WCDMA / UMTS. Principle of Spectrum Spreading. Frequency used

WCDMA / UMTS UMTS (Universal Mobile Telecommunications System) is a new mobile standard. We talk about the third generation of telephony or 3G. Purists prefer the term W-CDMA (Wideband Code Division Multiple Access) which takes the name of the technology deployed in Europe and some Asian operators. Its principle: use a frequency band wider in order to transmit more data and thus obtain a higher datarate. In theory, it can reach 2 Mbps, a transmission speed equivalent to that proposed for the Internet reached by ADSL or cable. Principle of Spectrum Spreading Frequency used 1

FDD Technical summary Frequency band:1920 MHz -1980 MHz and 2110 MHz - 2170 MHz (Frequency Division Duplex) UL and DL Minimum frequency band required: ~ 2x5MHz Frequency re-use: 1 Carrier Spacing: 4.4MHz - 5.2 MHz Maximum number of (voice) channels on 2x5MHz: ~196 (spreading factor 256 UL, AMR 7.95kbps) / ~98 (spreading factor 128 UL, AMR 12.2kbps) Voice coding: AMR codecs (4.75 khz - 12.2 khz, GSM EFR=12.2 khz) and SID (1.8 khz) Channel coding: Convolutional coding, Turbo code for high rate data Duplexer needed (190MHz separation), Asymmetric connection supported Tx/Rx isolation: MS: 55dB, BS: 80dB Receiver: Rake Receiver sensitivity: Node B: -121dBm, Mobile -117dBm at BER of 10-3 Data type: Packet and circuit switch Modulation: QPSK Pulse shaping: Root raised cosine, roll-off = 0.22 Chip rate: 3.84 Mcps Channel raster: 200 khz Maximum user data rate (Physical channel): ~ 2.3Mbps (spreading factor 4, parallel codes (3 DL / 6 UL), 1/2 rate coding), but interference limited. Maximum user data rate (Offered): 384 kbps (year 2002), higher rates ( ~ 2 Mbps) in the near future. HSPDA will offer data speeds up to 8-10 Mbps (and 20 Mbps for MIMO systems) Channel bit rate: 5.76Mbps Frame length: 10ms (38400 chips) Number of slots / frame: 15 Number of chips / slot: 2560 chips Handovers: Soft, Softer, (interfrequency: Hard) Power control period: Time slot = 1500 Hz rate Power control step size:, 1, 1.5 and 2 db (Variable) Power control range: UL 80dB, DL 30dB Mobile peak power: Power class 1: +33 dbm (+1dB/-3dB) = 2W; class 2 +27 dbm, class 3 +24 dbm, class 4 +21 dbm Number of unique base station identification codes: 512 / frequency Physical layer spreading factors: 4... 256 UL, 4... 512 DL Scheme of WCDMA transmitter 2

1. Spread spectrum and CDMA multiple access technique 1.1 Principles of spread spectrum The spread spectrum can be defined as a technique used to transmit an information signal over a bandwidth many times higher than the minimum bandwidth that the signal demands. For this reason, the spread spectrum is also considered as a modulation. In a spread spectrum system, the transmitted signal is "spread" with a code that is independent of the information message. After synchronization with the transmitter, the receiver must use the same code to "despread" the signal and to recover the information message. To understand why the spread spectrum has been so successful, the work of Claude Shannon should be reread, he was the first that formalized the concept. Starting from the famous expression that determines the capacity of a channel: C = B*log 2 (1+(S/N)) where C is the channel capacity in bits per second, B the bandwidth of the transmitted signal in Hz, S is the signal power in Watt, N the noise power in Watt. By developing this function in series, we can easily show that: C B/ln(2)*(S/N) S/N C*ln(2)/B (S/N <<1) The expression shows that there is an inverse link between the bandwidth B occupied by the transmitted signal and the signal to noise ratio S/N, which is measured at the receiver. Specifically, we observe that if B increases, a smaller signal to noise ratio is needed to have the same channel capacity C. Thus the virtues of spread spectrum are demonstrated: the spectral bandwidth is increased in order to reach good performances at the reception side, because the signal to noise ratio is reduced to a minimum. 1.2 CDMA and WCDMA techniques The CDMA (Code Division Multiple Access) is a multiple access technique that allows users to communicate simultaneously in the same frequency band. The distinction between different users is done with an dedicated code which is only known by the transmitter and receiver. 3

Transceiver Channel Receiver Spreading C1 Despreading C1 Bandwidth Baseband Channel noise RF frequency Baseband Spreading C2 Despreading C2 Figure 1: CDMA system The CDMA is used by the second-generation American standard CDMA One. It is also the access mode chosen for the radio access network of UMTS. Depending on the bandwidth occupied by the spread signal, a CDMA system is "wideband" (wideband CDMA) or, conversely, "narrowband" (narrowband CDMA). The two radio access technologies of the UMTS UTRA-FDD and UTRA / TDD are wideband CDMA systems, while CDMA One is an example of narrowband CDMA system. CDMA One bandwidth is 1.25 MHz while UTRA-FDD and UTRA / TDD bandwidth is 5 MHz. 1.3 Direct Sequence Spread Spectrum technique: DS-CDMA The spread spectrum is considered as a kind of modulation, because the original message is transformed so that the spectral width after processing is several times larger than the original message. This transformation can be performed in several different ways (FH-CDMA, TD-CDMA, DS-CDMA). In a DS-CDMA (Direct Sequence CDMA) system, the information signal is directly modulated by a sequence or code that has particular statistics properties. Spreading : Data Code with length L Rate Rate Spreading signal Baseband RF frequency Figure 3: Direct Sequence Spread Spectrum 4

The chips of the spreading code have a duration Tc, and are usually pulses +1 and -1. The chip rate Bc is equal to 1/Tc and expressed in chips per second (cps). The length of the information symbols is Ts and the symbol rate Bs = 1/Ts is expressed in symbols per second (sps). We define the "spreading factor" SF as the ratio between chip rate and symbol rate: SF = Bc/Bs = Ts/Tc equals to the length L of the code used. It is important to note that the symbol rate Bs is variable and depends on the service (voice, video, data), while the chip rate Bc = 1/Tc is constant. Thus, a high datarate service will require fewer chips to encode a information symbol than a low-rate service. For FDD and TDD modes of UMTS, the chip rate is Bc = 3.84 Mcps. Once the code is generated, the spreading is done in a very simple manner: simply multiply the message with the chips that take +1 and -1 as value. A narrowband signal is transform in a wideband signal, because Bc> Bs or, equivalently, Ts> Tc (see Figure 3). Finally, the spread signal is transposed at the carrier frequency fc. At the reception side, the signal is first converted into baseband signal with the RF demodulator, and then multiplied by the same spreading code used at the transmitter side. Since the spreading code is composed of 1 and -1, the result of multiplying the received signal with the same sequence has the effect of removing the code contribution in the received signal and keeps only the information message. Because the information message has the symbol rate, and not the chip rate, the signal is despread. This procedure of "despreading" is only possible if the receiver is perfectly synchronized in time with the transmitter with an accuracy less than a chip (see Figure 4). Code with length L Received signal Jammer Spreading signal RF frequency Baseband Figure 4: Spectrum despreading / Narrowband jammer effect The "spread" signal resists well to interferences, which have a smaller spectral width. This robustness comes simply because the information is "spread" on a frequency band rather important: only a small part of the spread signal spectrum is affected. After dispreading the useful signal, the jammer is spread, making the reception more efficient. 5

1.4 Spreading codes used in CDMA The choice of the spreading codes is directly influenced by their correlation properties, and specifically by their autocorrelation and crosscorrelation properties. From a statistical point of view, the autocorrelation is a measure of the correspondence between a code and a time shift version of it. Let s have a code ci with a length M and ci = [ci(0), ci(1),..., ci(m-1)]. We define its autocorrelation function Rci (t) as follows: Rci(t) = ci(0)*ci(t)+ci(1)*ci(1+t)+...+ci(m-1-t)*ci(m-1) Furthermore, the crosscorrelation represents the degree of correspondence between two different codes. The mathematical expressions of the crosscorrelation function Rcicj(t) of two codes ci and cj with a length M such that ci = [ci(0), ci(1),..., ci(m-1)] and cj = [cj(0), cj(1),..., cj(m-1)], is written: Rcici(t)= ci(0)*cj(t)+ci(1)*cj(1+t)+ +ci(m-1-t)*cj(m-1) If the crosscorrelation function between two different codes ci and cj (i j) satisfies Rcicj (0) = 0, we say that the two codes are orthogonal. When the codes are not completely orthogonal, a crosscorrelation residue appears between the useful user code and other users codes. This "residue" implies a kind of interferences which is called multiple access interference. In most radio systems based on CDMA, the codes used to perform the spreading process are two kinds of codes: the Walsh-Hadamard or orthogonal codes and the "pseudo-random" codes. 1.4.1 Walsh-Hadamard codes Orthogonal codes used in CDMA are chosen among those proposed by Walsh in 1923. There are different ways of generating Waslsh sequences. The generation of Hadamard matrix is one of the most famous. Hadamard matrixes are always square and can be recursively generated: H 2M = H M H M H M -H M where M is of power of 2. For example, if we want to have 4 orthogonal codes c1, c2, c3 et c4, we have to generate an order 4 Hadamard matrix. So we get: H 1 = +1 H 2 = +1 +1 +1-1 H 4 = +1 +1 +1 +1 +1-1 +1-1 +1 +1-1 -1 +1-1 -1 +1 where c 1 = [+1 +1 +1 +1], c 2 =[+1-1 +1-1], c 3 =[+1 +1-1 -1], c 4 =[+1-1 -1 +1]. Because of this method of generation, the Walsh codes are sometimes called the Walsh-Hadamard codes. 6

A Walsh-Hadamard code ci has the following characteristics: 1) the first element ci is always +1, for all i ; 2) ci has the same number of + 1 and -1 (excepted c 1 ) ; 3) The intercorrelation function of two codes satisfies: Rcicj (0) =0, for all i j ; 4) The crosscorrelation function Rci(0) of the code ci satisfies: Rci (0) = M Drawback of orthogonal codes : We assume that the receiver knows the beginning of the code to perform the corresponding crosscorrelation function Rcicj(0). But in practice, the receiver must perform a synchronization phase before proceeding the extraction of the signal. Furthermore, in a multipath channel, even if the receiver is perfectly synchronized in time, for example on the most powerful path, the other paths are not received at the same time and the orthogonality between different copies of the signal is lost. The result is a kind of interference which may be more or less inconvenient and is called inter-path interference. In these conditions, a major drawback of Walsh-Hadamard codes is that they have poor autocorrelation properties. Furthermore, we can see that the same problem occurs between two distinct Walsh-Hadamard codes, which means that these codes do not have good crosscorrelation properties. For example, if we consider the code c3 = [+1-1 -1 +1] one chip shifted to the right, c'3 = [+1 +1-1 -1], the crosscorrelation function between this code and c2 = [1 + 1-1 -1] gives Rc 2 c 3 (0) = (+1 +1)+(+1 +1)+(-1-1)+(-1-1) = 4 In other words, c2 and c'3 codes are not orthogonal. 1.4.2 Pseudo-random codes A pseudo-random code is a binary code whose statistical properties are similar to those of a random binary sequence, but a pseudo-random code is periodic and the larger is its period, the more similar to a true random binary signal it is. A binary shift register is one of the most common ways to generate pseudo-random codes (see Figure 6). Its operation is the following: once initialized the different states of the register, the bit output is calculated at each clock time by adding modulo 2 all bits present in each state. The bits are then circular shifted to restore the states and calculate the next output bit. The generated sequence depends on the initial state of the register, the number of registers n and the values (αi): either 1 or 0. Input Output Figure 6: Shift register scheme with 4 states feedback (1+X 2 +X 4 ) 7

This register whose full name is linear feedback shift register (LFSR) is periodic because, whatever the initial values, ie the values (αi), we find the same values after a finite number of clock time. So the sequence period is not greater than 2 n. A binary sequence with a period P = 2 n -1, generated with an LFSR register is called a msequence or maximum length sequence. The generator of m-sequences can also be represented using a polynomial notation: f(x) = X n +α n-1 X n-1 + α n-2 X n-2 +...+α 1 X+1 The interest of m-sequences in their use CDMA derives the following properties: the m-sequences are balanced. The number of 1 is equal to the number of 0 plus one. In the frequency domain, this property provides a symmetric spectrum with a DC value close to zero. the autocorrelation function of an m-sequence (binary) ui can have two values over a period P = 2 n 1: Rvi(t) = P if t = 0 et -1 if t 0, where vi = 1-2ui Autocorrelation peak becomes more pronounced when the period P is increasing and therefore the autocorrelation function becomes closer to that of the autocorrelation function of a white noise (this function is zero for t 0). The crosscorrelation properties between two m-sequences are not good because some quite pronounced peaks may appear. So, two m-sequences generated or not by the same primitive polynomial are not orthogonal. To improve the crosscorrelation properties of m-sequences, another kind of codes was proposed: the Gold codes. The main quality of Gold codes is related to the fact that the crosscorrelation function between two codes is uniformed and bounded. Gold proposes to add modulo 2 two m-sequences with the same length, generated from two different primitive polynomials. If the period of the two sequences is 2 n -1, the resulting sequence will also have a period 2 n -l. Considering all the shifts between the two m-sequences, we can generate 2 n -1 different codes, and therefore, with two shift registers, we can generate a total of 2 n +1 codes (the 2 n -1 previous codes plus the two generating m-sequences). Gold shows that some pairs of m-sequences have crosscorrelation peaks whose amplitude is equal to or less than the minimum crosscorrelation value between all pairs of m-sequences with the same length. This pair of m-sequences is called preferred pair. Gold also demonstrates that the crosscorrelation function of a preferred pair is predictable and can take three values: -f (n), -1 and f(n)-2, where f(n) = 1+2 (n+1)/2 for n even and 1+2 (n+2)/2 for n odd. 1.5 Spreading codes used in UMTS In UTRA / FDD and UTRA / TDD, the spread spectrum is performed using two codes: one of Walsh-Hadamard and the other of Gold. The function of each code depends on the link (up or down) (see Figure 7). Chips [cps] Chips [cps] Data [bps] Channelization code (Walsh-Hadamard) Scrambling codes (Gold) Figure 7: Channelization code et scrambling code 8

1.5.1 Channelization codes Channelization codes, also called orthogonal codes or spreading codes, allow the transition from symbols to chips, ie from a narrowband signal to a wideband signal. The number of chips per information symbol is called the Spreading Factor SF. The spread function is performed by multiplying, at the chip rate, the real (I component) and imaginary (Q component) of each symbol with the allocated channelization code (length SF). In the downlink, channelization codes identify the information symbols of each user in a cell. In the uplink, they help to differentiate the information symbols from the channel or dedicated physical data channels (DPDCH) and dedicated physical control channels (DPCCH) belonging to the same user. In the downlink, it is possible to use orthogonal channelization codes to identify users, because the network decides when the information is transmitted to the mobile (the spread signals are synchronous). In the uplink, the transmission from the mobile to the base station is not synchronized and signals that have been spread with the channelization codes are no longer orthogonal from one user to another. Also, we prefer to use scrambling codes to distinguish users due to their crosscorrelation properties. Moreover, in the uplink, different users in the same cell can use the same channelization code to transmit: it is the scrambling code that identifies them. The symbol rate Rs is: Rs = Rc/SF where Rc is the chip rate which is constant and equal to 3.84 MHz. Because of the wide variety of services and therefore, dataflow, it is necessary to generate channelization codes with a variable SF while preserving their orthogonality. To do this, we use the OVSF method (Orthogonal Variable Spreading Factor). The codes generated are called OVSF codes. OVSF codes have the particularity of being an orthogonal codes family in the sense of correlation. They are defined by a generator tree. The principle of this tree is similar to the Walsh-Hadamard matrices of (Figure 8): Root 2 branches Figure 8: OVSF codes generation One root voice generates two branches. The code of a branch is composed of the code of the root and its complementary. This principle is used to generate the OVSF code tree used for the UTRA (see Figure 9). 9

Used spreading factor Figure 9: OVSF tree In direct sequence spreading (DS-CDMA), the code length is equal to the spreading factor SF. Furthermore, the factor k, which determines the number of bits in the DPCH frame, verifies the following equation: SF = 256/ 2 k with k = 0.. 6. This means that SF can take the values: 4, 8, 16, 32, 64, 128, 256 in FDD. This tree (see Figure 9) shows the direct link between the number of codes available for a given spreading and the spreading factor. Indeed, the factor SF simultaneously determines the code length but also the number of codes available for a spreading SF. To use codes, it is necessary to respect certain rules. In a single cell, the OVSF codes can be used all at the same time because they are not all orthogonal to each other. The code of one branch is strongly linked to its root. So, a rule for the use of CDMA codes is defined: when a code is allocated, all the codes from its branches can not be used (see Figure 10). This code can not be used in the cell Figure 10: OVSF code allocation In figure 10, the code C 16,1, 1 is allocated, thus all the codes from it may not be used: C 32,1, C 32,2, etc.. But root codes of C 16,1 1 in the tree can not also be used. Indeed, if we used C 4,1, for example, the codes from these branches would no longer be usable, so C 16,1 could not be used. This rule for the use of CDMA codes provides a strong constraint on the availability of high rate channels. This rule indicates that the number of simultaneous users in downlink for a service of 2 Mbps (SF = 4) can not exceed 4 users. 10

1.5.2 Scrambling codes The signal spread by the channelization codes is multiplied chip by chip by another scrambling code. This code is a pseudo-random codes whose crosscorrelation properties are higher than those obtained by the channelization codes. Scrambling codes, which are not orthogonal, are derived from Gold sequences (see Figure 11). For both uplink and downlink, each scrambling code chip is represented by a complex value. Figure 11 : Scrambling code generator In the downlink, the scrambling codes are used to identify a cell. There is therefore a unique scrambling code per cell. With the scrambling code, it is possible to reuse the OVSF tree in the same geographical area. Furthermore, the use of the scrambling codes makes the statistical properties of interference close to those of a Gaussian noise, (this interference is caused by neighbouring cells that emit on the same carrier frequency). The scrambling code can therefore help the receiver to detect the transmitted symbols. In the uplink, the scrambling codes are used not only to improve the statistical properties of the channelization codes but also to differentiate each user in the same cell. A summary of the use of codes is given in Figure 12. Figure 12: Use of channelization codes (OVSF) et scrambling codes (SC) in UMTS cells 11

2. UMTS radio transmission chain 2.1 Radio characteristics The main radio characteristics of the UMTS/FDD system are given in the following tab. Multiple access technique Duplexing mode Frequency hopping Chip rate Spreading Factor Carrier frequency spacing Frequency spectrum in Europe (MHz) Carrier duplex space between UL and DL Data modulation Radio modulation Power control frequency Frame duration Slot duration FDMA/CDMA FDD Not neccessary 3.84 Mcps From 512 to 4 in DL and from 256 to 4 in UL 5000 khz 1920-1980 in UL and 2110-2170 in DL 190 MHz BPSK in UL and QPSK in DL QPSK 1500 Hz 10 ms 10/15 = 0.667 ms The time unit in a UMTS network is given in number of chip time which is constant and equal to 1/3.84 MHz = 0.26 µs. The following metrics are also referenced to determine the physical channels duration (see Figure 13): A radio frame is a period of 15 slots (or 38,400 chips or 10 ms). This temporal metric is fundamental for the coding and the channel interleaving. A slot is a fixed length time interval equal to 2560 chips. The number of bits per slot is variable. Figure 13: Frame and slot 12

2.2 Modulation of Dedicated physical channels 2.2.1 Dedicated physical channels in the uplink There are two kinds of dedicated physical channels in the uplink: the Dedicated Physical Data Channel DPDCH and the Dedicated Physical Control Channel DPCCH. In a single radio link, there may be zero, one or more DPDCH and always a single DPCCH. Several DPDCH is used by the mobile in case of multicode transmission. The DPDCH transmits the dedicated transport channels information, while the DPCCH carries the signal or control information generated by the physical layer and the TFCI bits provided by the MAC layer. The control information contained in the bit fields of the DPCCH included (see Figure 14): The pilot bits known to the receiver of the base station. The BS uses these bits to estimate the impulse response of the propagation channel. The pilot bits are also used to estimate the signal to interference ratio of in the radio link. An identifier of the combination of transport formats represented by the bits TFCI (Transport Format Combination Indicator). By using these bits, the base station determines the attributes that the data bits conveyed by the DPDCH were processed by the transport channels (multiplexing and channel coding operations). The feedback bits FBI (FeedBack Information). Power control commands represented by TPC (Transmit Power Control) bits. The mobile uses these bits to request the node B to increase or decrease the transmission power of DPCH in order to guarantee a certain level of quality in the received signal. There may be 1 or 2 TPC bits in a slot, bits that are transmitted and estimated slot by slot, so base station power control is performed with a frequency of 1500 Hz Figure 14: Dedicated physical channels in the uplink The DPDCH and DPCCH are transmitted in parallel, the first on the I component, the second on the Q component. The information on each component is spread with a different channelization code allowing the UTRAN to distinguish the information transmitted by the data channel and the information carried by the control channel. The upper layers of the mobile choose the SFvalue and choose the channelization code in the OVSF code tree. It is then used to spread each symbol. The selection of the OVSF code of DPDCH is restricted to C SF,SF/4 while the DPCCH channel is always spread with code C 256,0. The complex scrambling operation is applied with the sequence Cn (i), i = 0... 38399 (see Figure 15). 13

Figure 15: Example of an uplink processing chain The modulation that consists of transmitting data and control information in the independent I and Q components is sometimes called dual-channel QPSK modulation. It's equivalent in practice to the combination of two BPSK modulations - the one being applied to the binary information of the DPDCH and the other binary information of the DPCCH. In most of cases, the number of bits per slot of the DPDCH or DPCCH depends on the binary datarate that each of these two channels transmits, and it depends on the spreading factor SF following the relationship: N = 2560/SF 2.2.2 Dedicated physical channels in the downlink In contrast to the uplink, there is only one type of dedicated physical channel in the downlink called DPCH (Dedicated Physical Channel). This channel transmit the transport channel DCH information (see Figure 16) - information which may be data traffic or signaling traffic generated by the upper layers. It also transmits control information generated by the physical layer itself, and it can be considered as a time division multiplexing of a dedicated physical data channel (DPDCH) and a dedicated physical control channel (DPCCH). The data bits are separated over two fields of bits inside the DPDCH slot: Data 1 and Data 2. In the control (DPCCH) part, we find the same bits of information as in the uplink with the exception of FBI bits. 14

Figure 16: Dedicated physical channels in the downlink The number of bits in the DPCH slot is computed by the expression: N = 5120/SF Figure 17: Example of a processing chain in the downlink The even bits of the DPCH are distributed along the I component, while the odd bits are transmitted on the Q component (see Figure 17). The symbols are then created whose rate is twice less than the bit rate. The same channelization code is used to spread the symbols on the two components 15

2.3 Signal pulse shaping After spreading and scrambling, the resulting complex signal, composed of chips, is distributed on a real and an imaginary branch known as I and Q components. Then it is filtered with a digital filter and converted into an analog signal with a digital-to-analog converter. The signal is filtered again before being modulated in quadrature and transposed on a carrier frequency f 0. The digital filter is a raised cosine (RRC for Root Raised Cosine) filter with a factor β = 0.22 called roll-off factor. This factor determines the shape of the output spectrum, because the energy of the filtered signal is concentrated in a band BW=Rc(1+β), where Rc is the chip rate. The impulse response of the RRC filter is an approximation of the Nyquist filter whose aim is to reduce the problems of inter-symbol interference duringthe signal sampling. RRC filter impulse response as a function of β RRC filter frequency response as a function of β Once filtered, the analog signal is then modulated and transposed on a carrier using a QPSK modulation. This kind of modulation allows the transmission of two bits per symbol, or two chips per symbol (the binary signal has already been spread). The phase of the modulated signal takes different values that represent a single pair of chips. 16

Glossary 3GPP ADS ADSL BCH BER BLER BPSK BW CCPCH CDMA CPICH DL DPCCH DPCH DPDCH DS FDD GSM IMT MMS MS/ME NODE B OVSF QPSK RF RRC SNR SC SCH SF TDD TDMA UL UMTS UTRAN VSA WCDMA 3rd Generation Partnership Project Advanced Design System Asymmetric Digital Subscriber Line Broadcast Channel Bit Error Ratio Block Error Ratio Binary Phase-Shift Keying Bandwidth Common Control Physical Channel Code Division Multiple Access Common Pilot Channel Down Link Dedicated Physical Control Channel Dedicated Physical Channel Dedicated Physical Data Channel Direct Sequence Frequency Division Duplex Global System for Mobile communications International Mobile Telecommunications Multimedia Messaging Service Mobile Station / Mobile Equipment equivalent to the Base Station in GSM Orthogonal Variable Spreading Factor Quadrature Phase - Shift Keying Radio Frequency Root-Raised-Cosine Signal to Noise Ratio Scramble Code Synchronization Channel Spreading Factor Time Division Duplex Time Division Multiple Access Up Link Universal Mobile Telecommunications System UMTS Terrestrial Radio Access Network Vector Signal Analyzer Wide band Code Division Multiple Access 17

Simulation 1 1: UMTS Spreading code Q1 Open the simulation scenario Spreading_Code (File - Open Project UMTS_prj - networks Spreading_Code.dsn ), pay attention first to the different variables of the system and justify the choice of these parameters: TimeChip = 1/3.84e6, SF = 4, TimeSymbol = TimeChip * SF Detail now the functionality of the two blocs Bits and ' GPPFDD_OVSF, and explain the scenario. Q2 Run simulation ( Simulate or F7), the results can be viewed by opening a window New Data Display Window. It is possible to recharge a predefined data representation (File - Open Spreading_Code.dds ). Compare the shape of the digital signal before and after spreading, which spreading codes are used? Simulation 1 2: UMTS orthogonal spreading codes Q1 Open the simulation scenario Spreading_Code_Ortho. What are the characteristics of OVSF codes? Q2 What is the rule for allocating OVSF codes in a single cell? Q3 Run simulation and observe the results. What is the chip rate used in the transmission? What is symbol rate? Q4 Watch now in the receiving part, explain the use of blocks Correlator, Downsample and Quant. Q5 Try now to add a fourth user in the scenario. Give at least two possible OVSF codes for this new user. Run simulation and check the good implementation of the system. Simulation 1 3: UMTS orthogonal spreading codes Q1 Open the simulation scenario Spreading_Code_Autocorr. At the reception side, we see that a signal is mixed with several time shifted versions. What is the name of this phenomenon in wireless communications? Q2 Launch simulation and compare the results: why can t we recover the right information in the presence of multipaths? Which bad property of OVSF codes is linked with this problem? 18

1.5 1.5 Data_Send, V Data_Chip, V - - 1.5 1.5 Data_Receive0, V Correlator0, V - Data_Receive1, V Correlator1, V - One tap Add of a second tap with a delay of 2 chips 1.5 1.5 Data_Receive2, V Correlator2, V - Data_Receive3, V Correlator3, V - Add of a third tap with a delay of 4 chips Add of a fourth tap with a delay of 6 chips Q3 What codes are used in UMTS to improve the statistical properties of OVSF codes? Simulation 1 4: UMTS scrambling code Q4 Open the simulation scenario Spreading_Code_Scrambler_LFSR. What is the new element introduced in the chain of transmission? Q5 The length of used register is equal to 18 while the initial state is equal to 1185. What is the primitive polynomial of the register? What is the period of the generated sequence? Q6 Run simulation and compare results with those of 1-3, comment. 19

1.5 Data_Send, V - 1.5 1.5 Data_Chip, V - Data_Chip_Scramble, V - 1.5 1.5 Data_Receive0, V Correlator0, V - Data_Receive1, V Correlator1, V - 1.5 1.5 Data_Receive2, V Correlator2, V - Data_Receive3, V Correlator3, V - 20

Simulation 1 5: Codes and Multi-user in UMTS Q1 Open the simulation scenario Multi_Access_DL, what key parameter is used to multiplex several users in UMTS downlink? Justify your answer. Q2 Open the simulation scenario Multi_Access_UL, what key parameter is used to multiplex several users in UMTS uplink? Justify your answer. Q3 Run the two simulations and compare the error rate in the presence of noise. Explain this difference. Simulation 2 1: UMTS Up Link Constellation Q1 Open the simulation scenario Constellation_UL and start the simulation. Watch the constellation of a single DPDCH or DPCCH channel. What are the possible states (a gain of is applied to both channels to maintain the 0dB signal)? Is it a BPSK or QPSK modulation in respect of each channel? (the scrambling code used in a UMTS system is in the complex form ± 1 ± j ) Q2 Watch now a constellation of a two channels DPDCH + DPCCH multiplexing. By making a small calculation, find all the possible states in the case of a combination of two signals with an equivalent power gain. Q3 In the case of a combination of the two signals with different power gain, a 6dB attenuation is applied to the DPCCH channel (Q component). What is the effect of a complex scrambling on the modulated signal? Simulation 2 2: UMTS Down Link Constellation Q1 Open the simulation scenario Constellation_DL. In order to better visualize the states distribution on the constellation, a background noise is added to the signal. Compare the three signal sources and explain their differences. Q2 Launch simulation, what is the modulation mode used in DL? Q3 Compare the constellations achieved with the combination of the correlated and non-correlated sources, pay attention to the probability of occurrence of different states, what do you see? Q4 Explain this difference. 21

Simulation 2 3: UMTS pulse shaping filter Q1 Open the simulation scenario Filter and locate the two blocks of QPSK modulation, what is their difference? Q2 Launch simulation and watch the impulse response of this filter, which advantage presents a RRC filter compared to classical low-pass filter? Q3 Watch now the modulated signal in the VSA. Compare the temporal signal with and without filtering. Then measure the length of a chip Tc with a marker. (What signal do you choose to make such a measure?) Signal without filter Signal with filter Q4 Now use a digital demodulation to study the signal. Click on MeasSetup Demodulator Digital demod to activate this element. Then go in MeasSetup Demod Properties Format Preset to Standard and choose Cellular W-CDMA. Place in a first window the signal constellation ( IQ Meas Time ). Compare the result with the two signals, what do you see? (No background noise in this scenario.) Q5 Display in a second window the signal spectrum (Spectrum). What is the factor β used by the RRC filter? Compute the theoretical frequency band of the filtered signal, and compare with the value measured by the VSA (Occupied Band Width). Why do we need to filter the signal noting that the bandwidth of a UMTS channel is 5 MHz? Q6 Change now the factor β from 0.22 to 1 and restart the simulation. Do not forget to change the configuration of digital demodulation in VSA ( MeasSetup Demod Properties Filter Alpha / BT ). Measure the new frequency band occupied by the signal and compare with the constellation of the previous questions. What is the influence of this parameter on the modulated signal? 22