WCDMA: Wideband Code Division Multiple Access
The Universal Mobile Telecommunications System (UMTS) was designed for high-speed mobile data, video calls, and internet services. Serving many users required a radio-sharing method based on wide channels and codes. The selected technology was Wideband Code Division Multiple Access, or WCDMA. Compared with 2G/GSM, it changes how transmissions are separated among users sharing radio resources.
In GSM, the radio resource was shared through a combination of frequency division (FDMA) and time division (TDMA). Users received different frequencies or successive time slots. WCDMA allows many transmissions to occupy the same wide channel at the same time. A receiver separates the desired signal with the code assigned to its channel and treats the remaining transmissions as interference.
The Core Principle: Code Division Multiple Access (CDMA)
WCDMA is a specific implementation of a broader technology known as CDMA. At its heart, CDMA applies a code to each user's data. The code is multiplied chip by chip with the data symbols and "spreads" them across a wide frequency band, mixing them with other users' signals. A receiver that knows the code and timing can reverse the process, "despreading" the signal to extract the original data while other signals remain as interference.
A crowded room provides a useful analogy: several pairs are speaking, and each pair uses a different language. A listener who knows one language can focus on that conversation while the other voices form background. In CDMA, the code supplies the identifying property of a conversation, while the actual receiver uses correlation, synchronization, orthogonality, and interference levels. The analogy explains the selection process; it does not replace the receiver operation.
1. Spread spectrum: transmission and reception
The technical process behind CDMA is known as . It takes a low-bandwidth user data signal and produces a signal occupying a much wider frequency channel. The receiver restores the data by correlating it with the assigned code.
Original data, such as digitized speech or internet data, is a relatively narrowband signal. WCDMA uses a channel with a bandwidth of . This is 25 times wider than the channels used in GSM. The process multiplies the narrowband data stream by a high-speed code, often called a or a "chipping code." WCDMA uses a 3.84 Mcps chip rate, so this code has a much faster rate than the user data.
Multiplying the data stream by the fast code produces a wideband signal. The energy of the original signal is distributed across the entire channel. To an outside observer, such a signal can resemble noise; its power in one narrow frequency slice is small.
The receiver takes the incoming wideband signal, which is a mixture of several transmissions, and multiplies it by the synchronized code used for the selected channel. In the practical sequence, the receiver applies the relevant scrambling sequence and then the channelization code. This despreading operation has two effects:
- It restores the selected signal to a narrowband form, concentrating symbol energy in the correlation result.
- Transmissions spread with other codes remain as wideband interference, with a reduced contribution to the correlation result.
The receiver can then distinguish the selected signal from the other components. The increase in signal-to-interference ratio obtained during despreading is called . Its value depends on bandwidth and data rate; increasing data rate at a fixed chip rate reduces the spreading factor.
2. WCDMA codes: separating channels and cells
WCDMA performance follows from the mathematical properties of its codes and the rules used to assign them. The system uses code sets with defined orthogonality, while separate coding layers serve separate purposes. WCDMA has two main coding layers. Channelization separates physical channels within the cell, while scrambling separates configured sources and supports signal identification.
To separate channels for users connected to the same cell (Node B), WCDMA uses codes known as . Two codes from a valid assignment are orthogonal when they are not in an ancestor and descendant relationship in the code tree. The sum of the products of their chips over the code period is then zero. This property lets a receiver reduce the contribution of other channels during despreading when the code assignment follows the tree rules. Reserving a branch for one code therefore restricts which descendant codes can be assigned to another channel. In a multipath channel, delayed signal copies can disturb this orthogonality at the receiver input. Rake fingers and channel estimation separate useful components and limit the remaining interference.
The "Variable Spreading Factor" part is also important. The spreading factor is the ratio of the chipping code's rate to the user's data rate. In the OVSF tree, code lengths support channels with different rates, including the SF 4 to SF 512 range used in described configurations. A low spreading factor can support a high data rate, while a high spreading factor corresponds to a lower data rate. The network can use this relationship to manage resources and assign different rates to users or services.
OVSF codes separate channels within one cell when the allocation follows the code tree. A phone can also receive signals from several nearby cells, so a second coding layer helps distinguish their transmissions. This layer uses .
In downlink, a cell uses a primary scrambling code selected from the groups defined in TS 25.213, and a configuration can also contain secondary scrambling codes. The specification describes 512 primary-code groups, each containing a primary code and secondary codes, with the sequence repeated over a 10 ms frame. In uplink, the scrambling code is associated with the UE and link configuration. The code is applied on top of the channelization codes. The receiver applies the sequence for the selected source, while transmissions using other sequences remain as interference. Scrambling is not cryptographic encryption, and its selection belongs to radio configuration rather than subscriber confidentiality.
WCDMA physical-layer techniques
The use of spread spectrum and dual-layer coding enables several techniques that shape UMTS performance and reliability. The physical layer provides transport through channels, while transport formats specify coding, interleaving, rate matching, and frame timing. Together with power control and multipath processing, these operations describe the practical WCDMA radio path.
A central design constraint in CDMA systems is the . Since many users share a frequency, an illustrative scenario can place a phone near the base station at a signal level thousands of times stronger than a phone at the cell edge. Without regulation, this "loud" signal can mask the distant transmission and make it difficult for the Node B to detect.
WCDMA limits this effect with a fast power-control mechanism. The goal is to keep uplink signals at levels that allow the Node B to detect them together. The procedure uses two loops:
- Open-Loop Power Control: This is an initial rough estimate. When a phone first wants to transmit, it measures the strength of the downlink signal it receives from the tower and estimates the power for its own transmission. The estimate accounts for the observed path loss before closed-loop commands refine it.
- Closed-Loop Power Control: This is the fast, fine-tuning process. The Node B measures the power received from the phone. When the level is too low, it sends a "power up" command; when it is too high, it sends a "power down" command. TPC commands are carried in slots, and selected closed-loop modes use a 1500 Hz update rate. During soft handover, the UE can receive commands from several radio links, and the procedure must account for their combination. This limits the near-far effect, energy use, and interference in the network.
In wireless environments, especially cities, radio signals bounce off buildings, trees, and other obstacles. This creates a phenomenon called , where the receiver gets multiple copies of the same signal arriving at slightly different times. In traditional narrowband systems, these echoes interfere with each other and degrade the signal. A wideband receiver can separate components by delay and assess their quality before combining them.
WCDMA uses a . Its wideband signal allows echoes with different delays to be distinguished. A Rake receiver has several sub-receivers called "fingers". Each finger can track a different strong multipath component. The receiver independently despreads each finger and combines the results, which can improve detection reliability in a multipath channel. The number of fingers and the combining method depend on receiver implementation and channel estimates. Channel estimation determines which delayed components are useful and how their phases and weights contribute to the combined result.
In a same-frequency WCDMA scenario, neighboring cells use the same wide channel and distinguish transmissions with scrambling codes. A mobile phone can then maintain radio links with more than one cell at the same time. This enables . Radio-link changes use an ACTIVE SET UPDATE procedure that adds or removes links on the basis of measurement reports and RNC decisions. Other procedures, such as inter-frequency handover, have a different sequence.
In GSM, handovers are "hard," meaning the phone releases its connection with the old cell before it establishes a connection with the new one. In WCDMA, as a user enters overlapping coverage, the phone can add a radio link to the new cell while still maintaining its connection with the old cell. For a period, the phone communicates with both base stations. The network controls the active set and can remove the old radio link after conditions stabilize. This "make-before-release" approach limits the risk of an interruption during handover. An active-set update changes the serving access points without requiring the whole session to be released immediately.