1 aa byte status description: Alarms Table G.6: Alarms (ODU status) bit Status name Values 7 Self test alarm 0 : No Self test alarm 1 : Self test alarm 6 PLL status 0 : Lock PLL 1: Unlock PLL 5 Power supply status 0 : No Power supply Alarm 1 : Power supply Alarm 4 Faulty password counter [0.. 6] faulty password(s) (bit 4 is msb) 3 2 1 Software Download alarm 0 : No CRC alarm on downloaded file 1 : CRC (see note) alarm on downloaded file 0 0 : No other download error 1 : Other download error NOTE: This CRC corresponds to the CRC of the whole downloaded program (it does not refer to the CRC performed on each data packet - refer to 0xC2 command). G.5.7.1.2.2 bb byte status description: ODU state Table G.7: State (ODU status) Bit Status name Values 7 Reserved 0 6 Reserved 0 5 Reserved 0 4 Reserved 0 3 SSPA Status 0 : Off 1 : On (see note) 2 ODU main 0 : Not in Running state 1 : Running state 1 Reserved : 0 0 0 : Not in Software download state 1 : Software download state NOTE: This CRC corresponds to the CRC of the whole downloaded program (it does not refer to the CRC performed on each data packet - refer to 0xC2 command). G.5.7.1.2.3 cc byte status description: Reserved for future use Reserved, all bits set to zero. G.5.7.1.3 ODU Identification (0x54, 0x55, 0x56, 0xD5) These commands (see Table G.8) allow the factory or an authorized installer to collect the different ODU product information: ODU manufacturer's information (using EUI-64 standard from IEEE [i.5]), ODU software & hardware version/release, ODU type and ODU serial number, etc. After power up or reset, the IDU needs to issue this command to the ODU to move to on-line mode. The IDU should wait at least 10 s after an ODU power cycling to send this command (ODU boot time). ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 298 Table G.8: ODU manufacturer's identification (0x5C) Direction Message Comment IDU ODU E2 82 5C IDU sends Manufacturer's identification command to the ODU. ODU IDU E5 Request rejected, not supported by ODU (should never occur). ODU IDU E6 Request rejected, parity error during transmission. ODU IDU E7 Request rejected, message format not recognized. ODU IDU ED FF Command rejected, ODU locked (due to use of faulty password in at a previous stage, ≥ 6 faulty passwords used). ODU IDU E4 gg gg gg Request accepted. ODU should return the Manufacturer's OUI-24, first three bytes of EUI-64. Table G.9: ODU product identification (0x5D) Direction Message Comment IDU ODU E2 82 5D IDU sends Product identification command to the ODU. ODU IDU E5 Request rejected, not supported by ODU (should never occur). ODU IDU E6 Request rejected, parity error during transmission. ODU IDU E7 Request rejected, message format not recognized. ODU IDU ED FF Command rejected, ODU locked (due to use of faulty password in at a previous stage, ≥ 6 faulty passwords used). ODU IDU E4 hh hh hh hh hh Request accepted. ODU should return the Product ID, remaining 5 bytes of EUI-64. Table G.10: ODU firmware version (0xD6) Direction Message Comment IDU ODU E2 82 D6 IDU sends firmware version command to the ODU. ODU IDU E5 Request rejected, not supported by ODU (should never occur). ODU IDU E6 Request rejected, parity error during transmission. ODU IDU E7 Request rejected, message format not recognized. ODU IDU ED FF Command rejected, ODU locked (due to use of faulty password in at a previous stage, ≥ 6 faulty passwords used). ODU IDU E4 aa bb cc dd ff Request accepted. ODU should return the ODU firmware version. Table G.11: ODU serial number (0xD5) Direction Message Comment IDU ODU E2 82 D5 IDU sends serial number command to the ODU. ODU IDU E5 Request rejected, not supported by ODU (should never occur). ODU IDU E6 Request rejected, parity error during transmission. ODU IDU E7 Request rejected, message format not recognized. ODU IDU ED FF Command rejected, ODU locked (due to use of faulty password in at a previous stage, ≥ 6 faulty passwords used). ODU IDU EF ee ee ee ee ee ee CS EF ee ee ee ee ee ee CS E4 ee ee ee ee ee ee CS Request accepted. ODU should return the ODU serial number. NOTE: CS = Check Sum All the following values are considered hexadecimally coded. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 299 Table G.12: ODU identification codes Bytes Bits Status name Values aa 7..4 Current Software Major version 0..F 3..0 Current Software Minor version 0..F bb 7..4 Backup Software Major version 0..F 3..0 Backup Software Minor version 0..F cc 7..4 Hardware Major version 0..F 3..0 Hardware Minor version 0..F dd 7..3 Reserved 0 0..2 ODU type Gives the ODU type (1 to 4, depending on transmit symbol rate) ee ee ee ee ee ee ee ee ee ee ee ee ee ee ee ee 127..0 ODU Serial Number 0..FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF ff 7..0 ODU Boot Firmware version 0..F gg gg gg 63 .. 40 Company ID of the Manufacturer allocated by IEEE (OUI-24), [i.5] Identifies Manufacturer hh hh hh hh hh 39 .. 0 Unique Product ID allocated by Manufacturer according to [i.5] 0 .. FF FF FF FF FF G.5.7.2 Operational commands G.5.7.2.1 SSPA ON (0xC6) This command forces the ODU to enable its amplification output. Table G.13: SSPA on Direction Message Comment IDU ODU E2 82 C6 PWD IDU sends the SSPA output enabling command to the ODU, for ODU requiring use of password. IDU ODU E2 82 C6 IDU sends the SSPA output enabling command to the ODU, for ODU not requiring use of password. ODU IDU E5 Request rejected, not supported by ODU (should never occur). ODU IDU E6 Request rejected, parity error during transmission. ODU IDU E7 Request rejected, message format not recognized. ODU IDU ED Fn Request rejected, password used not valid (< 5 faulty passwords used). ODU IDU ED FE Request rejected, 5 consecutive faulty passwords used. ODU IDU ED FF Command rejected, ODU locked (due to use of faulty password in at a previous stage, ≥ 6 faulty passwords used). ODU IDU E4 Command accepted. ODU should turn on the SSPA. G.5.7.2.2 SSPA OFF (0xC7) This command forces the ODU to disable its amplification output. Table G.14: SSPA off Direction Message Comment IDU ODU E2 82 C7 IDU sends the SSPA output disabling command to the ODU. ODU IDU E5 Request rejected, not supported by ODU (should never occur). ODU IDU E6 Request rejected, parity error during transmission. ODU IDU E7 Request rejected, message format not recognized. ODU IDU ED FF Command rejected, ODU locked (due to use of faulty password in at a previous stage). ODU IDU E4 Command accepted. ODU should turn off the SSPA. By default, after a power on or a reset, the SSPA should be turned off by the ODU. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 300 G.5.7.2.3 Transmitter disable (0xCE) This command forces the ODU to power down the transmitter circuitry. This command will be effectuated by the ODU also when it has locked itself due to repeated use of incorrect password. Table G.15: Transmitter Disable Direction Message Comment IDU ODU E2 82 CE IDU sends the transmitter disable command to the ODU. ODU IDU E5 Request rejected, not supported by ODU (should never occur). ODU IDU E6 Request rejected, parity error during transmission. ODU IDU E7 Request rejected, message format not recognized. ODU IDU E4 Command accepted. ODU should disable the transmitter. This command is issued by the IDU whenever the RCST is put in Hold State. The transmitter should not be re-enabled by the IDU as long as the RCST is in Hold State. In case of error (including internal fault conditions such as PLL unlock and/or DC powering problem) or alarm, the ODU should automatically disable the transmitter; the ODU should be unconditionally stable. G.5.7.2.4 Transmitter enable (0xCF) This command allows the IDU to re-enable the transmitter, e.g. when the RCST Hold State is removed. At the completion of this command, the transmitter is again powered on, but the transmitter is still in the off state. Table G.16: Transmitter Enable Direction Message Comment IDU ODU E2 82 CF PWD IDU sends the transmitter enable command to the ODU for ODUs requiring use of password. IDU ODU E2 82 CF IDU sends the transmitter enable command to the ODU for ODUs not requiring use of password. ODU IDU E5 Request rejected, not supported by ODU (should never occur). ODU IDU E6 Request rejected, parity error during transmission. ODU IDU E7 Request rejected, message format not recognized. ODU IDU EA Command rejected, ODU locked (due to use of faulty password in at a previous stage, ≥ 6 faulty passwords used). ODU IDU E4 Command accepted. ODU should power on transmitter. G.5.7.2.5 Set Power level (0xC8) (optional) This command allows the IDU to adjust the output power level of the ODU in at least 1 dB or less steps. The command instructs the ODU by indicating how many "steps" up or down encoded by one signed data byte PWR_ADJ (±128 steps), this byte will follow in a separate block. Table G.17: Set Power Level Direction Message Comment IDU ODU E2 82 C8 PWD PWR_ADJ IDU sends the Set Power Level command to the ODU followed by the value in the PWR_ADJ byte, for ODUs requiring use of password. IDU ODU E2 82 C8 PWR_ADJ IDU sends the Set Power Level command to the ODU followed by the value in the PWR_ADJ byte, for ODUs not requiring use of password. ODU IDU E5 Request rejected, not supported by ODU (optional function). ODU IDU E6 Request rejected, parity error during transmission. ODU IDU E7 Request rejected, message format not recognized. ODU IDU EA Command rejected, ODU locked (due to use of faulty password in at a previous stage, ≥ 6 faulty passwords used). ODU IDU E4 Command accepted. ODU should change power level. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 301 G.5.7.2.6 Mod ON (0xC9) (optional) In the case when the modulation is applied within the ODU and a co-axial IFL is still used, then this command allows the ODU to re-enable the modulation (for future implementations). Table G.18: Modulation On Direction Message Comment IDU ODU E2 82 C9 IDU sends the Mod On command to the ODU. ODU IDU E5 Request rejected, not supported by ODU (optional function). ODU IDU E6 Request rejected, parity error during transmission. ODU IDU E7 Request rejected, message format not recognized. ODU IDU EA Command rejected, ODU locked (due to use of faulty password in at a previous stage, ≥ 6 faulty passwords used). ODU IDU E4 Command accepted. ODU should modulation on. G.5.7.2.7 Mod OFF (0xCA) (optional) In the case when the modulation is applied within the ODU and a co-axial IFL is still used, then this command allows the ODU to disable the modulation (for future implementations). Table G.19: Modulation Off Direction Message Comment IDU ODU E2 82 CA PWD IDU sends the Mod OFF command to the ODU for ODUs requiring use of passwords. IDU ODU E2 82 CA IDU sends the Mod OFF command to the ODU for ODUs not requiring use of passwords. ODU IDU E5 Request rejected, not supported by ODU (optional function). ODU IDU E6 Request rejected, parity error during transmission. ODU IDU E7 Request rejected, message format not recognized. ODU IDU EA Command rejected, ODU locked (due to use of faulty password in at a previous stage, ≥ 6 faulty passwords used). ODU IDU E4 Command accepted. ODU should switch modulation off. When the modulation is switched off the ODU will transmit a "continuous wave" i.e. a clean carrier. G.5.7.2.8 Set Rx Freq(0xD7) (optional) This command allows the IDU to set Rx carrier frequency in the ODU for the ODU to track the satellite used for certain service when the ODU is (re-)initialized. This command can be used optionally when the IDU/ODU are operated in moving environment. At the completion of the command, the ODU is locked to the specified satellite and ready to receive the FLS. Table G.20: Set Rx Freq Direction Message Comment IDU ODU E2 82 D7 aa aa aa aa IDU sends the Set Rx Freq command to the ODU with 4bytes of frequency value in MHz. ODU IDU E5 Request rejected, not supported by ODU (optional function). ODU IDU E6 Request rejected, parity error during transmission. ODU IDU E7 Request rejected, message format not recognized. ODU IDU E4 Command accepted. ODU should track the satellite. G.5.7.2.9 Set Beacon Freq(0xD8) (optional) This command allows the IDU to set Beacon frequency in the ODU for the ODU to track the satellite used for certain service when the ODU is (re-)initialized. This command can be used optionally when the IDU/ODU are operated in moving environment. At the completion of the command, the ODU is locked to the specified satellite and ready to receive the FLS. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 302 Table G.21: Set Beacon Freq Direction Message Comment IDU ODU E2 82 D8 aa aa aa aa IDU sends the Set Beacon Freq command to the ODU with 4bytes of frequency value in MHz. ODU IDU E5 Request rejected, not supported by ODU (optional function). ODU IDU E6 Request rejected, parity error during transmission. ODU IDU E7 Request rejected, message format not recognized. ODU IDU E4 Command accepted. ODU should track the satellite. G.5.7.2.10 Set Tx Freq(0xD9) (optional) This command allows the IDU to set Tx carrier frequency in the ODU for the ODU to transmit the return link signal. This command can be used optionally when the IDU/ODU are operated in moving environment. And, Tx carrier frequency can be obtained from the FLS (e.g. superframe centre frequency in SCT). At the completion of the command, the ODU is ready to send user data via return-link. Table G.22: Set Tx Freq Direction Message Comment IDU ODU E2 82 D9 aa aa aa aa IDU sends the Set Tx Freq command to the ODU with 4bytes of frequency value in MHz. ODU IDU E5 Request rejected, not supported by ODU (optional function). ODU IDU E6 Request rejected, parity error during transmission. ODU IDU E7 Request rejected, message format not recognized. ODU IDU E4 Command accepted. G.5.7.2.11 Set Satellite_ID(0xDA) (optional) This command allows the IDU to set Satellite ID to the ODU so that the ODU can select target satellite among the several searched satellites. This command can be used optionally when the IDU/ODU are operated in moving environment. This command may be used on the premise that ODU has all satellite information such as satellite position, channel configuration, and so on. This command has to be sent by IDU in the initial step of the ODU if mobile antenna is used. Table G.23: Set Satellite_ID Direction Message Comment IDU ODU E2 82 DA aa aa aa IDU sends the Set Satellite_ID command to the ODU with 3bytes of ID value(satellite_ID : 2bytes, beam_ID : 1byte). ODU IDU E5 Request rejected, not supported by ODU (optional function). ODU IDU E6 Request rejected, parity error during transmission. ODU IDU E7 Request rejected, message format not recognized. ODU IDU E4 Command accepted. G.5.7.2.12 Track OFF(0xDB) (optional) This command has to be issued by ODU whenever ODU detects missing of the satellite. This command allows the IDU to stop sending user data and start buffering. IDU can only resume sending data when it receives Track ON command from ODU. Table G.24: Track OFF Direction Message Comment ODU IDU F0 81 DB ODU sends the Track OFF command to indicate its tracking status(OFF). IDU ODU F1 Command accepted. IDU ODU F2 Request rejected, error during transmission. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 303 G.5.7.2.13 Track ON(0xDC) (optional) This command has to be issued by ODU whenever ODU re-acquires tracking of the satellite. This command allows the IDU to resume sending user data. Table G.25: Track ON Direction Message Comment ODU IDU F0 81 DC ODU sends the Track ON command to indicate its tracking status(ON). IDU ODU F1 Command accepted. IDU ODU F2 Request rejected, error during transmission. G.5.7.3 Download commands G.5.7.3.1 Download start (0xC1) (optional) This command allows the ODU to enter into download mode. This command can only be issued after and ODU power cycle and prior to identification status request. The IDU should wait at least 10 s after an ODU power cycling to send this command (ODU loader boot time). The DL_FL_SIZE corresponds to the Download File Size expressed in bytes on 24 bits. Table G.26: Download start Direction Message Comment IDU ODU E2 82 C1 DL_FL_SIZE IDU sends the Download start command to the ODU. It includes the number of bytes of the complete software to download and the CRC on the file size. ODU IDU E5 Request rejected, not supported by ODU (this answer may occur if the download start command is sent out of the allowed time after the ODU power ON event). ODU IDU E6 Request rejected, parity error during transmission. ODU IDU E7 Request rejected, message format not recognized. ODU IDU E4 Command accepted. ODU should enter into download mode immediately and store the download file size. EC nn Command accepted. ODU should enter into download mode, check status after nn seconds (1 to 127 binary) and store the download file size. Note that the DL_FL_SIZE may handle a value of 0 (zero). If the password given is the default one, the download start command should be refused except if the ODU is locked. In this case, the ODU should enter the download mode so that a new software version can be loaded to clear the faulty password counter. This command can take up to 6,5 s to execute. IDU timeouts should account for this delay. G.5.7.3.2 Download data (0xC2) (optional) This command allows the IDU to transfer ODU program bytes to the ODU divided in 256 bytes per command (message) in 43 blocks of 8 bytes. If the program code is longer than 256 bytes than multiple messages each starting with 0xC2 will be used. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 304 Table G.27: Download data Direction Message Comment IDU ODU E2 82 C2 L 248 data bytes IDU sends the length of message in terms of 6 byte block of data, up to 258 bytes including 2 byte CRC - i.e. max. number of data blocks is 43. Alternative framing byte (E8 or EA) in this first block will indicate the exact reply strategy implemented. ODU IDU E5 Request rejected, not supported by ODU (should never occur). ODU IDU E6 Request rejected, parity error during transmission. ODU IDU E7 Request rejected, message format not recognized. ODU IDU E4 Command accepted (first block OK) continue with data download. IDU ODU Block identifier + 6 data bytes + CheckSum L × blocks of 8 bytes (see clause G.5.5). ODU IDU E4 Command accepted. ODU should store the checked data until the download validation. Any failed packet should be ignored by the ODU. The complete program should be stored into not sensitive memory until the validation of the complete downloaded software. The IDU timeout period should be increased to at least 500 ms to allow for complete ODU processing of the command prior to sending the next message. G.5.7.3.3 Download abort (0xC3) (optional) This command allows the IDU to abort the downloading process when communication problems have occurred or on major trouble. Table G.28: Download abort Direction Message Comment IDU ODU E2 82 C3 IDU sends the Download abort command to the ODU. ODU IDU E5 Request rejected, not supported by ODU (should only occur if software downloading is not supported). ODU IDU E6 Request rejected, parity error during transmission. ODU IDU E7 Request rejected, message format not recognized. ODU IDU E4 Command accepted. ODU should remove the previous downloaded data bytes and exit the download mode in order to restore the normal running mode. Once the abort command has been acknowledged, the IDU has to perform an ODU reset (reset or power cycling). The current software should still be active. The IDU timeout for this message should be increased to 3,5 s. G.5.7.3.4 Download validate (0xC4) (optional) This command allows the ODU to check the received software and store it if the received data is correct. This command may be shown as indicating the end of the download procedure. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 305 Table G.29: Download validate Direction Message Comment IDU ODU E2 82 C4 IDU sends the Download validate command to the ODU. ODU IDU E5 Request rejected, not supported by ODU (should never occur). ODU IDU E6 Request rejected, if the new software is not valid (wrong CRC file) or the ODU is not able to store the new downloaded software or parity error during transmission of the latter command. ODU IDU E7 Request rejected, message format not recognized. ODU IDU E4 Command accepted. ODU should check the complete program validity and store it in order to activate this new software as the current software before acknowledging the command - if completed within 115 ms. ODU IDU EC nn Command accepted. ODU should check the complete program validity and store it in order to activate this new software as the current software before acknowledging the command, check status after nn seconds (1 to 127 binary) if validation is complete (e.g. IDU resends 0xC4 command until E4 is received). Before responding positively the command, the ODU should: • Check the new software validity (CRC). • Save the current software into the backup section. • Save the new received software into the current software section. • Restore the running bit into the main ODU status field. The new program should be active only after a reset command or a power off and on. The timeout for the ODU response should be increased to as much as 8 s to accommodate required processing. The SW version will be updated in the ODU status register once the reset has been launch by the IDU. The IDU has to check the ODU status and ODU identification registers to be aware of the software download result. G.5.7.3.5 Download toggle (0xC5) (optional) This command allows the IDU to toggle to the previous software. This command should be sent only if the ODU is in download mode. To do so, the IDU should use the "start download" command with a DL_FL_SIZE set to 0. The current software should be transfer to the backup non-volatile memory and the "old" program becomes the current one. Table G.30: Download toggle Direction Message Comment IDU ODU E2 82 C5 IDU sends the Download revert command to the ODU. ODU IDU E5 Request rejected, not supported by ODU (should never occur). ODU IDU E6 Request rejected, parity error during transmission. ODU IDU E7 Request rejected, message format not recognized. ODU IDU E4 Command accepted. ODU should toggle the current software with the previous one (in the backup section) if completed within 115 ms. ODU IDU EC nn Command accepted. ODU should toggle the current software with the previous one (in the backup section), check status after nn seconds (1 to 127 binary) if reversion is complete (e.g. IDU resends 0xC5 command until E4 is received). Before responding positively the command, the ODU should switch current and "old" software program. It has to be noticed that if this command is sent twice, the ODU status will not be affected. The "old" program should be active only after a reset command or a power off and on. The SW version will be updated in the ODU status register once the reset has been launch by the IDU (this status reflects the version of the effective running software). The timeout for the ODU response can be as large as 12 s to support the processing of this command. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 306 G.5.7.4 Password commands (optional) G.5.7.4.0 Overview The procedure to change a password is divided into 2 parts: the password change command (using the current password (PWD_cur) and the new one (PWD_new)) and the password validate command. Immediately after the acknowledgement of the password validate command, the new password becomes the current valid one. If any other command or request is inserted between the 2 password commands, the password error has to be raised, increasing the "Faulty password counter". Furthermore, the password modification procedure will have to be re-initialized. G.5.7.4.1 Change password (0xCB) (optional) This command enables the ODU password modification. This function only changes the password but does NOT change the "current" password validity. This command required the message to be split into two blocks as shown below. Table G.31: Change password Direction Message Comment IDU ODU E2 82 CB PWD PWD_new CRC IDU sends the change password command to the ODU with the current password value in the first block. New password calculated with CRC is sent in the second block. ODU IDU E5 Request rejected, not supported by ODU (optional function). ODU IDU E6 Request rejected, parity error during transmission. ODU IDU E7 Request rejected, message format not recognized. ODU IDU ED Fn Request rejected, password (current) not valid (< 5 faulty passwords used). ODU IDU ED FE Request rejected, 5 consecutive faulty passwords used. ODU IDU ED FF Command rejected, ODU locked (due to use of faulty password in at a previous stage, ≥ 6 faulty passwords used). ODU IDU E4 Command accepted. ODU should store the new password and wait to the next password command: validate password. The old password is still valid at this point. As already noticed, the passwords are coded on 4 bytes. G.5.7.4.2 Validate password (0xCC) (optional) This changes the current password to use the new password. Table G.32: Validate password Direction Message Comment IDU ODU E2 82 CC PWD_new CRC IDU sends the validate password command to the ODU in the first block. The new password calculated with the CRC is sent in the second block. ODU IDU E5 Request rejected, not supported by ODU (optional function). ODU IDU E6 Request rejected, parity error during transmission. ODU IDU E7 Request rejected, message format not recognized. This reply is the one sent by the ODU if the "validate password command" is not sent immediately after the "change password command" (see note). This should never occur. The IDU is in charge of sending the right command sequence. ODU IDU ED Fn Request rejected, password (old) not valid (< 5 faulty passwords used). ODU IDU ED FE Request rejected, 5 consecutive faulty passwords used. ODU IDU ED FF Command rejected, ODU locked (due to use of faulty password in at a previous stage, ≥ 6 faulty passwords used). ODU IDU E4 Command accepted. ODU has compared the 2 new passwords and should activate the new password. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 307 NOTE: This reply is also used if the "change password" command is sent twice consecutively, meant that 0xCA command will be followed by 0xCA command. If the "validate password" command is not sent immediately after the "change password" command, an error is generated by the ODU, and the "faulty password counter" will be incremented. The process to modify the password exits. The new password is valid if and only if the 2 commands "change password" and "validate password" are correctly sent with the current password and the new password. If the current password in the "change password command" or the new password in the "validate password command" is not correct, the faulty password counter should be incremented. This command should be sent immediately (consecutively) after the acknowledgement of the "change password command", otherwise the "change password command" should be discarded by the ODU (the process to modify the password exits). This command can take up to 3,5 s to execute. The IDU timeouts should account for this delay. G.5.7.4.3 Reset ODU locked (0xCD) (optional) This command allows authorized personnel to reset the "Faulty password counter" and reset the default password. Table G.33: Reset password Direction Message Comment IDU ODU E2 CD PWD_dft IDU sends the Default Password to the ODU which resets the faulty password counter to zero and sets PWD_cur = PWD_disable. ODU IDU E5 Request rejected, not supported by ODU (optional function). ODU IDU E6 Request rejected, parity error during transmission. ODU IDU E7 Request rejected, message format not recognized. One implementation of this command could be as follows: • ODU is delivered with both current_password and default password set to 0000. This password should be changed before the ODU will transmit. • During installation the Hub/IDU forces installer/user to enter the first "user" password. • Hub records this first_user_password and ODU changes default_password and the current_password to this value. • All subsequent changes to the current_password by the user are not recorded by the hub nor do they change the default_password in ODU. • When ODU becomes locked: out of band request (e.g. by telephone call to hub) for reset Hub authorizes IDU to send reset command "CD" using default_password, faulty password counter is reset and ODU changes current_password to and default_password to 0000 (i.e. unable to transmit) Hub either forces installer/user or itself to change password, new value recorded as default_password (both at hub and in ODU) and current_password. • ODU is unlocked. NOTE: For added security the hub can at any time change the default_password in the ODU by using the reset command. G.5.7.5 Other functions (optional) G.5.7.5.1 ODU calibration table (0xD0) (optional) This request allows the IDU to retrieve the ODU calibration matrix following the frequency and temperature curve. The format of the calibration matrix will be system and ODU dependent to account for frequency differences (e.g. Ka vs. Ku-Band), temperature variations, etc. This command is optional depending upon the implementation of the ODU. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 308 Table G.34: ODU calibration table Direction Message Comment IDU ODU E2 82 D0 IDU sends the calibration matrix request to the ODU. ODU IDU E5 Request rejected, not supported by ODU (optional function). ODU IDU E6 Request rejected, parity error during transmission. ODU IDU E7 Request rejected, message format not recognized. ODU IDU ED FF Command rejected, ODU locked (due to use of faulty password in at a previous stage). ODU IDU E4 aa … Request accepted. ODU should return the output power calibration matrix. NOTE: For a manufacturer specific table of less than 7 bytes the reply can use the simple message structure. The power calibration matrix is ODU manufacturer dependent. G.5.7.5.2 ODU measured temperature (0xD1) (optional) This command allows the IDU to obtain the measured temperature of the IDU. Table G.35: Measured temperature Direction Message Comment IDU ODU E2 82 D1 IDU request measured temperature from the ODU. ODU IDU E5 Request rejected, not supported by ODU (optional function). ODU IDU E6 Request rejected, parity error during transmission. ODU IDU E7 Request rejected, message format not recognized. ODU IDU EA Request rejected, ODU locked. ODU IDU E4 aa Command accepted. ODU provides internal temperature in degrees Celsius (2's complement encoded on 1 byte). G.5.7.5.3 ODU output power level (0xD2) (optional) This request allows the IDU to retrieve the measured output power of the ODU. Table G.36: ODU output power level Direction Message Comment IDU ODU E2 82 D2 IDU sends the output power level request to the ODU. ODU IDU E5 Request rejected, not supported by ODU (optional function). ODU IDU E6 Request rejected, parity error during transmission. ODU IDU E7 Request rejected, message format not recognized. ODU IDU ED FF Command rejected, ODU locked (due to use of faulty password in at a previous stage). ODU IDU E4 bb Request accepted. ODU should return the output power level (encoded on one byte). The output power level coding is ODU manufacturer dependent. G.5.7.5.4 ODU location (0xD3) (optional) This command allows the IDU to get the location information from the ODU. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 309 Table G.37: Get location data Direction Message Comment IDU ODU E2 82 D3 IDU request to get geographical location data. ODU IDU E5 Request rejected, not supported by ODU (optional function). ODU IDU E6 Request rejected, parity error during transmission. ODU IDU E7 Request rejected, message format not recognized. ODU IDU EF xx xx xx xx, yy yy CS E4 yy yy, zz zz zz zz CS Command accepted. ODU sends back its position co-ordinates as defined below. NOTE 1: CS = Check Sum - x_co-ordinate: This 32 bit field defines the x co-ordinate of the RCST location in metres; - y_co-ordinate: This 32 bit field defines the y co-ordinate of the RCST location in metres; - z_co-ordinate: This 32 bit field defines the z co-ordinate of the RCST location in metres. NOTE 2: The position of the satellites will be expressed as Cartesian co-ordinates x, y, z in the geodetic reference frame ITRF96 (IERS Terrestrial Reference Frame). This system coincides with the WGS84 (World Geodetic System 84) reference system at the one metre level. NOTE 3: These 32 bit fields are encoded in the same way as the satellite position data as spfmsbf = single precision floating point value, which is a 32 bit value formatted in accordance with ANSI/IEEE 754 [i.70]. The most significant bit (i.e. the most significant bit of the exponent) is first. G.5.7.5.5 Set ODU location (0xD4) (optional) This command allows the IDU to send the location information to the ODU in the case it is stored in the ODU. Table G.38: Set location data Direction Message Comment IDU ODU E2 82 D3 PWD A0 xx xx xx xx yy yy CS A0 yy yy zz zz zz zz CS IDU command to set geographical location data. The format of the position coordinates is the same as for the GET command. ODU IDU E5 Request rejected, not supported by ODU (should never occur) ODU IDU E6 Request rejected, parity error during transmission ODU IDU E7 Request rejected, message format not recognized ODU IDU E4 Command accepted. ODU stores geographical location data NOTE: CS = Check Sum. G.5.8 Command compatibility when SSPA ON IDU/ODU communications should not be performed during actual return channel transmissions by the RCST to avoid the introduction of spurious signals on the transmitted carrier. In addition, some commands are not available when the SSPA is powered on. The compatibility of commands with the SSPA on is shown in Table G.39. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 310 Table G.39: Command activity when transmitting Hex. Byte Command SSPA ON 00 Reset Not compatible 0A Soft reset Not compatible 12 Monitoring Compatible 5C Manufacturer's ID Compatible 5D Product ID Compatible C1 Download start Not Compatible C2 Download data Not Compatible C3 Download abort Not Compatible C4 Download valid Not Compatible C5 Download toggle Not Compatible C6 SSPA ON -- C7 SSPA OFF Compatible C8 Set power level Compatible C9 Mod ON Compatible CA Mod OFF, transmit Continuous Wave (CW) Not Compatible CB Change password Not Compatible CC Validate password Not Compatible CD Reset ODU locked Not Compatible CE Transmitter Disable Compatible CF Transmitter Enable Not compatible D0 Get calibration data Not Compatible D1 Get Temperature Compatible D2 Get power output value Compatible D3 Get Location Compatible D4 Set Location Not Compatible D5 Serial Number Compatible D6 Firmware version Compatible G.5.9 Use of extended message structures For commands with more than 4 bytes of additional data sent immediately after the command it is necessary to use either the fixed extended message structure (see clause G.5.5.2) or for very long messages (e.g. software downloading) the full extended structure (see clause G.5.5.1). For replies with more than 7 bytes of data then it is necessary to use the extended message structure (see clause G.5.5.3). ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 311 The number of data bytes and which extended message structure to use is indicated in Table G.40. Table G.40: Data bytes and use of message structures Commands Reply Hex. Byte Description No of Data Bytes Message structure No of Reply Data Bytes Message structure 00 Reset 0 simple 0 simple 0A Soft reset 0 simple 0 simple 12 Status 0 simple 3 simple 5C Manufacturer's ID 0 simple 8 simple 5D Product ID 0 simple 5 simple C1 Download start 3 simple 1 simple C2 Download data up to 256 full extended 0 simple C3 Download abort 0 simple 0 simple C4 Download valid 0 simple 1 simple C5 Download toggle 0 simple 1 simple C6 SSPA ON 4 simple 0 simple C7 SSPA OFF 0 simple 0 simple C8 Set power level 5 fixed extended 0 simple C9 Mod ON 0 simple 0 simple CA Mod OFF, transmit Continuous Wave (CW) 4 simple 0 simple CB Change password 10 fixed extended 0 simple CC Validate password 10 fixed extended 0 simple CD Reset ODU locked 4 simple 0 simple CE Transmitter Disable 0 simple 0 simple CF Transmitter Enable 4 simple 0 simple D0 Get calibration data 0 simple < 7 simple D1 Get Temperature 0 simple 1 simple D2 Get power output value 0 simple 1 simple D3 Get Location 0 simple 12 fixed extended D4 Set Location 16 fixed extended 0 simple D5 Serial Number 0 simple 18 fixed extended D6 Firmware version 0 Simple 5 simple ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 312 Annex H: Bibliography • IETF RFC 4306: "Internet Key Exchange (IKEv2) Protocol". • IETF RFC 4303: "IP Encapsulating Security Payload (ESP)". • IETF RFC 4307: "Cryptographic Algorithms for Use in the Internet Key Exchange Version 2 (IKEv2)". • IETF RFC 4835: "Cryptographic Algorithm Implementation Requirements for Encapsulating Security Payload (ESP) and Authentication Header (AH)". • FIPS PUB 186-2: "Digital Signature Standard (DSS)". • FIPS 186-3: "Describes keys greater than 1024 bits". • FIPS PUB 180-2: "Secure Hash Signature Standard (SHS)". • ESA Study: "Security for DVB-RCS at Control and Management Planes". • J. G. Proakis, 4th ed. McGraw-Hill, New York, USA, 2001: "Digital Communications". • D. Divsalar and F. Pollara: "Turbo codes for PCS applications" in. Proc. ICC, Seattle, Washington, pp. 54-59, June 18-22, 1995. • P. Moqvist and T. Aulin: "Serially concatenated continuous phase modulation with iterative decoding", IEEE™ Trans. Commun., vol. 49, no. 11, pp. 1901-1915, August 2002. • P. Moqvist and T. Aulin: "Trellis termination in CPM", IEEE™ Electronics Letters, vol. 36, no. 23, pp. 1940-1941, November 2000. • U. Mengali and M. Morelli: "Decomposition of M-ary CPM signals into PAM waveforms", IEEE™ Trans. Inform. Theory, vol. 41, pp. 1265-1275, September 1995. • ETSI TS 102 771: "Digital Video Broadcasting (DVB); Generic Stream Encapsulation (GSE) implementation guidelines". • Juan Cantillo: "Cross-Layer Optimization Techniques for Satellite Communications Networks", PhD Thesis, ENST, May 2008. • N. Benvenuto, R. Dinis, D. Falconer and S. Tomasin: "Single Carrier Modulation With Nonlinear Frequency Domain Equalization: An Idea Whose Time Has Come - Again," Proceedings of the IEEE™, pp. 69-96, January 2010. ETSI ETSI TR 101 545-4 V1.2.1 (2026-01) 313 History Version Date Status V1.1.1 April 2014 Publication V1.2.1 January 2026 Publication |
08abf9e8c303077e4292141faf8222a0 | 101 543 | 1 Scope | The present document describes some RFID Evaluation Tests that were carried to evaluate the characteristics and performance of RFID equipment operating at their three principal frequencies of use. The information derived from the tests is directly relevant to the work of STF 396 in preparing their response to EC Mandate M/436 [i.1]. |
08abf9e8c303077e4292141faf8222a0 | 101 543 | 2 References | References are either specific (identified by date of publication and/or edition number or version number) or non-specific. For specific references, only the cited version applies. For non-specific references, the latest version of the reference document (including any amendments) applies. Referenced documents which are not found to be publicly available in the expected location might be found at http://docbox.etsi.org/Reference. NOTE: While any hyperlinks included in this clause were valid at the time of publication ETSI cannot guarantee their long term validity. |
08abf9e8c303077e4292141faf8222a0 | 101 543 | 2.1 Normative references | The following referenced documents are necessary for the application of the present document. Not applicable. |
08abf9e8c303077e4292141faf8222a0 | 101 543 | 2.2 Informative references | The following referenced documents are not necessary for the application of the present document but they assist the user with regard to a particular subject area. [i.1] EC Mandate M/436: "Standardisation mandate to the European Standardisation Organisations CEN, CENELEC and ETSI in the field of Information and Communication Technologies Applied to Radio Frequency Identification (RFID) and Systems". [i.2] ETSI TR 187 020: "Radio Frequency Identification (RFID); Coordinated ESO response to Phase 1 of EU Mandate M436". [i.3] ISO/IEC 14443: "Identification cards -- Contactless integrated circuit cards -- Proximity cards". |
08abf9e8c303077e4292141faf8222a0 | 101 543 | 3 Definitions, symbols and abbreviations | |
08abf9e8c303077e4292141faf8222a0 | 101 543 | 3.1 Definitions | For the purposes of the present document, the following terms and definitions apply: batteryless transponder: transponder that derives all of the power necessary for its operation from the field generated by an interrogator eavesdropping: illicit reading of the response from a tag that is being activated by an authorised RFID interrogator effective radiated power: product of the power supplied to the antenna and its gain relative to a half wave dipole in the direction of maximum gain global scroll: mode in which an interrogator is able to read the same tag continuously for test purposes only ETSI ETSI TR 101 543 V1.1.1 (2011-04) 7 interrogator: equipment that will activate an adjacent tag and read its data NOTE: It may also enter or modify the information in a tag. radiated measurements: measurements which involve the absolute measurement of a radiated field tag: transponder that holds data and responds to an interrogation signal |
08abf9e8c303077e4292141faf8222a0 | 101 543 | 3.2 Symbols | For the purposes of the present document, the following symbols apply: A Amps dB decibel dBm power level relative to 1 mW kHz kilo Hertz m metres MHz Mega Hertz mm millimetres σ standard deviation uA/m micro Amps/metre |
08abf9e8c303077e4292141faf8222a0 | 101 543 | 3.3 Abbreviations | For the purposes of the present document, the following abbreviations apply: CCTV Closed Circuit Television DST Digital Signal Transponder e.r.p. effective radiated power EAS Electronic Article Surveillance HF High Frequency IR Infra Rred LF Low Frequency R&TTE Radio and Telecommunications Terminal Equipment RFID Radio Frequency IDentification SRD Short Range Device STF Special Task Force UHF Ultra High Frequency |
08abf9e8c303077e4292141faf8222a0 | 101 543 | 4 Introduction | |
08abf9e8c303077e4292141faf8222a0 | 101 543 | 4.0 General | The present document describes some RFID Evaluation Tests that were carried out at Nedap on 6th to 8th September 2010. The purpose of the tests was to evaluate the characteristics and performance of RFID equipment operating at their three principal frequencies of use. The information derived from the tests is directly relevant to the work of STF 396 in preparing their response to EC Mandate M/436. Additionally it provided a number of the experts in the STF with the opportunity to witness at first hand the operation of RFID in a number of defined scenarios. The reason for these tests arose from the apparent lack of any documented work on practical tests that determined the risk of illicitly reading or writing data to tags. The absence of this information has sometimes led to unrealistic claims by the media and other bodies on what is possible. These tests provided information on the capabilities and limitations of RFID when used in a typical operational environment. ETSI ETSI TR 101 543 V1.1.1 (2011-04) 8 The absence of any controlled experiments regarding illicit reading and writing of data is seen by the STF as a major obstacle for consumer acceptance. The uncertainty regarding what is referred to as the actual read and write range and the potential risks has lead to confusion and distrust among the public. This particularly includes concerns over the illicit reading and writing to tags. There is also confusion around the various RFID technologies, their capabilities and intended use. The scenarios specified in the tests were intended to address these particular concerns. |
08abf9e8c303077e4292141faf8222a0 | 101 543 | 4.1 Test Area | The majority of the tests were performed in the meeting area at the Nedap premises. This is a large open plan space with conditions that were considered typical of many environments where RFID might be used operationally. In addition tests were carried out in a mock up of a room in a house, which was also located in the Nedap premises. The test programme included visits to a working library equipped with RFID and to the Metro Future Store, which uses RFID in its daily operations. |
08abf9e8c303077e4292141faf8222a0 | 101 543 | 4.2 Equipment | The tests were carried out at the three principal frequencies of use using the equipment listed below: Low Frequency (< 135 kHz) 1) Nedap 120 kHz interrogator XS Accessor III 2) DC 1000 Loop antenna 3) General purpose LF cards 4) TPU Write unit 5) TI interrogator RI-TRP-251B-30 and antenna RI-ANT-G01E-30 6) Animal tag RI-TRP-0983-30 7) Key fob tag RI-TRP-RFOB-30 High Frequency (13,56 MHz) 1) Nedap 13,56 MHz Interrogator 2) Loop antenna (40 × 150 cm) for library use 3) General purpose HF vicinity cards 4) Handheld interrogator Quick Scan 5) NXP CL RD 701 interrogator driven by Golden Reader Software 6) Passport fitted with RFID card 7) Transportation card UHF (865 MHz - 868 MHz) 1) Nedap uPASS Reach interrogator 2) Nedap Handheld reader 3) Prototype interrogator 4) Three different designs of retail label tag 5) Airline label tag ETSI ETSI TR 101 543 V1.1.1 (2011-04) 9 Test equipment 1) Rhode and Schwartz Measurement receiver Type EB200 2) Rhode and Schwartz loop antenna Type HFH-Z2 3) Rhode and Schwartz spectrum analyser Type ZVL3 4) DC Coil for magnetic field DC 190 STF 396 appreciates the assistance given by manufacturers in making their equipment available for these tests. They also wish to thank the staff at the public library at Doetinchem and at the Metro Future Store for hosting conducted visits to their premises. For the purpose of the tests, RFID tags are defined as batteryless transponders, which only send a response when they are within range of the energising field of an interrogator. |
08abf9e8c303077e4292141faf8222a0 | 101 543 | 4.3 Test Supervisors | The tests were supervised by the Chairman of ETSI_ ERM_TG34 (John Falck) who was assisted by the leader of STF 396 (Scott Cadzow) and Dr. Christian Schenk. Members of STF 396 who were present at the event were invited to participate in a number of the tests. |
08abf9e8c303077e4292141faf8222a0 | 101 543 | 5 Overview of the Tests | The tests were divided into six different sections covering each of the main areas of concern. These sections are separately summarised below. |
08abf9e8c303077e4292141faf8222a0 | 101 543 | 5.1 Range tests | The purpose of these tests was to determine the maximum range at which it was possible to read a tag and to estimate the variability in performance between different tags. Measurements were made at LF, HF and UHF. For the LF and the HF tests, all of the tags had a form factor similar to a credit card. Two variants of the HF tag were supplied, which were the vicinity card and the proximity card. These were tested separately. Three different designs of tag were tested at UHF. They were of different physical sizes and intended predominantly for use as labels in retail applications. All of the tags tested were batteryless (passive) and were fitted with air cored coils. The tests at UHF included an assessment of the degradation in reading performance of tags when applied to certain materials or affected by the environment or rotated from their optimum orientation. |
08abf9e8c303077e4292141faf8222a0 | 101 543 | 5.2 Write Tests | These tests measured the maximum distance at which it was possible to write data to a tag. The tests were carried out at all three of the principal operating frequencies. The same tags used in the reading tests were also used for measuring the maximum write range. |
08abf9e8c303077e4292141faf8222a0 | 101 543 | 5.3 Illicit Reading | These tests covered a range of scenarios that had been raised by the experts during their discussions within the STF. They each represented situations that could arise during the normal course of people's daily lives. They included such situations as monitoring tagged items in shopping bags, as well as plastic bottles/ cartons of pills in handbags. In view of the results showing the limited reading range at LF and HF, it was decided to perform these measurements only at UHF. In addition tests were carried out to assess the ease with which the data in both a passport and a transport card (e.g. Oyster card) could be read. Further measurements were made to determine the range at which it was possible to read an airline tag fitted to a person's suitcase. Finally tests were undertaken to determine the reading range of an animal tag and the ease with which it might be possible to compromise the security of the RFID tag embedded in a key fob. ETSI ETSI TR 101 543 V1.1.1 (2011-04) 10 |
08abf9e8c303077e4292141faf8222a0 | 101 543 | 5.4 Eavesdropping | Concern had been expressed by some members of the STF about the ability of a person with criminal intent to monitor remotely the response from a tag while it was being read by an interrogator. In order to quantify the extent to which this was possible, a tag was continuously activated by an adjacent interrogator. Using a measuring receiver set to high sensitivity, the signal from the tag was repeatedly read at increasing ranges until it could no longer be detected. The results were believed to be indicative of the maximum ranges that eavesdropping would be possible. These measurements were carried out at each of the three main operating frequencies. |
08abf9e8c303077e4292141faf8222a0 | 101 543 | 5.5 Detection inside buildings | Claims had been made by the press that it was relatively easy for a person to read all of the tags that were inside a person's home. The experts were clearly interested to assess the extent to which this was possible. Tests were performed in a mock-up of a room inside a house. A tagged object was placed at different positions inside the room while an interrogator, which was immediately outside the room, was moved along the 20 cm thick brick wall in an attempt to read the tag. Due to the limited reading range at LF and HF, this test was carried out at UHF only. |
08abf9e8c303077e4292141faf8222a0 | 101 543 | 5.6 Combined EAS/RFID systems | It had been suggested that a likely spot for an eavesdropper would be at the exit of a shop equipped with a combined EAS/RFID system. The experts wished to know whether the handheld reader might adversely affect the performance of the EAS/RFID equipment located at the exit as shoppers left the premises. Similarly there was also a concern that the EAS/RFID equipment would influence the performance of a handheld reader being used to read tags illicitly. Tests were therefore carried out to determine if either of these effects was evident. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.