On Distributed Communications Series
VI. Mini-Cost Microwave
The nominal length of the mini-cost microwave link, or "span," which is composed of relay stations, shall be on the order of 100-200 statute miles. Each span shall comprise about ten tandem, simultaneously bi-directional (full duplex) links, and shall be designed solely to transmit digital data. The data rate of the spans shall be about 4.5-million bits/sec and the total overall span timing-jitter shall be less than
The span error rate shall be less than one error in l05 bits--measured on a long-term basis for at least 90 per cent of the time. If possible, an overall error rate of one bit in 106 bits will be sought, but is not mandatory.
The target up-time of a span shall be greater than 90+ per cent, assuming minimal routine maintenance and a mean 24-hour delay in the arrival of maintenance personnel and servicing. If possible, higher up-times are sought, but are not mandatory. (A 90 per cent figure corresponds to a mean time between failures (MTBF) of at least ten days per span, or 100 days per relay station.)
The prime design criterion shall be for the (minimum) ten-year system cost, and shall be considered to be even more important than the up-time goal. In computing the ten-year system cost, all assignable costs shall be considered, including those for rights-of-way, maintenance, power, etc. Research and development costs for the system components have already been factored out and are carried separately (see ODC-X).
All relay points shall be capable of operating without dependence upon a civilian power source for at least one week. The power requirement for each relay station shall be kept small (about 50 watts), to permit the use of a standby thermoelectric generator during emergencies. Such generators must be capable of full-time operation when power lines are not accessible, and be able to go several months between refueling. Fallback to the emergency supply shall be transient-free, if possible, and switchback to primary power shall be automatic.
Wherever possible, and as a measure to conserve thermoelectric generator fuel, primary electric power shall be obtained from commercial mains. Such power may be transmitted via a 117+ to 24 volt step-down transformer, allowing the use of an inexpensive, unprotected, buried cable.
Since some 120,000 air line miles of these bi-directional links shall be required, mass production techniques must be used both in the construction and in the siting and maintenance of the stations in order to achieve the low overall system price sought. The establishment of component standards shall be left entirely to the jurisdiction of the supplier. Inapplicable standards (JAN, MIL, etc.) shall not be specified. Lower-cost components are perfectly satisfactory provided they fall within the desired system up-time requirements at the lowest overall system price.
The transmitter/receiver shall be a pre-tuned-assembly and shall be mounted as a complete, lightweight, weatherproof package, easily removed by a single repairman. Replacement units may be interchanged without need for field alignment or adjustment. The transmitter and receiver shall mount integrally with the antenna reflector to avoid the cost and losses of long waveguide runs. An efficient Cassegrainian reflector antenna is suggested.
The antenna feed horn shall be designed so that excess heat dissipation from the transmitter/receiver unit is used to aid in the melting of ice from the front of the antenna feed, where such a problem exists.
All towers shall be of minimum height and be capable of being climbed in safety under most weather conditions. Such towers may be guyed to minimize cost and to provide improved antenna stability. The tower should not be high enough (generally, 170 ft or less) nor located in such a spot as to constitute an aircraft obstruction; if possible, aircraft warning lights will not be used. In such situations where a tower must be painted and lighted, the lights should be powered from the commercial mains only and not from the emergency power supply.
Wherever possible, the towers shall be sited within government rights-of-way alongside existing roads where access to power lines is available (unless, of course, it can be shown that savings will accrue through use of purchased or leased sites). An allowable cost-minimization method in certain situations would be to use a pulse-regenerative cable in combination with single-direction antenna towers. This might be necessary in the case where there is no site available having good visibility to the preceding and following repeater stations. (For example, two sites located on opposite sides of a hill but working as a pair often provide better visibility to their contiguous relay stations than a single site on top of the hill.)
The carrier frequencies used shall be either in the high X-band or low Ku-band. (Because the Ku-band is a relatively unpopulated portion of the spectrum, prices of X-band equipment are generally more meaningful in estimating costs.) The increased two-way antenna gain as the frequency is increased roughly offsets the loss of signal generation efficiency, plus free-space loss, plus increased crystal mixer noise at the higher frequencies. Therefore, many calculations can be more conveniently made at 10 Gcs, although we prefer to actually use a higher frequency, such as 13 kilomegacycles, because of the congestion on the lower frequencies. All techniques used shall be feasible at either the high X-band or the low Ku-band.
As a duplex binary stream will pass through each relay station, a simple supervisory control means can be included to pinpoint a defective relay link by interrupting the normal data stream. This emergency circuit may also be used as an "order wire" for field maintenance. No other supervisory channel need be provided.
Nothing in these specifications should be interpreted as discouraging the designer from considering new and untried methods and devices, especially during the earlier phases of the design cycle. Sufficient time shall be allocated to the design phase to permit research into whether or not such potential savings can be incorporated safely, while maintaining a margin for falling back onto more expensive traditional practice if unforeseen difficulties occur.
 Walker, A. P., NAB Engineering Handbook, 5th, ed., McGraw-Hill, New York 1960, pp. 1-312 to 1-372.