AW139 lost tail taxying DOH

[heliski22]

Noooby - did you mean P/N ..00234 or ..00134

Just been having a closer look at BT139-159. It does seem like an awful lot of bits and man hours for an Optional BT?

What is a little worrying, in light of this Middle East incident, however, is the question of just why would we want to "improve" (that's the word in the BT) our tail boom if there's nothing wrong with the one we've got, eh?

Answers on a postcard, etc., etc.....................

[TheOldTimer]

After a little time and I expect all operators of the 139 looking at the tail boom with large magnifying glasses and tapping like mad looking for any sign of a problem, I was wondering what the results are and the location of any failings??? A friend tells me the latest high tech tool is a little hammer rather than the coin?

[blakmax]

What is that thing?

OK, I'm not an Augusta expert (I'm not even a helicopter expert.) but what is that great phalic symbol attached to the separated residual boom and pointing at the sky in the recent pictures. It looks like some kind of strake which may have been bonded to the skin of the boom and failed as part of the event. I make that judgement on the basis that I can not see any fastener holes and also that anything that length would have bent if even a few fasteners were effective. A loss of such a length of stiffening structure would have made a very significant difference to the ability of the structure to survive any loads (wheel locked or not, and even if the pilot was wearing spotted underwear on a Monday afternoon two days after the full moon).

If it really was bonded and if the owners of this or other aircraft with similar failures can examine the surface to see if it has failed interfacially (areas with no adhesive on one surface) then there may be grounds for warranty action.

If this structure was not bonded, tell me and I will get back in my box.

It really is time that operators stopped accepting glib "you did xyz wrong" responses from manufacturers when expert assessment can demonstrate that the operator was not responsible for a failure. IF THE ADHESIVE HAS FAILED TO STICK, IT IS A MANUFACTURER'S PROCESSING PROBLEM. IF THE MANUFACTURER COMPLIED WITH THE REGULATIONS THEN IT IS A REGULATORY PROBLEM.

As I have already stated in this stream and others, curent FARs do not require assessment of bond durability other than static and fatigue testing, neither of which will prevent interfacial failure which is time dependent. Just do a search for "blakmax" postings.

This applies for other manufacturers, not just Augusta, as well as for non-rotary structures. If you overload the structure because the pilot forgot the brake or sneezed at the wrong time then the failure will separate the adhesive with residual adhesive on both surfaces at any location. That probably is a design or operational issue which could be debated. But interfacial failures are directly attributable to the manufacturer's process selection or the regulator's failure to exclude an important and common adhesive bond failure mode. There are very few things an operator can do which will result in interfacial failure of an adhesive bond unless the interface was already degraded due to the processes used by the manufacturer which probably met the regulator's requirements.

Perhaps action against the manufacturer to recover lost income due to reduced aircraft availability as well as cost of part replacement under warranty will actually address bond failure issues. Hopefully it will not be legal action on behalf of dependents after a loss of life incident. And before anyone asks, no, I am not a lawyer. I am an expert in composite and adhesive bonding technologies who wants to see best practice adopted by regulators and manufacturers so that we do not see failure of composite and bonded structures.

It is actually possible to design adhesive bonded joints such that they NEVER fail under any loads. That must be combined with validation of bonding processes that assure interfacial failure will never occur. So, if it is possible to eliminate bond failure by design methods, combined with validated processes which prevent interfacial failure then why do we continue to see bond failures? Because the current regulatory requirements for bond validation and common design practices do not preclude either failure mechanism.

After all, if temperature has an effect on durability of your bonded structures, then global warming is not your friend.

Regards

Blakmax

[heliski22]

blakmax

The "phallic symbol" you're looking at is a strake that is an aerodynamic add-on. It is fixed with fasteners (rivets) rather than bonded in any way. I can't remember if there's a sealant of any kind added to the join or if it's just a filler that is painted over. If you look at the second picture down in the group of photos on page 5, you'll see the holes where it was torn from the boom itself during the failure.

[sycamore]

212 man,I don`t think that it is necessarily suggested that the failure on this aircraft was as a nosewheel locked case; however,it would be interesting to know from `those who do` if the nosewheel is linked to the yaw pedals on the ground for ground manoeuvring or not at all. The nosewheel assembly looks to be directly attached to the oleo with very little `trail`,ie no trailing link, therefore all castoring has to come from the tail,or/in addition to any `inside` wheel braking. So all the yaw loads on the ground ,which will also include a rotary twisting torque,from the tail-rotor above the tail-boom will be felt in the `weakest` part of the boom.
Blakmax, the strake along the fuselage is for aerodynamic purposes to reduce yaw pedal transient reversals during sideways/crosswind hovering, by causing a loss/breakaway of the airflow around `deep`/heavily curved tailbooms due to the `Coanda` effect( teaspoon in running tapwater ),reducing sideloads in the low-speed areas.Developed in late `70`s by WHL as the Seakings got heavier,and pedal reversals at high torques moving on and off ships decks were becoming limiting. Early `breadboard` models were a piece of angle-iron,bolted on,cut to length as tests progressed,later production mods were pieces of angle-alloy bolted on....Works though !!

[ShyTorque]

however,it would be interesting to know from `those who do` if the nosewheel is linked to the yaw pedals on the ground for ground manoeuvring or not at all.

In thirty years of flying, mainly rotary wing, I don't know of any helicopter where the nosewheel is linked to the yaw pedals, unlike some fixed wing aircraft. The locking pin is designed to lock the nosewheel fore and aft i.e. straight ahead.

I can't think of a situation where having yaw pedals linked to the nosewheel would be advantageous (or to the tailwheel, where fitted). I can think of some where it would possibly make the aircraft uncontrollable, for example a slow rolling takeoff in a crosswind.

Larger aircraft, such as the CH-53 and the Chinook do have steerable wheels. For example, the CH-53 nosewheel, and one of the two Chinook "tailwheels" are steerable (the other side castoring). However, this isn't done via the yaw pedals but my other mechanical means, the pilot's control being a left/right control knob on the console.

[chopjock]

Ok I've just seen the pics on p5. Goodness me It appears to me than heat from the exhaust was blown down with rotor downdraft on to the boom, combined with little or no airspeed (no cooling air) may well have softened the epoxy laminate enough to weaken the structure. Has anyone stood behind an exhaust on a Hughes 500 when the engine is idling? It's jolly well hot!
Are there any pics of this ship before the accident? if so look for heat / soot exhaust stains as a tell tale sign!

Epoxy (depending on type) tends to weaken above a temp of 150 ~ 180 deg C. I'm sure a turbine exhaust at 2m distance will be hotter than that.

I assume It's alright to guess what went wrong because no one was hurt.

[Um... lifting...]

The nose wheel steering system gives the pilot control of the aircraft
during ground movement. After take-off, centering mechanisms will
automatically align the wheels in a fore and aft direction for retraction
in the NLG bay.
For high speed rolling, the nose wheels need to be held directionally
aligned. The nose wheel center-lock is controlled by a switch on the
Landing Gear Control Panel (LGCP). Pressing the switch once
engages the center-lock, locking the nose wheels in a directionally
aligned fore and aft position. Pressing the switch a second time will
disengage the center-lock, leaving the nose wheels free for ground
maneuvering.

While bearing the weight of the aircraft, the nosegear (so I have been told by the engineers) can caster 360 degrees. The nosegear has a small amount of trail. It is not directly steerable, but is not difficult to get to turn from a dead stop with moderate tail rotor pedal input, with or without differential braking, unless it's locked.
There is a centering cam mechanism with automatic locking pin that engages when weight is removed from the weight on wheels switch. What the above quote from the RFM doesn't exactly spell out for you is that in practice you can only disengage the lock on the ground.
The locking pin is very strong. If not unlocked before towing (the external warning flag is not readily apparent), the locking pin is not the first point of failure in the gear assembly. The nose gear is designed to keep the entire aircraft aligned with the runway during a 60-knot running landing (it also tends to shimmy at moderately high taxi speeds).
At 6400kg, 1400kg of the aircraft's weight is on the twin nose gear, and 5000kg is divided between the two single main gear. Nose gear tire pressure is 137 psi. Main tire pressure is almost double that, as is the contact area of the main tires. The aircraft is supposed to be taxied on paved surfaces with the nosewheel unlocked no faster than 20 knots and no faster than 10 knots on grass surfaces. With the nosewheel locked, those speeds double. For emergencies, you can add 20 knots to each of those for maximum touchdown speeds of 60 and 40.

[noooby]

Blakmax and heliski, the tailboom strake is secured with screws to the tailboom, not rivets. The screws locate into inserts bonded into the honeycomb. Material is not very thick at the edges, so would probably tear sideways and leave the screws behind in this type of incident.
heliski, sorry if I was getting my tailboom numbers mixed up. the 00134 tailboom assy, has the 00234 tailboom as part of that assy. The 00235 tailboom is part of a different P/N tailboom assy. It can get confusing!
Um...lifting is correct. The nose gear can rotate through 360 degrees on the ground. Tunr any sharper than the towbar limit lines on the aircraft, and you may pivot the aircraft on one tyre, not good for the tyre and sure to leave rubber behind!

[heliski22]

Ok, Noooby, so we have the 234 - this does not give me a nice warm feeling just now, despite nearly 400 hours on the bird and the recent completion of BT139-134.

By the way, does anybody know if this aircraft was still operating at 6400kg MTOW or had it had the 6800kg mods? Just a thought?

[9Aplus]

Epoxy (depending on type) tends to weaken above a temp of 150 ~ 180 deg C

Epoxy weakening begins at 82-90 deg C on 150 all is down to plastic features....

[spinwing]

Mmmmm ....


... By the way, does anybody know if this aircraft was still operating at 6400kg MTOW or had it had the 6800kg mods?

Just a guess ... but from the pics above the aircraft still seems to be sitting quite low to the tarmac ... I think therefore that it was still in 6400Kg config.

I have noticed that after the 6800Kg mods they tend to sit higher up on the main gear ... prob due to the extra prx in the u/c struts and especially in this case with the boom nicely folded up against the airframe ....

[amicus]

Composites inspection and repair

To Rigga, blakmax and all,
All thanks for your input and I am a composites engineer with a few decades in composite experience. I have found most QA inspection techniques in all composites and adhesives to be sadly lacking, e. g. you can check that there is a bond, but that tells you nothing re strength of the bond.
My experience led me to invoking proof loading to limit load, (if you are initially worried re incurring damage, then proof load five times to limit load on early proof tests to prove no damage), as most reliable way as it solves the issues re both poor Q.A. or poor in-process control and is a lot cheaper than a bunch of questionable Q.A. techniques. If temperature levels are issue do proof test at max certified operating temperature via heat blankets.
I would follow same mandatory proof load procedure after any tail boom repair also, of course. Either it works or it doesn't with no doubts or questions, that is a benefit of proof loading and it catches lots of in-process faults that Q.A, doesn't catch. And given the tail boom and its problems, I would invoke 100% proof loading requirement and have a cheap permanent rig built for fast testing, no instrumentation needed, just a load vs. deflection accept./reject curve to measure.
Anyhow that it what i would insist on if I were at Augusta as clearly their Q.A. isn't catching defects and their in-process controls are questionable too.

[JOURNIOO4]

AW139 torsim and bending coupling

I think that the designer think that to avoid torsion and bending coupling in the fin root the tail gearbox has been located on the top of the fin and the transmition shaft is inside rather than outside the structure, in this way the shape continuity between the fin aerodynamic profile and the tail cone cannot prevent stress concentrations in the fin root section which result fatigue problems this is what I think happen in AW 139

[blakmax]

Proof testing

Thanks everyone for explaining what that phalic symbol was and how it was attached.

For Amicus, proof testing will certainly help improve detection of production defects and is actually mandated in the FARs for bonded and composite structures where NDI is unreliable. However, even proof testing will not prevent in-service degradation of bond interfaces. Similarly, any QA strength tests on coupons will not prevent degradation of bond interfaces. Strength and proof tests are only a snap-shot of bond strength at the time of testing. Interfacial degradation is time dependent, with moisture and temperature accelerating that degradation. If you test before the interface has had time to hydrate, then you get a false positive result.

The secret to success is to use tests that actually interrogate the resistance of the interface to hydration BEFORE construction. You simply can not test for this after construction. Demonstrate by wedge testing that the proces you are using can actually develop a hydration resistant interfacial bond, and then duplicate that process in production and you will have a structure which will not fail interfacially in later service. The test is ASTM D3762, but ignore the stated acceptance criteria in that ASTM. The correct critieria are in DOT/FAA/AR – TN06/07, Apr 2007 which can be obtained by email from the FAA Tech Center Library. US nationals can download the document, but because it is a .gov website, we foreigners can not access it for homeland security reasons. Too hard? send me a PM and I'll send a copy to you.

Regards

Blakmax

[Skin King]

Having been directly invovled now with at least 15 AW 139 Tailboom disbonds I would like now to share my observations. Please remember I'm not a composite engineer but simply the poor guy with the tap hammer who has to make the call as to whether or not the Tailboom Assy is serviceable.
Firstly, all but one of the delaminations I have seen started on the curve section of the R/H side of the Tailboom structure aft of where the top longeron finishes. The longerons are a hat section sheetmetal part that are rivetted into a recess on the internal side of the bonded panel and are approximately 28 inches in length. There are 3 longerons per side. Cut outs of the delaminated areas show that where the longeron finishes there is a 1 inch internal doubler strip between the outside skin and the Nomex core that I assume is suppose to transfer loads from all 3 longerons to the bonded panel they are rivetted to. As I said before I am not a composite engineer but this 1 inch internal doubler strip appears a little insufficent to carry the loads between the longerons and the bonded panel.
I can liken the 139 Tailboom build as to having a VW swinging off a pool que which is attached to a tree stump by masking tape.
Cut out samples have shown that in some cases that there is a lack of adhesive to the outside skin (aluminiun skin is still visible in parts) while in other cases failure of the core to adhere to the outside skin i.e there is adhesive on the outside skin and not the Nomex core.
No evidence of obvious water contamination to the core was visible in any samples taken.
Teletemps where placed in various locations on the Tailboom Assy and the average skin temps while on the ground where around 80 degrees C. Yes it's hot here. Temps whilst in flight where unconclusive so I would not like to speculate.
Ok having said all, it is my personal belief that there is a combination of adverse factors at work here the worst being the sudden transition of loads through the longerons to the Tailboom bonded structure. Couple this with poor bonding processes and constant heat soaks then disbonds are inevitable.
My advise to all engineers on 139's is to visually inspect the area Aft of the R/H upper longeron for skin ripples or waviness on your daily inspections and after each flight. But I guess the majority of us know that already.
I do hope Agusta comes up with a fix soon as I'm really getting bored with changing Booms.

[blakmax]

Loads and disbonds

Thanks for the details skin king. However, let me assure you that there is no way structural loads of any sort can cause interfacial failures i.e. where there is an absence of adhesive on one adherend. Of importance, was there an impression of the core in the adhesive sample taken from over the disbond where the adhesive was absent from the core? If there is no impression, then that was not in contact at manufacture. If there is an impression, then it is probably inadequate drying of the core before bonding.

The disbond from the aluminium is more of a problem. If the surface of the adhesive was glossy, then there is probably another processing issue such as inadequate pressure application during bond or poor fit-up of the parts. If the surface of the adhesive replicates any features of the aluminium it was probably in contact but has disbonded due to hydration of the interface. The regulators need to ask Augusta what the process was, and not accept the usual "proprietary" crap. Lives are at stake and the regulator should step up and take control.

I certainly hope that the booms you are fitting are being replaced uner warranty because there are very few things an operator can do which will cause interfacial failure. To address the temperature theory, heat applied to adhesives reduces the strength of the bond but also increases plastic behaviour. Such failures should exhibit elongation marking on the bond, not interfacial failures.

[Skin King]

Thanks Blakmax for the reassurance about loads and disbonds. As I mention before the samples taken showed two types of failures.

1) There is adhesive and core impression on the outside skin. No adhesive on core cells.

2) There is adhesive on the core cells and only random areas or no adhesive on the outside skin. There is just a slight core impression in the areas of random adhesive.

It appeared to me that there maybe should have been an additional layer of unsupported film adhesive applied during layup particularly in the curved areas.

Anyway I'll leave it to the composite experts at Agusta to sort out their process woes.

And YES the replacement booms are under warranty even though some only last about 2 days before they disbonded as well.

The 139 is a good A/C and as with all new A/C has some growing pains. Luckily for us history dictates these pains get sorted out fairly quickly and I'm sure we will see a new improved Tailboom Assy sooner than later.

[blakmax]

The next step

Thanks skin king, your description basically confirms what I said. Maybe Agusta could do with a consultant expert to help sort out the problem. I'm not cheap, but neither is replacing tail booms on a regular basis. I'd be glad to help if they PM me.

Regards

Blakmax

[Skin King]

As mentioned in a previous post Agusta Optional BT 139-159 which replaces the Nomex core lateral panels of the Tailboom with aluminium core is available but before everyone gets too excited it does require specialised tooling and an approved jig of which there is only ONE..... Yep!! Italy.

It's good to see that Agusta are well aware of the problems and lets hope that along with a change in core material came a change in the manufacturing process of these panels. Time will tell but as we are approaching our cooler months we may have to wait until next year to find out.