|
Getting your Trinity Audio player ready...
|
Many pilots chuckled when an obscure Swiss company introduced its mockup of a large single engine turboprop at the 1989 NBAA show in Atlanta. Sure enough, it was five years before the Pilatus PC-12 was certified, but since then the skeptics have all but vanished. The “turbine Suburban” has earned a loyal following for its unique mix of large cabin size, long range, and impressive runway performance, with nearly 2000 delivered over the last 28 years.
Besides the specs, another key reason for the PC-12’s popularity has been its stellar safety record, one of the best in all of general aviation. So any effort to answer the question, “what’s wrong with Pilatus PC-12 pilots?” has to begin with: not much. But accidents do happen, including a recent one off the coast of North Carolina that killed eight people on the way home from a hunting trip, making headlines and rocking the local community.
As we typically find in this article series, the cause of such accidents usually has more to do with the pilot than the airplane (hence the last word in the title). That means it’s worth our time to examine the safety record of a specific airplane type, even if you don’t fly that model. There are lessons here for all pilots.
The safety record
As usual, we’ll start with the numbers, limiting ourselves to fatal accidents in North America. That keeps the data clean and keeps us focused on the most serious accidents.
The PC-12 has accumulated an excellent safety record over the last 28 years.
Searching the NTSB database shows very few such accidents: only five in the last 10 years. (There was a fatal accident in 2017 involving a U-28A, the military version of the PC-12, but I’m excluding it due to the unique nature of their operation.) With nearly 1300 PC-12s flying in North America, that means just 0.38% of the relevant fleet has been involved in a fatal accident during this time period. In fact, there was a period of over four years (January 2013 through April 2017) when there were no fatal PC-12 accidents—an incredible feat given how many hours these airplanes fly, often in challenging conditions.
For comparison, the TBM series of single engine turboprops has suffered seven fatal accidents in the last decade. When spread over the smaller fleet of 800 aircraft in North America, that leads to 0.88%, still excellent but more than double the Pilatus number.
In addition to fleet size, another popular way to evaluate safety is the fatal accident rate per 100,000 flight hours. This number can vary dramatically based on the estimated flight hours, which are often nothing more than a guess, but in this case it’s worth considering. Turbine airplane manufacturers tend to track hours flown much more carefully than piston OEMs, so the denominator here is fairly accurate. For this calculation, we’ll use total worldwide accidents.
There have only been 22 fatal accidents worldwide (including military) since the PC-12’s first delivery, during which time the airplane has racked up roughly 9 million hours flown. Doing the quick math leads to a lifetime fatal accident rate per 100,000 hours of 0.24, and probably lower for US-only operators. That compares quite favorably to the TBM’s lifetime rate of around 1.1 or the overall GA fatal accident rate, which was 1.02 in 2019 (for non-commercial fixed wing).
Five fatal accidents
At first glance, those low numbers might mean it’s case closed for the PC-12, but every fatal accident is a tragedy so it’s worth reviewing the details to determine whether there are common themes. Let’s quickly examine all five fatal accidents from the last decade.
Lake Wales, Florida. A new PC-12 pilot was climbing through 25,000 feet in IMC when he began a right turn to avoid potential weather. During the turn, the autopilot disconnected for undetermined reasons—possibly due to turbulence, possibly because the airplane got too slow, or possibly due to simple failure. Instead of flying the airplane, the pilot attempted to re-test the autopilot (presumably to clear a fault) and lost control in the process. Bank angle increased past 70 degrees and airspeed increased 175 knots past redline. The pilot eventually tried to recover but pulled the wings off during the dive and the airplane broke up in flight.
Burlington, North Carolina. A professional pilot flying lab samples was climbing through 3000 feet, at night and in clouds, when ATC asked him to reset his transponder code. Shortly thereafter, the pilot lost control and crashed. The NTSB report mentions possible confusion about the status of the autopilot—there is a scenario where it might appear to be engaged but in fact is not—but it’s impossible to know for sure. The end result was a classic spatial disorientation accident.
The ground track for the Lake Wales accident, like many others, points to in-flight loss of control.
Amarillo, Texas. An air ambulance pilot with two crew on board departed on a night IFR flight, with a 700 ft. ceiling and moderate turbulence. At 800 ft. AGL the controller asked the pilot to reset his transponder, which he did, but then the airplane made erratic pitch and roll changes, including a rapid climb (6000 ft./min.). Roughly a minute later, the PC-12 crashed in a steep descent.
Chamberlain, South Dakota. The airplane, with 12 people on board, crashed shortly after takeoff, probably due to a stall—airspeed never got over 100 knots and altitude never got over 500 feet. Weather was poor, with a low overcast and ½ mile visibility in snow, and the airplane had been sitting outside overnight. The pilot and a passenger arrived early to clear the snow, and apparently spent at least a few hours doing so. The NTSB’s report is still preliminary, but this accident has all the hallmarks of a low level stall/spin, probably exacerbated by contaminated flight control surfaces and possibly by weight and balance issues.
Beaufort, North Carolina. The airplane took off VFR for a 60-mile flight to Beaufort, but quickly contacted ATC for an IFR clearance—weather was marginal VFR over land and solid IFR closer to the coast. The pilot in command, with over 3000 hours logged, had his student pilot son in the right seat, along with six passengers in the back. After stumbling into a restricted area, the airplane turned around, but was eventually cleared for the RNAV 26 approach into MRH. Short of the initial approach fix, the airplane started to make erratic changes in speed and altitude, lost control, and crashed into the ocean. (This report is still preliminary.)
What can we learn?
Before we draw any sweeping conclusions, we should acknowledge that we are, to some extent, sifting through the statistical noise. Five accidents over millions of hours flown is not much to go on. However, there are some trends that are easy to spot.
First, notice what you don’t see here: engine failures. This was the big worry when the PC-12 was first introduced, and many a King Air pilot has turned up his nose at the idea of a 10,000-lb. airplane with only one engine. And yet that is clearly not a significant risk. There certainly have been occasional engine problems, including some in-flight shutdowns, but a new fuel control unit in the early 2000s addressed the few problems that popped up. Still, the takeaway is impressive: there has not been a single fatal PC-12 accident in the US caused by engine failure—ever. There are plenty of reasons why, including the incredible reliability of the Pratt & Whitney PT6 engine, the good gliding characteristics of the Pilatus wing, and a dead simple fuel system that is resistant to pilot mistakes.
On the other hand, weather seems to be a common thread. All five fatal accidents occurred in actual IMC and two were also at night. That’s significant, considering that the majority of time an airplane flies is in VMC and during the day. It’s a reminder that—no matter how capable the airplane—flying in low IFR conditions or at night brings additional risk. We have to be on our game as pilots for these flights.
The third theme, perhaps related to the previous one, is that every accident involved a single pilot (excluding the student pilot in the right seat for the Beaufort crash). That’s not uncommon for the PC-12, where the vast majority are flown by either the owner or a single pro pilot, but an airplane this sophisticated does demand proficiency. If something goes wrong—equipment malfunction, bad transponder code, or momentary confusion—a single pilot can quickly get overwhelmed. For particularly bad weather or long days, it might be worth having a second pilot in the right seat.
The PC-12 has a sophisticated autopilot, but like all avionics it can fail.
That’s especially true for less experienced pilots, and these accidents prove once again that the key number to track is time in type. No matter how many total hours you have logged, airplanes like the Pilatus have a learning curve that can only be surmounted with relevant experience—ideally in the same model with the same avionics. (Note that the accidents here suggest no difference in safety record between “pro pilots” and “owner pilots.”)
The Lake Wales pilot was notably inexperienced, with less than 30 hours in type and this after not having flown anything at all for seven years. This is an outlier—the pilot almost certainly wouldn’t be insured in today’s market—but some of the other accidents suggest the time in type issue is a factor. For example, the pilot on the Amarillo flight was a highly experienced ATP with almost 6000 hours total time, but he had logged fewer than 75 hours in type. That’s not brand new, but at night, in the clouds, bouncing around, that might not be enough to cope with any type of anomaly.
And there’s the final thread to pull on: the autopilot. Single pilot operations in a high performance airplane—especially in bad weather—mean the autopilot is a critical component, essentially an unnamed second crewmember. Fortunately, the Honeywell autopilot found on the Pilatus is a very capable and dependable model, but as the examples above prove, it is not foolproof. At least two and maybe as many as four of these accidents featured some type of autopilot anomaly. These can be caused by normal limitations (turbulence can make any autopilot kick off), pilot error (inadequate preflight testing, suspected on the Burlington accident), poor maintenance, or outright failure. We have to be prepared to both recognize and react to a variety of scenarios.
It’s not enough just to say, “more manual flying!” If you’re flying a PC-12 at night in weather, you can and should use the autopilot. The solution is not necessarily to use the autopilot less, but to make it an integral part of training so that you understand exactly how it works—and how it can fail. I know during my initial training on the PC-12, I spent far more time practicing engine failures than autopilot failures. Based on the statistics above, that is exactly backwards: the PT6 hardly ever quits, but the pilot-autopilot interface seems to be less bulletproof.
I now have over 1,000 hours in the Pilatus, and my anecdotal experience supports the story seen here in the accident reports. The airplane is built like a tank and is surprisingly easy to fly, especially for a larger turboprop: there is no prop lever, the fuel is either on or off, and the airplane can happily fly at 220 knots or 100 knots indicated. But it’s still a high performance airplane, and thus it requires proficiency and a relentless commitment to training. The airplane can probably handle whatever mission you have planned; can you?
- Weather flying means learning to read clouds - November 11, 2024
- What’s wrong with the teardrop pattern entry - September 25, 2024
- What matters for VFR proficiency: better landings - August 26, 2024
There was an incident in the SFO area where a complete electrical failure in icing IMC and subsequent failure of the battery backed AH did not result in a crash. Probably caused by the internal failure of a guarded master switch on the overhead panel.
Yes, that was a very interesting incident, one that led to a modification to the master switch. There are plenty of lessons to be learned from the non-fatal events, but they are beyond the scope of this article. Perhaps that’s an idea for part 2…
Thanks for a very informative article. Where I am based they service the PC-12 and there is a charter service that uses them based there so I see them all the time. When you see one in person the size is the first thing you notice, it is huge. A very neat plane.
This is why I teach my instrument students that the pilot is flying the plane and the autopilot is the assistant. I routinely flip it off during maneuvers, disable some automation like auto CDI switching, and create scenarios where pilot either fails or is misdirected by the gps. In single pilot IFR the autopilot is a critical piece of equipment, but needs to be address with respect and suspicion.
I guess that flight that originated in CA and crashed in ID(?) due to fuel imbalance caused by pilot not adding Prist was before your cutoff. I assume it’s still necessary to add that to the fuel? Seems like that would be an annoyance. That accident was a real mind bender.
Yes, that accident was in 2009. Prist is required for most PC-12s, although the new NGX model does not require it.
You can read more about that particular accident in an earlier Air Facts accident: https://airfactsjournal.com/2014/04/the-error-chain-in-action-pilatus-crash-at-butte/
Is that because the new PT6E-67XP has the fuel pre-heating from the new design that passes the heated air over the fuel supply? I don’t understand the tech fully, but just read somewhere that the XP has that feature
I read on and saw my question confirmed 2 posts down. Great article BTW
All PC12’s up to the NGX require Prist. The NGX has an oil to fuel heat exchanger that keeps the fuel in the tanks from getting below freezing. The accident you referenced occurred in Butte Montana. A very sad story and totally preventable.
Haha. Late on the push. What John said. I agree with everything in the article. We fly easier/shorter missions single pilot and longer and IMC flights with two. Just make more sense. Our owners are great about letting us decide when to crew flights.
I think you’re referring to the PC12 crash at Butte, Montana. I lived there at the time
Can you elaborate on “autopilot—there is a scenario where it might appear to be engaged but in fact is not—but it’s impossible to know for sure.”? I fly the PC12 professionally, haven’t heard about that behavior, can you shed additional light please?
It gets complicated in a hurry, and I’m pretty sure this only applies to the legacy PC-12 (the NG and NGX have a different autopilot setup). But here’s the relevant part from the final NTSB report:
“Examination of the KCP-220 flight computer revealed no physical damage to the circuit cards. A return to service test was conducted for the applicable airframe, which required replacement of the personality modules. The unit powered up and passed the self-test; however, the “AP CLU” lamp indicated there was no drive voltage to the Autopilot Roll and Pitch Servo clutches. Subsequent troubleshooting revealed that the R-259 resistor, which did not contain any obvious signs of physical damage, was open. The resistor was manufactured by Ohmite. It could not be determined if the open condition existed during the flight or was the result of impact forces. It could also not be determined if the autopilot was engaged at the time of the accident.
“According to Honeywell, during autopilot operation, a drive voltage is applied to the “AP Clutch Engage” solenoid when the autopilot is activated. This drive voltage enables the roll and pitch servos by engaging the clutches. If the autopilot is not engaged, the open R-259 resistor would have no effect on the flight control system. If the resistor is in an open condition at the time of autopilot engagement, the autopilot will appear to engage with a mode annunciation indicating engagement, but the pitch and roll servos will not engage. If the R-259 resistor becomes open while the autopilot is engaged, the pitch and roll servos will disengage and an aural warning would sound. The unit passed all return to service tests after the R-259 resistor was replaced.
“According to Honeywell, any failure of the R-259 resistor would not affect a pilot’s ability to manually control the airplane. In addition, the before taxiing checklist of the airplane flight manual (AFM) included checks of the autopilot system to verify autopilot function prior to takeoff, and section 4.20.1 Autopilot Operation Summary, included a warning which stated, in part: “The pilot in command must continuously monitor the autopilot when it is engaged, and be prepared to disconnect the autopilot and take immediate corrective action – including manual control of the airplane and/or performance of emergency procedures – if autopilot operation is not as expected or if airplane control is not maintained….”
“During March 2015, Honeywell issued service bulletin KCP 220-22-A0017, which included an inspection and replacement of the R-259 resistor on certain KCP-220 Flight Computers, if the resistor was manufactured by Ohmite or if the manufacturer could not be determined.”
Essentially, if you do the preflight test (turn on the autopilot, verify it engages, verify you can disconnect it) you should be all set.
It’s also worth pointing out that this airplane might have had a flaps fault after takeoff. I’ve had this multiple times and it’s not a safety issue but it is annoying. Hard to know for sure, but this might have been another issue the pilot was dealing with.
OK, I admit that I am an “analog ager,” but WHAT is a “personality module?” Sounds like something I could use…
Good Article ,
This airplane has a great following in Alaska as a tough, high performance airplane.
That is fast , comfortable , fuel efficient and has a great safety record.
The big cargo door in the back adds to its popularity.
I fly a legacy PC12 in Alaska , charter and corporate. I’ve got 3600 hrs over the last 7 years in this airplane flying over very inhospitable water and terrain.
I’m often flying into small gravel strips in the mountains, lots of IMC , Ice, turbulence with heavy loads. The more I fly it the more respect I have for this aircraft.
The autopilot will kick off in turbulence,( seems it’s usually at night in low weather :) but always lets you know and is easily reengaged.
I’ve never experienced any autopilot failures or issues in 3600 hrs. +
December 24 2019 I experienced a complete failure of the PT6 during climb out in remote SW Ak . Once I was able to get the prop feathered the airplane performed awesome in the quiet glide mode.
With lots of positive conditions, VFR, late eve light, We ended up gear up on a snow covered frozen lake with no injuries to 3 occupants and minimal damage to the aircraft. Self rescued with a helicopter and we’re home in our own beds late that eve. We salvaged the aircraft by Helicopter as well and it is back on line flying over 1000 hrs. post incident. I don’t feel unsafe flying it now as the failure rate on the PT6-67B Is minuscule and I’ve got mine behind me.
Note: This was not classified an accident , off airport emergency landing incident due to lack of airframe/ flight control damage and no injuries.
The successful recovery of a PC-12 from an off-field landing (on a frozen lake) was the subject of a television special. That has to be the aircraft that you flew and suffered the engine-out. The show actually featured the helicopter capable of the recovery. Congratulations to you.
Great article John. Any pilot should consider making upset prevention and recovery training a part of their training regimen if not annually, then every other year.
Your article also reminded me of this video. The success of the PC-12 even surprised Pilatus.
https://youtu.be/3PPbC7eeJRk
I am a passionate fan of the PC-12, in fact of Pilatus in general and in my work I am fortunate enough to encounter the 12 regularly, she is incredibly well built – this is often said but you have to see it to really understand the Swiss precision and engineering, belying the intricate engineering and precision is simplicity that seems almost contradictory. That said, I lost a friend to a PC-12 crash in the mountains but this was undeniably CFIT and unfortunately pilot error, whilst one could argue that some of these fatal incidents could have been avoided by a dual crew it is testament to the versatility of the type that it doesn’t have to be flown dual crew and that so many crashes were non fatal.
I am not a pilot, but I have several family memebers that are. I know one thing, if I here them talking about a particular type, I will refer them them to you.
One was totaled in a landing in Whitehorse last Monday, April 17 2023. It was a RCMP aircraft and only the pilot was on board. He survived.
Nice to hear PC-12 is a workhorse! Good to be aware of Auto pilot
may disconnect at a bad time. This is the one aircraft I would upgrade
to, given an opportunity. Thanks again, JCP
I totally agree with the assessment. I’ve flown with several PC-12 pilots who were not prepared for the unique demands of the aircraft, and it shows. Lack of situational awareness, poor planning, and a general sense of complacency are just a few issues I’ve observed. It’s essential for pilots to understand the PC-12’s quirks and nuances before taking the controls.
Several years ago, I worked for Mitsubishi Aircraft International (MAI).
In response of pressure on the FAA, they conducted a recertification review of the MU-2 as there were 22 fatal crashes in a few years. Needless to say, MHI was the butt of our competitor’s sales pitches, including a letter to prospects from the president of a Wichita airframe manufacturer that the MU-2 was a flying coffin.
For those that don’t remember the “Rice Rocket”, the two final variants, the Solitaire and Marque (stretched fuselage) were high wing, twin-turboprop, very high-performance aircraft. The inspection found that the majority the accidents were situations such as Pilot landed short, Pilot landed long, aircraft encountered an unexpectedly high mountain, and fuel imbalance. We actually had an aircraft where the fuel was so out of balance that the aircraft cartwheeled on landing (note: the wings did not separate from the fuselage and the only injury was that a passenger injured his elbow). Many of these accidents occurred under IMC. The result of their investigation resulted in an additional inspection (not an AD) of a bolt, and a minor wording change of the Flight Manual regarding operation in icing. The FAA also recommended further and recurring simulator training for all operators.
An airframe manufacturer suffers from any accident, even non-fatal ones, but this is especially true with high-performance aircraft.