FlightGlobal.com
Home
Premium
Archive
Video
Images
Forum
Atlas
Blogs
Jobs
Shop
RSS
Email Newsletters
You are in:
Home
Aviation History
1963
1963 - 0176.PDF
164 FLIGHT International, 31 January J 963 AIRCRAFT ELECTRICAL MACHINES . . Fig 4a Unsatisfactory fitting of potentiometer T*?' Fig 4b Modified fitting of same component actuators fitted with electric connectors traditionally sensitive to moisture are mounted on the undercarriage where they are exposed to spray from the runway; a fan-cooled rotary invertor draws air from a passenger compartment, ingesting in the process cigarette ends, carpet pile and other paraphernalia, the air inlet having been mounted too near to floor level. Scapegoats are all those components called upon to act as buffers between incompatible sections of equipment. Such scapegoats may be deliberately assigned to this function; or they may be dis covered, after several failures, to have unsuccessfully tried to meet excessive demands. They may be components like micro limit- switches, which on a particular unstable cabin-pressurization system were caused to be operated far in excess of design requirements; or a valve-operating actuator, which was made to oscillate just before the cut-out position thus inducing flutter and finally burning of switch contacts. Designing for Reliability An essential of any design aiming at high reliability is built- in latitude. The ideal design is traditionally one in which, after the expiry of a set life, all components fail simultaneously; but this cannot stand serious scrutiny where aviation is concerned. If a machine is to have a reliability of 95 per cent after time expiry of x hours, its initial reliability must be appreciably higher, since the failure rate of each component due to ab initio causes (material flaws, faulty assembly, etc) will increase as a result of general de terioration through wear, storage and exposure. If components were now to fail after x + 1 hour due to deterioration, then at x — 1 the nearly worn-out parts (still subject to the ab initio causes) would have an alarmingly low survival rate. Built-in latitude is therefore important: to allow for improvement if die product does not meet the specification; to meet the application if the specification has fallen short of actual demands; to contain a cer tain reserve to cater for an unforeseeable peak condition in the equipment or application; and to facilitate production, by pro viding a satisfactory product with the widest possible manu facturing tolerances. Despite adopting this policy of built-in latitude, firms with wide experience with certain types of equipment still sometimes produce unsatisfactory designs. Why? Making the same mistake twice is unlikely; inadequate liaison between different engineers dealing with the same problems is more probable. But one of the greatest sources of trouble is the inability of designers to read across fully from one installation to the next. For instance, the PV (unit bearing load • peripheral speed) values for a combination of bearing materials—a nickel-chrome steel and a phosphor-bronze alloy, using a certain grease—were estab lished after a number of designs. An increase in working tempera ture demanded a change to a special lubricant. Whilst this lubricant showed itself satisfactory in ball bearings, and in laboratory tests on the above pair of metals, trouble arose in service. Since an actual life of 3,000hr implies continuous testing over 125 days— which is not always practical—duty cycles are frequently condensed to provide the equivalent number of operations. This abbreviated test, however, did not reveal the slow corrosive action of the grease on the well-tried bearing combination, and the resultant loss of machine output. In this case two well-proven "Rights" produced a third "Right" in the laboratory, only to be proven "Wrong" in service. In view of the nature of the trouble, which is active during the whole installed life rather than only during flight, a full year's tesi would have been more representative. Even so, for a required life of 3,000hr, such a test would represent a daily aircraft utiliza tion of 8ihr, a rather low figure for airliners today. As such pro longed tests are clearly impractical, equipment tests must therefore be backed by long-term material and environmental testing, if be haviour in service is to be predicted with a high degree of certainty. Despite built-in latitude, and despite prolonged testing, "rogue" units still enter service. To ensure that vital services are not inter rupted, duplication or multiplication is now common in Class 1 application. Such equipment is defined in Av.P.24 Ch 101 as one whose failure during take-off, flight or landing might be the direct cause of: structural collapse; loss of control; failure of motive power; unintentional operation of, or inability to operate, any service or equipment essential to the safety or operational functions of the aircraft; and injury to any occupant. Duplication can be limited to certain components of relatively low reliability standard; or complete duplication of, say, an actua tor with its separate supply wiring and switching may be specified. In either case no single failure of one unit or channel must be allowed to affect the stand-by unit or channel. Motors, limit switches, potentiometers, design, internal wiring and connectors must be so designed that neither failure to operate, nor inadvertent operation nor a malfunction can interfere with the normal working of common items such as gearing, bearings and screwjack, which are powered by either the main or auxiliary channels, or by both simultaneously. In the "back-to-back" actuator installation men tioned previously, special precaution must be taken to ensure tnat, following a single failure which arrests one machine in a partly extended position, the unimpaired actuator cannot exceed the combined full stroke. Every development department has its own rogues' gallery. No single function—design, production or servicing—is whoDy responsible for turning such a rogue into a casualty; for instance, if servicing instructions were strictly adhered to, even a poor design would often get by. But a poor design not backed up by meticu lous care in servicing leads to trouble. For example, a trimming potentiometer (Fig 4a) was fitted with a tapered nut N and collet securing the trimming spindle S. To adjust the spindle, the tapered nut required slackening. If no special effort was made in servicing, the torque applied to N overcame locking nut L, with the result that the whole potentiometer body was rotated and wiring damaged. Hidden from view, this defect went unnoticed. A modified version (Fig 4b) dispenses with the tapered nut, and its tendency to seize, and relies on the spring loading of four tabs to prevent the spindle from altering its setting. Linear actuators are normally set for stroke length and attach ment centre-distance by the makers before dispatch. Where the application is provided with external mechanical stops or buffers, it is the duty of maintenance personnel to ensure that these stops lie outside the limits of ram travel. Fig 5 (left) illustrates damage resulting from this omission. Where setting difficulties in the air craft are likely, a cushion drive fitting (Fig 5, right), permitting overrun in two directions, will often overcome this problem. The terminal post in Fig 6 was secured by two locknuts in a syn thetic terminal board. An accidental overload in ground testing excessively heated the terminal post, and caused the board to ex pand and retain a permanent imprint of locknut and washer upon cooling. The loosening of the terminal tag and its failure was due to arcing. The revised design of this detail made the electrical con tinuity independent of the terminal-post fixing arrangements.
Sign up to
Flight Digital Magazine
Flight Print Magazine
Airline Business Magazine
E-newsletters
RSS
Events
