Primary Distribution Of Audio Programmes By Wire.
Associated Companies in over 175 towns in Great Britain and in Barbados, Bermuda. Canada, Ceylon, Guyana,
Hong Kong, Jamaica, Malaya, Malta, Singapore, South Africa, Trinidad and West Africa etc.
REDIFFUSION ENGINEERING LIMITED
Carlton House, Lower Regent Street
LONDON SW1Y 4LS
A link is the communication channel between the point where the actual signal is received and the feeder distribution point, as opposed to the feeder system itself, and it may be High Level, Service Level or Low Level depending upon the voltage at which the link is driven as compared with the Service Level of the feeder system.
High Level Links
In the days before transistor feeder amplifiers, such as the A.623 and AA.105, Rediffusion links were usually High Level Links carrying a voltage of up to 660v supplied by large amplifiers, typically the A.126 and A.40. The object of the high voltage was to minimise the
copper loss and so permit the use of wires of moderate gauge. The problem with this was the capacitive loading of the link cables and the feeder cables coupled via the H.L.L. transformers. It was possible to reduce the capacity by the use of open wire links but this was not a practical proposition in built-up areas because such systems tended to be bulky and unsightly, particularly if the distribution of many programmes was required. The result of the capacitive loading was three fold.
First, the frequency response tended to be poor (unless the length of the link was severely restricted) and it varied with loading and so could not easily be corrected.
Second, the amplifier had to be larger than was necessary for supplying sufficient audio level to the subscribers: in fact the capacitive loading could exceed the subscriber loading if the subscriber density was low.
Third, the efficiency of any amplifier driving a capacitive load could be very low indeed and so the operating cost tended to be high: in fact the mains consumption and cost of valve replacements for the type of high-power amplifier required to drive a High Level Link was higher, for a given audio power, than a number of smaller transistor amplifiers and, in addition, High Level Links were themselves rarely more than 70% efficient.
Added to this there was the problem of resonance which on a long, lightly-loaded link could result in standing waves (similar to an
un terminated vision pair) causing peculiar frequency response and which, in severe cases, could effectively short circuit the amplifier at
certain frequencies. It could be possible to use inductive loading to maintain a constant, non-reactive link impedance analogous to the characteristic impedance of the vision cables (of very low Z) but this would only be possible if the load were constant and evenly distributed (and also matched to the link). This was the technique that was used but it was fraught with difficulties and the loading
coils themselves were expensive and unless carefully manufactured tended to reduce the reliability of the system.
Service Level Links
A logical development of the High Level Link system was the development and introduction of small, high quality line amplifiers (AA.106 and AA.107) used to drive the link system (instead of the high-power A.126 amplifiers), and the H.L.L. transformers were replaced by feeder amplifiers (A.623 or AA.105).
Given proven reliability, this system had several advantages over the previous High Level system of distribution and so the use of Service Level Links feeding transistor amplifiers superseded the use of High Level Links in new developments, for the following reasons:
First, the feeder load and capacitance did not affect the performance of the link and so subscribers at the far end of the link received
exactly the same service as those at the beginning regardless of feeder loading.
Second, the power consumption of the system was very low indeed, resulting in a considerable saving in operating cost.
Third, although the capital cost of the amplifiers was comparable for the two systems, there was a considerable saving in cable cost because the link could be a light gauge multipair cable instead of the heavy gauge quad cables necessary for the H.L. L. system.
There were, however, some operational problems associated with the S.L.L. system.
The main drawback of using transistor amplifiers was that they required a mains supply and they occupied more space than the step-down transformers of a High Level Link system. When first introduced, transistor amplifiers were less reliable than valve amplifiers used in High Level Links. Developments in transistor technology resulted in the transistor amplifier becoming more reliable and its advantages
began to outweigh other considerations.
There was still the effect of link capacitance and this had to be allowed for in link design. It was possible to use inductive loading as the load on the link could be made constant and this would give the better results in that the longest link that could be energised from a single amplifier being the best solution if a very large, complex system was required. However, even in such a system the induction coils could be greater than the saving in amplifier cost. The alternative method was to swamp the capacitive effect by loading the link resistively, so attenuating all frequencies to the extent that, proportionally. the loss at high frequencies was less; although attenuation was high, the system had a better frequency response than an unloaded link Figure 1
This resistively loaded arrangement was the one adopted for Rediffuslon Service Level Link systems.
The link, which was usually 26 s.w.g. multipair cable, was loaded at intervals of 1 mile with a 600 ohm resistor.
In practice, it was rare for actual loading resistors to be needed because the input impedance of the feeder amplifiers (or their input attenuator pads) was usually arranged to provide the necessary loading. The recommended circuit is as in Figure 2 which shows the CN.100/A pad coupling the feeder amplifier to the link. The attenuator part of the pad, R1, R2 and R3, was arranged to drop the level and also to load the link to the necessary degree, whilst R4, C1 and R5 provide frequency correction where necessary. At the beginning of
the link, for example, R4 will be short-circuited and R5 and C1 omitted. R1 and R2 will add up to nearly 600 ohms whilst R3 will be low, giving a 600 ohm input-impedance, high-attenuation pad with no frequency correction. At the end of the link, R1 and R2 could well be zero. R3 would be slightly over 600 ohms and R4, R5 and Cl would be arranged to give 3dB of "treble boost" at 8 khz.
Setting up levels was by no means a casual process and required great care and experience. Although there was a.g.c. on
both line and feeder amplifiers, these were only designed to cater for variations between programmes and also, to some extent, between
different operators’ ears; they were not really adequate to cope with an inaccurate gain setting.
The a.g.c. on all these amplifiers only worked at or near the full output voltage: the AA.106, for example, needed to be adjusted to give its full output (the adjustment being made with the a.g.c. "Out") otherwise, if the level was too low, the a.g.c. would have no effect.
It was necessary to set up the link-driving amplifier. This may have been a modified AE23 in older installations where the feeder amplifiers were A.623 amplifiers, or in later installations it may have been an AA.106 or even an AA.107 if the system was a large one.
With the type of amplifier and sending level decided and the feeder amplifiers adjusted, the programme was then be applied to the link-driving amplifier and the gain increased until, with the a.g.c. "out" and the amplifier driving either the link or the internal 600 ohm load in the amplifier.
Both the AA.107 and modified A623 were 30W amplifiers and could drive several links in parallel at 55V. The AA.106 was a 3W amplifier and had to be tapped down to 27V to drive one link otherwise be overloaded at the high frequencies by the capacitance of the
link. Even with this limitation, the AA.106 was the preferred link amplifier because of its very high reliability.
A126 3KW Amplifier
AA.105 Transistor Amplifiers
Click thumbnail for larger view