New Technology for Cellular Phones Challenges Test Procedures

Code division multiple access (CDMA) is the newest technology for cellular telephony to be fielded in North America. While it promises to provide excellent performance, the key advantages of CDMA also make it very difficult to test–a difficulty sorely compounded by the fact that multiple users share a single channel.

CDMA became a cellular standard when the Telecommunications Industry Association (TIA) adopted IS-95 in July 1993. It is based largely on the CDMA system developed by QUALCOMM, Inc., and is currently being tested in several field locations using equipment from multiple vendors.

A channel in CDMA is the collection of users who share one frequency. Each user is assigned a code that differentiates his or her transmission from those of other users on the same frequency. This assignment is called the code channel.

In the CDMA design, each user’s signal experiences high levels of interference, dominated by the signals of all other users on that CDMA frequency. A spectrum analysis of a CDMA transmitter provides important measurements of occupied bandwidth, adjacent channel power and total power in the channel; and can be used for time-domain measurements of power transitions as well. A spectrum measurement cannot provide insight into the modulation or the information being transmitted.

Test Equipment Capabilities

A CDMA phone must be on a traffic channel for much of its testing. Test equipment must emulate a real base station in a cellular system. A minimum of five code channel transmitters are needed to achieve this:

(o) Pilot: This is a timing and phase reference sent from each cell that acts like a beacon for the mobiles to find.

(o) Sync: This channel sends the time-of-day information to the mobiles. This is used as the basis for the codes used in CDMA transmission.

(o) Paging: This is the digital control channel from the base station. It is used to page the mobile or to answer a mobile origination and make an assignment to a traffic channel.

(o) Traffic: Conversation takes place on this channel.

(o) Orthogonal Channel Noise Source: This channel is the largest of the code channels used in testing. It simulates the noise from all the other users on that cell. The codes for the forward link (base to mobile) are designed to be orthogonal, or noninterfering.

Two code channels are required to emulate the base-station receiver:

(o) Access: This is the control channel for the

reverse link (mobile to base).

(o) Traffic: This is the other half of the conversation link.

In a real base station, these channels are implemented with separate hardware. In test equipment, it is possible to use one set of hardware for these code channels because they are used in sequence and never simultaneously. This, however, requires more complex software.

Receiver Testing

CDMA receiver testing has many parallels in analog (Figure 1). The frame error rate (FER) is the fundamental measure of link quality in CDMA. In analog, SINAD (signal + noise + distortion) divided by noise + distortion is used as the performance measure. Sensitivity testing is performed in both systems by setting the signal to a low level at the antenna port and measuring the link quality.

Co-channel interference is more important in CDMA than in analog. Analog FM typically needs 18-dB freedom from interference on the same channel to achieve good quality. Generally, CDMA operates in the presence of noise that is 15 dB larger than the traffic channel signal.

To simulate the noise from other users, two types of noise are used: CDMA noise and additive white Gaussian noise (AWGN). The CDMA noise from a base station is designed in the coding scheme to be orthogonal to all other signals. This means it is invisible to the other users of that cell.

AWGN is used to simulate the CDMA traffic from nearby cells and all other background noise in the channel. Both analog FM and CDMA systems also measure link quality with interfering signals injected at close offsets in frequency.

The measure of frame errors for testing is a statistical process. The specification requires a 95% confidence of a FER under specified conditions.

CDMA Transmitter Testing

Waveform quality for analog FM is defined in terms of distortion and frequency response of the recovered audio. Waveform quality was not defined for CDMA until the TIA standards committee developed the mobile station performance document IS-98. Waveform quality has been defined as the normalized cross-correlation of the actual transmitted signal and an ideal version of the same signal. This is an important measure because CDMA uses correlative receivers. A signal that is 95% correlated to its ideal has 5% less energy for its own link.

Probably more important, however, is that the 5% lost energy is interference to all users on the band. It reduces capacity of the system. The limit for mobile transmitters has been set to 94.4%, which is the equivalent of 0.25 dB extra energy to have the same quality as a perfect signal.

Power control on the mobile stations is vital to CDMA performance. Each mobile should be received at the base station at the minimum signal level just to maintain a link. This means that every mobile should be received at exactly the same level at the base station.

To achieve this, two forms of power control are used: open loop and closed loop. Open-loop power control is based on the premise that the loss from the base to mobile will match the loss from the mobile back to the base. The two signals are 45 MHz apart in the 800-MHz band, so this assumption holds fairly well.

Closed-loop power control is achieved by using up or down power commands sent every 1.25 ms by the base station. Each step is 1 dB. Test procedures exist for both forms of power control.

It is important to measure the spectral content of a CDMA transmitter because it shows the interference level that is transmitted outside the assigned frequency band (Figure 2). For compatibility with existing analog systems, this must be minimized, and has prescribed limits.

Base-Station Test

The CDMA base station shares much of the test concepts of the mobile, with some twists to the test methods. The modulation format is different from the mobile and base transmitters, so the modulation accuracy measurement has different definitions.

The reverse link also does not use orthogonal modulation codes, so the base-station receiver is tested only with AWGN added to the signal. The specified signal level for sensitivity on the base station is -117 dBm/1.23 MHz, a level 4 dB lower than the noise bandwidth at room temperature. Coding gain allows the system to work at this level.

Conclusions

CDMA test is substantially more complex than the analog test procedures. New measurements have been defined, and parameters specific to the system require new test procedures.

About the Author

David Whipple is an R&D Technical Leader at Hewlett-Packard Co. He joined HP in 1973, and for the past several years, has been working on test solutions for CDMA. Mr. Whipple graduated from Purdue University with an M.S.E.E. degree. Hewlett-Packard Co./Spokane Division, 24001 E. Mission Ave., Liberty Lake, WA 99019-9599, (509) 921-3226.

SIDEBAR

The Fundamentals of CDMA

CDMA is a class of modulation that realizes its channelization by using codes that are shared by the transmitter and receiver. The receiver is a correlator, and it must be timed to the incoming signal. Multiple users share the same frequency.

CDMA uses two levels of coding. The first level is convolutional coding, sometimes called forward error correction. The matching decode function for this is maximal likelihood detection, of which the Viterbi Algorithm is optimal.

The second level of coding involves the use of Walsh Codes, or Hadamard Codes, which are each 64 bits long. Different coding schemes are used in the forward (base to mobile) and reverse (mobile to base) links because the base-station transmitter can time-align all the code channels being transmitted. Thus, orthogonal modulation can be used so each user does not interfere with others.

Multipath signals, however, do cause interference. The reverse link does not have the capability to time-align with sufficient precision to allow the use of orthogonal codes. For this reason, the reverse link has less capacity than does the forward link. The reverse link power is monitored by the base station and dynamically adjusted to keep it at the bare minimum needed to maintain the link, which lessens the interference to other users.

CDMA radios receive on their assigned code in the presence of substantial interference. This is the sum of all other users on the same frequency, both within the same cell and from nearby cells. Under typical conditions, the interference is 15 dB larger than the desired signal.

With CDMA, every cell can use the same frequency. In a typical analog cellular deployment, only 1/7 of the available frequencies are used in any cell. This provides a distance of two cells before the channel is reused and 18-dB roll-off of the reused channel. By improving frequency reuse, the capacity of CDMA increases greatly. In fact, CDMA has about 10 times the capacity of the analog system.

The test of a CDMA radio must simulate the actual environment. For this reason, AWGN is added to the test signal to simulate the interference. It is also important to test the power control functions of the mobile station.

April 1995