Drive Testing From Telephony
Drive Testing From Telephony
Drive Testing From Telephony
TI envisions phone becoming central repository for apps, media and info Camiant supports LTE in new policy server Supercomm shifts to October to capitalize on broadband stimulus spending CTIA: The year of mobile ads CTIA: AirWalk linking the enterprise PBX to the femtocell WiMax from Sacred Wind blanketing Navajo lands The daunting task of Sacred Wind
More Related Articles The old joke about the subscriber being the wireless service provider's primary test tool lost its punch long ago. With so many wireless choices today, maintaining high-quality service is essential. Drive-test systems have become vital weapons in the war against poor network performance and churn. Drive-test systems, as the name implies, are test tools that characterize the call-handling performance of a system simply by driving around and making measurements. These systems usually have relied on a wireless phone to make service-quality measurements from a subscriber's perspective. The primary drawback of this approach is that the phone is dependent on the network it must analyze. Thus, its ability to find problems is limited. A better approach is to use a probe that is independent of the network to collect data. This is the approach taken by some of the latest CDMA drive-test tools, which employ a digital receiver, rather than a phone, as the probe. Because the receiver is independent, it is able to measure pilot pollution, missing neighbors and base-station timing errors. Although the phone still is needed to determine what kinds of problems the subscriber encounters, such as dropped and blocked calls, the receiver provides more complete measurements to determine why these problems are occurring. A Balanced Approach One of the greatest contributors to dropped calls and poor service quality is "pilot pollution." This occurs when pilot signals (Walsh code 0) from too many base stations appear at the wireless phone and cause interference. Although wireless phones perform well when being served by as many as three base stations, service quality degrades rapidly when the number increases. To determine which base stations are serving an area, optimization engineers monitor pilot channels and determine how many of these pilots are significant. Your goal is to optimize network performance so that at any given location, no more than three strong (and appropriate) pilots will be present.
The closest base stations are usually the most desired to provide service. Although common sense dictates that the closest base stations always will be the strongest, the vagaries of RF and microwave propagation over varying terrain make it possible for distant base stations (10 miles or more away) to have a high received power level at the mobile unit. One way the CDMA network deals with optimizing hand-off is by limiting the number of base stations that the phone looks for when in a particular cell. The more accurate the "neighbor list," the better the system will perform, because only the most appropriate (usually the closest) base stations will be on the list. Phone Shortcomings The phone-based system can determine that there is a problem (calls are dropped at the intersection of Mendo St. and Hillside Ave.), but it offers little insight about the cause of the problem. The receiver-based approach helps engineers determine the cause quickly. Figure 2 on page 48 compares phone-based and receiver-based measurements at the same location and time. The top of the illustration shows those measurements made by the phone. There are four "active" pilots (PN 220, 152, 396, and 468) and 14 "neighbors" (PN 96 through 372) at this location. Because the network determines which pilots to look for, the phone measures only 18 of the 512 possible pilot signals. The receiver measures any of the 512 possible pilot signals received independent of what the network dictates. Optimization engineers generally refer to this as "PN scanning." Base station is the strongest and should be communicating with the phone. However, the phone does not detect it and cannot perform a soft hand-off with it because it is not in the neighbor list provided by the network. At this location, the power received from 428 is purely interference, only increasing the noise floor. The phone-based system provides frame-erasure-rate (FER), which indicates that voice quality is poor. What it does not provide is a clue as to why this is the case. Due to the nature of CDMA signals, a network-independent digital receiver is needed to see this type of interference, which could not be measured by an analog scanning receiver or spectrum analyzer. In the previous example, the distant base station 10 miles away may not be expected in a certain area and may not be included in the neighbor list. Because the phone rarely monitors its remainder set, the interference caused by this distant site could go undetected. The Plot Thickens In the example, calls are being dropped at the corner of Mendo St. and Hillside Ave. because 428 is not included in the neighbor list. PN 428 actually could be a distant base station. If it is, it would not belong in the neighbor list even though it has a strong relative signal strength. Instead of adding 428 to the neighbor list, the solution may be an antenna down-tilt or turning down the transmit power at the base station. The receiver-based system calls this situation to your attention through its ability to identify absolute delay (a combination of propagation delay and base station timing errors)
from the base station to the mobile. If base station 428 is not having problems with its timing and has a long propagation delay, then it is not close by and deserves further scrutiny. The receiver-based system also can identify one of the phantom problems that plagues carriers, the "island cell." CDMA systems are reliant on the precision timing reference provided by the GPS. When a base station's timing is skewed, system operation in its area is dramatically degraded. The ability to identify island cells has eluded carriers using phone-based systems, resulting in days of drive testing and visits to numerous base stations to find the offender. Engineers can identify the cause with the receiver-based system by driving past base stations and measuring the delay. The receiver uses the 1-pulse-per-second-signal from GPS as its timing reference, whereas the phone derives its timing by decoding the sync channel (Walsh code 32). Assuming base-station timing is correct, delay should be low (six chips of propagation delay corresponds to approximately one mile) at the site. If it is not, the base station is having timing problems. The receiver-based approach also can be used to optimize newly constructed CDMA networks, even before the switch comes on-line. As long as the base stations can transmit pilot signals, the system can measure them. The receiver-based solution does not negate the validity of its phone-based counterpart. The receiver system makes network-independent RF measurements. It does not provide information that the phone-based system does about FER, which is necessary in optimization. The phone-based system also can perform all of the functions of a regular phone, such as initiating calls and testing and compiling data about blocked and dropped calls. Once identified, data collected by the receiver-based system can be analyzed to determine why the problems are occurring. The phone-based approach gives the network engineer a good feeling for how well the network is performing from the subscriber's perspective. The most comprehensive test system combines both phone- and receiver-based approaches. The result is a system that diagnoses what is wrong with the network and determines why these problems are occurring.