. 63
( 132 .)


timing of the cell. There are three steps to the above:

1. slot synchronization
2. frame synchronization
3. scrambling code identi¬cation.

The structure of the SCH and P-CCPCH is shown in Figure 6.32. The P-CCPCH is not
transmitted during the ¬rst 256 chips as the SCH is sent during this period.

6.9.2 Slot synchronization
The SCH consists of a primary SCH (P-SCH or SCH-1) and secondary SCH (S-SCH or
SCH-2). To obtain slot synchronization the UE listens to the P-SCH for a synchronization
code, which consists of 256 chips and is referred to as the primary synchronization code
(PSC). It is a system constant common to all BTSs and is easily identi¬ed (Figure 6.33).
A matched ¬lter is used, with the PSC as input, across the 5 Mhz frequency band. The
output of the ¬lter will be a series of peaks where there is a match between the PSC


P-CCPCH Tx off Broadcast data

1 slot (2560 chips)

Figure 6.32 Multiplexing of physical layer SCH and P-CCPCH



Slot 0 Slot 1 Slot 14

One frame

Figure 6.33 Format of synchronization channels

the UE is expecting and the P-SCH signal being transmitted. These peaks in the ¬lter™s
output, as illustrated in Figure 6.34, are therefore the start points of each slot, thus slot
synchronization has been achieved.

6.9.3 Frame synchronization
To obtain frame synchronization, the UE needs to listen to the S-SCH. Unlike the P-
SCH, which continuously sends the same 256 chips at the beginning of every slot, the
S-SCH rotates through a prede¬ned sequence of different chip groups, known as the
secondary synchronization codes (SSC). These sequences not only indicate the frame
synchronization but are also used to provide the UE with a guide for determining the
cell-speci¬c scrambling code. Table 6.9 shows the ¬rst two groups (0 and 1); there are

band Matched

Figure 6.34 Output of matched ¬lter

Table 6.9 SSC codes
Group Slot number
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Group 0 1 1 2 8 9 10 15 8 10 16 2 7 15 7 16
Group 1 1 1 5 16 7 3 14 16 3 10 5 12 14 12 10
Group 63 9 12 10 15 13 14 9 14 15 11 11 13 12 16 10

64 groups in total. Each UE will have this table at its disposal. Using the unique codes,
it is possible for the UE to obtain frame synchronization. It also completes the ¬rst step
of ¬nding the cell-speci¬c scrambling code.
There is a de¬ned set of possible 256-chip sequences that are used to represent each
of the 64 groups, and all sets of sequences are unique with respect not only to each
other, but also for every possible offset. As an example of how this mechanism works,
consider that there are 15 slots, and that instead of a 256-chip sequence, rather a letter is
transmitted in each slot, for example the word SYNCHRONIZATION. Then regardless of
which slot the UE starts decoding the S-SCH, it can calculate where the ¬rst slot occurs.
For example, consider the UE begins at slot 5, and then reads for 15 slots. It will then
have the sequence: HRONIZATIONSYNC. It will then compare this to the stored table
and therefore know where the ¬rst slot is positioned. Instead of just using one word, there
are 64 groups of different letter sequences de¬ned, for example.
Slot: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Group 0: S Y N C H R O N I Z A T I O N
Group 1: U E M O B I L E D E V I C E S
Group 63: M A N C H E S T E R U T D F C
Again, regardless of starting slot, the position of slot 1 can be found and the sequence
of letters that is yielded is not equal to any other of the 64 letter sequences, at any slot
position. Hence since the ¬rst slot is known, frame synchronization is achieved.
In a multi-sectored BTS site, a timing delay de¬ned by the variable T cell may be
used as an offset between the SCH channels in different sectors. This avoids overlapping
and interference of the SCH channels. T cell is of resolution 256 chips and can have a
value of 0“9.

6.9.4 Scrambling code identi¬cation
After frame synchronization, the UE also then knows which of the 64 words is being
used in this cell. There are 512 possible primary scrambling codes used in the system,
and these are broken up into 64 groups, each containing 8 codes. Therefore the code
group has narrowed the scrambling code down to 1 of a possible 8. The UE will now

look for the strongest correlation across the CPICH. The code which gives the best ¬t is
assumed to be the cell-speci¬c primary scrambling code. It is now possible for the UE to
decode the BCCH and thus the cell- and operator-speci¬c information can be read.

6.9.5 Random access procedure
It is now possible for the UE to receive and decode general information broadcast on the
downlink from the BTS. However, at some stage it will need to transmit on the uplink to
the BTS. Two questions then arise:

1. How does the cell know the scrambling code of the UE?
2. From where does the UE obtain its scrambling code?

Each UE in the cell must use a separate scrambling code. Although there are millions of
scrambling codes available, it is impractical to provide each UE in the world with a unique
code and place this in the SIM card. The scrambling codes used by the UE are actually
allocated by the RNC. However, since the UE has not begun to communicate, what
scrambling code should it now use? To overcome this problem, the UE must initially
use a scrambling code that the BTS understands for the initial request for connection.
The request is made via the RACH and thus this channel must use a code that the BTS
understands. The BTS will broadcast a preamble scrambling code that should be used,
as well as a number of signatures, which can be used on the RACH uplink channel by
the UE. The signatures are 16-chip sequences, and there is a maximum of 16 possible
signatures de¬ned. The UE selects one of these signatures at random and uses it to code
its request message to the BTS.
When this initial access is made, clearly there is a possibility of a collision with other
UE requests. By having a number of signatures, with one chosen randomly, this is mini-
mized somewhat; however, this alone is not enough. The RACH channel for initial access
consists of a number of access slots, based on a slotted ALOHA mechanism. In addition
to the signatures, the UE will also randomly pick one of these slots, further reducing the
probability of collision. There are 15 possible access slots across two frames, spaced 5120
chips apart, as shown in Figure 6.35.
The initial access is in the form of a 4096-chip preamble, which consists of the selected
signature repeated 256 times. The preamble is sent repeatedly until the network acknowl-
edges that it has heard it. It does this on the acquisition indication channel (AICH)
associated with the RACH, which will echo back the 4096 chips to the UE. The broad-
cast channel provides the UE with a power level for this initial access and a maximum
time, within which the access will be acknowledged if the network has heard the UE. If
there is no indication on the AICH within this time, the UE will power up, and send the
preamble again. Once an indication has been received on the AICH, the UE will then
send a 10 or 20 ms message part, which will typically be an RRC message requesting a
connection. The process is illustrated in Figure 6.36.
It is also possible for the RACH resources to be divided into a number of different
access service classes (ASC). This allows the resources to be given different priorities

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14

RACH initial access

RACH initial access

RACH initial access

one frame - 10ms one frame - 10ms
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14

Figure 6.35 RACH initial access

RACH preamble message

no indication
on AICH so
power up
AICH indication


Figure 6.36 RACH initial access procedure

for different types of access. This would enable, for example, emergency call access to
be given a higher priority level.


For handover measurements, the rake receiver has the necessary resources to simultane-
ously make measurements of neighbouring cells, while the user is on an active connection.
However, it can only perform this for cells that are at the same frequency as the current
connection. For inter-frequency measurements, the UE must temporarily cease transmis-
sion and reception to allow the measurements to be made, but the connection must be
maintained. Another alternative is for the UE to implement a second receiver for such
measurements, but this is somewhat impractical. UMTS allows for this pause in trans-
mission for inter-frequency measurement through compressed mode. This is used for

performing measurements of both cells within UMTS that are at a different frequency, as
well as inter-system measurements, for example to acquire GSM system parameters.
Compressed mode opens up periodic gaps in the data ¬‚ow during which such measure-
ments can be made. The RNC will instruct the UE and the BTS on the mechanism for
providing a gap, and the position of the gap. As the name suggests, compressed mode
takes the information to be sent within a 10 ms radio frame and compresses the time it
takes to send. There are three methods that can be used to perform this compression:

1. Puncturing: rate matching is used to further puncture the data to achieve the
necessary transmission gap.
2. Reduction of SF: the spreading factor is reduced by a factor of two during a
compressed frame. This has the effect of doubling the data rate. However, since the
amount of data has not changed, it halves the time in which it is sent, opening up
a gap.
3. Scheduling: higher layers can permit only some of the de¬ned transport formats to
be used when compression is required, thus generating a gap. Clearly this method is
only appropriate for variable-rate services.

Of these three, all are available for downlink compression, but method one, puncturing,
is not permitted in the uplink. The largest allowed gap that can be created is seven slots
in one radio frame.
Figure 6.37 shows an example of compressed mode. When the transmission gap is
opened up, if SF reduction is used, it will result in an increase in transmission power, as
illustrated. The gap can span two radio frames as long as there are at least 8 out of 15
slots in each frame containing transmitted data.
Compressed mode can be de¬ned to be used in downlink, uplink or both. When used
in both uplink and downlink simultaneously, the transmission gap must be coordinated to
occur simultaneously in both directions.
For each of the physical channels, it was seen that there are prede¬ned slot formats to
describe the size of the DPDCH data and DPCCH control bits. For each de¬ned format,
there are associated formats for use in compressed mode. Two formats are de¬ned for
each: format A is to be used where the puncturing method is employed, and format B
for SF reduction. As an example, consider the downlink slot format 9 that was shown


. 63
( 132 .)