All hell broke loose… then, only wind noise
A recent accident involving a vacuum failure in a V35B Bonanza and subsequent loss of control and airframe failure made me recall that this was a really substantial problem in the 1980s and early 90s. It also made me recall one of the better private flying war stories – from a pilot who survived an airframe failure.
This pilot was testing a Piper Seneca for a company that was seeking approval for airframe modifications on that aircraft. The flight on the day in question was to check flutter margins. (To appreciate the importance of flutter, watch this video before continuing to read. Plum scary.)
As I understood it, the proper procedure was to increase airspeed in small increments and pulse the elevator control after each increase. As the airplane came close to the flutter limit, an airframe anomaly would develop, more often than not in the horizontal tail, telling the pilot that was enough already. This would be done at speeds in excess of Vne, thus the term flutter margins. The faster an airplane is flying, the more susceptible it is to flutter. The margin would be the difference between the never exceed speed and the speed at which flutter started.
This pilot anticipated no problems and was in a hurry so he was increasing the airspeed in larger increments than recommended before pulsing the wheel.
That was when everything came unglued. He related incredibly loud noises as the airframe disassembled itself. Extremely high g-forces were also felt before serenity returned and only wind noise remained. The pilot had a good four-point restraint system and a crash helmet and was unscathed by the airframe breakup violence that probably lasted only a few seconds.
The pilot also had a parachute firmly strapped on and, when reality set in, he exited what was left of the airplane, popped his chute, and descended to a mountainside.
They knew right where he was and sent a helicopter, a really minimalist helicopter. When the Brantly B-2 came into view, the pilot wondered if the little two-seater would be up to the task. It was, the happy ending followed, and he was telling me the story a few weeks later.
People have survived airframe failures in flight without the help of a parachute. I actually remember one case, but just one, in the over 50 years I have been reading accident reports. In other words, if the airframe fails, the rest of the day doesn’t look good.
The Bonanza that was lost recently was being flown by a 4,000 hour ATP. The pilot reported a vacuum system failure and added that he was VFR on top and would continue to his destination. Then the airplane flew back into IMC and the pilot reported losing control of the airplane. It broke up shortly after that.
The debris path of the airplane was less than half a mile long. The right ruddervator was at the beginning of the debris path followed shortly by other wing, tail and fuselage parts. The engine and instrument panel were at the end of the debris path.
The NTSB is descriptive in outlining the debris path in such accidents for good reason. Any airframe will fail if operated far enough outside the envelope and it’s useful to know what fails first. You haven’t been tuned in over the years if you didn’t know that it’s the V-tail on a Model 35 Bonanza. The tails were required to be strengthened a while back but from that most recent event, it looks like the tail still fails first.
Accidents like this used to be commonplace, so much so that the NTSB/FAA went through three periods of near-panic on the subject. One was caused by vacuum pump failures.
The first accident posted in an NTSB safety recommendation came in early-1982. It was the in-flight breakup of a Cessna P210 following a vacuum system failure in IMC. The reason this accident made the subject pop to the surface was the fact that four prominent professional people were lost in the accident.
A Cessna T210 was listed next. There were two instrument-rated pilots in the aircraft and they made it through 20 minutes of partial panel flying after the vacuum failure only to lose control on an attempted ILS approach.
In another Cessna T210 accident the pilot was flying at Flight Level 190, above the clouds, when he was cleared to 13,000 feet. The pilot lost control of the airplane shortly after entering the clouds and the aircraft broke up in flight. The vacuum pump had failed.
A Mooney M20F with four on board was flying above clouds when the pilot reported a vacuum failure. The pilot continued on toward his destination and was subsequently cleared to descend into the clouds. Shortly thereafter the airplane crashed in a steep, high speed, nose down attitude.
Still another Cessna T210 was lost after the pilot reported a vacuum failure while flying at 12,000 feet, above the clouds. He was cleared direct to a requested airport and given a lower altitude. Shortly after entering the clouds at about 10,000 feet the pilot lost control of the aircraft. The right wing and empennage separated in flight. Four lost. (FYI, after this Cessna tested a 210 wing to destruction. It reportedly failed at above 7-g which is well in excess of the 5.7-g ultimate load factor requirement for this airplane.)
These accidents happened in about one year.
The NTSB took aim at vacuum systems, especially on Cessna 210s with deicing equipment. Because I had one, I did test flying in connection with this and you can read about that in my post about my experiences with the P210 and what was done to improve the reliability of the vacuum pump used on that airplane.
The NTSB came up with a laundry list of recommendations for the FAA to address this problem. I’ll sum them up.
An Emergency AD requiring dual vacuum pumps on Cessna 210s with deicer boots was called for plus a design certification review of the pneumatic portion of the deicing system on that airplane.
There were extensive other recommendations related to engineering evaluations of vacuum systems as well a new requirements for certification of airplanes with boots as well as some new requirements for turbocharged airplanes.
There were also recommendations for more rigorous pilot training and recent experience requirements on partial panel operations.
As is usually the case, the FAA followed some of the NTSB recommendations and pretty much ignored others.
Things settled down for a while but up jumped the devil in the late 1980s. This time vacuum pumps were not the culprit and only one type was involved. From May 31, 1989, to March 17, 1991, Piper PA-46 (Malibu or Mirage) airplanes were involved in seven airframe failure accidents worldwide. Five happened in the U.S., one in Japan and one in Mexico.
This activity got the undivided attention of the NTSB and the FAA. The NTSB did a special investigative report (NTSB/SIR-92-03) on the subject and the FAA did a special certification review on the PA-46 which is, in effect, a complete review of the certification of the airplane.
First, the wrecks. Because more complete information was available on those that happened in the U.S. the NTSB report concentrated on those.
In every case, the pilot lost control and the airframe failed. The vacuum system was not implicated in any of the accidents. One originated in a thunderstorm. That happens in all types so the airframe failure was all that accident had in common with the other four.
The NTSB detailed the debris pattern and in most failures it appears that the horizontal tail failed first. A possible exception was one where control was lost at high altitude, the pilot had the foresight to extend the landing gear to limit the speed buildup, and the airplane emerged from the clouds at 2,000 feet, apparently in one piece. Despite the gear being down the speed was still high. The nose was down 15 to 20 degrees and when the pilot saw the ground he apparently pulled too hard to get the nose up. A witness said that both wings broke off simultaneously.
In the special certification review the FAA obviously looked at the airplanes’ ability to withstand the required 5.8-g ultimate loads. As Cessna had previously done, Piper went beyond that to see what loading would eventually cause the wings to fail. The result was about the same as for the Cessna 210: 7.7-g.
No smoking gun could be found in or about the PA-46 so they turned to the pilots and, guess what, they identified the culprit. A common thread was that ice was likely and the pitot heat had not been activated by the pilot. There was also a strong suggestion that the autopilot had not been operated in a proper manner. These items led to a review of the training programs for the airplane.
There was one incident (as opposed to accident) in the report and it is of special interest because the pilot lived to tell the story. This well illustrates how there can be a great difference between what a disoriented pilot thinks is going on and what is actually going on.
The Malibu was flying normally and level at FL180 when the pilot noticed some cumulus buildups ahead and requested a higher altitude. The controller cleared the flight to FL200.
The pilot programmed the autopilot for the climb and as soon as it was engaged in the climb mode the airplane pitched up sharply to a much more nose-up attitude than the pilot expected. He disengaged the autopilot as the airspeed reached zero. The pilot reported that the airplane then stalled. He also remembered observing the manual trim wheel complete a move to the full nose-up position.
The pilot said he attempted to push forward on the controls but that the airplane entered a left spin or spiral into the clouds. The pilot said that the uncontrolled descent in a spin or spiral continued until he was able to recover when the airplane exited the clouds at 3,000 feet. The pilot said he might have recovered near 7,000 feet but the airplane resumed the spin or spiral.
The pilot was talking to the controller the whole time and there was apparent confusion. Once he was in VFR conditions he first said he was in visual conditions but the airplane would not respond to any power setting. Then the pilot said he was at 3,000 feet, flying level.
The NTSB reconstructed the flightpath and once the airplane departed from controlled flight at FL180 it showed wide speed variations, to a groundspeed as high as 250 knots (Vne is 203 knots). The descent slowed briefly at 14,000 feet but then continued rapidly to 9,000 feet after which the airplane climbed back to 11,600 feet.
For the next five minutes the airplane flew through a series of climbs and descents and sharp turns left and right. Then the pilot reported level at 3,000 feet.
At some point in the proceedings, an airline pilot suggested that the Malibu pilot turn the pitot heat on. The NTSB speculated that this was probably sufficient to alert the pilot that his airspeed indication was erroneous.
The pilot estimated he had about 650 hours which included 20 hours of instrument flight time, about five or six of which was in actual instrument conditions. The pilot had attended the Piper school on the PA-46 and had flown the airplane for about 70 hours.
After the pilot landed, numerous loosened rivets were found in the aircraft structure, the elevator trim tab was bent and the right wheel well door was missing. Nobody knows how close the airframe was to failing but it was probably at least nibbling at the edges.
As has been often true, the event began with a low-speed loss of control which evolved into a spiral dive (graveyard spiral) that results in speed well in excess of Vne.
In the investigation of the factory training program for the PA-46 it was found that it provided minimal instruction on the use of the KFC 150 flight control system especially including the vertical speed and altitude selection features.
In the training the student’s instrument flying skills were evaluated to some extent and instructors made recommendations on whether further training was needed. There was no training on recovery from unusual attitudes in IMC or on partial panel instrument flying.
In the heat of battle the FAA issues the obligatory airworthiness directive on the P-46. As you might imagine, it was a touch Draconian.
Flight in instrument meteorological conditions was prohibited.
Use of the autopilot or associated devices for altitude changes was prohibited and to back this up it was required that the control panel for these functions be removed from the aircraft.
It was required that the pitot heat and alternate induction air be on at all times except during takeoff and landing.
It was recommended that flight into area of known or forecast moderate or severe turbulence be avoided.
Within a month the FAA lifted the prohibition on flight in IMC and added that flight into known or forecast ice, thunderstorms, moderate or severe turbulence was prohibited. Virtually all PA-46 airplanes were equipped and approved for flight in icing conditions; this effectively removed that approval.
About ten months later, the FAA rescinded the AD.
One thing the NTSB recommended but that was rejected by the FAA related to the requirement for special training before operating a pressurized airplane certified for flight above 25,000 feet. The NTSB wanted the altitude lowered to 18,000 feet, to include the PA-46 and P210 airplanes and if the FAA had put any thought into this that would have been done. There must have been a strong lobby against the change.
The V-tail Bonanza has also undergone a special certification review prompted by a number of airframe failures in that type over a lot of years. (The V-tail started coming to the fleet as a 1947 model.) One thing that might have added to interest in this was the fact that the Model 33 Bonanza (originally the Debonair) didn’t have a history of airframe failures and the primary difference between the two airframes is the tail.
Every airplane has some flying surface that will let go first and on the V-tail it was indeed found to be the tail.
Under some combination of high speed, turbulence induced air loads, and pilot control input the twisting loads that developed on the tail could lead to a failure. The ground and flight tests that revealed this were more thorough than those found in normal certification work and it might be said that the V-tail Bonanza airframe is the most thoroughly tested in the private aviation fleet.
There were ADs, mainly restricting the speed at which V-tails could be flown. There were fixes to restore the original limitations and this action did confirm that under extreme circumstances the airplane might not have met the certification standards even though the FAA and Beech had held that it did.
Ironically, in more than half of the approximately 230 V-tail airframe failures that occurred up to the late 1980s, the wings failed first. If the number of failures seems high, over 10,000 V-tail Bonanzas were built and were flown in this 40-year period
Even though the Mirage and V-tail were the airplanes where the structure was questioned, the airframe failure rate in those types is about the same as in the Cessna 210s and Piper PA-32 retractables. The simple truth is that if you lose control and the speed and/or g-loading goes outside the envelope some part of the airframe will likely fail, leading to the failure of other parts.
If this type accident is declining, and it does appear to be doing so, there are two main reasons this might be so. Training has become many times better in the past 30 years and more pilots now have a better understanding of autopilots and airplane limitations. The other reason is a negative. There is simply far less flying being done and pilots pushing weather to get utility out of these airplanes is being done far less often. Certainly the airplanes have not grown stronger and the weather has not become more forgiving.
In closing, a technique note. We all know that reducing the angle of attack will greatly lessen the likelihood of a low speed loss of control. The solution to avoiding a high speed event is just as simple. If things are going crazy, just keep the wings level. This can be done with the most basic instrument (turn and bank or turn coordinator) or with attitude indication. Some autopilots have a panic button that will do this for you. If you ever let a cat out of a bag and tried to put it back in, you know how much easier it is to just keep the cat in the bag.