MIB Processing

 BCH physical-layer processing, such as channel coding and resource mapping, differs quite substantially
from the corresponding processing and mapping for DL-SCH outlined in Chapter 10.
As can be seen in Figure 14.4, one BCH transport block, corresponding to the MIB, is transmitted
every 40 ms. The BCH Transmissions Time Interval (TTI) thus equals 40 ms.
The BCH relies on a 16-bit CRC, in contrast to a 24-bit CRC used for all other downlink transport
channels. The reason for the shorter BCH CRC is to reduce the relative CRC overhead, having the
very small BCH transport-block size in mind.
BCH channel coding is based on the same rate-1/3 tail-biting convolutional code as is used for the
PDCCH control channel. The reason for using convolutional coding for BCH, rather than the Turbo code used for all other transport channels, is the small size of the BCH transport block. With such
small blocks, tail-biting convolutional coding actually outperforms Turbo coding. The channel coding
is followed by rate matching, in practice repetition of the coded bits, and bit-level scrambling. QPSK
modulation is then applied to the coded and scrambled BCH transport block.
BCH multi-antenna transmission is limited to transmit diversity – that is, SFBC in the case of two
antenna ports and combined SFBC/FSTD in the case of four antenna ports. Actually, as mentioned in
Chapter 10, if two antenna ports are available within the cell, SFBC must be used for BCH. Similarly, if
four antenna ports are available, combined SFBC/FSTD must be used. Thus, by blindly detecting what
transmit-diversity scheme is used for BCH, a terminal can indirectly determine the number of cell-specific
antenna ports within the cell and also the transmit-diversity scheme used for the L1/L2 control signaling.
As can also be seen from Figure 14.4, the coded BCH transport block is mapped to the first subframe
of each frame in four consecutive frames. However, as can be seen in Figure 14.5 and in contrast
to other downlink transport channels, the BCH is not mapped on a resource-block basis. Instead, the
BCH is transmitted within the first four OFDM symbols of the second slot of subframe 0 and only over
the 72 center subcarriers.2 Thus, in the case of FDD, BCH follows immediately after the PSS and SSS
in subframe 0. The corresponding resource elements are then not available for DL-SCH transmission.
The reason for limiting the BCH transmission to the 72 center subcarriers, regardless of the cell
bandwidth, is that a terminal may not know the downlink cell bandwidth when receiving BCH. Thus,when first receiving BCH of a cell, the terminal can assume a cell bandwidth equal to the minimum
possible downlink bandwidth – that is, six resource blocks corresponding to 72 subcarriers. From the
decoded MIB, the terminal is then informed about the actual downlink cell bandwidth and can adjust
the receiver bandwidth accordingly.
Clearly, the total number of resource elements to which the coded BCH is mapped is very large
compared to the size of the BCH transport block, implying extensive repetition coding or, equivalently,
massive processing gain for the BCH transmission. Such large processing gain is needed as it should
be possible to receive and correctly decode the BCH also by terminals in neighboring cells, implying
potentially very low receiver Signal-to-Interference-and-Noise Ratio (SINR) when decoding the
BCH. At the same time, many terminals will receive BCH in much better channel conditions. Such
terminals then do not need to receive the full set of four subframes over which a BCH transport block
is transmitted to acquire sufficient energy for correct decoding of the transport block. Instead, already
by receiving only a few or perhaps only a single subframe, the BCH transport block may be decodable.
From the initial cell search, the terminal has found only the cell frame timing. Thus, when receiving
BCH, the terminal does not know to what set of four subframes a certain BCH transport block is
mapped. Instead, a terminal must try to decode the BCH at four possible timing positions. Depending
on which decoding is successful, indicated by a correct CRC check, the terminal can implicitly determine
40 ms timing or, equivalently, the two least significant bits of the SFN.3 This is the reason why
these bits do not need to be explicitly included in the MIB.

PBCH: How does the MIB tell the UE how many antennas are used in the cell?

Before answering this question let's review the contents of the Master Information Block (MIB). You will find the definition of the MIB in the RRC specification, 36.331. It contains just three parameters: DL Bandwidth, PHICH Configuration and System Frame Number as well as ten spare bits for future expansion. In a previous blog, "PBCH: How quickly can a UE read the MIB?" I discussed at length the SFN. The presence of the system bandwidth in the MIB is a reflection of the PBCH being in the center 1.4MHz regardless of the actual bandwidth of the channel. Why does the UE need the PHICH Configuration information to be in the MIB? Not because it will be receiving ACKs or NAKs immediately, but it must know WHERE this channel is so that it can read the PDCCH.
You have probably noticed that there is no parameter in the MIB for the number of antennas in the cell. The MIB has a CRC however, which is scrambled with one of three sequences which maps to the number of antennas used in the cell. Perhaps then, when the UE calculates the CRC from the decoded MIB it can compare against each of the three descrambled CRCs looking for a match and hence discover the number of antennas - perhaps not. It is tempting to think of this scrambling sequence as a parameter. But it is not a parameter in the same way that a C-RNTI used to scramble the CRC of the PDCCH is a parameter. Ask yourself the question "How does the UE decode the MIB in each of the three possible scenarios, namely, one antenna, two antennas or four antennas?" In the one antenna case there is nothing special the UE has to do, but for both the two and four antenna cases the UE has to be in sync with the base station. LTE uses a specific type of transmit diversity in which both the transmitter and receiver are aware of the method and participate in its application.  For the two antenna case LTE uses Space-Frequency Block Coding (SFBC) and for the four antenna case a combination of SFBC and Frequency Switched Transmit Diversity (SFBC+FSTD) is used. The LTE UE will need to blindly detect the number of antennas by trying each possible antenna configuration in turn, decoding the MIB and descrambling the MIB's CRC with the corresponding antenna mask in order to compare the CRC.
So the next time you are told that the MIB tells the UE how many antennas are used in the cell you'll know exactly what this means.