Losing an engine is a recurring nightmare for many of us, but the actual risks are lower than most of us may imagine. In this video, AVweb’s Paul Bertorelli reports on his research project into engine-failure accidents and reveals that easily half of them—if not more—are caused by poor maintenance or pilot mistakes. Running the airplane out of gas is one of the distressingly common causes of engine failure.


  1. Thanks for what I expect was many hours of work to produce this. One point that I don’t think was emphasized in the comparison between automotive and aircraft powerplants is the difference in operating conditions. Automotive engines spend little time at maximum output where aircraft engines spend most of their life above 60% max output.

  2. My wife’s car has driven 110,829 miles at an average speed of 27.8mph. (according to the computer)
    That’s 3,987 hours with the only maintenance being oil changes and one replacement of the variable valve timing solenoids (about $30 and less than an hour to fix…)
    My car made it to 100,000 miles and then it needed a head gasket. That’s major maintenance but not a failure and it’s still going strong at 127,550 and 3,751hrs.

    Terry has a point about the amount of time at high output for airplane engines. But why can’t we just have more power? A 400HP skyhawk could be flown at 33% power with the same performance as the 180 at 75%. My cars have 205 and 260 HP respectively, far more than they need. The fact is that airplane engines haven’t changed much since the 50s, while car engines have grown ever more powerful and reliable.

  3. Great video Paul, as usual you hit the nail on the head! Regarding automotive comparisons, I don’t think the various commenters appreciate the real difference between auto use and aircraft. Imagine getting in your 350HP Dodge Challenger and driving for 10 minutes at max RPM for ten minutes (take off and climb in a plane). Then keep it at 75% power (5,000 RPMish??) for another 3 hours WITHOUT LETTING UP. Pretty sure the “modern” aircraft engine will fare better than a mass produced automotive engine in that comparison.

    As far as the 400HP derated Skyhawk, well thats really what it is. An engine the size of a Skyhawk could easily produce 400HP question is for how long. Todays 160 Cubic Inch, Six cylinder Indy Car puts out 900 HP, (admittedly though with OHC turning 12,000RPM) however if that Indy car engine manages to finish the race without grenading, it is completely rebuilt for the next race, so whats that, a 4 Hour TBO??

  4. As the pilot of the Mooney in this video, I do have recurring nightmares about an engine failure inflight. It happens when I feel a little vibration in the airframe or the engine sounds a little different (or when I’m over water). I would surely love to know WHEN the engine is going to fail BEFORE it fails. How many times have you heard of a pilot say, “ya I felt this little vibration so I made a precautionary landing only to discover….” Then he or she then reveals one or more of the litany of things that can go wrong. How do you know when that that slight engine vibration or the occasional ignition miss is actually foreshadowing a complete meltdown?

    It’s been 17 years since the engine failed on takeoff in the Mooney. We had a basic JPI engine monitor (if I remember correctly), measuring EGT and CHT. No indication in the previous 4 hours of flight that anything was amiss. How do we monitor our engines today? The same basic parameters! With the cheap electronic sensors available today and machine learning software, why can’t an engine failure be predicted? Actually it has been done in Locomotives and turbine engines. General Electric Aircraft and Transportation has been using sensor data to predict engine failures for years. Locomotive engine failures on the track do not happen any more. Infight failure of GE Engines is very rare. I’d be willing to share my engine performance data so that a machine learning tool could predict failures. What is preventing us GA pilots from getting good predictive engine failure analysis?

    • Hi Dana:

      So sorry about the loss of yours and Paul’s Mooney. There certainly is a lot of great engineering that goes into making turbine engines very very reliable. However, as two recent incidents show the complex and robust system built to maintain the airworthiness of turbine engines occasionally breaks down. The United 777 Pratt 4077 and Southwest Airlines GE/Snecma CFM56 were both fan blade failures. I’m not familiar with the exact inspection procedures for these parts but expect that interpreting the NDT results requires a high degree of human inspector expertise. Hidden fatigue inclusions evidently occasionally slip by. I don’t know if either of these engines had vibe sensors that more modern engines are equipped with to detect rotating machinery anomalies. I’m not sure if these type of sensors would have caught an imminent catastrophic fatigue blade rupture.

    • We installed CGR-30Ps in all three of our club aircraft a few years ago. Besides measuring individual cylinder CHT & EGT, it also measures fuel flow and pressure, RPM (from both mags), MP (in the case of our Dakota), oil pressure and temp, voltage, GPS altitude and ground speed, and carb temp. We upload the data roughly once a month (sometimes once every two – it depends when I get a chance to download the data) to Savvy Analysis, and they are building a predictive engine analysis database. We also get monthly “report cards” from them to keep an eye on engine trends. To date we have been able to troubleshoot and resolve numerous minor issues, including diagnosing a magneto that would only fail intermittently once it had heated up.

  5. Paul, I love ya, man. But, with all the time and effort that you put in to present all that data and and analysis, I find very little that will help me predict my future engine failure. For a miniscule price compared to a 100 hour or annual inspection, we could have a data recording device like your Mazda to acquire ops specs of the engine or entire a/c to truly learn what caused the failure. Brainstorming!

  6. Thanks for this great article. I was surprised that the carb ice incidents were not actually greater, but an O-470 Skylane driver might have presuppositions. One thing I would add is that the few bucks saved by installing reman cylinders is not worth the additional failure risk. Go new, especially with turbocharged engines.

  7. A great video as always. I like how Paul was very honest about what he know for sure, was pretty sure off, and didn’t really know. This IMO is rather rare in technical video presentations.

    I have had 2 engine failures caused by a mechanical failure. On the first an oil pump drive gear failed in a C150 Continental O 200. I was able to return to my home airport and land before the engine seized although it required an overhaul. The reason this incident did not become an accident was because early in my flying career a “been there done that” x bush pilot told me to pay attention to the engine gauges. In the green was not good enough, he demanded I should be able to tell him without looking what they actually indicating.

    Because of that as my student and I were climbing out after takeoff I noticed the oil pressure gauge was one needle width below the little white line in the middle of the green arc on the no numbers POS Cessna gauge. On every other flight the oil pressure needle had always sat right on the white line. I immediately sat up and started paying close attention. About 30 seconds later the needle was now 2 needle widths below the reference line and so I immediately told the student to head directly back to the airport. By the time we were on short final there was no oil pressure indication and the engine had a slight vibration so I shut the engine down and we glided to an uneventful landing. The cause of the oil pump failure was a basically undetectable metallurgical flaw in the drive gear.

    Not paying attention to what the engine gauges were telling me, even the decidedly unimpressive Cessna gauges, would have caused me to have had the engine quit in an area without a lot of good options. Thank You for the great advice Bush Pilot Bob !

    The second one was a catastrophic turbo failure in a Navajo. This happened in IMC on climb out at an isolated airport at gross weight and just at sunset. The resulting single engine NDB circle to land approach to minimums on the unlighted runway constituted a “significant emotional event” for me the pilot.

    I then experienced an extreme sense of humour failure when I found out 2 other pilots had put a total of 4 quarts of oil in that engine in less than 6 hours of flying including topping it up for my flight in an effort to be helpful, and did not think to tell anyone. This engine was screaming “ Look At Me” but nobody was paying attention.

    So personally FWIW I am 1 for 2 on preventable engine failures caused by a mechanical failure.

    One thing that Paul alluded to, but did not explicitly mention is pilot training. This is a personal crusade. Flight training for the engine failure is binary. The perfectly good engine is running fine and then it suddenly completely stops. That is a scenario but not IMO the most common.

    Nowhere in training do we talk about the fact that the best way to deal with an engine failure is to not have it fail in the first place. That shopping list of no brainer “don’t be stupid” that he presented at the end of the the video is not IMO covered in an organized way in most flight training.

    I hope none of my students will ever run out of gas. To help prevent this from the very first lesson I ask what is the time in the tanks and what time do we have to be on the ground. This conversation happens for every flight we do.

    Personally I don’t see a lot of this kind of emphasis on the boring unsexy flight discipline that will prevent pilot caused engine failures happening in flight training….

  8. Here’s one for you:

    I had an avionics shop in another state install a JPI EDM 790 in my new-to-me ’73 Piper Aztec. The installation was delayed, delayed, delayed, but finally finished. On the flight home with my instructor, one to the TIT sensors failed; I discovered that it had been dutifully installed adjacent to a 6″ long crack in the manifold, and the hot gases had cut the wire. Conversation with the shop owner by phone ended with, “That crack happened on your flight home. I’ll send you a sensor, but it is not our fault.” The fact that crack’s edges were beveled, the split 1/8″ wide and the turbocharger mount blistered and rusted from chronic heat cycles didn’t matter. It wasn’t their fault.

    Fast forward. Life intervenes and some time passes. Finally I have everything ready to go my multiengine rating. We are doing the usual maneuvers, and the “left” engine begins to overheat. I try full mixture, lower power; nothing works. Temps are over 500°F. One is 710°. I tell the instructor, “I’m going to shut that engine down.” His response is, “No – if it’s still running, keep it spinning. I cannot make any sense of the indications, and I don’t know what’s wrong.” Saved our lives.

    We landed and the left engine showed no signs of any damage. Just for completeness, I checked the right engine. Aluminum in the cylinder. Two jugs showed 30# compression. What the hell?

    It turns out that the avionics shop had reversed the thermocouple leads. The ones marked “Left” went to the “Right” engine, and vice-versa. The reason that none of the readings made sense when compared to the native gauges – which remain with the 790 – is that while the EGT and CHT readings were on the wrong engines, the others went to a different plug on the instrument and were correct.

    I nearly had shut down my good engine and run the other at full throttle to get us home; of course it would have failed. 200 gallons of avgas makes a big fire. I can see it now: “Pilot error.”

    Once my mechanic discovered the discrepancy, I called the shop again. “It’s all your fault; you should have known something was wrong.” No offer to defray the $14,000 top-overhaul cost. Nothing. I don’t sue people, but that was a very expensive lesson well-learned.

  9. At one of my first flight lessons 47 years ago, my instructor chastised me for what I was wearing, telling me to always dress as if I might have to walk home. That was some of the best flying advice I ever received, and I pass it along to my students. As we begin cross-country flight training, I have my students assemble and carry along a bag containing a sleeping bag, snacks, and signaling gear, just in case they end up spending the night with an airplane after an off-airport landing. It makes little sense to survive an engine failure, and not be prepared to survive the night until rescued.

  10. Great overview, Paul–Thanks.
    Ditto Charles Dickinson’s insight on being prepared, In my case, I was crop dusting in The Sudan–flying an R1340 powered Thrush when the engine quit–gas line failure.
    IT was close to a million degrees in the cockpit , so I was in my underwear, no shirt, my helmet, respirator and sockless Adidas running shoes.
    I landed next to a small village , which population instantly swarmed and circled the plane, offered me hot tea and no doubt were amused by my attire.
    It took 6 hours for the Land Rover to show up with gas and a fuel line.

  11. I have my own personal theory on why the Rotax engine failure numbers top the Continental and Lycoming numbers. First of all, I assume that we are talking about the 4-stroke Rotax 912’s and 914’s.

    A couple years ago I visited a LSA manufacturer in Eastern Europe. I asked the owner why he didn’t design a small door that made it easier to check the oil level of the engine, just like most GA aircraft did. On his aircraft, it was necessary to remove the top cowl to do this very routine check. His simple answer – “because it is not a Continental or a Lycoming.”

    A couple years before that I was flying a weight-shift-control trike in Thailand. The owner showed me a bucket full of old carburetor rubber flange assemblies. With all of the engine vibration, these flanges are prone to cracking or tearing. When that happens, additional air gets introduced and the mixture goes lean. Worst of all, these cracks can be invisible unless the flange is squeezed, revealing the cracks.

    Secondly, a lot of time gets spent troubleshooting Bing carburettors. Sinking floats, needle clip or jet problems – there are a lot of headaches that can keep owners busy. Otherwise, the Rotax 9-series engines are great, and a big improvement over the 2-strokes that many of them replaced in the LSA world. My first engine failure (in a piston aircraft) was on a 912-powered weight-shift-control aircraft… at night (why do engines seem to know its night or you are over water?). In fact, it was the beginning of a night cross country training flight. Fortunately, it was immediately after takeoff and we landed straight ahead on the remaining runway. Spooked, we returned to the hangar for troubleshooting. We drained the fuel (again), and the instructor/LSA repairman pulled the carburetor bowls to check for anything that shouldn’t be there. Nothing was found. After exhausting all possibilities, we fired it up, and tried again. No repeat performances. We continued from Zephyr Hills all the way Ocala in a stiff headwind and c-c-c-c-c-cold winter day. Landed, took a leak, and flew back. Still here to talk about it.

    • You flew 150 miles at night in a weight shift ? What was the license / rating you were seeking By the way ? I am also an LSRM for LSA with the Rotax endorsement. Mechanics love those engines because they are bullet-proof, until they aren’t. It can be difficult with those engines to figure out what’s going on without rebalancing the carburetors and a host of other interventions. The ASTM is not the FAA when it comes to documenting failure modes.

  12. It seems to me these comments are developing into good list of all the things that can go wrong with engines. Perhaps that could be a thing. A list of tell tale signs that an engine is going to quit.

    I also had a partial power loss with a brand new Rotax 582 (after break-in of course). The idle jet fell out of it’s socket in the Bing carb and got lodged under the float. The result was that cylinder flooded. The landing back at the airport was uneventful. Trying to find the problem was a challenge however. The idle jet was hidden under the float and I could not see it on initial inspection. I didn’t know what was going on.