The Savvy Aviator #58: Why Mechanics Make Mistakes
During the century since the Wright Brothers first flew, the predominant perpetrator in aircraft accidents has shifted dramatically from machine to human. Today, human error is responsible for 90 percent of aircraft accidents and incidents. It's not that people have become more careless, forgetful, inattentive or reckless. It's that aircraft and aircraft components have become much more reliable. As component failures become fewer and fewer, human failures represent an ever-increasing percentage. Most of the efforts of the aviation research community have focused on errors committed by pilots. This is appropriate, since about 75 percent of serious aviation accidents are due to pilot error. However, there have been a significant number of serious, even fatal, accidents caused primarily by maintenance errors. While there has been increased focus on maintenance errors by the airlines, particularly in the wake of the Aloha and ValuJet crashes, not nearly enough attention has been given to maintenance errors in General Aviation (GA).
Kinds of Maintenance Errors
Less-than-adequate maintenance can be divided into two broad classes:
- Introduction of a problem that was not there before the maintenance began; and
- Failure to detect a pre-existing problem during maintenance inspections.
Errors of omission seem to be the most prevalent kind of maintenance errors. An analysis of 122 maintenance errors detected by a major airline over a three-year period revealed the following breakdown:
- Omissions: 56 percent
- Incorrect installation: 30 percent
- Wrong parts installed: 8 percent
- Other errors: 6 percent
When the 56 percent of errors attributed to omissions was further examined, the breakdown was:
- Fasteners left undone or incomplete: 22 percent
- Items left locked or pins not removed: 13 percent
- Filter/breather caps loose or missing: 11 percent
- Items left loose or disconnected: 10 percent
- Spacers, washers, etc., missing: 10 percent
- Tools, spare fasteners, etc., not removed: 10 percent
- Lack of lubrication: 7 percent
- Access panels left off: 3 percent
- Miscellaneous: 11 percent
The Reassembly Problem
Clearly, most maintenance errors occur not when taking something apart, but rather when putting that something back together. There's a good reason for this. Consider a bolt (figure below) onto which eight nuts have been assembled, each one labeled with a unique letter A through H.
Assume that the mechanic's task is to disassemble the nuts from the bolt, clean them, and then reassemble them in the original order. There is really only one way to take this assembly apart, but there are 40,320 different ways in which it could be put back together -- and 40,319 of them are wrong! This simplistic example illustrates the fact that the task of disassembly usually constrains the mechanic to one particular sequence, with each succeeding step being prompted by the last. The mechanic doesn't require much guidance, because the disassembly procedure is usually obvious. In contrast, correct reassembly usually requires knowledge -- either in the mechanic's memory or in written form. Human memory being as imperfect as it is, reassembly based on memory is error-prone. Reassembly based on written guidance (such as a checklist or service-manual instructions) is far more reliable, but people doing a hands-on job tend to be reluctant to consult written instructions. (Watch your A&P work on your airplane, and note how rarely he consults the service manual or any other form of written guidance.) Reassembly-by-memory is probably adequate for a task that the mechanic does every day. Most maintenance tasks aren't like this, however, and we all know how easily we can forget the details of a task after even a short period of time. To make matters worse, a wrongly-assembled component is not always obvious on later inspection. The absence of washers, bushings, fasteners, seals, O-rings, caps, lubrication, etc., are often concealed once the component has been reassembled. Thus, reassembly errors often create the opportunity for double jeopardy: a high probability of forgetting something important during reassembly, and a low probability of detecting the error once the job is completed.
Slips, Mistakes, and Violations
Failures by a mechanic to perform a task as planned are commonly termed slips, lapses, trips or fumbles. A slip occurs when the mechanic is trying to do the right thing, but screws it up somehow. Slips can be caused by:
- Omitting some necessary action;
- Performing some necessary action in a clumsy fashion;
- Performing some unwanted action; and
- Carrying out the right actions in the wrong order.
Such slips most often occur when doing tasks by memory -- often well-practiced tasks that are done frequently in an automatic fashion. Mistakes are higher-level failures caused by an error in the plan itself. These are usually caused by lack of knowledge, and occur most commonly when performing tasks that are not done very often. Often, mistakes are caused by trying to do something by memory that should have been looked up on the service manual. Forgetting to torque a cylinder hold-down nut is a slip; torquing it to the wrong torque value is a mistake. Violations are deviations from standard practices, rules, regulations, or standards. While slips and mistakes are unintentional, violations are usually deliberate. They often involve cutting corners in order to take the path of least resistance, and often become part of a mechanic's habit pattern. In a recent columnI wrote about an incident in which the pilot of a Cessna 340A launched into IMC on the first flight after maintenance, only to discover that his airspeed indicator, altimeter and VSI stopped working as the aircraft climbed through 3000 feet. The cause of the problem turned out to be a mechanic's failure to reconnect a static line that had been disconnected during maintenance to facilitate access. The mechanic's failure to reconnect the line was an inadvertant slip -- he forgot. On the other hand, the mechanic's failure to perform a static-system leak check (required by FAR any time the static system is opened) was a deliberate violation. Because of the violation, the slip went undetected and jeopardized safety of flight.
Distractions play a big part in many errors of omission. A common scenario is that a mechanic installs some nuts or bolts finger-tight, then gets a phone call or goes on lunch break and forgets to finish the job by torquing the fasteners. I have personally seen some of the best, most experienced mechanics I know fall victim to such seemingly rookie mistakes. I know of several fatal accidents and countless less-serious incidents caused by such omissions. Just as pilots need a "sterile cockpit" during high-workload phases of flight, mechanics need a distraction-free workplace when performing safety-critical maintenance tasks. Unfortunately, the typical piston GA shop is a distraction-rich environment. Phone calls come in. Customers drop by unexpectedly. UPS and FedEx drivers deliver anxiously-awaited parts. The Snap-On tool truck stops by. The shop's FAA principal maintenance inspector pays a surprise visit. The roach coach arrives with lunch. This is less of a problem in the big turbine shops, where there's usually a Parts Manager to deal with deliveries, a Customer Service Manager to handle customer visits and phone calls, and a Compliance Manager to interface with the FAA. But in the smaller shops that owners of piston GA usually use, employees usually wear multiple hats and must deal with these distractions as they come. That leads to mistakes. Big shops have their own issues. Shift changes cause lots of problems, when the first-shift technician assumes the second-shift technician will handle something, but the second-shift guy fails to do it because he assumes the first-shift guy handled it.
I've visited a half-dozen different GA aircraft and engine factories to watch how they build our flying machines. One of the fundamental work rules at these plants is that there must always be at least two sets of eyes that look at every step of the process: the technician that performs the work, and an inspector who verifies that the work has been done properly. Often, there are three sets of eyes: two technicians who work as a team and check one another's work, and then an inspector who re-checks the work. Large repair stations that work on turbine aircraft often have similar rules, where designated inspectors are required to check the work of each mechanic and sign it off. But the smaller shops where most piston GA maintenance is done seldom can afford the luxury of having dedicated inspectors on staff. One A&P will sometimes ask another to check a particularly critical or complex task, but most maintenance is checked by just one set of eyes belonging to the mechanic who did the work, and most scheduled inspections are done by just one IA. Fewer sets of eyes inevitably means that more slips, mistakes, violations and discrepancies escape detection.
The Owner As Final Inspector
Aircraft owners and pilots need to understand that maintenance errors create a significant hazard, and act accordingly. The most likely time for an aircraft to suffer a mechanical problem is on the first flight after maintenance. Prudence demands a post-maintenance test flight every time the aircraft comes out of maintenance. The test flight should be done in VMC, without passengers, and in a place where the pilot can easily put the airplane back on the ground if something isn't right. Prior to the test flight, the owner or pilot should conduct an extraordinarily thorough preflight. Make sure that all inspection plates and fairings are installed and secure, all cowling fasteners are tight, and all fuel and oil caps installed. Check that all flight controls and trim systems are free throughout their full range of motion and operating in the correct direction. Check that all instruments and avionics systems are functioning properly. Perform a ground test of the autopilot. Run up the engine thoroughly, then shut down and check for leaks. Be sure you don't smell fuel or anything burning. In short, be thoroughly skeptical any time an aircraft comes out of maintenance. Your pre-flight and test flight are the last line of defense against maintenance errors.