Base Station Testing Using a Spectrum Analyzer
The quality of service a wireless network operator can provide is dependent upon the performance of the base station. By Tom Rittenhouse, IFR Systems, Inc.
The quality of service a wireless network operator can provide is heavily dependent upon the performance of the base station, a large portion of providing this quality depends upon the performance of the antenna and coaxial feeder connected to the base station. A degraded antenna installation can result in dropped calls and poor coverage. This leads to lost revenue, and irate and lost customers.
When installing and maintaining an RF feeder and antenna, the key measurement to test the quality of the system is return loss (VSWR). If the return loss measurement fails to meet specification, a fault location measurement can be used to provide a precise identification of the faulty component's location.
The 2399 spectrum analyzer Distance to Fault (DTF) option provides a quick and convenient method of measuring return loss (VSWR) and fault location on coaxial transmission lines in real time. The high selectivity of the spectrum analyzer is a big advantage as this measurement is performed in a polluted RF environment. The 60 dB dynamic range (sensitivity) in DTF mode will aid in locating obscure items like connectors and cable clamps.
During commissioning, the system performance can be verified and stored. At a later date, the data can then be used for comparison purposes during future routine maintenance testing.
During the installation and commissioning stage of a site there can be many sources of reflection in the cable that will cause degradation of the wanted signal out of the antenna of which adapters, connectors, jumpers and antennas are just a few. Also, sometimes parts are simply bad out of the box.
The measurement of return loss (VSWR) is fundamental for the characterization of a system in the frequency domain. The measurement involves applying an RF signal over the operating bandwidth of the system and measuring the amount of power reflected by impedance discontinuities within the system. The source of the RF signal is the tracking generator, while the spectrum analyzer is the measuring system. The return loss is simply the ratio of reflected signal to input signal, expressed in dB.
As the RF signal moves from the transmitter through the cable to the antenna, any imperfection will cause some of the signal to be reflected back towards the transmitter end. The worse the imperfection, the greater the amount of power reflected back to the transmitter. Every system is different, but in cable, one would expect to see a return loss of 14 dB or better (1.5 VSWR). This equates to less than 4 percent of the transmitted power being reflected back into the transmitter. The more power reflected means less power is transmitted from the antenna. This equates to a smaller than planned footprint for the radio site. If just the feeder were to be tested, the antenna would be replaced with an RF load at the end of the cable. Generally, the numbers would improve over an antenna measurement.
Figure 1 displays a failing antenna feeder system with a maximum return loss of 12.8 dB at 876.60 MHz.
A directional device (return loss bridge) is required to separate the transmitted and reflected signals. Calibration is performed with an open or short connected to the test port. The directivity of the return loss bridge should be 10 dB better than the expected value to be measured for nominal accuracy.
Crushed cables cause the outer and inner conductors' separation distance to change, causing an increase in impedance. When water enters a connector, the relative permittivity increases and the characteristic impedance decreases. Lightning strikes usually take the form of a streak of metal running from the outer conductor through the dielectric towards the inner conductor, reducing the characteristic impedance. Exceeding the critical bend radius changes the dimensions between the inner and outer conductors causing an increase or decrease in impedance. A kink presents the greatest problem resulting in thinning the center conductor, which dramatically increases the impedance.
Figure 2 displays the return loss of an antenna that is tuned to 821.2 MHz and does meet a minimum specification of less than 14 dB return loss.
Single Ended Insertion Loss
Single ended insertion loss measures the insertion loss of a cable by connecting to only one end of a cable. It is measured by looking at the reflected power of a cable system with an open or short connected to the remote end of the cable, instead of the customary load while testing for return loss. Since all the tracking generator's power is reflected by the short or open, the level being measured by the return loss bridge is the total loss of the cable. The level measured is then divided by 2 because the signal traverses the cable twice, forward and then reflected back to the measurement port. This enables a relatively accurate measurement of the insertion loss of an installed cable. It has several uncertainties, but when used in conjunction with a return loss and DTF measurement its results can be confirmed.
Distance To Fault (DTF)
The third measurement necessary to establish the performance of feeders and antennas is some type of fault location. Return loss and insertion loss can partially characterize a feeder system but it is only by including a fault location measurement that one can have full confidence that the system under test is satisfactory. Fault location is a measurement of return loss (VSWR) versus distance.
Fault location measurements also involve applying an RF signal to the transmission line over the operating bandwidth. The reflected signals are recombined with the input signal to produce a ripple pattern. This ripple pattern has, encoded within it, amplitude and distance information for all the reflections occurring within the line. A Fourier transform is required to decode this complex waveform. Calibration is performed by connecting a matched load to the test port of a power divider and normalizing the response.
Figure 3 is a display of fault location. Marker 1 is a bad connector. Marker 2 is the end of the cable. Marker 3 is a kink in the cable. Marker 4 is a mashed section.
Once it has been determined that the return loss measurement is out of limits, the location of the problem must be found. Inconspicuous problems of crushed or kinked cable are difficult to find requiring a sensitive Fault Location system. With a dynamic range of 60 dB, the IFR 2399 shows connectors, ground straps, tight cable clamps or cable kinks. A previously saved trace showing all of the above can be readily compared with the present trace to see differences. For cable lengths exceeding 150 feet it is important to be able to see known phenomena the tight cable clamp, connector and other minor faults in relation to your suspicious problem. Knowing your problem is 6.5 feet up the tower from a known cable clamp is a real road map to the actual physical location of the fault. The fast processor speed will show the location of intermittent problems that could be caused by the wind whipping a cable. Its pinpoint accuracy will aid in determining the true length of the feeder and will help in figuring cable loss and effective radiated power (ERP).
Figure 4 is a computer screen shot of the EasySpan II display. In addition to all the spectrum analyzer information, it shows 5 moveable markers, remote front panel and a limit line for triggering a trace to be saved if it is exceeded.
IFR EasySpan II software operates on a PC in a Windows environment and enables traces to be extracted from IFR test equipment. EasySpan communicates remotely with the instrument through direct connection or a modem, using the RS-232 serial communication format. The instrument can be remotely located around the world and when used with a modem, will transmit trace information through telephone lines. With the use of limit lines and trace save features, it easily provides traces with time and date stamping for verifying interference problems as they occur. Applications can be written in text format to automate test procedures.