hash stringlengths 32 32 | doc_id stringlengths 5 12 | section stringlengths 5 1.47k | content stringlengths 0 6.67M |
|---|---|---|---|
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.9 Implementation Impact | There are a number of alternatives when applying higher order modulations with increasing degree of impact to consider:
a) Replacing 8-PSK with 16-QAM for MCS-8 and MCS-9 only with the same user data rate (case B in subclause 8.4.1.3). The impact is mainly on the RF receiver and transmitter.
b) Modify MCS-8 and MCS-9 a... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.9.1 Impacts on the Mobile Station | If higher order modulation is applied to DL only, then the main impact is the increased complexity of the receiver as described above.
The capability to receive and decode correctly QAM modulations need to be signalled in classmark 3 and MS-RAC, so the network know which coding schemes that could be used to each mobile... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.9.2 Impacts on the BSS | If higher order modulation is applied to DL only, then the main impact is the potentially more stringent requirements on PA and synthesizer for keeping EVM approximately constant for all modulations. However, if only 16-QAM is considered, there is fair chance that the HW impact is small or none. The impact on HW depend... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.9.3 Impacts on the Core Network | The impact on core network is negligible and only on SW. Addition of new signalling parameter is as simple as any other new feature.
8.9a Implementation Aspects of MORE
8.9a.1 Mobile Stations
- Implementation of higher order modulations (16 ary and 32 ary) as discussed in the GERAN Evolution feasibility study.
- Implem... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.10 Impacts on the Specifications | Following specifications will be affected:
- 3GPP TS 24.008: "Mobile Radio Interface Layer 3 specification; Core Network Protocols; Stage 3".
- 3GPP TS 45.001: "Physical Layer on the Radio Path; General Description".
- 3GPP TS 45.002: "Multiplexing and Multiple Access on the Radio Path".
- 3GPP TS 45.003: "Channel Codi... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.11 References | [1] AHGEV-050016, "Higher Order Modulations", source Ericsson.
[2] AHGEV-050023, "3GPP TR ab.cde vx.y.z Feasibility Study on Future GERAN Evolution", Output document from Adhoc on GERAN evolution, Copenhagen, May 18th - 19th.
[3] SMG2 EDGE 2E99-021 "Noise Model of Transmitter-Receiver Chain for EDGE", source Ericsson.
... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9 Dual symbol rate and modified dual symbol rate | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.1 Introduction | This paragraph proposes two alternative uplink concepts: Dual Symbol Rate (DSR) and Modified Dual Symbol Rate (MDSR) for Future GERAN evolution to double bitrates in uplink. DSR doubles the modulation rate in the transmitter of the mobile station and MDSR combines higher symbol rate (3/2) with 16QAM and optionally with... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.1.1 Technology outline | The transmitter power of Mobile Station is limited e.g. by multi slot power reduction, thus more effective method than adding uplink timeslots or carriers (7.) is needed to improve uplink throughput. Interference Rejection Combining diversity algorithm is widely used in EDGE BSS and it has potentially some unused gain ... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.1.2 Service outline | The EGPRS uplink bit rate evolution is needed to support e.g. imaging feature evolution in EGPRS mobile phones. Camera phones have couple of Mpixel resolution, high quality optics and integrated flash producing decent pictures for family use. In consequence camera phones are replacing point-and-shoot cameras - the bigg... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.2 Concept description | The dual symbol rate and modified dual symbol rate double up link bit rates with minimal impact to mobile stations. The transmission bandwidth is widened and needs appropriate receiver in BTS. According to simulations both spectral efficiency and coverage can be enhanced significantly. With widened signal bandwidth it'... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.2.1 Comparison with MIMO | A MIMO receiver needs at least as many Rx antennas as there are data streams occupying the same timeslot and channel. With DSR, spatial multiplexing is proposed without ensuring that this MIMO requirement is fulfilled. A BTS has typically 2 Rx antennas, but there may be more than two data streams in a given timeslot an... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.2.2 Modulation | The Dual Symbol Rate could apply the existing 8PSK parameters excluding symbol rate and shaping filter. The Modified Dual Symbol Rate uses 16QAM modulation at 3/2 times higher symbol rate compared to GSM. QPSK is considered optionally for coverage extension. MDSR modulator could produce 100 kHz frequency offset to loca... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.2.3 Multiplexing | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.2.3.1 Burst format | In DSR burst one 8PSK modulated symbol corresponds to three Gray mapped bits as defined in 3GPP 45.004, subclause 3.2 Symbol mapping. In MDSR burst, one 16QAM modulated symbol corresponds to four Gray mapped bits, thus with 3/2 symbol rate MDSR burst carries double amount of bits (as in DSR) compared normal 8PSK. A par... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.2.3.2 Blind symbol rate and modulation detection | A BSS needs to detect which symbol rate and modulation were used in the received burst. Detection may be enabled by orthogonal training sequences as in blind modulation detection between 8PSK and GMSK modulations as illustrated fin figure 239.
Figure 239: Illustration of blind symbol rate and modulation detection for r... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.2.3.3 Multi slot classes | Current 8PSK multi slot classes or DTM multi slot classes should apply for DSR and MDSR. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.2.4 Channel coding | The channel coding of dual symbol rate should be carried out in a similar way as with existing 8PSK modulated coding schemes of EGPRS (MCS5-5 to 9), so that incremental redundancy (IR) can be supported between 8PSK and DSR or MDSR blocks.
Table 100 illustrates possible new modulation and coding schemes. The coding rate... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.2.5 RLC/MAC | The RLC/MAC header need to carry information of 4 RLC blocks thus new header type is needed, but the EGPRS uplink RLC/MAC header type-1 could be re-used for 2 lowest DSR MCSs, without adding new bytes. Three highest DSR MCSs carrying 4 blocks would need the following additions to the EGPRS uplink RLC/MAC header type-1:... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.2.6 RRC | Introduction of new Radio Access Capability is needed. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.2.7 Radio transmission and reception | It could be assumed that Dual Symbol Rate has quite similar properties as 8PSK and the same approach as used for specifying properties 8PSK could be applied, but some considerations are needed due to wider spectrum.
It is assumed that BTS uses IRC diversity allowing interferes to overlap from adjacent carriers. BTS per... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.2.7.1 Transmitter output power and power classes | No changes expected for DSR and existing E-power classes could be applied due to similar linearity requirements with 8PSK. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.2.7.2 Modulation accuracy | Current EVM figures should likely apply for DSR with a note of different symbol rate and shaping filter. EVM for MDSR is FFS. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.2.7.3 Power vs. time | No major changes are expected for DSR, since PAR is similar with current 8PSK and burst structure is specified according to the current 8PSK modulated normal burst. Due to the shaping filtering the lower limit during random data symbols may need to be removed due to possible zero crossings. PVT for MDSR is FFS. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.2.7.4 Spectrum due to modulation | Spectrum due to modulation mask needs to be changed to apply for dual symbol rate. As an initial starting point the current spectrum mask for 8PSK could shifted by 200 kHz for DSR and 100 kHz for MDSR and relative amplitude normalized to correspond the same absolute power with 8PSK. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.2.7.5 Spectrum due to transients | Spectrum due to transients needs to reflect changes in spectrum due to modulation. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.2.7.6 Receiver blocking characteristics | Channel filtering of BTS transceiver is assumed to meet existing blocking characteristics for GMSK despite being wide enough to pass through DSR or MDSR signal. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.2.7.7 AM suppression characteristics | No changes expected. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.2.7.8 Inter-modulation characteristics | No changes expected. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.2.7.9 Nominal Error Rates (NER) | Similar limits as for 8PSK could be applied. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.2.7.10 Reference sensitivity level | Adding diversity cases need to be considered. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.2.7.11 Reference interference level | Adding diversity cases need to be considered. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.3 Modelling assumptions and requirements | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.3.1 MS transmitter modelling | Ideal transmitter was used in coverage and interference scenarios, but power amplifier model based on the GaAs HBT technology was used in spectrum due to modulation and adjacent channel power evaluations. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.3.2 BTS receiver modelling | Uplink Interference Rejection Combining diversity (IRC) was used in simulations and some reference simulations were also performed with Maximum Ratio Combining (MRC) or without diversity. The effective noise figure was 5 dB, antenna branches were uncorrelated and no other impairments were included to receiver simulatio... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.3.3 Simulation approach for interference modelling | As seen in figure 237 DSR spectrum overlaps over three normal 200 kHz carriers resulting in about 3 to 5 times more stringent interference situation for the BTS receiver. Thus conventional single interferer models (like CCI, ACI) or even the multi-interferer method used in SAIC cannot be used for DSR performance evalua... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.4 System level model | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.4.1 Network model and system scenarios | 4 different system scenarios were used to collect burst-wise interference data and wanted signal statistics. Network configurations and simulation parameters are listed in the tables 101 and 102. Frequency re-use 4/12 was studied for BCCH, and reuses 1/1 (1/3 load), 1/3 and 3/9 for hopping layer. It should be noted tha... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.4.2 Network interference statistics | In figures 241, 242, 243 and 103 cumulative co- and adjacent channel interference distributions are shown for scenario 1 and scenario 2. Carrier level shows Rx levels measured from EGPRS connections. The percentage value after the interference number displays a probability of an interferer. The complete list of the int... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5 Performance characterization | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.1 Spectrum due to modulation | Figure 246 shows simulated example of spectrum due to modulation with GaAs HBT PA model biased near to class-B resulting 35 % power added efficiency (PAE) for DSR. For comparison the spectrum due to modulation for 8PSK would kiss the existing limit line at 400 kHz offset with the same PA as depicted in figure 245.
The... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.2 Adjacent channel power | The adjacent channel power due to DSR transmission was evaluated by using GaAs HBT PA model biased near to class-B yielding to 35 % power added efficiency. Simulated spectrum is shown in figure 246. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.2.1 Adjacent channel power to GSM/EDGE uplink | Adjacent Channel Power (ACP) for different offsets was estimated through 180 kHz rectangular filter compared to the total transmitted signal power. The adjacent channel power limits were derived from the reference interference level limits for 3 lowest offsets and from spectrum due to modulation mask for higher offsets... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.2.2 Adjacent channel power to WCDMA uplink | Adjacent channel power (ACP) was estimated through 3840 kHz rectangular filter compared to the total transmitted signal power. The impact to adjacent WCDMA uplink was estimated by determining ACP at 2.7 MHz offset and comparing it to allowed ACP of WCDMA transmitter at 5MHz offset.
As a result modelled PA has 19 dB mar... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.2.3 Spectrum mask and spurious emissions | At first glance, the spectrum of MDSR looks as wide as the spectrum of two adjacent GMSK modulated carriers. However, the adjacent channel performance is worse. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.2.3.1 Adjacent channel protection | Today, the attenuation of the first adjacent channel interferer (±200 kHz) through reference, 180 kHz wide RX filter amounts to 18 dB. However, for MDSR, the attenuation in -200 kHz/+400 kHz offset would be only 16 dB as shown in table 117. 2 dB lower adjacent channel protection correspond to almost 60 % more power lea... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.2.3.2 Spectrum after PA and spectrum mask | The simulated spectrum after a PA model and the comparison with a spectrum mask, shifted by 100 kHz as shown in figure 247, shows that the mask is violated at 300 kHz and continuously between 500 kHz and 1 000 kHz carrier offset despite a 2 dB higher output power back-off than for 8-PSK. It can be concluded that:
- eit... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.3 Coverage | The DSR and MDSR coverage was modelled with the following assumptions, resulting to level of -108 dBm at cell border (95 %).
Noise floor of BTS with NF=5dB - see note)
-115 dBm
Required Eb/No for EFR (FER < 1 %) with diversity
2 dB
Body loss difference between talk and data positions
3 dB
Power decrease for 8PSK relate... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.4 Performance at Hilly Terrain | The receiver performance was evaluated also at Hilly Terrain to ensure receiver's capability to cope with delay spreads at least up to 20 µs. As a result the DSR provides about 2 times higher average throughput than 8PSK at HT3 iFH conditions.
Figure 250: Throughput at Hilly Terrain 3km/h
The throughput at cell border ... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.5 Performance at interference scenarios | Throughputs versus carrier level are shown in figures 251, 252, 253 and 254. Vertical lines mark 95 % and 50 % signal levels. Figure 109 summarises average throughputs at different system scenarios.
Figure 251: Throughput at scenario 1 (4/12)
Figure 252: Throughput at Scenario 2 (1/3)
Figure 253: Throughput at Scenario... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.6 Spectral efficiency | The spectral efficiency of DSR was estimated only for BCCH re-use 4/12 (Scenario 1) providing 522 kbit/s average throughput per cell. Thus applying Dual Symbol Rate at BCCH layer may be attractive option.
Table 109: Spectral efficiency for Scenario 1 (BCCH 4/12)
Modulation
Spectral Efficiency
8PSK
124 kbits/s/MHz/Cell
... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.7 Impact to voice users with 1/1 re-use | Simulation results presented later in this paragraph are with random resource allocation e.g. without any DSR specific RRM optimisation. Simulations show that DSR has similar impact to voice quality as EDGE at lower data load and smaller impact than EDGE with higher data load.
Simulations are performed at the worst cas... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.8 Impact to voice users C/I distribution with 1/3 re-use | The wider frequency spectrum of DSR will spread interference over a larger frequency range an increased interference on neighbouring channels. This is shown in figure 246 and table 111. The interference in the adjacent channel is just about 1 dB less than in the centre channel.
Table 111: Power distribution over 3 chan... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.8.1 Simulation results for uplink | In order to investigate this effect, system simulations have been run. A mixed traffic scenario with 80 % speech traffic and 20 % data traffic was considered. The simulation parameters are summarised in table 112.
Table 112: Summary of system simulation parameters
Parameter
Value
Reuse
1/3
Traffic mix
80% speech, 20% d... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.9 Uplink/downlink balance | It is earlier assumed that a degradation of uplink speech performance is not a problem, since the downlink performance is anyway limiting the overall performance. While this may be true in some scenarios, it is not always the case.
For instance, shadowing from buildings and other obstacles will impact the uplink-downli... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.10 Real Time service coverage | The coverage for real time data service was evaluated by using RLC un-ACK mode and next MCS exceeding 64 kbps with 2 uplink slots for both the EDGE and DSR resulting to 89.6 kbit/s. At 0.1 % target BLER the DSR gain was about 6.4 dB. It could be possible to develop optimised DSR coding schemes for real time.
Figure 259... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.11 DSR and Speech Performance in Legacy MRC Network | Dual Symbol Rate performance in mixed voice and data interference scenario, with Interference Rejection Combining (IRC) in all the BTS receivers show 1.7 to 1.9 fold data capacity gain in uplink. In this subclause DSR performance is evaluated assuming legacy Maximum Ratio Combining (MRC) receivers for voice time slots ... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.11.1 DSR interference model for system simulation | Used dynamic system simulator calculates total received interference (for C/I definition) as a sum of co- and first adjacent channel interference levels through a reference, 180 kHz wide, RX filter, which provides 18 dB adjacent channel attenuation for EGPRS and GMSK interferers.
In the case of DSR interferer co-channe... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.11.2 UL speech performance in legacy MRC network | Dynamic system simulations were run with the frequency reuse of 1/3, which is basically the same as the DSR scenario 2.
At first, the reference EGPRS simulation was run (20 % EGPRS FTP data and 80 % AMR 5.9 kbit/s voice). Speech service quality was evaluated with the following criteria:
- Relative number of bad quality... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.11.3 DSR performance in legacy MRC network | EGPRS and DSR data throughputs were studied with link level simulations based on recorded bursts resulting similar voice performance i.e. DSR TX power was limited by 2 dB. Median interference level and variance for 3 strongest co‑channel interference levels and 2 strongest adjacent-channel interference levels are shown... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.12 Impacts to the signalling | As shown in subclause 9.5.11, the DSR power control can be used to remove voice impacts due to DSR, thus it's expected that the same would apply also for signalling. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.13 MDSR performance at interference scenario 2 | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.13.1 Modelling assumptions and requirements | The modelling assumptions were the same as for DSR, excluding the following:
- MS use 2dB lower maximum power for 16QAM.
- Spectral properties, e.g. adjacent channel power levels of MDSR, were taken into account. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.13.2 System level model | The system model and simulation approaches were the same as for DSR in purpose of compare MDSR to DSR. Only 1/3 re-use (the scenario 2) was studied assuming MRC for voice as in chapter 9.5.11. Power control was applied for MDSR so that voice performance is not reduced but rather improved related to EGPRS.
Two differen... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.13.3 Performance at mixed voice and data interference scenario 2 | Throughputs versus signal level at interference limited scenario are shown in figure 263. The following MDSR power control scheme was used to maintain the same or better voice quality than with EGPRS:
- Maximum MDSR power was 2 dB lower than EGPRS or DSR due to increased PAR.
- 4 dB power reduction related to EGPRS UL ... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.13.3.1 Impact of antenna correlation | The impact of receiver antenna correlation to MDSR throughput was studied by repeating scenario 2 by assuming RX antenna correlation of 0.7. This inclusion of interference distribution from system simulator instead of link level analysis should give realistic outcome. The results are plotted in figure 264.
Figure 264:... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.13.4 Network performance at data only interference scenarios | MDSR network level performance was studied with dynamic system simulations. Wide MDSR carriers were modelled in the simulator e.g. carriers were shifted +100 kHz so that MDSR signal was overlapping with two basic 200 kHz carriers.
Presented simulation results are uplink results assuming 2-antenna diversity. MRC and IRC... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.13.4.1 Performance at scenario 5 (4/12 re-use) | Figure 268 depicts reuse 4/12 simulation results. 10th percentile of connection bit rates are shown for EGPRS with MRC, EGPRS with IRC and MDSR versus offered FTP load. Then, MDSR capacity gains are compared to EGPRS with MRC and IRC are shown for different 10th percentile session bit rate requirements. Network capacit... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.13.4.2 Performance at scenario 6 (1/3 re-use) | In case of re-use 1/3 capacity gain at 100 kbps session bit rate is 120 % to 190 % for 4-TRX implementation and 90 % to 150 % for 2-TRX implementation. For the 144 kbps bit rate requirement gains are 180 % to 420 % and 140 % to 340 %, respectively. Reuse 1/3 was found to be spectrum limited and therefore 2-TRX implemen... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.13.4.3 Performance at scenario 7 (3/9 re-use) | In case of re-use 3/9, capacity gain at 150 kbps session bit rate is about 150 % compared to EGPRS with IRC and 300 % compared to EGPRS with MRC.
Figure 268: System simulation results for scenario 7 (3/9) |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.13.5 Performance of two transceiver implementation | This paragraph presents a dual TRX performance at DTS-2 scenario with impairments.
Two conventional receivers e.g. like used in the first generation GSM BTS over decade ago were assumed. Both receivers have independent impairments as listed in table 120. The used channel filter has 200 kHz bandwidth and in addition to ... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.13.6 Voice and higher symbol rate in asynchronous interference scenario | This subclause presents some results for voice performance with asynchronous MDSR interference that has no 100 kHz offset. Offset was removed since two TRX implementation option found to have low interest. In addition IRC performance at low SNR conditions is also evaluated.
Simulations shown, 1.5 times higher symbol ra... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.13.6.1 Simulation assumptions | General modelling assumptions were:
- MRC and IRC receivers were used for AMR 5.9.
- Asynchronous and synchronous DTS-2 scenarios were used.
- Interfering signal were 8PSK and 16QAM with 1.5 symbol rate.
- Low SNR simulation was modelled with noise floor at 0dB instead of -17dB from I1.
- Receiver impairments were incl... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.13.6.1.1 Receiver impairments | Table 122 lists used receiver impairments that were used in the receiver performance simulations.
Table 122: Receiver impairments
Impairment
Value
I/Q gain imbalance
0.125 dB
I/Q phase mismatch
1°
DC offset
-30 dBc
Phase noise
1.2° RMS
Total receiver noise figure
5 dB |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.13.6.1.12 DTS-2 interference scenario | The DTS-2 link level interference scenario is illustrated in figure 272 and table 124. Note that DTS-2 provides a link level approximation of system interference of a typical interference limited network for speech, which was agreed in TSG-GERAN for testing SAIC performance. The interference distribution for mixed spee... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.13.6.2 Performance characterization | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.13.6.2.1 Voice performance at asynchronous and synchronous scenario | It can be seen from DTS-2 results with MRC that higher symbol has a bit smaller impact to voice than EGPRS interference at 1 % FER.
In asynchronous scenario MDSR interference has about similar impact to voice as EGPRS. But with IRC and in synchronous scenario, voice performance was degraded by 0.3 dB with MDSR interfer... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.13.6.2.2 Voice performance at low SNR with MRC | Simulations with DTS-2 scenario at low SNR conditions show that higher symbol rate has 0.7 dB lower impact to voice than EGPRS.
Figure 274: Voice performance at low SNR (noise power at 270 kHz equal to I1 power) |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.14 High Symbol Rate (HSR) performance | Evaluated concept uses 1.2 times higher symbol rate i.e. 325 ksymbol/s and 32QAM modulation with convolution channel coding. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.14.1 Modelling assumptions | Legacy EGPRS receiver SAW filter with 240 kHz BW (3GPP TR 45.912, subclause 9.7.1.1.1.1) was assumed. Two different transmitter shaping pulses were used with bandwidths of 240 kHz and 325 kHz to determine the performance impact of shaping filter.
The receiver performance modelling assumptions were:
In the receiver simu... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.14.1.1 Receiver impairments | Table 124 lists used receiver impairments that were used in the receiver performance simulations.
Table 124: Receiver impairments
Impairment
Value
I/Q gain imbalance
0.125 dB
I/Q phase mismatch
1°
DC offset
-30 dBc
Phase noise
1.2° RMS
Total receiver noise figure
5 dB
SAW filter bandwidth
240 kHz |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.14.1.2 DTS-2 interference scenario | The DTS-2 link level interference scenario, illustrated in figure 275 and table 125, was used 1.2 times higher symbol rate with different shaping pulses and receiver filters.
A random time offset was used for asynchronous interferers per radio frame. Two consecutive bursts with independent fast fading were generated to... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.14.2 Performance characterization | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.14.2.1 Coverage | Throughput versus received signal level is depicted in figure 276 for 1.2x32QAM with different shaping and receiver filters at TU3iFH conditions with receiver impairments given in table 124.
Table 126 shows average throughputs and throughput gains at -98 dBm median RX level. 1.2 x 32QAM used additional power due to hig... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.14.2.2 Data performance in synchronous DTS-2 interference scenario | The throughput of 1.2 x 32QAM was evaluated in Figure 277 for the following cases:
- 1.2x32QAM with 325 kHz shaping filter and 325 kHz receiver filter.
- 12x32QAM with 325 kHz shaping filter and 240 kHz receiver filter.
- 12x32QAM with 240 kHz shaping filter and 240 kHz receiver filter.
It can be seen that wider shapin... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.5.14.2.3 Voice impact in asynchronous DTS-2 interference scenario | The AMR 5.9 FER was simulated with asynchronous DTS-2 scenario. Three curves are presented:
- AMR 5.9 FER with EGPRS interference.
- AMR 5.9 FER with 1.2x32QAM interference with 240 kHz shaping filter.
- AMR 5.9 FER with 1.2x32QAM interference with 325 kHz shaping filter.
Figure 278: AMR 5.9 FER versus C/I1 in asynchro... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.6 Impacts to the mobile station | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.6.1 DSR | Dual symbol rate has small impact to terminal e.g. HW changes could be limited to the modulator. Linearity requirements due to peak to average ratio are similar as for 8PSK. Requirements due to peak to minimum may be higher. Modulation spectrum mask at 800 kHz offset may need to be optimised allowing reasonable transmi... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.6.2 MDSR | MDSR has small impact to terminal e.g. HW changes could be limited to the modulator. Generation of 100 kHz offset is considered to be done by the modulator on burst basis thus additional synthesizer settling time is not needed for it. Linearity requirements, e.g. due to peak to average ratio are a bit higher than for 8... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.6.2.1 MS implementation issues | To generate the MDSR modulation in the MS, it was proposed to apply an offset of 100 kHz in the baseband modulator. This means that the required baseband bandwidth will double although the symbol rate is only 50 % higher. The two different symbol rates and the doubled bandwidth may require a different, higher DAC sampl... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.7 Impacts to the BSS | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.7.1 Impacts to the transceiver | The BTS receiver is required to have sufficient channel bandwidth and also should have sufficient processing power for double amount of uplink data. The sampling rate should be at least equal to the symbol rate. So DSR is not compatible with all legacy BTS hardware. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.7.1.1 Two transceiver implementation for DSR | The secondary BTS transceiver implementation option may be based on the use of pair of legacy transceivers instead of one. This option utilises half band sampling where sampling rate can be half of DSR symbol rate and channel filter half of the BW of DSR i.e. like in EDGE. Indeed it is possible to share the DSP process... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.7.1.1.1 Performance impact of two TRX implementation | Uncoded DSR performance has been evaluated by simulations. A conventional DSR receiver is compared to a two TRX receiver. The DSR equalizer has the following characteristics.
- 11 channel taps.
- Synchronization window that handles time offsets comparable to the EDGE 8PSK equalizer.
- DFSE equaliser with 2 MLSE taps. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.7.1.1.1.1 Impact of decreased receiver bandwidth | With a two TRX receiver, the signal is received through two separate receiver filters, each having a bandwidth optimised for a conventional GSM/EDGE carrier. Since a DSR signal has a bandwidth of 540 kHz, two GSM/EDGE receiver filters typically cannot receive the entire signal. In the simulations, the two receivers are... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.7.1.1.1.2 Impact of other analogue components | When two receivers are used to receive one wideband signal, a tight match of the analogue components of the two receivers is essential; otherwise the reconstructed signal will be distorted. The analogue parts of a GSM/EDGE receiver are designed for optimal reception of a single GSM/EDGE carrier and the tolerances of co... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.7.1.2 Processing complexity | The complexity increase due to DSR varies depending on the type and architecture of used receiver and performance requirements for DSR. Next is characterised the complexity of the receiver used for DSR simulations.
Table 127: Processing complexity estimation for dual symbol rate
Function
Processing demand relation to s... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.7.1.2.1 Evaluation of receiver complexity | Above it stated that the receiver complexity for DSR is about up to 50% higher per bit under the condition that the simulated receiver has 11 taps. We think the complexity is different for different propagation models. For TU channel model, the maximum RMS delay is 5.0 μs, its influence to the taps number of channel es... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.7.1.3 MDSR impacts to the transceiver | In simulations a Frequency Domain MMSE equalizer was used. The channel estimate used in the simulated receiver has 9 taps i.e. about 1.5 times as much as in EDGE. The decoding complexity is 2 times higher per timeslot than for 8PSK due to doubled amount of bits. Total processing complexity of MDSR is in order of 2 time... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.7.1.4 Network implementation issues | The IRC performance which was assumed in the MDSR simulations was based on synchronous networks. However, most networks are asynchronous. IRC performance in asynchronous networks may be lower. Since IRC is a prerequisite for full MDSR performance, simulations would be appreciated which prove that MDSR provides similar ... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.