Arrow PA-28 Experienced In-Flight Break-Up

[IO540]

You can bend or break any plane if you hit something with it. That doesn't mean anything.

The question is how easy it is to bend when flying.

Most common types have suffered in flight breakups. A few (very few) types haven't. The Socatas have a single piece aluminium spar, machined from a solid lump.

The Commander landing gear is the sturdiest I have ever seen, but it will only be as good as its attachment points. The TB20 is a very similar trailing link setup though not apparently as strong.

Deice- Yes.

Fuji- did you read about that Irish King Air which got bent to the point where upper skin was flaking off, but "nobody noticed"?

[Three Yellows]

Another pearl of wisdom from Sternone

And if you are hour building, please fly a real twin, like a Seneca V, Baron, ... not a DA42... i disapprove commercial student pilots learning in FADEC controlled twins...

This was posted on Nov 9th, just one month after his first solo!

Sternone, do you think that, for example all new train drivers should learn on the Flying Scotsman before they are allowed to take out a start of the art Pendelino?

By the way, you don't need an armrest on the DA42 as the autopilot does all the work from about 800ft after take off until about 300ft above the runway (ILS equipped ones anyway). The armrest would get in the way of pouring out the coffee.

What did you solo on by the way?

[ronnie3585]

Fuji- did you read about that Irish King Air which got bent to the point where upper skin was flaking off, but "nobody noticed"?

Elabourate - please?

[wsmempson]

Quote "What did you solo on by the way?"

Concorde, I shouldn't wonder....

[englishal]

Any plane can break up in flight, it depends on the noodle at the controls.

Saying that, there was a B747SP which hit severe turbulence over the pacific some time ago. It went out of control and during the recovery they estimated that the pilots pulled > 5G ....they bent the airframe. Still managed to fly it to the USA, several hours away....

5G is a lot. But that is not the whole story. If I grab the stick and yank it aft (at Va), I may well get away with pulling 5g and surviving to tell the tale. If however, I yank the stick full aft and also wack it over full left or right, the wings may pop off well before we even get near to stalling.......

[SkyHawk-N]

If you want a plane with no known case of in-flight failure, get a TB20

N110U?

N163GT?

[SkyHawk-N]

cjboy, yes I think you missed the point, try reading the post, no sign of the word 'structural' in IO540s quote.

[421]

Irrelevent. Engine failures are down to the engine manufacturer (Lycoming in this case). Engine failures happen to any aircraft. This thread is specific to in flight break up. IO540 was refering to the fact there has been no case of an inflight break up of a Socata.

[SNS3Guppy]

All aircraft, without exception, have a Va limit, which depends, to a large extent, on the strength of the wing spar. In fact, in order to certify, the aircraft structure has to be strong enough to survive certain "g"'s at MTOM, and Va is a direct result of that. (For a normal category aircraft, I think the positive g limit is 3.8. Not a lot actually.)

Not all aircraft have a Va assigned. Transport category aircraft, for example, don't use Va. Other speeds are used, such as recommended turbulence penetration speeds, though sometimes these are minimumum speeds, and in some cases, maximum.

No, VA doesn't depend on the strength of the spar, but it does have to do with loads on the controls as well as the primary aircraft structure. VA is predicated on stalling before structural damage will occur, and is therefore predicated on on a stall relieving the load before any structural margins are met. Meeting Va also depends on movement at a given rate; more or less than this does not provide protection, and one can indeed break an airplane below Va. Va is also predicated on a single movement of the controls (single event deflection of the control surface), not control reversals or multiple deflections; again, an airplane may break well below Va. It's not the magic number many instructors seem to believe it is.

The G limit of the airplane isn't tied to Va. Maneuvering limits are not the same as the loads imposed by Va, and trying to think of them together only confuses the issue. Certification G load limits are assigned per the aircraft type certification, whereas loads considered for maneuvering speed, turbulence penetration speed, and other such speeds (which can vary depending on aircraft configuration and weight) are design limits. One discusses where you can go legally and *safely,* whereas the other discusses where things will start to break. Two different subjects.

Ever wonder why maneuvering speed varies with weight? It's a stall margin...stall at higher airspeeds at higher weights, and maneuvering speed increases accordingly. Again, you can overstress an airplane below maneuvering speed.

A good example of that is American Airlines Flight 587, which experienced a separation of the vertical stabilizer owing to multiple control inputs by the pilot at a relatively low speed during climb.

If you fly faster than Va (and Va changes with actual mass) and apply full and abrupt control deflections, you will exceed the design parameters of the airplane. If you also exceed the safety margin that manufacturers built in, then you will break off the wings (or the tail section). Doesn't matter whether the aircraft is built by Mooney or Piper.

The assertion that the wing or tail section will break off is erroneous. Not necessarily. But you are operating outside the parameters of the aircraft and again, structural damage or failure can occur below Va.

If you challenge me, I would be willing to take up a Mooney and demonstrate that it is indeed possible to fly its wings off. Mind you, I'd want to do an emergency egress course and a parachute free fall course first. But then I'd just dive a Mooney to Vne (or beyond) and pull up sharply. (On second thoughts, never mind.)

Of course your'e saying this tongue in cheek, but as a skydiver and parachutist, and one who has experienced structural failure in flight in aircraft, I'll tell you it's a bad idea, anyway. I'd also like to add that exceeding Vne doesn't mean you're going to break up. A bigger issue is passing beyond a region of flutter testing for the airplane...and that's what should really be being discussed here regarding this accident...not maneuvering speed and not wing spar failures. Flutter.

The ailerons were separated. Not the wings. It wasn't the wings that broke up. It was the ailerons that separated from the airframe. Overloading a wingspar can break a wingspar, or break a wing. I have had an entire wing crack due to low level maneuvering (and a pre-existing unknown stress riser) at low level in mountainous terrain, in a large airplane...and at speeds well below cruise (no maneuvering speed for the airplane). Not in this case...something caused not the wing to break, but the ailerons to separate. What does that? Flutter.

Control flutter can occur suddenly. Its onset is rapid, and in a true flutter situation, violent. It can remove the control surfaces in less than a second, it it nearly always results in the loss of the aircraft. It can also predicate, or lead to an aircraft breakup.

I was only a few miles from the crash site of this aircraft when it happened. As I turned out, I knew nothing about it. The wind was calm, the weather fair, not particularly cloudy, no mountain obscuration. It's an uncontrolled area, little radar coverage (what you do have isn't useable for ATC purposes), nobody to talk to once you reach the elevation of the mountain tops. The entire area is very mountainous, with elevations ranging from 4,500' to 11,000'. When any wind is present, turbulence can be strong, ranging from moderate to severe or greater. On this particular day, there was very little wind.

Airplanes are built with minimal structure to save weight. Certainly one should remember that while limitations and numbers are given, airplanes can break at much smaller numbers, and care should always be taken flying the airpalne. I the case of this Arrow, we hae precious little information to go by. Guesswork does not become a professional attitude in the cockpit (which should be the hallmark of the private pilot, too); we don't speculate; we know. In this case, we don't know, and shouldn't speculate. Certain observations can be made regarding the facts, but there are far too facts upon which to base any reasonable assertions. For now, the best that can be said is this should serve as a reminder to operate the airplane well within the parameters given you, and not push your luck.

[IO540]

The G limit of the airplane isn't tied to Va

I thought the two were tied, by the fact that if you arrange for the wing to stall and thus unload, its design limit cannot be exceeded.

Flight at/below Va ensures that the wing will stall and unload before the design limit is reached.

That is why Va varies with weight. At lower weights, Vs is lower and therefore you have to fly slower to maintain the said margin. If it was not for the main spar stress, Va would not vary (much) with weight because the aircraft weight is carried primarily by the main spar. The elevator, in fact, tends to apply a downward force...

As you suggest, there are other ways to break things, without exceeding the main spar limits. But e.g. control (or other) surface flutter is not related to Va, it's related to Vne.

[modelman]

Anybody there?

Sternone seems to have gone a little quiet

[Fuji Abound]

Sternone seems to have gone a little quiet

Quietly listening perhaps.

[Three Yellows]

Quote "What did you solo on by the way?"

Concorde, I shouldn't wonder....

Surely it was one of the Wright Brother's early aeroplanes as we know that Sternone doesn't approve of training on Technically Advanced Aeroplanes.

[SNS3Guppy]

Flight at/below Va ensures that the wing will stall and unload before the design limit is reached.

No, it doesn't. Again, that's a myth that's sadly perpetuated by instructors today. The airplane can be broken well below Va.

Further, a stall does not ensure that the wing has unloaded, in fact, it can be stalled and in excess of its ultimate loading.

A stall does not automatically unload a wing, or control surface.
Va provides a margin under specific conditions with a specific rate of control deflection in which the greatest chance under those specific conditions exist that the surface will stall before becoming damaged. The structure can be pushed beyond it's capabilities long before that speed is achieved however. See my previous post.

The elevator, in fact, tends to apply a downward force...

Not in all aircraft, not in all conditions.

As you suggest, there are other ways to break things, without exceeding the main spar limits. But e.g. control (or other) surface flutter is not related to Va, it's related to Vne.

No. Flutter can occur at speeds well below Vne, and well below Va, too. Flutter is very much a control balance issue.

In the C-130, for example, we were restricted to 250 knots if anything at all had been done to the ailerons or other control surfaces, including painting a stripe. That restriction remained in effect until the control surfaces were removed, and rebalanced.

Steve Whitman, a prolific aircraft designer and homebuilder, died with his wife in a Whitman tailwind enroute to the Oshkosh airshow some years ago. The aircraft broke up as the result of aileron flutter due to a control imbalance when some pinking strips of fabric delaminated along the trailing edge. The ensuing flutter separated the ailerons and it cost him his life. His crime? Mixing two fabric processes (Stits and Ceconite, I believe). The incompatibility lost adhesion between the different fabrics and treatments, imbalancing the controls, causing flutter (well below Vne, mind you, and below Va), and abruply terminating the flight as the aileron separated from the airframe.

Normal flutter tests are done as dive tests above Vne; an airframe normally shouldn't flutter until those speeds. However numerous things can affect the flutter characteristics of an airfoil, including the control cable tension, pilot input on the control stick or yoke (including pulsing the controls by input, or stiffenign the controls by applying a death grip on the yoke), atmospheric conditions, air density, temperature, altitude, counter balances, stuctural stiffness, control stiffness, control geometry, etc. Flutter is dynamic instability, which can, once again, occur well below Vne. Or Va.

The G limit of the airplane isn't tied to Va
I thought the two were tied, by the fact that if you arrange for the wing to stall and thus unload, its design limit cannot be exceeded.

No. The G limits previous discussed involved certification limits, such as those applied to a given category of airplane (normal category limits, for example). These are not at all the same as design limit criteria, nor do they relate in any way to the ultimate airframe limits (the point at which the airframe will actually break).

Limits and flutter and even Va can become somewhat complicated subjects, as there's a lot more than stalling before the airplane breaks. Suffice it to say for now that one should understand that the airplane can indeed be broken at speeds less than Va.

Years ago I worked on a lot of single Cessnas. In particular the 200 series (206, 207, etc). I found numerous vertical stab attach brackets broken from slipping. Pilots don't often realize just how little is holding the parts of the airplane together, and many would be absolutely shocked to learn that some entire surfaces are attached with as few as two bolts. The single Cessna clan has the vertical stab supported with attach brackets, which came from the factory as aluminum brackets. These have mostly all been replaced with steel brackets (which sets up the subject of another separate discussion; dissimiliar metals corrosion, or electrolytic corrosion). The aluminum brackets were often found during inspections to have failed. Did this occur under high loads above Vne? Nope. Normal slips to landings as many instructors teach.
I used to throw big airplanes (large four engine airplanes) and small airplanes alike around in full slips, as my different employment required After enough experience in the shop I learned to stop doing that a long time ago. I've seen the results from simple, low speed slips. An enormous stress is placed on the airplane as a result of those sideloads. I've seen the same cracks and breaks on large airplanes during depot-level inspections when we've pulled the vertical stab as part of the inspection process. It can happen just as easily to the light airplane you own or rent locally, just from normal flying...and it's not something you'll be finding in the checklist or pilot operating handbook, either. Certainly something to keep in mind.

Your airplane most definitely can break at speeds less than the rough air, turbulent penetration, or maneuvering speeds. The presence of those speeds in your aircraft handbook doesn't mean you can't break the airplane below; it only means you should avoid abrupt control motions above those speeds. It means you can break them above those speeds, it doesn't mean you can't break them at a lesser velocity.

[Mark1234]

Seems I misread the initial post(s) to suggest that large control movements were the cause. However, it does state that 'outboard portions' of both wings had detatched, (as well as the ailerons).

I am curious however as to how a stalled wing can be producing more force than an unstalled, and exceeding the load limitations, through a reasonably feasible flight manouver up to and including Va?

At VERY high speed I can see it, but surely the whole definition of aerodynamic stall precludes there being more force stalled vs unstalled?

As for breaking it <Va - well, Va is for the full deflection of ONE control (usually pitch plane I believe), so if you start throwing in large deflections of other controls at the same time, that would seem reasonable.

[Pilot DAR]

Quote: :Hihi, Also i feel quit sorry for you that the only argument you can tell me is that i'm low houred not yet PPL rated.. comon, please do better than that, i could actually learn something from you guys.

Com'on Steronon,

You're really a super experienced, highly qualified pilot and maintainer, just stirring the pot for fun here right? I do find your posts really entertaining, in a laugh at the ridiculous kind of a way. If you really wanted to learn from those of us with a lot of experience, you'd ask intelligent questions, and not attach silly statements, with no basis of fact behind them.

The Arrow, as the other Piper aircraft of that lineage, are fine aircraft, quite deserving of their well established place in our industry, and longevity.

There was in past times, an AD against PA28 wings for structural inspection at the attachment. I helped pull a few for this inspection. Eventually the AD was withdrawn, because apparently (as we certainly learned) you generally did more damage getting to wings off for inspection, than the damage (none ever found by us) that you were supposed to look for.

Any aircraft can be broken if you abuse it (either deliberately, or accidentally) badly enough. Poor aircraft designs are uncompetitive, and just don't survive in our industry, much less prosper as well as the PA28 series.

Sternone, You're welcome to read more, and comment less. There is a whole lot of wisdom here, why not learn all you want, without annoying the contributors?

Pilot DAR

[SNS3Guppy]

I am curious however as to how a stalled wing can be producing more force than an unstalled, and exceeding the load limitations, through a reasonably feasible flight manouver up to and including Va?

A stall has nothing to do with unloading a wing. It has everything to do with exceeding the critical angle of attack for a wing or airfoil, to the point of airflow separation. This does not mean the wing is no longer loaded. The mere act of increasing it to that point may overstress the wing, stall or not.

A rapid pitch up for example may cause an accelerated stall, as could a turning stall with a load factor causing the stall to be higher than normal...as you understand, a stall may occur at any attitude and at any airspeed.

Rapidly causing a surface to stall, be it a wing or othewise, places a high load on one side of the surface and not on the other; the pressure distribution shows a rapid increase on the side against airflow. While the airplane may now be stalled, or may be rotating deeper into the stall, the load doesn't magically disappear, and may continue to increase, depending on the dynamics of the airplane. If you're doing local stall practice, straight ahead, poewr off, on a calm clear day, you may not perceive this...but you can certainly break an airplane in the process of stalling it...high speed or low. All one need do is exceed the ultimate load capabilities for any given component, and it need not be the spar.

One airplane I used to fly routinely lost 27,000 lbs during the flight in a matter of one and a half to four seconds, as part of it's mission assignment. It was a very stout airplane which underwent regular maitnenance and detailed inspections. I performed many of them myself. Four years ago on a normal drop, as the aircraft finished unloading, well below any design limits or maneuvering limits, the right wing folded back,the airplane rolled left and the left wing came off. Three seconds later it exploded as it struck the ground, killing three of my associates. A short time later, three weeks, another airplane I was regularly flying did the same thing. Both in undemanding conditions, both below design maneuvering speeds (no published for either one, but below the turbulence penetration speeds. The second crash killed two.

I previously identified American Airlines Flight 587 as having lost it's vertical stabilizer during a normal climbout, well below penetration speeds. It encountered some wake turbulence; the turbulence did not separate the flyin surface; the pilot did that with multiple control inputs. These inputs were within what was considered to be a safe flying speed, and were done in an aircraft which automatically determined how much was too much based on the airspeed and it's control logic. It theory, aerodynamically the airpalne couldn't break, and operationally it wasn't supposed to allow itself to be broken. But it was, killing a large number of people. Do some research.

Last year I did a brief assignment in a learjet. I had to attend a recurrent training in the type, and one of the things I noted was the addition to the Aircraft Flight Manual of statements regarding the ability to break the airplane at low speeds.

One should also remember that Va is predicated on a single event; one application of the controls in one direction under very specific conditions. In a typical turbulence encounter, we may move the controls both directions to counter a roll, yaw, or pitch excursion. Youv'e doubtless experienced this yourself...flying down final approach during a blustery, gusty day, for example. You may be very busy on the rudders and very busy on the control yoke as you fly...that's normal. However, Va doesn't account for anything but the first control input. After that, you're no longer on the charts, and this is something that's not taught...most pilots eroneously believe that so long as they're below maneuvering speed, they're okay. Not at all true.

At VERY high speed I can see it, but surely the whole definition of aerodynamic stall precludes there being more force stalled vs unstalled?

No the definition of a stall is exceeding the critical angle of attack for the flying surface...and again, has nothing to do with unloading the surface.

When I teach a student about basic aerodynamics, I like to ask the student if he or she has ever put his hand out the car window and "flown" it in the slipstream. Of course, everyone has. This then allows the student to understand much of what he or she needs to know, just from experiencing it with the hand. By slowly increasing angle of attack, the hand will experience an increase in lift, requiring less and less effort from the muscles in the arm to hold it up. At some point, past the maximum coefficient of lift when drag exceeds lift and we enter into a region where airflow separation over the hand increases, we see an enormous rise in induced drag and a drastic reduction in lift. The hand has stalled. Lift isn't there to hold the hand up, but now you have a lot of retarding force pulling the hand back toward the rear frame of the car window. Do it rapidly, of course, and you may strike your hand or arm on the window frame. The force hasn't suddenly quit; it's changed velocity (direction) and been given new vectors.

In a steep turn the airplane may be made to stall at a higher speed than one sees while straight and level. Pulling harder in the turn works up to a point when the angle of attack is exceeded, and the airplane stalls. But at that point, the wing has not unloaded. Rolling the wings level helps, if you're able, but pushing forward unloads the wings...even in the turn, and increases the ability to help roll out of the turn.

What we're not seeing in this mishap is a case of the wings being pulled off, but a separation at the outboard portions, with loss of the ailerons.

Again, do a little research on aileron flutter, which can occur at speeds considerably lower than Va, as well as high speeds. It's a complex subject, but flutter is very often a fatal event. The time from initial onset to loss of the control surface and loss of control is typically well under a second, and it's violent. Many things can cause it from a loose control to a control imbalance to air density to pilot input to excess speed and a combination of these things. Aircraft are flutter analyzed, and design tested and dive tested, but each aircraft is subject to it's own individual condition, and individual operating conditions. Since no detail is yet available regarding this particular mishap, that' about as far as we can reach and go with the topic...until more information is given.

As for breaking it <Va - well, Va is for the full deflection of ONE control (usually pitch plane I believe), so if you start throwing in large deflections of other controls at the same time, that would seem reasonable.

Mark, you raise a valid point there, except it's one control input...not just one control. That is, one movement to the left of the rudder, for example, or one movement right of the ailerons. Moving them back the other way, reversing the control input, changes everything, and may cause the airplane to break without any other inputs. Even below Va.

Your point is well taken that the use of other controls may put additional forces on the structure that take it outside it's design envelope. Rolling and pitching at the same time, for example, can put some severe uneven bending and twisting forces on the wing and on the fuselage, as well as on the control surfaces themselves. Some aircraft don't have a whole lot holding the control surface on in the first place...also something to consider. The structure is only ever as strong as it's weakest point, and that's seldom the control cable to which the pilot is roughly attached. Always keep that in mind.

[kiwi chick]

Oh my god! This is the most interesting thread I've read in my entire life I think!

I started reading it 'cos I fly a PA28... and was ready to quip that if doesn't break the way I fly it, it must be pretty sturdy

But my goodness, SNS3Guppy I was riveted to my seat reading all that! And yes, it is all stuff I should know but I felt like I learnt a whole lot more!

Thank you!

(how did you LEARN all that?! I thought I was a PoF geek but now I feel quite humbled, hahahaha!! )

[SNS3Guppy]

how did you LEARN all that?!

Unfortunately, the hard way.

[deice]

Interesting reading as always SNS3Guppy, but what you are saying is that normal flight can cause an aircraft to fall apart around us. We should all remain on the ground then!

Are you referring to the C130 that lost its wings unloading over a fire? There's a video of that on the web and it looks absolutely terrifying! But, aren't those aircraft subjected to abnormal conditions and flying in extreme environments, hot, low level and abrupt maneuvring? I always thought it looked exceptionally dangerous...

On another note, perhaps you've heard of the CASA 212 from the Swedish coastgaurd that lost a wing at low level flying just off the coast last summer. They grounded the fleet and I have yet to hear of any actual cause, but you have hinted on a few reasons. Very interesting indeed!

Keep up the lecturing please!