By: E. Mayberry

ABSTRACT

Various approaches to slotted cylinder antenna electrical design are available. Choices made at the design stage impact antenna performance parameters important for DTV use. The antenna output response performance, radiated signal amplitude and phase frequency response across the channel, is a major concern for digital TV. Most critical to antenna output response are the aperture illumination and the feed method: end or center feed. The effect of these design choices on the frequency response performance of the beam tilt and null fill are examined.

Specifically, the gain variation across the DTV channel versus depression angle for both end-fed and center-fed designs are evaluated. A comparison of calculated and measured results for a center-fed design is presented. Beam tilt variation across the DTV channel as a result of end feeding the antenna is considered The impact of end-fed antenna output response variations on the selection of antenna gain is discussed. Also, the performance of adjacent channel antenna designs using end and center feeding is compared.

INTRODUCTION

U.S. broadcasters planning their digital transmission facilities must choose a new antenna for DTV. A variety of antenna designs are available to the broadcaster from the many antenna manufacturers located around the world. Complicating the selection issue for the broadcaster is the necessity of locating and operating the DTV antenna system simultaneously with their NTSC antenna system. The broadcasters’ goal is to configure the DTV and NTSC antenna systems to provide good signal coverage while minimizing tower wind loading.

The vast majority of UHF antennas currently used in NTSC service are slotted cylinder designs. Performance characteristics that made slotted cylinder antennas the antenna of choice for NTSC UHF service are also desirable for DTV, i.e., excellent omnidirectional azimuth patterns, low wind loads, and smooth null fill. However, the digital TV transmission system will require more stringent performance with regards to output amplitude and phase frequency response. As a result, the antenna output response performance, which was given little consideration in NTSC service, is an important consideration for DTV.

WHY SLOTTED CYLINDER ANTENNAS?

UHF slotted cylinder antennas gained prominence in NTSC broadcasting due to their combination of low wind loading and superior omnidirectional performance. Their small diameter construction, most within the 8" to 14" outside diameter range, provides the minimum wind load reducing the cost of tower structures. Small physical diameter also translates into small electrical radius (R/l ) for the slot radiators, which results in excellent circularity of the azimuth pattern.

Figure 1 shows an overlay of two azimuth patterns. The smooth, nearly circular azimuth pattern, is typical for a slotted cylinder antenna with a circularity of ±0.5 dB. Compare it to the typical azimuth pattern of a panel antenna with a circularity of ±2.0 dB. When both are normalized to unity, the amplitude difference between the patterns at the minimum of the panel pattern reaches 3 dB. Considering approximately 1 mile of coverage loss per dB, this could mean 3 miles of service reduction in some directions for the panel.

Of course, if the patterns were normalized to the same RMS value, the difference would reduce, but would still amount to 2 dB. In the NTSC domain, normalization to the same RMS value is the rule, however, it is not currently clear the same applies to DTV. The FCC has assigned a directional reference ERP pattern for every digital station allotment that sets the maximum radiation at each azimuth heading. The FCC DTV rules suggest that the peaks of the patterns must stay below the reference ERP at those azimuth directions, which would effectively prohibit use of the RMS gain. Further FCC clarification on this issue is expected.

The other principal advantage of the slotted cylinder antenna relative to a panel, low wind load, is demonstrated in Figure 2. This comparison illustrates two approaches to providing an existing UHF broadcaster with DTV & NTSC service from the same tower top. The slotted cylinder stack is typically 2 to 3 times lower in wind area than the wide band UHF panel [1].

Figure 1

APERTURE ILLUMINATION

All the parameters that describe the elevation pattern are determined by the amplitude and phase illumination of the antenna aperture. The important parameters are beam width, gain, beam tilt, and null fill. Two antennas with the similar number of layers can have greatly different elevation pattern results depending on the aperture illumination design as demonstrated in Figure 3.

While the type of elevation pattern shown in Figure 3a was widely used with success in NTSC, the following investigations into the antenna output response performance will demonstrate why the elevation pattern of Figure 3b, with its smooth null fill response and wider beam width, is far superior for DTV.

Figure 2

Figure 3a

Figure 3b

ANTENNA OUTPUT FREQUENCY RESPONSE

The antenna output frequency response received virtually no attention for NTSC applications, but is of major importance for digital TV. Consider that the visual carrier, color sub-carrier, and the aural carrier dominate the NTSC RF spectrum with signal energy falling away rapidly from the carriers. NTSC antenna optimization was often performed concentrating on the visual carrier + 2 MHz. The effective luminance bandwidth is less than 4 MHz. By comparison, the DTV RF spectrum is flat across of the 6 MHz channel, except for the last 0.3 MHz. The entire channel is of equal importance and of larger effective bandwidth than NTSC. The antenna can no longer be optimized around a 2 to 3 MHz of the channel; DTV antennas should exhibit flat output response over a larger bandwidth.

END-FED ANTENNAS

Many slotted cylinder antennas are designed to feed the RF power from the bottom end of the antenna. This is mechanically convenient, especially for antennas mounted on the tower top. Broadcast slotted cylinder antennas must produce the main beam perpendicular to the vertical axis of the antenna. This requires that each slot level be nominally in phase. With the signal fed from the bottom and traveling towards the top, the end-fed antenna is made with a nominal one wavelength spacing between slots at the design frequency. As the signal progresses upward from one slot level to the next, a phase rotation of 360¡ occurs putting each successive slot level in phase.

However, the one wavelength spacing is only obtained exactly at the design frequency. As the signal frequency scans above or below the design frequency, the electrical spacing changes causing the beam tilt to vary. The end-fed configuration is depicted in Figure 4.

Click here for Figure 4

Consider a 30 slot, end-fed design with a smooth pattern, a calculation of the elevation pattern at the center (design) frequency, at the lower edge, and at the upper edge is plotted in Figure 5. Note that the beam tilt varies ± .25¡ from the design tilt.

Click here for Figure 5

The detrimental effect of this beam tilt sway with frequency is the variations it produces in the antenna signal amplitude (gain) and phase output responses. Figure 6a demonstrates the maximum calculated gain deviation over the DTV channel for depression angles 0¡ to 10¡ below the horizontal.

Figure 6a

Likewise, the maximum calculated group delay variation at depression angles from 0¼ to 10¼ is plotted in Figure 6b.

Figure 6b

CENTER-FED ANTENNAS

An alternative feed design employs electrical center feeding. This is simply done with side-mount antennas by using an input ‘T’ between the two antenna halves. Center feeding of top mount antennas is mechanically more complex, but is accomplished by using a triaxial configuration in the bottom half of the antenna to deliver the RF power to the center. These two configurations are shown in Figure 7.

Figure 7

With the center-fed design, the signal travels up the top half antenna and down the bottom half antenna. The beam tilt varies in each half as frequency scans across the channel, but the bottom and top half tilt in opposite directions which produces a constant beam tilt for the complete antenna. Figure 8 demonstrates the electrical considerations of center feeding.

Click Here for Figure 8

Returning to the center-fed illumination design of Figure 3b, a calculation of the elevation pattern at the center (design) frequency, at the lower edge, and at the upper edge is plotted in Figure 9. Note that the beam tilt variation is insignificant and variations in the nulls are minimal.

Figure 9

Figure 10a demonstrates the maximum calculated gain deviation over the DTV channel for depression angles 0¡ to 10¡ below the horizontal.

Figure 10a

Likewise, the maximum calculated group delay variation at depression angles from 0¼ to 10¼ is plotted in Figure 10b.

Figure 10b

All of the previously shown elevation patterns and output responses were calculated. Like most calculations, these results are better (demonstrate less output response variations) than will occur with actual antenna hardware. The calculations do not include the frequency response effects of individual slot radiators that have a specific "Q" factor. To demonstrate the effect of the radiator "Q" and other hardware factors, actual measurements of the center-fed illumination design of Figure 3b are presented in Figure 11a & 11b. As expected, the mesaured reaults show an increase in gain variation in the nulls. The beam tilt still has insignificant variation.

Figure 11a

Figure 11b

FREQUENCY RESPONSE COMPARISON

End vs. Center Feed

The above maximum gain variation plots allow us to compare the worst case signal variations versus depression angle; however, they do not reveal the frequency response shape. It is instructive to compare the end-fed and center-fed designs with regards to the frequency response plots for a specific depression angle.

For this comparison, the depression angle of the gain variation maxima nearest the design beam tilt angle for each antenna is used. This represents the worst signal variation that will occur at the greatest distance from the transmitter site for each antenna. The end-fed gain frequency response at a depression angle of 2.1¼ is plotted in Figure 12a. Figure 12b shows the end-fed group delay frequency response at the same depression angle. Note that the min-to-max differences correspond with the plots in Figures 6a and 6b.

Figure 12a

Figure 12b

The corresponding frequency response plots for the center-fed antenna are shown in Figures 13a and 13b. Note that the end-fed response is a straight line while the center-fed response is bell shaped which will result in a lower frequency response distortion penalty [2]. The calculated min-to-max gain variation of the end-fed design is 4.1 dB, while it is 1.4 dB for the center-fed.

Figure 13a

Figure 13b

The other significant point of difference between end and center feeding is the location of this first gain variation maximum. The center-fed antenna peaks at a higher depression angle, 3.6¼ compared to 2.1¼ for the end-fed antenna. The end-fed maximum at 2.1¼ occurs on the slope of the main beam as a direct result of the beam tilt sway with frequency.

ANTENNA GAIN & SYSTEM DESIGN

The above analyses clearly show that for a given number of layers the center-fed design is superior to the end-fed design due to lesser antenna output response variations and beam tilt sway. One approach often used to minimize these detrimental performance effects of the end-fed antenna is to design systems with lower antenna gains. Low end-fed antenna gain results in less beam tilt sway and increases the null fill levels to mitigate the impact of signal variations at depression angles below the main beam.

However, the low gain system designs can have significant economic impact on the transmission system costs. Compare two system designs for the following scenario:

DTV Ch. 40 @ 1 MW ERP
Line length = 1800’

A: Antenna: 27.5 Gain Center Feed

6"-50 ohms rigid (58.3% eff.; 70 kW rated)
TPO= 62.4 kW (3 tubes)

B: Antenna: 20 Gain End Feed

1) 8"-75 ohm rigid (68.6% eff.; 102 kW rated)
TPO=72.9 kW (4 tubes)
Transmitter and line cost difference to "A" ~ $650k Wind Area> 1.3 x tower load

Or

2) DTW-1500A (77.3% eff.)
TPO=64.7 kW (3 tubes)
Line cost difference to "A": ~ $200 k
Wind Area > 2.5 x tower load

The ability to use higher gain with the center-fed antenna allows the use of smaller rigid line and transmitter. Choosing circular waveguide to maintain the transmitter size increases the tower load 150%. Waveguide adds another source of signal distortion that requires compensation by the transmitter manufacturer due to the signal dispersion (group delay) across the channel.

ADJACENT CHANNEL DESIGNS

Although adjacent channel slotted cylinder antennas have been made using both end-fed and center-fed designs, the performance characteristics described above apply across 12 MHz as well giving the center-fed antenna the advantage.

The calculated elevation pattern performance for a center-fed antenna across the NTSC and DTV channels is shown in Figure 14.

Click here for Figure 14

An example of calculated elevation patterns for an end-fed, adjacent channel design is presented in Figure 15. Note the beam tilt variation and differences for each channel. Using the end-fed design produces a different beam tilt for the DTV channel compared to the NTSC channel.

Click here for Figure 15

CONCLUSIONS

Comparing slotted cylinder antenna performance of end-fed and center-fed designs, it is apparent that center-fed antennas provide significant advantages for DTV applications.

• Center-fed antennas have insignificant beam tilt sway vs. frequency.
• Center-fed designs with smooth null fill have low amplitude and phase response variations throughout the main beam and null structure.
• Frequency response shape of the center-fed antenna yields lower penalty due to frequency distortion.
• Systems designs can use higher gain, center-fed antennas to reduce transmission line and transmitter sizes resulting in lower cost transmission sites.
• Adjacent channel, center-fed antenna designs produce the same beam tilt angle specification for both the NTSC and DTV channel.

References

1. Mayberry, E. H., "Stacked Antenna Considerations for HDTV Conversion", Presented at IEEE Broadcast Technology Society 46^th Annual Broadcast Symposium, September 27, 1996.
2. Bendov, O., "A New Approach to the Analysis of Adjacent Structure Effects on HDTV Antenna Performance", Presented at the 1995 NAB Convention.