CDPD presents new challenges for RF testing: Even when cellular channels are dedicated exclusively to use for cellular digital packet data (CDPD), RF performance tests should be conducted on the CDPD systems to assure that it is not degrading the AMPS sys

Even though cellular digital packet data (CDPD) service is generally overlaid on an existing advanced mobile phone service (AMPS) system, it still requires a battery of RF performance tests independent of those conducted to evaluate its AMPS counterpart. It is the complementary nature of CDPD that necessitates effective RF measurements, because the CDPD system has the potential to degrade AMPS system performance. Although these tests are not complex, they vary somewhat from those required to evaluate an AMPS system.

Considerable attention has been paid to these RF measurements as more CDPD systems have been deployed, because CDPD performance has often been less than optimum. Some of the most important deficiencies are related to the performance of the CDPD transmitter and receiver. A brief description of how CDPD works will demonstrate the importance of RF testing.

CDPD is a scheme in which a packet data service can be added to an existing AMPS voice cellular system for an incremental increase in base station cost. CDPD was originally conceived to transparently complement an AMPS system, making use of AMPS channels that are not being used by voice traffic as shown in Figure 1 below. This concept allows many users to occupy a single channel and has the potential to supplement the income from the AMPS system, at little additional cost to the service provider.

CDPD does not require much additional hardware. It essentially “bolts onto” the existing AMPS base station equipment. Both the AMPS and CDPD systems have full-duplex capability and 30kHz channels, and they generally share frequency reuse, coverage and power levels.

The Search for Acceptance CDPD has always been touted as a simple, inexpensive, rapidly deployable way to provide mobile data service. Its appeal has been for applications such as order entry, inventory control, remote utility reporting, package tracking, electronic trading and other applications that require transmission and reception of small amounts of data.

This simple scenario would lead to the assumption that CDPD would gain immediate acceptance. This has not been the case for several reasons. First, pricing structures have not been low enough to drive competitors with dedicated data services from the market. These competitors have developed nationwide networks that cover as much as 95% of the United States. CDPD is the lowest-cost service, but it is still not low enough.

Technical hurdles have made it difficult to develop this seemingly simple system into a commercially viable, reliable service. Some services providers have been successful in deploying CDPD systems through hard work and determination. Their efforts have been focused on a core set of end-user apllications.

The development of end-user applications has not kept up with the deployment of CDPD, which is probably the most important reason why CDPD has not been more successful. Because few CDPD software applications have been written, customers must develop their own_a major hindrance to market acceptance.

Lack of an accepted standard has also hindered growth, although this will no longer be the case after adoption of System Specification 1.1 by the Telecommunications Industry Association (TIA) and the American National Standards Institute (ANSI). The CDPD Forum, a group sponsored by service providers that created the specification, has presented its completed work to the TIA and ANSI for consideration and ultimate adoption as a national standard.

Even with these shortcomings, the appeal of CDPD has lead to its implementation in many of the top U.S. markets, a testament to its inherent viability. Still, there are currently fewer than 10,000 CDPD users nationwide, compared to more than 43 million cellular and PCS users for voice service.

How CDPD works There are two methods of deploying CDPD. The first is through “sniffing and hopping.” In this technique, a dedicated receiver in the CDPD system called a “sniffer” samples each 30kHz AMPS channel every few milliseconds and determines whether voice traffic is present.

The most important role of the sniffer is to detect the AMPS system powering up as a call is assigned to the channel that CDPD is using. When a call is detected, the CDPD system looks for a CDPD-assigned channel that is not in use. (See Figure 2 on page 14.)

If voice traffic is detected on one channel, the system seeks out another CDPD-assigned channel. When an unoccupied channel is located, the system hurries to transmit packets of data until AMPS traffic is detected. The CDPD system then immediately seeks out and switches to another unoccupied channel. CDPD was originally designed to work this way because it requires nothing from the AMPS system but an empty channel on which to transmit and receive.

It is important to note that proper operation of the sniffer receiver is essential if this method is to be successful because accurate detection of AMPS traffic is the most direct determinant of CDPD performance. For example, one of the problems that has occurred with CDPD systems is a phenomenon called “channel sealing,” in which CDPD traffic impedes AMPS traffic on supposedly vacant channels. Channel sealing occurs when AMPS thinks a channel is too noisy to be used. (It does not differentiate CDPD traffic from interference and noise). As long as the CDPD channel dwell time is less than a certain period (the milliseconds required by the AMPS system to determine whether a channel is usable), AMPS will not seal the channel. (See Figure 3 on page 14.)

If the “noise” level is too high, the AMPS system will shut down the voice channel, making it unavailable for communication. It is possible that a malfunctioning CDPD system could shut down most or even all AMPS channels, with catastrophic results.

The other CDPD deployment technique is used when sniffing and hopping provides unacceptable results. This most often occurs in metropolitan areas where there is heavy AMPS traffic. In this situation, transmission latency occurs in CDPD service because AMPS traffic has priority. If the AMPS traffic is nearly continuous on every channel (an increasingly common occurrence in some areas), the CDPD system must wait for it to clear. This can take a long time at peak traffic periods. As a result, subscribers experience significant delays.

To combat this problem, some service providers have dedicated one or two AMPS channels to CDPD service. This eliminates the need for sniffing and hopping because there is no need to search for vacant channels. On the other hand, it also negates the original appeal of CDPD, its ability to be overlaid on existing AMPS channels without requiring dedicated spectrum. However, because of the aforementioned problems, channel dedication is increasingly becoming the method of choice.

CDPD tests Because the performance of both the CDPD system and the AMPS system that it complements depends in large measure on the proper operation of RF functions, comprehensive RF testing is essential. This applies not only to system commissioning, but to periodic maintenance as well. CDPD base station tests can be divided into four areas: transmitter tests, receiver tests, sniffer tests and system tests (which are reports of protocol data being exchanged by the mobile data base station (MDBS) and mobile end station (MES)). Each test is specified and defined in CDPD Systems Specification 1.1.

Transmitter Tests _ Six tests must be performed to verify proper transmitter operation. Transmit frequency-error tests are conducted to determine the difference between the measured center frequency of the CDPD modulated signal and the desired frequency. This measurement is required to ensure that the transmitted signal does not interfere with adjacent channels and to facilitate proper demodulation of the signal by the receiver.

Some instruments designed to make this measurement use digital signal processing to measure average peak positive and negative frequency deviation. This method has the advantage of being insensitive to the amount of time the signal spends at peak plus-or-minus frequencies.

Both total RF power and channel power must be measured to provide a complete picture of transmitter output. Total RF power is broadband output power and is measured with an RF power meter. Channel power is the power in a 30kHz bandwidth centered on the frequency specified for a given channel. It determines the area of coverage for the individual radios of the base station and interference to distance cells.

Modulation accuracy, or the ability of the Gaussian minimum-shift keying (GMSK) modulator to maintain an accurate modulation index and peak deviation, is specified in System Specification 1.1 as a modulation index of 0.5 with 61% accuracy. Measurement of adjacent and alternate channel power is required to ensure that power spread into these channels from the CDPD signal does not interfere with accurate voice or data communications.

Adjacent and alternate channel power is defined as the amount of transmitted power that falls into the adjacent, first alternate and second alternate channels. An adjacent channel is specified as a 30kHz bandwidth centered 630kHz from the center frequency. The first and second alternate channels are defined as a 30kHz bandwidth centered 660kHz and 690kHz, respectively, from the center frequency.

Finally, phase noise, or incidental FM, is the amount of frequency fluctuation accompanying the positive and negative deviations of the modulated waveform. Incidental FM directly correlates to modulation quality. Although it is not defined in the CDPD system specification, incidental FM is a useful method for predicting impending component failure. For example, high incidental FM can be caused by high oscillator phase noise (which is defined in the specification) or GMSK modulator distortion, both of which can be detected by measuring incidental FM.

Receiver Tests _ The receiver tests required of a CDPD system are similar to those employed in testing an AMPS system, with some additions and modifications. Receiver sensitivity is the minimum RF power level at which the receiver block error rate (BLER) is equal to or less than 5%. This measurement verifies that the MDBS has the sensitivity to adequately receive weak signals from a distant MES. Power level is set, at the antenna input terminals, at a nominal frequency, and BLER is based on results from a Reed-Solomon error-correction decoder and is calculated as:

where N = the number of blocks sent and C = correctable blocks received.

Forward-channel control flag performance ensures that the MDBS is generating the correct decode status and busy/idle bits in the forward-channel stream. These bits are collectively called the control flag. Busy/idle performance measures the specific RF levels at which the MDBS sets the busy flag. An iterative measurement technique is used to find the RF levels at which setting and resetting occurs.

Sniffer Receiver Tests _ As discussed earlier, some service providers use the sniffer and hopping configuration to implement CDPD. Consequently, proper functioning of the sniffer receiver is imperative if the CDPD system is to perform acceptably. Two tests, sniffer threshold and sniffer activation time, are performed to assess this performance.

Sniffer threshold testing measures the RF-received signal strength required to cause the MDBS to shut down its transmitter and move to another channel (if available) when AMPS voice traffic is detected. The interval between the sniffer’s detection of AMPS traffic and when the CDPD transmitter is shut down is called sniffer activation time. System Specification 1.1 requires this interval to be less than 40ms after voice traffic is assigned to the channel. If the CDPD transmitter takes longer than this to shut down, AMPS voice channels will experience interference.

In addition to these RF measurements, a host of protocol-based measurements determine the ability of the base station to perform the proper operations in conjunction with the mobile terminal. These measurements, along with RF tests, can completely characterize the performance of a CDPD base station.

Instruments are now available that perform these measurements. Because of the nature of the service environment, test equipment manufacturers originally chose not to incorporate both RF and protocol measurement capabilities in a single instrument. Potential users of these instruments do not always need to make both RF and protocol measurements, but rather one or the other. This scenario is likely to continue in the future, with test and measurement advancements coming in areas such as speed, automation and integration with other instruments via a network.

Summary CDPD, as a complementary service to AMPS, has the potential to degrade the performance of the system to which it is attached. Properly functioning RF components and subsystems can reduce this possibility, especially when the sniffing-and-hopping mode of operation has been chosen by the service provider.

Service providers who have chosen the dedicated-channel method of operation (who are growing in number) still must ensure that the RF performance of their CDPD system does not degrade the performance of the host AMPS system in terms of adjacent and alternate channel interference and other parameters. As a result, periodic monitoring of RF performance remains a necessity.