Wideband CDMA: Talk is cheap

Third generation digital wireless formats now being developed rely on that old standby-voice-as their ace in the hole.

While the two leading contenders-both versions of wideband code division multiple access (W-CDMA)-differ in implementation details, both are recognized for their ability to bring high-speed data to mobile users.

With less pizzazz but more marketplace impact, they also promise wireless voice at significantly higher capacity and lower cost than existing systems can provide.

In today's crowded airways, it is important to maximize the traffic capability of available bandwidth. At the same time, it is becoming increasingly important to support the transport of user data.

First generation wireless telecom systems relied solely on frequency division to achieve capacity. The allotted RF spectrum was subdivided into individual 30 kHz channels that were assigned as users required.

This scheme had two drawbacks: It limited the number of voice users to one per channel, and because the narrow channels were optimized for analog voice, it could support only low data rates.

Second generation wireless systems expanded overall capacity by adding time division multiple access (TDMA) or CDMA to frequency division channels.

With TDMA, three users are assigned the same 30 kHz channel on a fast-switching timeshare basis. Through modern signal processing, the system provides digital transport that can represent either voice or user data.

CDMA, based on IS-95, is more complex. It uses 1.25 MHz channels. In the CDMA channel, the narrowband, digitized signal for each user is combined with a fast binary code sequence before transmission. This spreads the user signal over the full 1.25 MHz channel bandwidth.

In the smart CDMA receiver, the incoming 1.25 MHz wide signal is combined with the same binary code sequence that was used during transmission, regenerating the user's original digital signal (Figure 1). Because each user is assigned a unique code sequence, multiple users can occupy the same bandwidth simultaneously.

When extracting a particular user's signal from the composite, the pseudo-random nature of each code causes the other users' coded signals to be regenerated as noise.

IS-95 CDMA has experienced rapid growth in the U.S. and Korea but has received spotty acceptance elsewhere.

Standardization issues The International Telecommunication Union is soliciting proposals from regional standards organizations for a global third generation wireless standard to be called IMT-2000 (Figure 2). Performance requirements for the standard-with a data rate of up to 2 Mb/s-mandate a wide bandwidth.

Diverse proposals have been narrowed down to two leading versions. Both use code division techniques to spread users' signals over a wide bandwidth.

One is from the Telecommunications Industry Association in the U.S. and is called cdma2000. The CDMA Development Group has been a leading force in defining the TIA's proposal. It is based on the CDMA group's recommendation of a family of capabilities called cdmaOne, which offers upward compatibility with existing IS-95 technology.

The proposal would minimize the changes to embedded CDMA infrastructure by using compatible coding formats and existing procedures for code synchronization, soft handoff and roaming. The wideband member of the proposal has a basic chip rate three times that of IS-95-3.686 megachips per second (Figure 3). Including guard bands, the occupied bandwidth becomes 5 MHz.

The other version is being proposed by the European Telecommunications Standards Institute and the Association of Radio Industries and Business in Japan. They plan to submit separate proposals, but the content should be similar because they are working to harmonize them. Unlike the TIA proposal, there is no compatibility incentive to an existing CDMA system because no CDMA systems operate in Europe and only limited implementation exists in Japan.

ETSI and ARIB have proposed wideband-5, 10 and 20 MHz-capabilities with a number of interesting characteristics, including the ability to support indoor base stations without access to the signal from global positioning system (GPS) satellites.

IS-95 and its relatives need the GPS signal to assure timing synchronization between base stations. Freedom from the GPS signal eases the installation of indoor base stations, supporting the vision of seamless personal communications indoors and out.

NTT DoCoMo (the wireless arm of NTT) has been a leading force defining the ARIB proposal. It is committed to early W-CDMA implementation and is projecting field trials in 1999 and first deployment in 2000.

Compatibility between the two versions is unlikely because their chip rates are not compatible. The basic chip rate of the wideband capability of the TIA version is 3.686 megachips per second, while the ETSI and ARIB's version is 4.096 megachips per second.

With no resolution in sight, other players in the larger picture such as the test and measurement community-including Hewlett-Packard Co.-now are supporting exploratory development efforts in all W-CDMA standards.

Voice payoff With limited worldwide deployment of CDMA, why the push for W-CDMA?

The answer is that W-CDMA will deliver higher transport capacity that will support high-speed data and low-cost voice.

Of these, data is the more glamorous-but voice is the more important. And to the extent that a single international standard can be achieved, wireless voice and data services could be compatible and seamless from local to global networks.

W-CDMA can provide lower-cost voice through lower per-user infrastructure costs-potentially as low as half that of IS-95-because of the increased sharing of the higher cost RF components. In an IS-95 system, a single base station transmitter/receiver might support 15 users. A W-CDMA system, under similar conditions and also with a single transmitter/receiver, should support more than 50 users.

Of course, the wideband radio will be more complex, but its cost increase will be less than its capacity increase. Although the wideband system will have added complexity in its digital processing, the volume price of the digital elements is dropping rapidly.

The net result is a much lower per-user cost for the wideband system. This advantage for the W-CDMA system is augmented because the larger number of users on the wideband channel yields more predictable traffic statistics, which enhances the effectiveness of the CDMA technology.

High-speed data? Some involved in W-CDMA development have promoted high-speed data as the salient characteristic of the new technology. On second generation systems, wireless data has been limited to the voice standard of 14.4 kb/s, though refinements support up to 64 kb/s.

Wideband systems will support data at rates up to 384 kb/s (and eventually to 2 Mb/s), but only by sacrificing substantial voice capacity. For example, the 5 MHz cdma2000 system will transport data at 307 kb/s, but uses 40% to 50% of a cell's capacity.

This data use has the cascading impact of further lowering the cell's capability for carrying voice traffic by reducing the statistical base of users, which CDMA depends on to achieve high bandwidth use. The remaining voice users are few, so they form a smaller base for statistical averaging.

This makes the voice/data dilemma clear: Although wideband CDMA can support a wide mix of voice and data users technologically, high-speed data transmission pre-empts a significant number of voice users. Is high-speed data worth the cost or will moderate-speed data serve as well?

In the consumer sector, all indications are that voice will continue to represent the majority of wireless use. Although significant growth is expected in wireless e-mail, the total number of bits needed for its transport is small-and modest data rates may be acceptable.

Thus, a key to making W-CDMA a success will be service providers' ability to offer voice at lower cost than current wireless formats. And because the cost savings derive from increased capacity, W-CDMA should be most attractive in dense urban markets.

Like most emerging telecom technologies, the evolution path for W-CDMA is not clear. Will a single international standard win out or will there be multiple, incompatible standards?

Two things seem certain. First, W-CDMA is coming and will arrive soon. Second, despite the hoopla surrounding high-speed data, it is W-CDMA's ability to deliver low-cost voice that will drive its success in the marketplace.