SonexBuilders.net

Results are in, the NTSB has released the result of its investigation into the fatal crash of Waiex N75654 where the "V" portion of the "Y" tail separated in flight.

Synopsis
http://www.ntsb.gov/aviationquery/brief2.aspx?ev_id=20111023X81705&ntsbno=ERA12FA018&akey=1

Entire Public Docket
http://dms.ntsb.gov/pubdms/search/hitlist.cfm?docketID=52467
There are many more photos in the docket than I have posted here in this thread.. I chose to omit some of them in this thread to respect some people's sensitivities to this type of thing.


This is the photo I have been waiting to see for a year and a half.. what exactly was the failure mode that caused that tail to rip off?

I had been thinking that the attach bolts had pulled through this whole time, clearly not the case. It is obvious that the lower attach bracket failed and then "unzipped" from right to left.

Quoting the Materials Lab Report

The profile was transversely fractured separating the horizontal and vertical flanges. Fracture features such as shear lips and rough matte grey surfaces indicated tensile and shear overstress fracturing. There were no indications of fatigue or other progressive crack growth.The profile fracture exhibited a 2.5 to 3 inches long tensile overstress in the vertical flange. The tensile region was at the right edge of the vertical flange, near the profile’s fillet. The rest of the profile was fractured by shearing overstress in the horizontal flange. Figure 2 shows these fracture regions.
Figure 3 shows the large out-of-plane deformation of the horizontal flange of the profile, along with clear longitudinal elongations of the bolt holes, as shown in figure 1A. The vertical portion of the profile did not show any out-of-plane or within-plane deformation, but the two rightmost bolt holes exhibited elongation in the vertical direction.

Everything to me looks like the forward failed, then as a result the aft attachment failed, but here is what they have to say about the aft attachment

Both plates showed similar overstress fracture characteristics. The fracture surfaces indicated a tensile overstress in the vertical direction, along the entire width of both plates. The fracture surfaces were more erratic (shifting planes) on the left side, as compared to the right side. There were no indications of fatigue crack growth.
Both plates exhibited similar plastic deformations at the left side of the fracture. As figure 4 shows, these deformations locally distorted the the left sides of the plates toward the left and aft, with respect to the aircraft.
The bolt holes in the right sides of the Y plates and in the short L profile, were elongated in the vertical direction, shown in figure 5. The portions connected to the fuselage also exhibited yielding around the right side holes at or near the fractures, as depicted in figure 6.

From the "Structures Study"

This airplane was climbing in a left hand circling pattern. It had reached about 4700 feet msl and started a slight descent and accelerated to about 100 to 110 mph groundspeed (derived from radar data)1. The airplane broke up in flight. The exact circumstances including airspeed and any maneuvering leading up to the breakup are unknown.

This to me means we will never know if it was the result of an overspeed condition caused by pilot incapacitation, etc. The Structure's study goes on to describe numerous builder "less than optimum" practices in this airframe.

The breakup appears to have initiated in the forward V-tail attach structure. The lower attach angle (XNS-T09-04) separated in tension along the right side and further separated with the fracture running from the right to the left. The lower attach angle (XNS-T09-04) is fastened to and is an integral part of the forward spar structure of the stabilizers. It provides the angle for attachment points to the fuselage. As the lower attach angle was breaking, the structure that forms the rear attach Y-fitting also fractured and separated.

The kit designer estimated that the Waiex fleet has accumulated between 5,000 and 10,000 flight hours that include many aerobatic maneuvers. The kit designer is not aware of any other inflight structural breakup of this model airplane.
Photos of less-than-optimum construction are provided.

Interesting to me, was how relatively in-tact the tail group was, despite the inflight breakup and subsequent fall from 4700 feet. It is clear where the weak point of an otherwise solid airframe was.




LESS-THAN-OPTIMUM CONSTRUCTION

The photo's speak for themselves, the builder was cutting corners in a lot of places. The report did not come out and say that there was a single definitive flaw that produced the breakup, however.

Here is the NTSB's description of the new FAR 23 load case that, when evaluated, prompted Sonex Aircraft to release the service bulletin on the tail:

NOTE: The following loads evaluation is used to demonstrate how loads applied to a V- tail design may be greater than that of a conventional tail airplane. The evaluation is based on theoretical values that are not derived from the specific events of the accident at Washington, GA. In addition, the simple buildup does not allow for the rapidity of changes in AOA or yaw during the abrupt application of flight control inputs.

The V-tail design is often considered more aerodynamically efficient that a conventional tail design since the profile drag of the two surfaces is less than a three-surface configuration. However, each surface of a V-tail may have to provide both vertical and lateral components simultaneously. In some combinations, the structural maneuvering loads of the V-tail design may be significantly greater than the loads sustained on a conventional tail design.

A conventional empennage design isolates the vertical and horizontal components of flight loads to the respective vertical and horizontal surfaces of the empennage. A conventional horizontal stabilizer would be loaded by the increase in AOA and decrease in aft stick. Yaw and pedal inputs would not add to the loads on the horizontal stabilizer.

A V-tail design does not isolate the loading to a single axis. While vertical loading is equally distributed to both surfaces, similar to a conventional empennage, the lateral loading will be additive to one surface, and subtractive from the other2.

The following buildup presents a simplified example of loading to a V-tail structure while maneuvering. As can be seen, increasing the angle of attack (AOA) will increase the up load on the stabilizers. The up deflection of the ruddervators necessary to increase the AOA would decrease the up loading on the stabilizers. A yaw to the left would decrease the up loading on the left stabilizer and increase the loading on the right stabilizer. Pushing the right pedal would further lower the up loading on the left stabilizer and further increase the up loading on the right stabilizer. The step inputs incorporated into the calculations result in slightly higher than actual values since the AOA and yaw angles would be responding immediately to the start of the movement of the flight controls.

The build-up below follows a simple maneuver – pull the stick aft to increase the AOA, yaw left, followed by moving the stick forward and depressing the right pedal. In this case, the yaw and pedal input approximately double the load to the right stabilizer.

And the following is my 100% my own humble opinion.

There was brief mention of some deformation in the wings that could have potentially been caused by high G loading (both negative and positive) but no mention as to if this may have been caused by the aircraft tumbling "end over end". There were several builder shortcomings documented, but none that were identified as a "smoking gun" to what caused the failure. There was mention in the toxicology report that the pilot had existing heart conditions, but there is no definitive evidence that he became incapacitated in flight.

Because of this, it does look to me like there was the possibility of a weak point in the design. Looking at the failure mode, and looking at the service bulletin, it appears to me that Sonex Aircraft has done the right thing and fixed it in an appropriate way. This happens in aviation, there is no perfect design out there.. but as I approach flight test in my own Waiex in the next few months, I am confident in the design as it stands with the service bulletin and can finally put this matter out of my mind.

And for one off topic observation.. look at this fuel tank. After seeing the wreckage and knowing what the tank went through, I feel pretty good about not experiencing a rupture during a survivable crash.




But most mystifying to me is that the tank was found bone dry, engine clearly stopped (based on eye witness reports and propeller damage). The pilot clearly ran out of gas.. but how did that contribute to this event?

Makes you wonder if some movement in the tail might have been detecable prior to the flight..?
Grabbing the tail tips and moving them up/down/forward/back to check for movement/noises at the root would seem like a good idea for any of the fleet?
Seeing the tank intact demonstrates the poly tank is pretty strong!

Richo

I can only speculate but the pilot was a fairly low time sport pilot, under 60 hours. Running out of gas with low flight experience could have contributed to overstressing the airframe in a panic mode.

Very troublesome. I was expecting he hit a large bird. Clearly not the case. But this really makes me glad I did the SB on the tail before I few.

This was only the factual information release. There will be a probable cause release soon.

I suspect it will be a root cause of running out of fuel with contributing factors of pilot failing to maintain control and poor construction techniques.

Soon after the accident, I was told by an inspector that the attatch bolts had "pulled through" and that was obviously not accurate. I do wonder if there is not more to the story. I think it's odd that there was no fuel and I still can't imagine the failure occuring under normal or even aerobatic maneuvering. I think it's worthy to note that after takeoff the pilot immediatley climbed to an altitude that is rather high for local sightseeing. Perhaps he was planning for more than straight and level flight. I did not know Wooten to be an aerobatic type, though.

While I'm glad to see the report finally complete I am saddened but not shocked of the quality of construction of Wooten's airplane by the original builder. He deserved better.

Interesting report. I had always assumed that the NTSB did not do detailed reports like that for experimentals.

One thing that still sticks in my mind a bit is what about the conventional tail Sonex? The report seems to indicate that the break up started with the forward attach point. It is pretty clear in the photos that the piece of angle holding the forward portion of the tail on came appart right at the radius of the angle. The Sonex uses a very similar piece of angle to hold the front of the horizontal stabilizer down. It looks like they issued a service bulletin that beefed up the angle in that area by using a piece of solid block but they didn't do the same for the Sonex.

Keith
#554

The report also indicates that there are a lot more loads on a V tail than a conventional tail in various conditions.
Even though there were indications of poor workmanship, it doesn't appear the failure was due to any of that. the angle simply tore apart at the radius. Hence the SB to install solid blocks at that point. The new configuration greatly strengthens that area.

If the failure was caused by an overspeed situation, that isn't supported by the GPS record. At least horizontal. I suppose he could have created an overspeed situation in a steep dive keeping the ground speed within normal operating ranges but exceeding VNE. Come to think of it, that's the only way an Aerovee powered Sonex can exceed VNE.