Guide to CDMA2000 Network PN offset Planning (V1.0)Internally
Unclassified
Huawei Technologies Co., Ltd.
Radio Network Planning DepartmentDocument No.Product
VersionSecret Level
Product Name:18 Pages in Total
Guide to PN Offset Planning of CDMA Networks (V1.0)
(For Internal Use Only)
Prepared by:Wireless Network Planning
DepartmentDate:2002-07-10
Reviewed by:Date:yyyy/mm/dd
Reviewed by:Date:yyyy/mm/dd
Approved by:Date:yyyy/mm/dd
Huawei Technologies Co., Ltd.
All Rights ReservedRevision Record
DateRevised versionDescriptionAuthor
2002/07/101.00 betaFirst Draft finishedSun Jintao
2002/08/061.00ModifiedSun Jintao
Table of Contents51 General Description
52 PN Offset Planning Related Knowledge
52.1 Code Types Used in a CDMA System
52.1.1 Orthogonal Code- Walsh Function
62.1.2 Pseudo-Noise Code, M-Sequence
62.2 Search Windows
73 Purposes of PN Offset Planning
94 PN Offset Planning Analysis
94.1 Determining on PILOT_INC
104.1.1 PN Offset Separation between Pilots
104.1.2 PN Offset Reuse Distance between two Co-Offset
Pilots
104.1.3 Calculating the Lower Limit of PILOT_INC
124.1.4 Calculating the Lower Limit of the PN Offset Reuse
Distance between two Co-Offset Pilots
134.2 Engineering Approaches to PN Offset Planning
144.2.1 Approach 1 to PN Offset Planning
164.2.2 Approach 2 to PN Offset Planning
195 Sum-Up
Guide to PN Offset Planning of CDMA2000 Networks
Key words: CDMA, PN offset planning
Abstract: This document describes the PN offset planning
principles to be followed during the network planning from the
point of view of PN offset reuse distance and adjacent offset
separation, as well as engineering approaches to PN offset
planning.
List of Abbreviations: Describe the abbreviations used in the
document. It is required to list both the full English form and the
Chinese form of each abbreviation.
List of References: Please list the name, author, title, number,
release date and publishing unit etc. of each reference used in the
document in the table below.Table of References
TitleAuthorDocument No.
Release DateLook-up Place or Channel
Publishing Unit (Fill in this field if the document is not
published by our company.)
PN Phase Offset Planning of CDMA2000 NetworkLi Yunzhi
Technologies of CDMA (Code Division Multiple Access) Mobile
Communications SystemsSun Lixin, Xing NingxiaPeoples Post and
Telecom Publishing House
Design and Optimization of CDMA SystemKyoung I1 KimPeoples Post
and Telecom Publishing House
1 General DescriptionIn a CDMA system, a pair of m-sequences
with length 215 are used as the spectrum spreading codes for
forward and backward links, which are referred to as I and Q PN
sequences. This pair of m-sequences are also the pilot codes for
pilot channels, and different sectors are assigned different pilot
code phases.
Different sectors employ different phases of the m-sequence with
length 215 for modulation, with the phase difference between
adjacent phases required to be at least 64 bits. Hence, the maximum
number of available phases is: 215/64=512.
Although all sectors employ different PN offsets, at a MS, some
undesired pilot signal may act as strong interference due to either
a propagation delay (adjacent offset interference) or not long
enough reuse distance between two pilots that are assigned the same
PN offset (co-offset interference). There would be a propagation
delay before a pilot signal arrives at a MS, and if the propagation
delay difference between two pilots of two different sectors just
compensate for the time shift from one another, the MS would err in
synchronizing to the intended pilot signal and may switch to a
wrong sector, or even drop calls. Therefore, a detailed PN offset
plan should be developed for a CDMA system.
2 PN Offset Planning Related Knowledge
2.1 Code Types Used in a CDMA System
Two types of codes are used in a CDMA2000 system, i.e.
orthogonal code and pseudo-noise code.
2.1.1 Orthogonal Code- Walsh Function
An orthogonal code used in a CDMA2000 system is a Walsh
function, named after Walsh, a mathematician, who proved it in 1923
to be an orthogonal function. It is represented in the form of Wal
(n, t), where n is the sequence number.
The following recursion formula can be used to obtain 2N-order
Walsh functions:
, where N is the power of 2.In a CDMA system, the 1.2288Mbit/s
64-order Walsh function is used to spread the spectrum of each code
multiplexed forward channel so as to make code multiplexed forward
channels orthogonal to each other. Each forward channel in a sector
is assigned a Walsh code. A code multiplexed forward channel spread
by using 64-order Walsh function n (n=0-63) is defined as code
multiplexed channel n, wherein Walsh function n refers to row n+1
in the above Walsh function matrix. The sequence number in the
Walsh function of the pilot channel is 0, namely, Wal (64, 0).
2.1.2 Pseudo-Noise Code, M-Sequence
A pseudo-noise code (called PN code in short) has properties
similar to a noise sequence. It is actually a regular periodical
binary sequence though looking like a noise one. M-sequences are
the most important and fundamental among all pseudo-noise codes.
M-sequence is the abbreviation for Maximal Length Sequence of
Linear Feedback Shift Register . The definition of such a sequence
is as follows:
If the period of an output sequence of a N-stage linear feedback
shift register is P =2N -1, this sequence is called a
m-sequence.
In a CDMA system, usually Galois generators are used to produce
m-sequences each of which is comprised of the sequence of period
2N-1 (not all-0 state) and a mask. M-sequences with different masks
have different output phases.
Pseudo-noise codes used in a CDMA2000 system are of two types,
namely, m-sequence with length 215-1 and that with length
242-1.
On the forward link, the m-sequence with length 242-1 is used to
scramble fundamental channels while the m-sequence of length 215-1
is used for quadrature modulation of the forward link (the period
of the m-sequence is 215 with an all-0 state being added into it).
Different sectors use different phases of the m-sequence for
modulation, with the phase difference between adjacent phases
required to be at least 64 bits. Hence, the maximum number of
available phases is 512.
On the backward link, the m-sequence with length 242-1 is used
for direct spreading. Each MS is assigned a m-sequence phase
calculated by the ESN of the MS. Backward channels for these MSs
are basically orthogonal to each other for the two-valued
self-pertinency of m-sequence.
2.2 Search Windows
A CDMA system employs the synchronized detection technique on
the forward link. In order to successfully demodulate the intended
signal, a MS must be able to estimate the system time with
precision. The MS is receiving a pilot signal and will get the
estimation from this reference pilot channel. Using this reference
pilot as the reference signal, the MS can send and receive signals
with any PN code by modulating and demodulating the carrier
wave.
However, the pilot intended for a MS may not arrive precisely at
the anticipated time because the system time estimated by the MS
also includes the propagation delay of the reference pilot, and
besides, the time shifts of other pilots arriving at it are also
based on their propagation delays. The MS cannot tell the
propagation delay of a random given pilot, and therefore it must
search the pilot within an appropriate delay window till it detects
the actual time shift of the reference pilot, and this window is
referred to as the search window.
A MS would search the pilot using the following 3 different
types of search windows.
SRCH_WIN_A: used to search the activated and candidate pilot
set;
SRCH_WIN_N: used to search the adjacent pilot set;
SRCH_WIN_R: used to search the remaining pilot set.
The measurement unit for each window is chip. As illustrated in
the figure below, the center of a SRCH_WIN_A window is located near
the peak of the first arrived multipath component.
With the SRCH_WIN_A as the reference, the center of a SRCH_WIN_N
or SRCH_WIN_R is located about at the PN offset from the moment
when the intended pilot arrives. As illustrated in the figure
below, suppose a sector is assigned PN offset 4, and an adjacent
sector is assigned PN offset 20.
With a search window, a MS can detect multipath components
either before or after the arrival of the intended pilot, and
therefore, it can maintain a coherent detection regardless of the
propagation delay. The size of a search window must be set to the
effect that the best quality pilot signal should fall into it. If a
search window is too small, the important pilot signal will be
lost, thus resulting in interference; if it is too large, the
search efficiency will decrease, and the conversation quality will
be compromised.
Sizes of search windows are listed in the following table:
SRCH_WIN_A
SRCH_WIN_N
SRCH_WIN_RSearch Window Size (chips)SRCH_WIN_A
SRCH_WIN_N
SRCH_WIN_RSearch Window Size (chips)
04860
16980
2810100
31011130
41412160
52013226
62814320
74015452
3 Purposes of PN Offset Planning
As we know, sector separation for a CDMA2000 system is performed
by two m-sequences with length 215, which are phase shifted by
PN-offset for different sectors. Because there is a limited number
of PN offsets, a maximum of 512, a PN offset plan must be developed
to avoid any PN confusion.
As described above, although all sectors employ different PN
offsets, at a MS, some undesired pilot signal may act as strong
interference due to either a propagation delay (adjacent offset
interference) or not long enough reuse distance between two pilots
that are assigned the same PN offset (co-offset interference).
There would be a propagation delay before a pilot signal arrives at
a MS, and if the propagation delay difference between two pilots of
two different sectors just compensate for the time shift from one
another, the MS would err in synchronizing to the intended pilot
signal and may switch to a wrong sector, or even drop calls.
We will illustrate it with an example.
If we use C(0)(t) to represent a pilot signal with PN offset 0,
the pilot signal with PN offset i can be then represented by:
Suppose there are two cells with two different PN offsets that
are respectively and ; and are time delays respectively from the
base stations to the MS; and are the radiuses of coverage areas of
the two cells; represents one half of search window SRCH_WIN_A
while represents one half of search window SRCH_WIN_N; represents
the chip width. The measurement unit for all the above values is
chip.
Transmit powers of the two cells are p1 and p2 (dB). See the
figure below.
Figure 1 Analysis of interference between pilot signals of two
cells
Then, pilot signal of cell 1 is represented as:
Now, is the time delay. Suppose is the propagation loss, then is
the propagation loss in dB.
Then, the propagation attenuation slope of the pilot signal
received at the MS from cell 1 is:
Likewise, the propagation attenuation slope of the pilot signal
received at the MS from cell 2 is:
When arriving at the MS, the two pilot signals may have the same
PN offset if
In other words, if , the two pilot signals with different
offsets would have the same PN offset when arriving at the MS, thus
causing interference.
Hence, a detailed PN offset plan should be developed for a CDMA
system.
The PN offsets of pilot signals are usually represented by
offset indexes ranging from 0 to 511.
4 PN Offset Planning Analysis
In a CDMA system, two adjacent PN offsets must be separated from
each other at an interval of at least 64 chips.
1 chip=3108/1.2288M=244.14(m)
64 chips=64244.14=15.6(km)
In reality, due to complicated radio propagation environments
and limited sizes of MS search windows, the 15.6km separation is
not enough for distinguishing between two adjacent PN offsets in an
actual PN offset planning. For this reason, we use parameter
PILOT_INC to set the number of available PN offsets.
Available number of PN offsets=512/PILOT_INC.
4.1 Determining on PILOT_INC
The value of PILOT_INC determines the phase difference between
pilots of different cells. The less the PILOT_INC is, the more the
available PN offsets, and the further the PN offset reuse distance
between two co-offset pilots. Though it can reduce co-offset
interference, it can also reduce the phase difference between
adjacent pilots as well, thus being likely to result in a pilot
confusion. Therefore, we can analyze this issue to get the lower
limit of the PILOT_INC.
The greater the PILOT_INC is, the less the available PN offsets,
the less the remaining pilots, and accordingly, the shorter time a
MS would spend on pilot searching. Hence, in practice, the
probability of losing a strong pilot signal decreases. But this
improvement is only to a limited degree, because the priority for
the remaining pilot set is the lowest in pilot searching. Moreover,
the greater the PILOT_INC is, the less the available PN offsets,
thus resulting in a decrease in the PN offset reuse distance
between co-offset pilots and an increase in co-offset interference.
Therefore, the reuse distance between two co-offset pilots must be
appropriate.
In reality, we can compare the PN offset separation issue to the
adjacent frequency isolation issue in GSM systems, and the PN
offset reuse distance issue to the frequency reuse issue in GSM
systems.
According to the above analyses, how to select an optimal
PILOT_INC value is the key to PN offset planning.
Here below is how we should calculate the lower limit of
PILOT_INC.
4.1.1 PN Offset Separation between Pilots
According to the above analyses, the minimum PN offset
separation between two pilots determines the lower limit of
PILOT_INC. So, the restriction on the minimum PN offset separation
between two pilots should be taken into consideration first.
PN offset separation between two pilots should be mainly based
on the following principle:
When a pilot of another sector with a different PN offset is
present in search window SRCH_WIN_A of the current PN offset, the
interference of the current pilot by it should be below a certain
lower limit.
4.1.2 PN Offset Reuse Distance between two Co-Offset Pilots
Likewise, the PN offset reuse distance between two co-offset
pilots of different cells should be based on the following
principle:
The interference of the current pilot by a co-offset pilot of
another cell should be below a certain lower limit.
4.1.3 Calculating the Lower Limit of PILOT_INC
As shown in figure 1, suppose that the strength of the pilot
signal of cell 2 received at a MS in cell 1 is T (dB) lower than
that of the pilot signal of cell 1. If the pilot signal of cell 2
is required not to jam the pilot signal of cell 1, then:
; where Ls are propagation losses.Suppose P1 is equal to P2,
then:
Considering the propagation loss to be:
Where d is the distance between the base station and MS in
question.
Then, we have:
In the case of a direct radiation signal, To make the above
inequality equation true under any circumstances, this inequality
equation must hold true even if assumes the maximum value and
assumes the minimum value at the same time.
Consider the worst case in which interference is most likely to
occur: a MS is located at the boundary of cell 1, in which case the
pilot signal of cell 1 arriving at it is the weakest, and
therefore, the interference likelihood is the highest, namely, in
this case. In addition, the MS search window is just large enough
to let the pilot signal of cell 2 enter into it,
Namely:
According to the above equation, we have:
Then:
Because:
Then:
; where is the propagation attenuation slope of a radio wave.
Generally speaking, =4.3 in a densely populated downtown area;
=3.84 in a suburb. is related to such system parameters as T_ADD
and T_DROP, and =24dB in most cases.
According to the above inequality equation, the lower limit of
PILOT_INC is related to the coverage area radius, r, propagation
attenuation slope as well as size of the MS search window. The less
the coverage area radius, r, is, or the greater the attenuation
slope is, or the smaller the search window ARCH_WIN_A is, the less
the lower limit of PILOT_INC would be.
Suppose that radiuses of coverage areas of all cells are
identical. We then have:
In a densely populated downtown area: In a suburb: ; where r is
the coverage area radius of the cells, and S1A represents one half
of search window SRCH_WIN_A (in chips).
For a densely populated downtown area, if r is 500m (about two
chips), we can draw the following conclusions from the above
inequality equation:
Coverage area Radius r (m)r (chips)SRCH_WIN_A
SRCH_WIN_A Size (chips)S1APILOT_INCLower Limit of PILOT_INC
Number of Available PN Offsets
5002.0480420.178938321512
5002.0481630.194563321512
5002.0482840.210188321512
5002.04831050.225813321512
5002.04841470.257063321512
5002.048520100.303938321512
5002.048628140.366438321512
5002.048740200.460188321512
5002.048860300.616438321512
5002.048980400.772688321512
5002.04810100500.928938321512
5002.04811130651.163313322256
5002.04812160801.397688322256
5002.048132261131.913313322256
5002.048143201602.647688323170
5002.048154522263.678938324128
For an open field, if r is 10Km (about 41 chips), we can draw
the following conclusions from the above inequality equation:
Coverage area Radius r (m)r (chips)SRCH_WIN_A
SRCH_WIN_A Size (chips)S1APILOT_INCLower Limit of PILOT_INC
Number of Available PN Offsets
1040.960423.3701076224128
1040.961633.3857326224128
1040.962843.4013576224128
1040.9631053.4169826224128
1040.9641473.4482326224128
1040.96520103.4951076224128
1040.96628143.5576076224128
1040.96740203.6513576224128
1040.96860303.8076076224128
1040.96980403.9638576224128
1040.9610100504.1201076225102
1040.9611130654.3544826225102
1040.9612160804.5888576225102
1040.96132261135.104482622685
1040.96143201605.838857622685
1040.96154522266.870107622773
According to the above statistics, if r=10km, SRCH_WIN_A=6 (the
size of SRCH_WIN_A is 28 chips), and PILOT_INC3.5576, namely,
PILOT_INC=4.
4.1.4 Calculating the Lower Limit of the PN Offset Reuse
Distance between two Co-Offset Pilots
According to the said principle, i.e. the interference of the
current pilot by a co-offset pilot of another cell should be below
a certain lower limit. In other words, the PN offset reuse distance
between co-offset pilots is required to be long enough.
As illustrated below in Figure 2, cells 1 and 3 use the pilots
assigned the same PN offset; D (A, N) represents the distance
between point A and cell N; D represents the distance between cells
1 and 3; is the coverage area radius of cell i; is the relative
offset of cell 1; is the time delay from base station i to the MS;
represents one half of search window SRCH_WIN_A while represents
one half of search window SRCH_WIN_N; represents the chip width.
The measurement unit for all the above values is chip.
is the transmit power of cell i; is the propagation loss
attenuation slope.
Figure 2 Calculation of the PN offset reuse distance between
co-offset pilots
Here below is how we should calculate the lower limit of D.
II. To prevent a MS from being unable to distinguish between two
co-offset pilots within search window SRCH_WIN_A
Consider the worst case in which interference is most likely to
occur: a MS is located at the boundary of cell 1 and aligned with
cell 3. Now the MS is about to be handed over for cell 2 to cell 1,
in which case the pilot signal of cell 1 arriving at it is the
weakest and the interfering signal from cell 3 is the strongest;
and therefore, the interference likelihood is the highest, namely,
D (A, 1)=r1 and D (A, 3)=D-r1. In addition, search window
SRCH_WIN_A is just large enough to let the pilot signal of cell 3
enter into it. Then, we have:
Namely,
(4.1.41)
III. To prevent a MS from being unable to distinguish between
two co-offset pilots within search window SRCH_WIN_N
To prevent co-offset pilot signals from both falling into search
window SRCH_WIN_N of the MS at point A in cell 2, one of pilot
signals of cells 1 and 3 should fall out of search window
SCH_WIN_N. In other words, one of distances from cells 1 and 3 to
cell 2 should be:
Consider the worst case, namely, cells 1, 2 and 3 are aligned,
and suppose r1=r3=r, then:
(4.1.4--2)
Compare two inequality equations (4.1.4-1) and (4.1.4-2);
obviously, (4.1.4-2) should be used to calculate the lower limit of
PN offset reuse distance.
4.2 Engineering Approaches to PN Offset Planning
In PN offset planning, we should first of all settle on
PILOT_INC.
According to the previous analyses of the PILOT_INC lower limit,
the PILOT_INC is usually set to 4 in practical applications, in
which case the number of available PN offsets is 128 and the
adjacent offset separation is 256 chips, equivalent to 62km; in
other words, a pilot signal must travel at least 62Km in order to
fall into search window SRCH_WIN_A for another pilot. This can work
in most of network applications. Moreover, to avoid a networking
complication, two adjacent networks had better employ the same
PILOT _INC so as to prevent MSs from searching the remaining pilot
set (a MS shall search the remaining pilot signals by integral
multiples of PILOT_INC).
Having determined the value of PILOT_INC, we can then proceed to
the PN offset reuse model. If PILOT_INC=4, the number of available
PN offsets is 128. When three cells are taken into consideration,
there can be a maximum of 42 cells in each reuse cluster.
4.2.1 Approach 1 to PN Offset Planning
If N is the number of cells in a reuse cluster, it should be
; where i and j are positive integers. We must follow the
procedure below to find out the adjacent co-offset cells nearest to
a certain cell: 1, move across i number of cells from each link on
the regular hexagon (each borderline of the hexagon); 2, turn 60
counterclockwise and then move across j number of cells. See figure
below, wherein i=3 and j=2(N=19)
Figure 3 Locating co-offset cells in a reuse cluster
If i=4 and j=3, N=37.
Figure 4 Approach 1 to PN offset planning
In this case, the number of PN offsets required is 111. Because
the number of available PN offsets is 128, the remaining PN offsets
are reserved for future use. In practical applications, some PN
offsets should be reserved for use in future capacity expansions or
of microcells. Or, if some antennas must be mounted on higher
levels because of environmental restrictions, to avoid interference
of other cells by them, PN offsets not used by cells in the reuse
cluster can be used by them. PN offsets are typically grouped as
follows:
128 PN offsets are divided into two groups, with 111 of them
being used by cells in the reuse cluster while the remaining 17 of
them being reserved PN offsets. The 111 PN offsets are further put
into three groups. Based on the principle that adjacent PN offsets
of sectors should be separated as far from each other as possible,
and to standardize data settings, PN offsets are assigned as in the
following table. Numbers in a column for a sector represent the PN
offset serial numbers assigned to different sectors in the cell in
question.
Cell No.Sector 1(Group A)Sector 2 (Group B)Sector 3 (Group
C)
Cell 14172340
Cell 28176344
Cell 312180348
............
Cell nn*4(n+42)*4(n+84)*4
............
Cell 37148316484
In approach 1, there are 37 cells in each reuse cluster. Since
unified cell numbering is complicated, the rule below should be
followed to number cells.
As illustrated below, cell n+1 is on the southwest (S60W) of a
given cell n with two cells on the south skipped. If cell n+1 is
not within the reuse cluster, the same rule should be followed to
number it.
Figure 5 Cell numbering rule
The above cellular distribution is an ideal one that does not
exist in reality; therefore, it is not easy to practice this
approach to PN offset planning. Here below is another approach,
used more often in engineering, to PN offset planning.
4.2.2 Approach 2 to PN Offset Planning
We put 42 cells in one reuse cluster, and divide this cluster
into several sub-clusters. There is no absolute rule on the number
of sub-clusters in each cluster (or, the number of cells in each
sub-cluster). Usually, in a densely populated downtown area, there
are relatively more cells in each sub-cluster. In network planning,
42 cells are usually put into 4 sub-clusters with each of them
being comprised of 10~11 cells.
Sub-cluster 1
Sector 142036526884100116132148164
Sector 2172188204220236252268284300316332
Sector 3340356372388404420436452468484500
Sub-cluster 2
Sector 182440567288104120136152168
Sector 2176192208224240256272288304320336
Sector 3344360376392408424440456472488504
Sub-cluster 3
Sector 1122844607692108124140156
Sector 2180196212228244260276292308324
Sector 3348364380396412428444460476492
Sub-cluster 4
Sector 1163248648096112128144160
Sector 2184200216232248264280296312328
Sector 3352368384400416432448464480496
The distribution of cells within each sub-cluster is illustrated
in the following figure (sector 1 is used as the example):
Figure 6 Distribution of cells in a sub-cluster
As shown above, cells in a sub-cluster are distributed in a
spiral (please refer to the application case at the end of this
section).
As shown below in the figure, several kinds of reuse cluster
layouts are available (sector 1 is used as the example):
Figure 7 Comparison between reuse cluster layouts
By comparing the above three reuse cluster layouts, we can
conclude that the PN offset reuse distance between co-offset pilots
in the second layout can reach a maximum of 15.2r (where r is the
cell coverage area radius). Hence, this layout is usually adopted
in engineering. See figure below (sector 1 is used as the
example):
Figure 8 Frequently used cluster layout
Here below is an application case of PN offset planning:
Figure 9 Approach 2 to PN offset planning
5 Sum-Up
An appropriate PN offset plan should be made at the initial
stage of the system design. An inappropriate PN offset plan may
result in network interference and modifications of a large number
of network data in future network optimization, thus making network
maintenance a difficult job.
Before an IS-95 system is upgraded (moved or expanded in
capacity) to a CDMA 1X system, the existing PN offset planning data
should be collected. It is recommended that the same principles be
followed for re-planning to avoid confusions caused by PN offsets
within search windows in adjacent cells, which may result in
interference between a large number of signals within the upgraded
network, and even drop calls.
A description on pages appended to a document or revisions of it
due to version upgrades, which is usually included in a design
document.
2002-9-3
2002-08-06All Rights ReservedPage 18 of 18
_1091537996.vsd
_1092206578.unknown
_1091538000.xlsSheet1
148164152168148164152168
1322036136244013220361362440
116452120856116452120856
1008468104887210084681048872
156160156160
1402844144324814028441443248
1241260128166412412601281664
1089276112968010892761129680
152168148164152168148164
1362440132203613624401322036
120856116452120856116452
1048872100846810488721008468
160156160156
1443248140284414432481402844
1281664124126012816641241260
1129680108927611296801089276
_1091537994.vsd