Eliminating that nasty ghost

The problem of ghosts in the picture resulting from impedance mismatches has been discussed in this column in the past. However, there still seems to be some misunderstanding concerning the cause of such ghosts. In addition, the transition to DTV has brought up some new problem areas.

Kerry Cozad of Dielectric and Bob Plonka of Harris have both presented papers discussing the problems of VSWR in the digital antenna system at IEEE Broadcast Symposiums. Copies of these papers can be obtained from IEEE.

In all television antenna systems, the problem of VSWR can be considered to be caused by at least three separate areas. First, and probably most significant, would be any mismatch between the transmission line system and the antenna at the input to the antenna. A signal reflected from that point will travel down to the transmitting equipment, be reflected in part and travel back to the antenna. This twice-reflected signal is then radiated as a signal delayed in time from the original. The radiated reflected signals appear in the receiver as ghosts of the original signal.

It is widely accepted that a VSWR at the input to the antenna of 1.05 or less, representing a reflection of no more than 0.03, will result in acceptable performance regarding ghosts.

The next problem would be the elbow complex connecting the transmission line to the base of the antenna. In some cases, as where the antenna is stacked or is a member of a group of antennas, there may be more than one set of elbows involved, along with another complex at the bottom of the tower. A significant mismatch in the elbow complex just under the antenna can cause a ghost just as significant as one caused by the antenna mismatch. Elbow complex mismatches at the tower base would not be expected to result in ghosting but will have an impact on the load seen by the transmitter.

The antenna and the elbow problems both have relatively simple solutions. The elbow complex(s) should be assembled at the factory and tuned for proper operation on the channel(s) involved. This often involves the installation of tuning slugs. Changing slugs in the field is usually not advisable. For a large complex, it is better to have the work done at the factory in a more controlled environment.

The third and final contributor to the system is the transmission line itself. Normally, that is not a problem if all connections are made properly and if the line hasn't been damaged during shipping or installation.

The mismatch problem at the antenna is resolved in two distinct ways. First, the antenna should be tuned at the factory for a VSWR at visual carrier of 1.05 or less. The visual carrier frequency is the most important, as there is the greatest power level at and immediately adjacent to that frequency. Then, after installation, variable fine matching slugs at the antenna can be adjusted when the antenna checkout is completed. This advice applies to analog systems. We will get to digital systems in a bit.

Modern network analyzers allow the VSWR at the antenna itself to be determined. The transmission line can be gated out of the measurement, allowing the antenna to be seen by itself without the effects of other VSWR contributors. Second, analyzers make use of a fast Fourier Transform to convert the frequency domain representation to a time domain response. The measurement is taken over an adjustable width of frequencies. For evaluating the antenna, a narrow band of frequencies (either 6MHz or 12MHz) is used. For the transmission line system, bandwidths of 100- to 400MHz are used for coaxial cable. The wider bandwidths are much more accurate in pinpointing the exact location of a discontinuity but are useless when looking at the bandwidth-limited match to the antenna. The narrow bandwidth tests accurately represent the antenna but show other discontinuities as broad peaks in the response, making it difficult to exactly locate the problem.

So, here is the normal drill. The initial measurement usually consists of looking at the VSWR at the input to the transmission line. The next measurement is normally a narrowband look at the system, hopefully showing a nice flat transmission line with only a mismatch at the antenna. The antenna tuners are then adjusted to optimize that match. Then, wideband measurements are made to confirm that the transmission line system has no problems with bad connections, dents, etc. If that first narrowband measurement shows a problem in the transmission line, that problem should be addressed before attempting to optimize the antenna match, since a bad spot in the line will affect any signal passing that point. In the case of the type of measurements being discussed here, that signal must pass the bad spot twice before a measurement can be made — causing significant errors. The final measurement is the VSWR at the input to the transmission line.

Up to now, the entire discussion has been about analog systems. In digital systems, there is no major energy contributor as in the analog visual or aural carriers. Rather, the energy is mostly spread across the entire channel within the limits specified by the mask filter. Also, there is no ghost in the DTV picture. As is well known, the signal is either there and perfect or it is gone. However, the reflections from the antenna are still terribly significant. VSWR problems contribute to the bit error rate. When the bit error rate exceeds the threshold, good-bye picture. Therefore, it still is necessary to reduce the VSWR and associated reflections to the greatest extent practical. Furthermore, the presence of VSWR in the system will impact the correction process in the transmitter itself, something unknown in analog systems.

The result is the need for some new standards of evaluation. The whole channel must be carefully treated. While it may be possible to have the VSWR exceed 1.1 or so at some point in the channel, the average value of the VSWR should be held to some reasonable value. In the meantime, lacking any other standard, the only reasonable approach is to continue adjusting and tuning for the best possible overall response.

Don Markley is president of D.L. Markley and Associates, Peoria, IL.