By now, I suspect most AVweb readers know that I am a strong advocate of overhauling piston aircraft engines “on condition” rather than “on time.” I have always considered it both irrational and uneconomical to tear down an otherwise healthy engine just because it has attained some arbitrary number of hours or years in service.In my Savvy Owner Seminars, I teach my aircraft-owner students to monitor the health of their engine using all the available tools — oil-filter inspection, oil analysis, monitoring oil-consumption and oil-pressure trends, compression tests, borescope inspections, spark-plug inspections, and digital engine-monitor data — and to overhaul their engines only when the engine says that it’s time.I practice what I preach. The engines on my Cessna T310R are both now at 750 hours past TBO (time between overhauls), and running so well that it would be a capital offense to tear them down. I monitor their condition very carefully, and have no plans to overhaul them any time soon.It’s perfectly legal for owners of Part 91 aircraft to operate engines past the engine manufacturer’s recommended TBO. Commercial operators under Part 135 often press on beyond TBO as well, although they need to ask permission to do so from their FSDO principal maintenance inspector (and such permission is routinely granted).Nevertheless, many aircraft owners and mechanics remain uncomfortable with the idea of operating beyond recommended TBO. There is a widespread belief that such “TBO busting” is unwise because it increases the risk of engine failure.
Risk vs. Engine Hours
When I conduct seminars for aircraft owners, I ask my students to perform a little “thought experiment” on this subject. I show them the following schema …
… and then I ask them to think about the shape of a curve that would portray graphically the relationship between the risk of catastrophic in-flight engine failure and time since engine overhaul. After the class has a chance to think about this a bit, I ask for volunteers to describe what they think the shape of this curve would look like, and why.The most common reaction I get from my aircraft-owner students is that the curve would look something like this …
… with the risk of catastrophic in-flight engine failure starting out quite low, but gradually increasing with engine time in service, and starting to rise more rapidly as the engine is operated beyond the manufacturer’s recommended TBO. Most aircraft owners seem to believe that the risk-to-hours relationship looks something like this, and most A&P mechanics do as well.But they’re dead wrong.
In most of my classes, I find one or two sharp folks who correctly point out that the highest risk of catastrophic in-flight engine failure occurs during the first few hundred hours after the engine is assembled. They describe a risk-to-hours curve that looks something like this …
… with a high risk of catastrophic in-flight engine failure during the initial “infant mortality” period, quickly dropping to a low level after the first few hundred hours, and then rising as the engine is operated beyond recommended TBO.These folks are spot on about the high risk of catastrophic engine failure during the initial infant mortality period. If the engine was assembled with any bad parts (like a defective crankshaft), or if any errors or omissions were made during engine build-up (like failure to torque a critical fastener properly), chances are such problems will show up rather quickly once the engine has been placed in service.For instance, the most catastrophic kind of engine failure I can think of is breakage of the crankshaft. We’ve seen a rash of such failures in recent years in both TCM and Lycoming engines, and a rash of Airworthiness Directives recalling the affected engines for crankshaft replacement. In some cases, the crankshaft failures were caused by improper manufacturing procedures, while in others they were caused by metal impurities in the raw crankshaft forgings. In all cases, the crankshafts invariably failed within the first 200 hours of engine operation.Nevertheless, I think these folks still have the shape of the risk-to-hours curve wrong.
Is TBO-Busting Risky?
While there’s no question that the highest risk of catastrophic failure occurs in the first few hundred hours after engine assembly, I have yet to find any evidence that the risk rises significantly as the engine is continued in operation beyond the manufacturer’s recommended TBO. In particular, I’ve never seen a scintilla of evidence suggesting that an engine at TBO+500 is any more likely to suffer a catastrophic engine failure than one at TBO-500. In my view, the risk-to-hours curve looks more like this …
… with maximum risk of catastrophic in-flight failure during the first few hundred hours of engine operation, then dropping quickly to a low level and then rising only very slightly during the remaining time that the engine is kept in service, with no pronounced increase in risk as the engine is continued in service beyond manufacturer’s recommended TBO.Please understand that I’m not suggesting for one moment that an engine can be continued in operation forever without being torn down for overhaul. That’s obviously not true.What I am suggesting is that when a “mature” engine (beyond its infant mortality period) ultimately does develop a problem that makes a teardown necessary, that problem is extremely unlikely to be one that results in a catastrophic in-flight failure that makes you fall out of the sky. In the overwhelming majority of cases, the event that ultimately necessitates an engine teardown will be a spalled cam, a cracked crankcase, worn or contaminated bearings, a worsening oil leak, or some similar old-age disease that may cause serious impact to the owner’s bank balance but not to life and limb.
Show Me The Data!
Over the years, I’ve often wished I could get my hands on enough actual data about catastrophic engine failures to create a definitive risk-to-hours curve that would prove this. TCM and Lycoming investigate almost every catastrophic in-flight failure of their engines, and have a lot of good data that could be used to plot such a graph.Unfortunately, both TCM and Lycoming have been stubbornly unwilling to make this data available to anyone … undoubtedly on the advice of their corporate attorneys. On some occasions, they’ve made such data available to the FAA in support of a petition for an Airworthiness Directive, but in such cases they’ve always demanded that the FAA treat the data as proprietary and keep it confidential (and the FAA is required by statute to do so). So we’ve never been able to get our hands on this data and see what the risk-to-hours curve actually looks like.About the best we can do is analyze NTSB data on aircraft accidents and incidents attributed to engine failure. The NTSB’s data isn’t nearly as complete as TCM’s and Lycoming’s, because lots of catastrophic engine failures don’t result in reportable accidents or incidents, and also because many NTSB accident reports don’t include information on engine time. Nevertheless, it’s probably the best data we can actually get our hands on.Recently, a brilliant friend of mine named Dr. Nathan Ulrich undertook such an analysis. Dr. Ulrich is a Ph.D. mechanical engineer who has spent the last 20 years designing, building, testing , and performing failure analysis on highly stressed parts and systems, including very high-performance internal combustion engines. He also owns and flies a Beechcraft Bonanza V35.Dr. Ulrich recently searched the NTSB’s accident data for the five-year period from 2001 to 2005 (inclusive) to see what he could learn about the correlation of engine failures and engine time-in-service. He examined all the accidents during that period that the NTSB attributed to engine failure.He limited his analysis to piston-powered aircraft weighing less than 12,500 pounds, and further eliminated accidents involving crop dusters and air show/race aircraft. Unfortunately, the NTSB standard accident report format lists airframe hours but not always engine hours or years, so he further limited his analysis to engine-failure accidents where engine time was reported. (Presumably, reports that didn’t mention engine time were ones where the NTSB did not consider it relevant to the cause of the accident.)Dr. Ulrich found 180 accidents that met these criteria. Here’s how they were distributed with respect to engine hours and years in service:
It’s important to understand that the NTSB data can’t tell us much about the risk of engine failure beyond TBO, because relatively few piston aircraft engines are allowed to remain in service beyond TBO (and we don’t even have good data on how many are). What the data does show quite clearly is that engines fail with disturbing frequency during their first few years and few hundred hours in service after manufacture, rebuild or overhaul.Of the 180 engine failure accidents that occurred during 2001-2005, only about 15 of them involved engines beyond TBO. Analysis of those 15 past-TBO engine failure accidents revealed that 80 percent of them were attributed by the NTSB to something other than high engine time — most commonly inadequate or incompetent maintenance. In one case, for example, several cylinders had recently been replaced and the crankcase had been split and then reassembled improperly. In another case, the engine failed catastrophically 30 minutes after a new vacuum pump was installed incorrectly. In yet another, the mechanic tasked with replacing a bad cylinder immediately prior to the accident flight replaced the wrong jug. In one case, the engine had 4,605 hours since major overhaul (SMOH) and turned out not to have had any inspections or maintenance during the preceding 530 hours!
Does TBO Make Sense?
Why should we overhaul at some fixed TBO? That’s a tricky question, because engine overhaul at TBO is a two-edged sword. On one hand, overhauling at TBO presumably helps to ensure that the engine is retired before it wears out (although the data suggests that a worn-out engine seldom causes safety-of-flight issues). On the other hand, overhauling puts the engine right back into the infant mortality window, where the data tells us clearly that the probability of engine failure is highest.Do we really want to do that before it’s absolutely necessary? I don’t think so.Or as my scholarly friend Dr. Ulrich phrases it, “There is no engineering basis for assuming a correlation between aviation piston-engine unreliability and high time in service.”See you next month.
Want to read more from Mike Busch? Check out the rest of his Savvy Aviator columns.