Figure 1. Spectrum analyzer display of adjacent TV channels, DTV channel 19 (center) and NTSC channel 20 (right). Note NTSC signal’s visual, chroma and aural carrier components. Image courtesy the Freberg Engineering Company. A copy of John D. Freberg’s NAB05 paper is available for purchase from the NAB publication office. Click here to see an enlarged diagram.
Like it or not, DTV is not only coming, it has fully arrived. With essentially everyone operating a DTV facility at some level, it's now possible to get a good idea of how successfully the new medium is performing. We aren't talking about a good controlled test environment, but some general observations, primarily the result of a lot of antenna measurements and the subsequent comments.
The VSWR considerations on DTV antennas aren't much different than the requirements that have been standard for NTSC systems for years. For NTSC, the basic goals have been an antenna VSWR around 1.05 with a system VSWR of under 1.0:1 across the 6MHz. band. Emphasis was always placed on the visual carrier, aural carrier and color frequencies because those were where the greater amounts of energy existed in the transmitted signal and where the most effect would be observed in the received signal.
Reflected signals at or very near the visual carrier will result in a good old-fashioned ghost in the received signal. This is probably the most objectionable result of high VSWR as far as the viewer is concerned, which brings up another point that engineers are likely to forget. The purpose of all the tuning isn't just to make ideal meter readings in the transmitter building. The goal is the best possible signal quality for the viewers — distortion- and ghost-free to the fullest extent practical. Luckily, that goes along with the good meter readings — usually.
With regard to VSWR at color and aural frequencies, the result is distortion. That distortion shows up in the audio as degraded frequency response, increased harmonic distortion, increased cross talk and decreased separation in stereo systems. At color frequencies, the result varies but can result in “smearing” of color information. It's the old “the blue of the eyes isn't supposed to appear on the lips” syndrome. The entire thing gets worse when it is realized that the television transmitter really wants to see 50 + j0Ω across the television channel. As a general rule, the more the channel impedance varies from that amount, the greater the degradation of the transmitted signal. However, the worst problems occur around the three more critical frequencies.
For DTV, there really isn't any small part of the channel where the power is significantly greater than the remainder of the channel and where significant information is carried. The information is essentially spread across the entire channel and not significantly susceptible to minor amplitude variations. Reflected signals show up at the receiver essentially the same as multipath signals would appear, that is, the same signal, reduced in amplitude, arriving slightly later in time. The current generation of DTV receivers will cope with multipath signals that are as large in amplitude as the direct path desired signal. Therefore, the total system can cope with minor reflections, even though larger amounts of VSWR may distort the signals so badly that the bit error rate (BER) is increased.
For DTV, the initial goals were to keep the antenna down to around 1.05:1 and the entire system under 1.1:1 across the channel. However, there is a significant amount of thought that DTV systems will perform quite adequately as long as the average value of VSWR is well below 1.1:1. In other words, some excursions slightly above 1.1:1 might be acceptable if the majority of the response across the channel is lower. As there is no critical frequency in the ATV signal, what were previously thought to be undesirable variations may not be as bad as originally feared.
More work needs to be done in this area involving real stations and on-air signals, not just simulations in the lab. There are a few problems involved in such testing. First, it requires taking a station off the air, detuning the antenna and taking measurements over a clean path — preferably short. Second, it is necessary to schedule the necessary engineers, riggers and equipment to do the tests. A third issue is getting someone to either pay the bills for all those people or getting them to simply absorb their costs in the interest of gaining knowledge.
So far, it has been difficult to meet all those problems at the same time, but it is being worked on. Until such research is completed and evaluated, the wise course is to attempt to meet the above criteria — that is, the antenna at or below 1.05 and the system at or below 1.1. This seems to result in good performance by the DTV transmitting system.
One desire of a lot of stations has been to diplex their DTV signal onto the same antenna as their NTSC signal when they are either first adjacent or they are only separated by a few channels. The separation by several channels is simplest to deal with. Unless the antenna was specifically designed for broadband operation, it won't work. Most NTSC UHF antennas are designed and tuned for a specific channel. All the initial tuning work is done to optimize the antenna over 6MHz with the understanding that everything outside of that bandwidth may go to the bad place as far as the designers are concerned. As NTSC signals on adjacent channels were not allowed, no one really cared about out-of-channel impedance values.
If one is lucky, and the antenna isn't too old, diplexing on n±1 may be possible. Antenna designs in later years have tended to have a little better bandwidth that extended outside the channel a bit. To determine if this is the case, a network analyzer can be used to evaluate the input impedance of the antenna on the additional channel. This is done by looking at the antenna in the time domain mode over the additional desired channel. If the VSWR at the antenna is fairly low, it may be possible to add a fine matching section at the antenna to achieve satisfactory operation. Such operation may even be possible in systems with round, truncated or rectangular waveguides. The transitions to and from waveguide will often have to be redesigned to allow additional tuning, but such hardware is far cheaper than having to replace the entire antenna.
In any case, the only reasonable way to attempt such diplexing is to obtain the existing VSWR data on the desired channel and then go directly to the manufacturer. Other problems may exist that rule out any combining signals. For example, if the transmitting antenna is directional, what will the pattern look like on the new frequency? However, it has been shown that diplexing on single channel antennas can often be done by accepting some slight increase in VSWR on the DTV channel. Again, it required careful measurement, adjustment and coordination with the manufacturer.
This article has primarily been concerned with the effects of VSWR on analog and DTV signals. The measurement of the transmitter output regarding distortion, BER and other variables is a much broader category. Readers are advised to review the excellent articles that were presented in that area at the recent NAB conference. In particular, “Understanding DTV Transmission Measurements” by John D. Freberg is an easy read with a lot of valuable information. The paper can be found in the “Conference Proceedings,” which are available from the NAB store.
Don Markley is president of D.L. Markley and Associates, Peoria, IL.
Send questions and comments to: firstname.lastname@example.org