Aircraft have very long operational lifespans. Most provide greater than 20 years of usable, high cycle service life. I have worked on plenty of aircraft that date back to World War II, or the 50s. From DC-3s to Bell 47s, they remain very serviceable today.
Interestingly (and no doubt to the extreme annoyance of pilots), it is most often the component subsystems and their interconnects that fail early in the life of an aircraft, not the fundamental structure itself. Virtually everyone understands that wear and aging occurs in purely mechanical systems like bearings, engines and tires. But this phenomenon is less understood in avionics or electrical systems. A quick review of the wear and aging processes may be helpful to anyone trying to evaluate their system’s lifespan and how it will behave over time.
It’s intriguing (or perhaps humiliating, depending on your perspective) to see how many subsystems simply cannot survive as long as the airframe. And that fact dictates some sound and workable strategy for their replacement or refreshing during the aircraft’s life.
Age issues start right at the component level. Each type of component has a specific pattern it follows to become unusable. In simple terms, as soon as anything is manufactured, it begins to deteriorate, whether in active use or not. There is rarely anything that can halt these processes, some of which occur within "sealed" components and at the molecular level.
Generally, the deterioration processes are all oxidation related, or involve electrochemistry with water or salts. Electronic parts are often bagged for protection, or Mil-Spec packaging is used like Mil-B-131G and the later electrostatically protected Mil-B-81705B, in an attempt to halt surface oxidation, static handling damage, and deterioration from gasses and humidity.
While quite effective compared to no packaging, even the best protective packaging cannot halt the internal degradation of components, such as contact oxidation inside relays and switches, or contamination at the chip or wafer level. Even the best packaging often can only maintain lead solderability and prevent damage to exposed contacts.
Many parts (like O-rings and seals) are tracked by cure date, or a manufacturing date (like capacitors) and retired because of age (representing presumed material deterioration) before they ever get used. Storage for most spare parts was considered practical for about seven to 15 years, although some agencies and manufacturers had their own age limits, which could be as short as three years from the date of manufacture, including ISO-9001 systems.
Interestingly, newer SMD (surface mount device) components have extremely short storage shelf life, sometimes measured only in days or weeks from package opening, and have special sealing and handling requirements. This is because oxidation on their solderable surfaces can result in total solder joint failure, especially when used with less active no-clean fluxes, which have very poor wetting characteristics compared to rosin fluxes. This deterioration of solderability can be a hidden manufacturing problem that leads to later intermittent or vibration-related failures. Because of this issue, and the other odd effects that show up in surface mount manufacturing, I am a big believer in 100% visual inspection by microscope of all SMD assemblies. With this technique, I have routinely caught many problems, which had yet to show up in electrical testing or by casual visual examination.
From an electronic viewpoint, we tend to see component parts, especially solid-state devices and passive parts, as essentially having unlimited life. We assume they will never fail except through over-stress or physical damage. Most statistical models assign some weighting to temperature, cycle life or other factors, and then pick some defensible "failure rate" per 1,000 hours to arrive at assessments of component life. This model can break down quickly when systems go into actual service, because of the multimode environmental forces that work in concert to bring about part failures.
A more useful and correct orientation is to start with the premise that all parts will fail. Then try and determine how you can prevent that from happening prematurely or in a dangerous manner, and establish what you consider to be a useful working life-span. Keep in mind that few avionics systems can last 10 to 20 years in service without difficulty, but most aircraft operate longer than that.
The specific mechanisms that occur with aging are often quite unintuitive, and this can sometimes lead to bizarre and unexpected system operation. A classic example is an electrolytic capacitor, one of the few parts generally recognized as having some definite lifespan. With time, virtually all wet electrolyte capacitors show dramatically decreased capacitance with life, regardless of all other factors. In addition, they will often show increased leakage, especially when operated at elevated temperatures. These changes cause large reductions in filtering capability and shifts in offset/bias voltage in many circuits. But they do not appear as a solid failure. Worse, they may go completely undetected during even stringent bench tests, especially in power supply circuits, because the box is not tested with impressed AC on the DC supply to truly test their functionality.
Solid tantalum capacitors exhibit other odd properties, especially with regard to voltage and leakage. They tend to become stabilized to the applied voltage (which is zero in storage). These capacitors then exhibit failure of the dielectric when a higher voltage is applied. This is one reason why the excessive voltage "over-rating" of these parts has little effect on long-term reliability.
Resistors tend to increase in value with time and loading, and may do so dramatically if seriously over-driven, while semiconductors exhibit increased leakage over time. The inevitable surface and material contamination present in semiconductors also has its own electrochemistry that leads to device failure. This is more significant as circuit details become finer on the wafer.
Package integrity plays a big role in semiconductor survival, as the long-term fault data shows that contamination at the wafer level leads to failure in virtually all parts. This can take 10 years or more to appear, but can then be system-wide if poor package choices, assembly, heat sinking, or sealing techniques were used.
Circuit boards that support these parts become leakier over time, as they absorb water and other contaminants. They may fail catastrophically if copper edges are not sealed with plating or masking, and are then exposed to water and heat.
Even the solder used plays a role in system life. The 60/40 solder starts out by producing more poor quality joints than the 63/37 solder, and often fails completely on parts (like resistors) that are operated too hot, as the heating and cooling cycle can produce granular, high resistance joints.
On the system LRU (line replaceable unit) level, it is worth repeating the landmark discovery some years ago at USAir that "bad or problem boxes" will eventually and inevitably gravitate to the spares pool over time. This makes effective field maintenance impossible and erodes confidence in the support system. The process occurs because many small, essentially undetectable problems accumulate as systems age, eventually leading the unit to be sent in for repair. These factors may not produce solid detectable faults, especially with ATE (automatic test equipment) testing. This causes the box to be marked "no fault found," and returned to the loaner pool, where it simply fails again when put into service, and returns for more "service" and heads back to the loaner pool.
Eventually, only problematic units are brought into the loaner pool, and from that point on, no effective spares support is possible. In the end, the only solution is to purge these units with new, known good replacements, and to do a thorough check of the related airframe interconnects. This can be hard to do with "irreplaceable" units no longer made, or high ticket items. Ugly as that purging process may be, it remains the solution.
Wire has become a focal point in the study of both civil airframe aging and the space shuttle program. It is now clear that the range and prevalence of wire and interconnect faults is far greater than first thought. Wiring does not merely open or short to ground due to vibration. Insulation cracks under temperature cycles, especially at bends, allowing unanticipated wire-to-wire shorts, ignition points for fires, and leakage paths never contemplated in the system design. In addition, there is loss of continuity, which has a vibration-related component with degraded contacts inside connectors, and changing conductance to ground in poor airframe ground connections.
Wiring may also begin life in a deformed state with inadequate insulation, due to clamping, wire strapping, stretching or high angular stress. This "precondition" makes the chances of failure due to secondary causes increase significantly. Wiring in outer unpressurized areas can easily drop to -40ï¿½ C or lower in flight, while simultaneously being exposed to vibration, flexing, stress, solvents, lubricants and fuel. Not a very attractive combination from the wire’s perspective.
Most modern airframe designs also have done away with wire ways and conduits for ease of assembly and weight reduction. So the chances of opportunistic wire damage through mechanical accidents, external forces and chemicals is inevitably increased because this protection is missing. These conditions all accumulate over time to bring about many unexplained upset events that go unresolved, and may eventually lead to solid, detectable failures.
In addition to other wire-related problems, coax cables can also absorb water over time, especially with polyethylene dielectrics. This alters the dielectric properties and greatly increases cable losses at higher frequencies.
Water also can seep into fiberglass antenna shells and housings, changing state (ice, water, steam) with temperature, making the antenna perform poorly, sometimes in different ways at different temperatures. Many installers have switched to Teflon coax cables, as PVC cables were identified as fire and smoke hazards. But Teflon’s poor cold flow characteristics are not always well understood, and tight clamping or strapping can seriously alter cable impedances or even induce failure with enough compression or angular stress.
Clearly, all wire insulation systems are not created equally. Some early insulation materials like Kapton or PVC may not give adequate service life or remain safe on-board the aircraft over time. Ultraviolet (UV) radiation from sunlight now is also understood to be a serious factor in the degradation of virtually all plastic materials. And exposed wiring, which receives considerable sunlight (in the cockpit, freight areas, doorways and landing gear), can be adversely affected over time due to this cause alone. Teflon also has been shown to be a real problem under mechanical pressure or stress, as its cold flow characteristics are poor, and the wire can literally pass through the insulation under these conditions over time, with catastrophic results.
A key aspect of these interconnect and wire related failures is that they often defy detection by the traditional one-path-at-a-time sequential mode of analysis. This not only fails to spot the problem under vibration (a time dependent failure), but ignores many combinatorial faults that occur between wires and other surfaces on an erratic basis. Only massively parallel and true analog analysis can even hope to detect and correctly identify this problem.
One ever-present age mechanism is the inevitable deterioration of all surface bonding and grounding over time. This results in increases in airframe resistance, reduced quality of ground connections for power and antenna returns, and a steady increase in static discharge noise and p-static. Composite surfaces and dissimilar metals seem to suffer the most over time. But all aircraft see a steady shift upwards in their background "noise signature" and ground loop noise floor.
The actual mechanisms are varied, however, a granularization of metal exists at all conductive surfaces where a ground bolt is installed. This occurs because of vibration-induced wear and initial assembly. Such material easily oxidizes due to increased surface area and loss of surface sealing, and the connection slowly increases in resistance. Physical tension or pressure is also lost due to compression and cold flow, as well as vibration. This leads to an inevitable increase in resistance, which may cause unwanted heating, and a further rise in surface oxidation and resistance. It can take years to appear, but then can surface as a problem at hundreds of points all over the airframe seemingly at once.
Breakers exhibit an especially interesting age-related problem. As they are cycled, contact resistance increases, leading to increased self-heating, which alters the trip current threshold. Since things have a bad habit of being added to existing protected lines over time, the system can creep closer to the trip point over the years. Eventually, this makes the breaker trip early, causing what may be a perfectly acceptable load to be opened.
It is worth noting, however, that some on-board fire related incidents seem to stem from attempts to reset breakers that have tripped "for no reason." This can easily occur if wiring additions from a large breaker to some supplemental circuit are done with lighter gauge wire. Such wire will not be able to stop a fire if it should short to ground while supplied by the large breaker. The primary load may be fine, but the "additional" circuit may have a serious problem, not obvious to the flight crew. As a result, it may burn if power is not removed.
Switches also exhibit contact deterioration and wear over time. They will eventually fail, regardless of the original rating or eventual derating. Even switches of very high quality have a mechanical life cycle rating, and when discussing DC loads of 5-10A, 25,000 cycles is exceptional, and 10,000 cycles or less is much more realistic. Above that current, or with inductive or lamp loads, cycle lifespan drops very quickly.
As an operational bonus, switches may exhibit intermittent behavior for a period prior to total failure, which can be especially awkward for flight crews to interpret, and technicians to repair. The result can be switches that only work when cycled repeatedly, jiggled or teased–a frustrating cockpit experience if it happens to be the landing-gear switch.
Relays have the same contact failure problems and life-cycle limits that switches have. But they also have additional age-related problems of vibration induced intermittent operation and the possibility of contact welding if arcing conditions occur. This may leave a circuit connected even when coil power has been removed.
The airframe interconnect has within it all of these items, from capacitors to connectors, as well as wiring, switches, relays, steering diodes and antennas. When you add together the progressive aging of these individual items, it is easy to see how the aircraft’s general behavior begins to shift with time. It becomes less predictable and more prone to upset events triggered by changes in temperature, vibration, altitude and use. This leads to more electrical noise, increased RFI, and a generally deteriorated operating environment.
It helps to understand all these factors whether you are an airframe designer or a systems designer. These aging trends affect everyone over time. One thing remains inescapable: the over-all airframe will shift its fundamental performance with time. More things, such as wire harnesses, switches, relays and surface bonding, will have to be included earlier in major overhauls to insure the best and safest performance. While a system may not burn out like a panel lamp, it can certainly go from being an invisible component to a problematic one.
Avionics systems have a long way to go to be even half as reliable as the airframe, a sobering thought for all of us in the avionics business.
Walter Shawlee 2 welcomes reader comments. He can be reached by e-mail at email@example.com.