REPEATER DESIGN — NEWFOUNDLAND-NOVA SCOTIA LINKS 2-17 



Manufacture and testing of the electrical units was again to be carried 

 out by a contractor in a temperature- and humidity-controlled produc- 

 tion building, and in order to enable manufacture to start as early as 

 possible, arrangements were made for the contractor to co-operate with 

 the Post Office at an early design stage so that engineering could follow 

 fast on the heels of the design. The outer housing and method of braz- 

 ing-in the bulkheads had all proved entirely satisfactory on the proto- 

 type, and therefore these operations could proceed according to previous 

 production. The Post Office assumed responsibility for the production 

 and testing of the glands, since no contractor had experience of this work. 



Forward planning in early 1954 scheduled the first electrical unit to 

 be completed in June, 1955, with units following at 5-day intervals. This 

 target was, in fact, delayed until August, 1955, but all 16 working re- 

 peaters were available fully tested before the commencement of the 

 alying operation in May, 1956. 



Distortion Monitoring Equipment 



The pulsed-carried supervisory method as used on previous systems^ 

 is employed primarily for measuring the distortion on repeaters. Under 

 normal operating conditions the distortion level may be only just notice- 

 able above the noise, but should appreciable distortion occur, e.g., failure 

 of one amplifying path in a repeater, it could be readily located, since 

 the pulse amplitude from the faulty repeater, would increase by about 

 12 db for second-harmonic distortion and about 18 db for third-order 

 distortion. 



Loop-Gain Monitoring Equipment 



For the loop-gain monitoring equipment' the repeaters have to be 

 designed to incorporate a second-harmonic generator operating at a 

 frequency unique to each repeater at 120-cycle spacing in the 260- 

 264-kc band. The second harmonics return to Sydney Mines in the 520- 

 528-kc band, and these two channels must be taken out of service during 

 the measurement. 



Levels and Equalization 



Controlling Factors. 



In practice, deviations from an ideal system wherein all repeaters 

 match the cable and operate at all times at the same levels require the 

 repeaters to be designed with specific margins against overload and 

 intermodulation to meet an agreed maximum noise figure for the system 



