3G Technologies Complicate CDMA Testing

Flexibility is particularly important because 3G standards aren’t set in stone. Moreover, the jump from 2G to 3G won’t be discrete—there will be detours through “2.5G” technologies, such as EDGE, which extends the data-handling capabilities of GSM. (See “Wireless Communications Glossary,” for an explanation of the abbreviations used in this article.) Some CDMA technology will undoubtedly play a big part in the standards because of its efficient use of spectrum, but two incompatible CDMA approaches—cdma2000 and W-CDMA—are competing. The cdma2000 approach is gaining ground in North America because of its cdmaOne heritage; cdmaOne in turn evolved to maintain compatibility with the US analog AMPS standard. W-CDMA is ascendant in Europe, where there is no need for AMPS compatibility.

The general frequency ranges you must test for in 2G and 3G technologies don’t differ, and you can use any general-purpose RF instrumentation to measure a cell phone’s RF output, for example. But to ensure accuracy and high test throughput, you’ll need instruments that can quickly measure digitally modulated RF signals. And because of the complexity of the CDMA signal, you would be well advised to choose equipment that has the computational smarts to derive meaningful CDMA parameters—such as ACPR, CCDF, and SNR—from raw data and whose software can be upgraded to support either emerging CDMA standard as well as interim standards such as EDGE. Our product survey chart lists manufacturers of test equipment for CDMA.

CDMA Basics

A look at how CDMA differs from analog FDM and digital TDM helps illustrate the complexity of CDMA measurements. An AMPS analog system, for instance, allocates about 416 30-kHz forward, or downlink (base station to mobile phone), communication channels within a 12.5-MHz bandwidth assigned to a cellular service provider. (A similar frequency arrangement—the uplink, or reverse channel—handles mobile-phone to base-station transmissions; uplink frequencies are 45-MHz lower than downlink frequencies.)

The service provider wants to accommodate as many phone calls as possible within that allocation. Ideally, the provider establishes a honeycomb pattern of cells (the limited ability to place antennas plus geographical and architectural vagaries degrade the ideal hexagonal pattern in real life) with a base station at the center of each cell (Fig. 1). With AMPS, a given cell cannot share any channels with its six adjacent cells, because of the likelihood of interference. For example, as a mobile user (A) communicating with base station 1 over channel 1 approaches the border between cell 1 and cell 2, he or she would receive unacceptable levels of interference from a user in cell 2 (D) communicating with cell 2’s base station over channel 1. Similarly, users D and A cannot share the same channel to communicate with their respective base stations, because user D’s signal (intended for base station 2) would be nearly as strong at base station 1 as user A’s. To make sure that users in one cell never share frequencies with users in an adjacent cell, each cell in a seven-cell cluster can use only 59 channels, approximately one seventh of the available 416 channels.

This approach has two severe limitations. First, it limits the call capacity for the network, because adjacent cells can’t share channels. Second, it forces a roaming mobile phone to switch frequencies when moving from one cell to another, resulting in call interruption or a lost connection. CDMA, in contrast, allows adjacent cells to share frequencies, increasing cell capacity and facilitating a soft hand-off that results in fewer interruptions.

Figure 1. Cellular service providers establish seven-cell configurations, with each cell containing a base station. A service provider can duplicate this seven-cell cluster throughout its service area. FDM and TDM technologies require that adjacent cells not share frequency assignments. CDMA overcomes this limitation through power control and the use of PN codes.

CDMA employs two ingenious methods to permit adjacent cells to share the service provider’s entire frequency allocation: PN coding and power control. In the first, instead of allotting a discrete frequency band to each user, CDMA assigns a pseudonoise signal to each user (Fig. 2) and uses a direct-sequence spread-spectrum technique to spread transmitted messages over a bandwidth much broader than the 30-kHz of AMPS. The PN signal operates at a spectrum-spreading rate called the chip rate.

Current-generation cdmaOne technology employs a 1.2288-Mcps chip rate—called SR1—to spread the message over a 1.25-MHz bandwidth. The transmitter modulates the user’s voice or data using the assigned pseudonoise signal, and the receiver uses the same pseudonoise signal to demodulate the transmitted signal. (The transmitter must tell the receiver which pseudonoise code to use.)

Initial 3G cdma2000 systems will employ the SR1 of cdmaOne but add QPSK modulation (on the downlink; the uplink employs HPSK) and other features to double cdmaOne’s capacity. Subsequent cdma2000 systems will employ SR3, in which a message will be spread over a single 3.75-MHz carrier (the direct-spread approach) or three 1.25-MHz carriers (the multicarrier approach, an interim approach that maintains compatibility with cdmaOne).² With CDMA, other users’ transmissions, modulated using other pseudonoise signals, appear as random noise. Users on different channels can share frequencies, but the pseudorandom noise from other channels can degrade a cell’s signal-to-noise ratio (SNR). To improve SNR, base stations employ orthogonal Walsh codes for message coding. With orthogonal codes, at any given time half the users in a cell are transmitting a +1 while the other half are transmitting a –1, so from a given user’s perspective, interference from other users tends to cancel. Mobile units don’t employ Walsh codes for message coding because of synchronization difficulties, but the long PN codes they do use have a similar effect in reducing overall SNR.

The second method employed by CDMA minimizes the total random noise—thermal noise plus pseudonoise from other users—through power control. In non-CDMA systems, transmitters operate with sufficient power to handle worst-case scenarios—for example, a mobile AMPS phone always transmits with the power it would need to reach a base station from a cell border. In Figure 1, users B and C communicate with their respective base stations at full power, despite their proximity to their base stations. In contrast, CDMA provides that all transmitters in a cell transmit at the lowest possible power to maintain a target SNR at the receiver. Because of this, cdmaOne provides better than a four fold capacity improvement in contrast to AMPS. Improved, faster power-control schemes will contribute to cdma2000’s doubling of cdmaOne’s capacity.

CDMA and Migration

W-CDMA³ is conceptually similar, although it employs different chip rates and doesn’t support simple migration from first-generation CDMA systems.⁴ The smoother migration path of cdma2000 is attractive to service providers operating cdmaOne and AMPS networks, but it can complicate test and measurement. For example, because cdma2000 can coexist with cdmaOne and AMPS, you’ll need to continue performing AMPS adjacent-channel power measurements on cdma2000 equipment (although the direct-spread SR3 versions of cdma2000 won’t permit interleaving of AMPS service). For CDMA, the ACPM determines the level (expressed as ACPR) of a CDMA signal’s power that leaks into a neighboring 30-kHz AMPS channel. You’ll need to choose instrumentation that can measure the signal power for interim “2.5G” products as well as early 3G cdma2000 products. In addition, for both 3G CDMA approaches, you’ll need instruments that can measure power in adjacent 1.25-MHz and wider CDMA channels.

For any power measurements on CDMA or other digitally modulated RF signals, you’ll need an instrument that can handle the carrier’s varying envelope amplitude, which results from the digital modulation. Calorimetric power meters can accurately determine RMS power levels of CW and constant amplitude FM signals, and they’ll indicate average RMS power of any RF signal, but they are too slow to measure a digitally modulated RF carrier’s instantaneous power.⁵ You’ll need that capability for deriving a signal’s CCDF, so you should choose a digital-sampling instrument that can provide adequate response.

In addition to measuring RF power, you’ll need to measure CDMA modulation quality. Error-vector-magnitude measurements can help you test a transmitter’s baseband filters, I/Q modulator, and IF and RF stages; to measure EVM, you can use an instrument that can generate QPSK and QAM constellations. For code-domain quality measurements of cdma2000 signals, you can use cdmaOne equipment, but for a CDMA forward channel, the multi-length Walsh codes corresponding to varying spreading rates and data rates complicate the measurement task, and you’ll benefit from an instrument that automates the process. T&MW

FOOTNOTES

1. “The Path to Third-Generation Mobile Systems,” Tektronix Wireless Test Backgrounder, Tektronix, Beaverton, OR, www.tek.com/Measurement/Products/backgrounders/3G/index.html.

2. “Performing cdma2000 Measurements Today,” Application Note 1325, Agilent Technologies, 1999.

3. “W-CDMA Measurement Challenges,” Rohde & Schwarz, München, Germany, 1999.

4. “CDMA Terminology and Definitions,” The CDMA Development Group, Costa Mesa, CA, www.cdg.org/tech/cdma_term.html.

5. Blackwell, Richard H., “Digital Sampling Power Analyzer for GSM and CDMA,” Boonton Electronics, Parsippany, NJ, 1999, www.boonton.com/digitalsampling.html.

You can contact Rick Nelson at rnelson@tmworld.com.