|
7.6
Downlink Dual Carrier7.6.1
Overall throughput considerations for dual carrier on the downlinkA preliminary assessment is that multi-carrier is most feasible for the downlink. Whether is can be applied also to the uplink depends on MS implementation constraints which are studied in further subclauses. However, even by just allowing multi-carrier reception in the downlink, it is possible to increase the uplink data rates since receiving more effective downlink time slots in a shorter period of time allows to accommodate more uplink timeslots. For instance, the definition of higher multi-slot classes with effective sum=9 could be studied for the case of dual-carrier reception, as shown in figure 50. Although fast frequency synthesizers are assumed, the monitoring slot will be a little bit shorter to allow for tuning from the Tx to the monitoring frequency and from the monitoring to the Rx frequency. As figure 50 shows, this concept is compatible with DTM. This allocation gives a gain of 80 % in the overall throughput compared with a state of the art multislot class 12 MS (sum = 5).
[attach]117419[/attach]
Figure 50: Example of higher multislot classes with effective sum=9
using a second receiver for downlink reception
If multi-carrier is not applied in the uplink, it would still be advantageous if the MS was capable of alternating between the uplink carriers corresponding to the allocated downlink carriers according to the dynamic allocation (see subclause 7.5.2.2 for detailed description).
The multi-carrier operation is illustrated in figure 51, which shows a dual-carrier mobile (4+1) multiplexed with two legacy mobiles (2+1). Note the multiplexing of the dual-carrier MS on two uplink carriers.
[attach]117420[/attach]
Figure 51: Dual carrier multiplexing
7.6.2
Inter-carrier interleavingThis is investigated in clause 10.
7.6.3
Dual-carrier diversityThe same baseband signal is transmitted over two carrier frequencies. At the receiver, the signals on the two carriers are converted to baseband, providing two diversity branches.
7.6.4
Adaptation between dual carrier and receive diversityIn many cases, the dual-carrier on the downlink would be deployed in a network that already supports the MS RX diversity. In order to guarantee the most optimal utilization of network resources, it should be possible to switch between the two modes. The performance evaluation of this scheme is studied in clause 12.
7.6.5
Impacts to the mobile station7.6.5.1
Multiple narrowband receiversThere are different options for the implementation of the multi-carrier RF in the MS receiver. One option, suitable mainly for a small number of carriers (e.g. dual-carrier), is to have separate receiver chains for each carrier. This means that the multi-carrier terminals exploit an architecture, where the receiver branches can be tuned to different frequencies (see figure 52). The receiver branches can use either the same antenna or separate antennas.
[attach]117421[/attach]
Figure 52: RF architecture for dual-carrier receiver with separate receiver chains for each carrier
7.6.5.2
Wideband receiverAnother option, mainly suitable for a larger number of carriers, is a wideband receiver. This option may have additional impacts to the network since it may be necessary to limit the carrier spacing of the multi-carrier assignment. Also, blocking requirements may be an issue.
7.6.5.2.1
Larger bandwidthSimultaneous reception of n carriers would obviously imply larger bandwidth for the receiver front-end. This is in itself a source for additional complexity. However, it is difficult to assess such complexity without a clear requirement on the width of the wideband front-end.
Given that most, if not all, of the GERAN carriers of the multi-carrier allocation will effectively be MAIO's, the receiving interval (from the lowest frequency carrier to the highest frequency carrier) might even be variable. Obviously the receiver shall be dimensioned for the worst case. Thus, it would be beneficial to establish some assumptions in that sense. In other words:
·
Can there be any assumption on the maximum interval between carriers for which the multi-carrier receiver shall be dimensioned for?
7.6.5.2.2
Channel separation
As mentioned in a previous contribution (see [2] in subclause 7.12), channel separation may be performed with known techniques, e.g. digitally.
However, it is important to note that the complexity of digital channel separation is also dependent on the width of the wideband receiver, which shall maintain the same C/N applicable today for GERAN (see note), which in turn is likely to have an effect on power consumption.
NOTE:
The C/N requirement for GERAN is 28 dB, while the C/N requirement for WCDMA is 16 dB.
7.6.5.2.3
Blocking requirementsBlocking requirements are described in 3GPP TS 45.005.
In-band blocking requirements are obviously defined assuming that there is one "useful" carrier, and the receiver has to fulfil some blocking requirements towards all frequencies higher and lower than the "useful" carrier.
This can be illustrated pictorially by figure 53, which refers to a "small MS" in the GSM900 band.
[attach]117422[/attach]
Figure 53: In-band blocking requirements for a Rel-6 "small MS" in GSM900
It is very unlikely that a similar blocking requirement structure can be maintained for a wideband multi-carrier receiver.
In essence, we would now have multiple "useful signals", around each of which we should depict a structure as in figure 53. This is obviously not a practicable option as we would end with drawing a blocking requirement on top of a "useful signal".
Thus, it seems that blocking requirements should be relaxed. A qualitative sketch of such relaxation is illustrated in figure 54.
[attach]117423[/attach]
Figure 54: Possible relaxation of blocking requirements for a multi-carrier "small MS" in GSM900
Note that the "grey area" between the "useful signals" corresponds to the area where the performance requirements for adjacent interference apply. A redefinition of these requirements may also be needed, depending on the respective spacing of the "useful signals".
Further, it is important to consider that, if the frequencies of "useful signals" are effectively MAIO's, then also the respective spacing are changing on a TDMA frame basis. Thus, it should be discussed
·
Whether any bound on the respective spacing of the multiple carriers can be assumed.
·
How blocking should be defined (qualitatively) for a receiver expected to receive multiple carriers at once (i.e. should it look like figure 54?).
7.6.5.3
BasebandOn baseband, the receiver is required to process multiple RLC/MAC blocks per time slot. This requirement may have an impact on meeting the timing requirements of baseband processing. The baseband complexity is directly proportional to the number of carriers.
The support for multi-carrier incremental redundancy may have an impact on the baseband design. In practice, it is required that the channel decoder of a multi-carrier mobile is able to store and retrieve soft decisions from a common pool of soft values.
|
关注楼主