 |
UK Broadcast Transmission |
        |
|
    |
|
|
|
|
|
||||||||
|
|||||||||||
ADAPTIVE AERIALS FOR UHF REBROADCAST
LINKS
From an original article by MD Windram, L
Brunt and EJ Wilson
PART THREE - OPERATION
a. Array
 |
|
Fig. 7: The SABRE array installed on the Alderney tower (the elements are in the flat panels). The masthead box containing the complex attenuators, filters and preamplifiers is mounted on the access platform immediately behind the array. |
For Alderney. the aerial requirements were not only the horizontal distribution of 16 elements at ~2/3 wavelength spacing, but also a gain of ~24 dB. This was achieved as a 16x4 dipole array constructed as a 2x2 array of 8x2 dipoles. The 8x2 dipole array was constructed with a solid backplane and used eight printed circuit panels, each consisting of two bat-wing dipoles with printed feeders to a single point. This design combined wide bandwidth with good reproducibility, performance and ease of construction. The dipoles are protected from the direct effects of weather and salt spray by a fibreglass front cover to the box structure holding the 8 x 2 dipole array. The aerial array on the tower at Alderney is shown in Fig. 7, and Fig. 8 shows the arrangement of a single module with the cover removed.
 |
|
Fig. 8: An 8 x 2 dipole array module shown with the protective front cover removed. In the operational model, the four modules are mounted as a 2 x 2 array at a height above ground of approximately 23m. To meet structural requirement, the array is in the form of a 17-element array with the middle element missing but measurements and computer simulation show no significant differences from that of a 16-element array. |
The four modules are mounted as a 2 x 2 array at a height of ~23 m up the tower at Alderney. Because of the modular construction, necessitated by mechanical and structural requirements, the array is in effect set up as a 17-element array with the middle element missing. Computer simulation and actual measurements show no significant difference in behaviour from that of a 16-element array.
b. Masthead box
The amplitude and phase adjustments, and combining, are all carried out in a box mounted behind the aerial. The block diagram is shown in Fig. 9.
 |
Fig.9: Block diagram of the masthead box which carries the amplitude and phase complex attenuators and the combining units together with selective filter to overcome any electromagnetic compatibility problems arising from the presence of a local transmitting aerial mounted on the same tower. |
The filters are required because there is a local transmitting antenna mounted on the same tower. The amplitude and phase adjustment is carried out by the complex attenuator, so called because it multiplies the signal by the complex number X+jY. The block diagram of an individual complex attenuator unit which is a single printed-circuit board is shown in Fig. 10.
| Fig. 10: Block diagram of a complex attenuator of which 16, one for each aerial element, are fitted in a masthead box. After amplification, the signal is split into two separate attenuators; subsequently the two processed signals are recombined in phase-quadrature. These complex attenuators have proved to be operationally very stable with a transfer characteristic approaching the ideal, permitting the constant-gain algorithm to function correctly. |  |
Use of the reflective type of attenuator is important in that it enables the gain in either the in-phase or quadrature path to be any value in the range - through 0 to +1 for a perfect attenuator. The performance of the complex-attenuator network with appropriate linearisers has been found to be very stable and to give a transfer characteristic closely approaching the ideal. This is essential for the constant-gain algorithm to operate correctly.
Because the box is mounted at masthead, complete testing facilities are provided so that, by means of p-i-n diode switches, a test signal can be routed into any one of the attenuators to enable the attenuator to be checked. If the test signal is switched to a particular attenuator during normal operation, then it appears as interference and the system automatically adjusts the appropriate control voltages of the attenuator to zero; this state can readily be checked.
The masthead box also incorporates a voltage- controlled attenuator which forms an extension to the receiver AGC. The drive to this AGC is a single digital line which causes the gain to increase or decrease slowly with time. This has been designed to operate not only in a system such as the adaptive aerial, but also for masthead pre-amplifiers and can be driven from any combination of receiver AGC system via a simple comparator unit. A photograph of the masthead box is shown in Fig. 12.
 |
Fig. 12: The masthead box. The RF input connections are at the top of the photograph with the sixteen selective filters mounted as two blocks of eight near the top. The two large enclosures contain the blocks of eight complex attenuators. |
c. Receiver
From the masthead box, the signal is passed down to the receiver or receivers. The type used is a modification of a receiver designed by the IBA and currently in use at the majority of its higher-power rebroadcast relay transmitter sites. The main feature of the unmodified receiver is synchronous detection which has been successfully used by the IBA for some years now to eliminate quadrature distortion which would otherwise result from envelope detection of a vestigial-sideband filtered signal. The detector used has automatic locking circuitry and a capture range >10 times the maximum input frequency error which can occur.
The main features of the modified receiver are:
This combination ensures a receiver capable of locking to the correct signal, and thus providing correct outputs from the measuring circuits. Without these three features, the aerial would not be able to optimise and hence reject the interference.
The co-channel interference is measured by sampling the synchronising pulse tips of the video, a signal which, in the absence of interference, should be constant in level. Interference appears as a beat at the synchronising pulse tips and is extracted by sampling to give a complex low-frequency signal in the frequency ranges 0-1 kHz or 2-4 kHz; these frequencies resulting from the aliasing of the possible offsets of the transmission frequencies used in the United Kingdom to reduce visual patterning caused by co-channel interference. These sample signals are fed into the measuring system.
d. Measuring system
Detailed theoretical and experimental work has shown that one of the major problems of adaptive systems, and the adaptive aerial in particular, lies in the area of measurement. A simple measuring system might operate by measuring the amplitude of the CCI beat, changing the aerial pattern slightly and remeasuring all amplitude of the beat. This was tried and worked well for CCI caused by unmodulated signals. It does not work satisfactorily, however, for real interference whose modulation is changing, even when the measuring period is synchronised to the frame period. For this reason, an alternative system was designed. This operates by defining two aerial patterns A and B with a very small difference in their control settings called the perturbation. The aerial is switched between these patterns on a line by line basis. The A alternate synchronous samples now provide a measure of the CCI for the A pattern and the B synchronous samples for the B pattern. Unfortunately, however, in the system described here, the output is in continuous use for rebroadcast and therefore any changes provided to the aerial must not be visible on the output. The perturbation size is set such that for CCI levels up to equal to the wanted signal level incoming to the receiving site, the CCI can be rejected to greater than ~56 dB down on the wanted signal on the final output (carrier-to-carrier ratio). This very small perturbation necessitates a very accurate measuring system capable of detecting changes greater than 0.2% in the CCI level, especially for multiple CCI. The method used is shown in Fig.14.
| Fig. 14: The CCI measuring system used in the receiver to provide adaptive control of the aerial array. The processed analogue signals are converted to digital form or further logic processing to provide the control signals. |  |
The analogue filters provide a carefully tailored response to eliminate for example aircraft flutter, and also to provide a suitable weighting of the nonoffset 0-1 kHz interference with respect to the offset interference (aliased to 2-4 kHz), since nonoffset interference is subjectively far more annoying for a given level than offset. The major feature of the above arrangement which is in essence filtering and detecting the A and B signals is the use of summing and differencing prior to the filters. This eases the filter component tolerancing from very difficult to simple and, in fact, abnormal differences in the filters result merely in desensitising the detector. This circuit has worked well in operation and indeed has a sensitivity ~0.1%, the limits being set by the initial samplers and the digitising accuracy. Digital detection and integration is used because it is very difficult to produce analogue detectors with the required accuracy and reproducibility.
e. Control unit
Adaptive control is provided by hardwired logic. Many algorithms were investigated during our computer simulations and it was found that the simple ‘hill climb’ modified to maintain constant gain, combined simplicity with a performance only about half the speed of very sophisticated and complex algorithms requiring large amounts of computation. Preset control values are incorporated so that during the night, for example, the aerial pattern is fixed ready for optimisation when the wanted signal is turned on. The controls are fed to the masthead box via analogue drivers which incorporate a degree of lightning protection.
Next >
|
|
|
||||||
|
