Helicopter down outside Leicester City Football Club
skadi
Join Date: Apr 2010
Location: U.K.
Posts: 33
Likes: 0
Received 0 Likes
on
0 Posts
Join Date: Apr 2010
Location: U.K.
Posts: 33
Likes: 0
Received 0 Likes
on
0 Posts
If the TR went into a neutral position, why did the rate of rotation speed up rather than slow down? There is no sign of a correction in the video, only a deterioration. The speed of rotation gains momentum which suggests (to me anyway) the yaw was fully extended with no way of it resetting/correcting. As it gained momentum it made the spin even more violent.
Are you suggesting drive to the tail rotor failed so the TR was no longer being driven or along the same lines of it being a control issue?
Are you suggesting drive to the tail rotor failed so the TR was no longer being driven or along the same lines of it being a control issue?
I think there are two different "neutral" concepts in play here now. Neutral can mean no TR blade pitch, which should be roughly the same as a TR drive failure, but neutral can also mean "provide enough torque to counter the torque from the main rotor".
I would think that if left to "adjust itself", the first kind of neutral would be most likely, while the second kind of neutral would be what one would want as a safety fallback for control failure.
I would think that if left to "adjust itself", the first kind of neutral would be most likely, while the second kind of neutral would be what one would want as a safety fallback for control failure.
Avoid imitations
Join Date: Nov 2000
Location: Wandering the FIR and cyberspace often at highly unsociable times
Posts: 14,573
Received 413 Likes
on
218 Posts
If a tail rotor control servo runs away to full negative pitch (some most definitely can do this and I’ve written about this before in previous threads) it will give a situation worse than a failed tail rotor driveshaft. The aircraft will continue to yaw away from the main rotor direction despite all main rotor torque being removed by fully lowering the collective and shutting down the engines. The pilot would be unable to correct this. A horrible situation.
Mitchaa wrote "The speed of rotation gains momentum which suggests (to me anyway) the yaw was fully extended with no way of it resetting/correcting. As it gained momentum it made the spin even more violent."
My speculation from experience of other designs; if the T/R control became disconnected, the blades would revert to a pre determined position by design. This position is a balance between aerodynamic and centrifugal turning moments which are usually designed to apply some positive pitch to the blades to allow a running landing. In a vertical climb OGE, this pitch setting would not be sufficient to prevent yaw developing (quite rapidly).
Now to address Mitchaa's observation of the rate of turn accelerating; as the yaw rate develops, the effective NR reduces, and more collective is required to maintain height, the torque reaction increases and yaw rate increases further, requiring increased collective pitch due to an even greater reduction in effective NR. It's a viscous circle. This would apply in the first few seconds before the descent was initiated. After descent is initiated, this theory falls down as the yaw rate would reduce noticeably and the video didn't really show this.
If the yaw servo did indeed travel to full deflection due to lack of feedback loop, then the rate of deflection, and consequent yaw, would be a function of how much the servo control valve was open at the time of failure. This could explain the relatively progressive increase in yaw rate initially.
If it was indeed a servo 'runaway', then deselection of the appropriate hydraulic channel should allow the blades to revert to the position mentioned above and permit some reduction in yaw rate and recovery of control.
I now refer you to a post I made https://www.pprune.org/rotorheads/528810-category-takeoff-background.html#post10298782
There I tried to explain that tail rotors don't tend tofail 'quietly'. However, servo runaways are 'quiet' and should not be mis-diagnosed as a t/r drive failure, despite similar yaw rates.
JJ
My speculation from experience of other designs; if the T/R control became disconnected, the blades would revert to a pre determined position by design. This position is a balance between aerodynamic and centrifugal turning moments which are usually designed to apply some positive pitch to the blades to allow a running landing. In a vertical climb OGE, this pitch setting would not be sufficient to prevent yaw developing (quite rapidly).
Now to address Mitchaa's observation of the rate of turn accelerating; as the yaw rate develops, the effective NR reduces, and more collective is required to maintain height, the torque reaction increases and yaw rate increases further, requiring increased collective pitch due to an even greater reduction in effective NR. It's a viscous circle. This would apply in the first few seconds before the descent was initiated. After descent is initiated, this theory falls down as the yaw rate would reduce noticeably and the video didn't really show this.
If the yaw servo did indeed travel to full deflection due to lack of feedback loop, then the rate of deflection, and consequent yaw, would be a function of how much the servo control valve was open at the time of failure. This could explain the relatively progressive increase in yaw rate initially.
If it was indeed a servo 'runaway', then deselection of the appropriate hydraulic channel should allow the blades to revert to the position mentioned above and permit some reduction in yaw rate and recovery of control.
I now refer you to a post I made https://www.pprune.org/rotorheads/528810-category-takeoff-background.html#post10298782
There I tried to explain that tail rotors don't tend tofail 'quietly'. However, servo runaways are 'quiet' and should not be mis-diagnosed as a t/r drive failure, despite similar yaw rates.
JJ
Last edited by jellycopter; 6th Nov 2018 at 15:13.
Trouble is JJ , the SB isnt really relevant to a servo runaway. Its relevant to a structural failure of the parts concerned or even more explicit of a missing split pin / nut/ wirelocking procedure. I hope they are also looking at security at Fairoaks.
Join Date: Apr 2009
Location: Hotel Gypsy
Posts: 2,821
Likes: 0
Received 0 Likes
on
0 Posts
Thoughts? A $2 nut (or thereabouts before being associated with an EASA Form 1), mitigated by a piece of locking wire, presents the potential of a catastrophic single point of failure.
Join Date: Apr 2010
Location: U.K.
Posts: 33
Likes: 0
Received 0 Likes
on
0 Posts
This is the case for many components, hence two locking methods, duplicate inspections and daily visuals.
Any MR Servo Mount bolt. Any MR Blade Bolt. Any TR Blade Bolt. The "Jesus" nut, to name but a few. There are certification limits for the probability of failure of these critical items, which has been mentioned here already.
This nut is double locked. A dirty fat Split/Cotter Pin and lockwire. No single locking method failure should induce a failure of the control. Both locking systems have to fail. And with the nut being upstream of the servo assist it does not have flight loads on it only the same loading that the pilot feels on his/her feet.
This "nut on the end of the servo" system has flown for years and years on the 139 and 189. Although the 139 nut is slightly different. The 189 also has an SB out for inspection of the nut (out this morning).
I anxiously await the next update from the AAIB.
Join Date: May 2010
Location: Boston
Age: 73
Posts: 443
Likes: 0
Received 0 Likes
on
0 Posts
Hopefully someone with knowledge of the system can comment on this.
Not clear that this is true, looking at the SB wording and diagram only (and with no knowledge of the system) it refers to the 'Servo feedback link". If it is indeed inside the control loop disconnecting it could very much result in a servo runaway. That said it would also seem to be a critical path and might well have a redundant path of some sort beyond what is shown.
Hopefully someone with knowledge of the system can comment on this.
Hopefully someone with knowledge of the system can comment on this.
Standard method of controlling servos.
Join Date: May 2010
Location: Boston
Age: 73
Posts: 443
Likes: 0
Received 0 Likes
on
0 Posts
The lever has three attachment points. The far end is where the control rod from the pedals connects. In the middle is the linkage that operates the pilot valve that then moves the servo. The end that is being inspected is connected to the servo itself. When this moves it resets the pilot valve to stop servo movement.
Standard method of controlling servos.
Standard method of controlling servos.
So Is this roughly correct?
If the control path disconnects the system could be rigged to settle on the 'aerodynamic neutral' mentioned above.
If the feedback path fails then the servo will likely continue to be driven to a stop since there is no (or at least incorrect) input at the pilot valve.
Last edited by MurphyWasRight; 6th Nov 2018 at 17:02. Reason: typo
Join Date: Mar 2008
Location: England... what's left of it...
Posts: 161
Likes: 0
Received 0 Likes
on
0 Posts
"3.2 Check the connection elements of the input lever (110) of the TR servo-actuator
taking particular care on the lever feedback link."
Was that photograph issued with the SB? If so, was there another photo? The position referred to in the above text is not shown, the way I'm interpreting (or trying to). Also, should the word after "(110)" be to the servo-actuator, not of? They also call component #110 both "Input lever" and "Lever feedback link". I'm thinking the intention of SB item 3.2 is to order a check of the connection to the servo from part 110, which can't be seen in the posted photo.
taking particular care on the lever feedback link."
Was that photograph issued with the SB? If so, was there another photo? The position referred to in the above text is not shown, the way I'm interpreting (or trying to). Also, should the word after "(110)" be to the servo-actuator, not of? They also call component #110 both "Input lever" and "Lever feedback link". I'm thinking the intention of SB item 3.2 is to order a check of the connection to the servo from part 110, which can't be seen in the posted photo.
Join Date: Apr 2009
Location: Hotel Gypsy
Posts: 2,821
Likes: 0
Received 0 Likes
on
0 Posts
Silly old me, I thought we were in the business of reducing/mitigating risk and not just accepting the status quo. Is a split pin and a locking wire really the best way? Clearly someone has recognised the importance of this nut. Perhaps the whole linkage could be redesigned.
Join Date: Mar 2008
Location: England... what's left of it...
Posts: 161
Likes: 0
Received 0 Likes
on
0 Posts
At some point, Mr. Cows, an assembly or artefact has to terminate. There are only so many ways to connect things together in a manner that can be dissembled and rebuilt. There is also only so much material and weight that can be used and included too. I think all methods have been tried by now!
If a split pin and a lockwired nut on an assembly of high quality design and material fail, on the back of proper maintenance and inspection, then one really is subject to the worst of probability.
If a split pin and a lockwired nut on an assembly of high quality design and material fail, on the back of proper maintenance and inspection, then one really is subject to the worst of probability.
Join Date: May 2010
Location: Boston
Age: 73
Posts: 443
Likes: 0
Received 0 Likes
on
0 Posts
I would think it would be nigh on impossible to have a failure of the pin and the wire locking considering as someone else has mentioned, there’s no loading there, they are simply to lock the nut to the shaft.
The release of the SB most probably points to neither the pin or the locking wire being present. The big question, were they ever fitted and locked and if so, how did 2 sets of eyes during a duplicate inspection miss it? Not impossible, but certainly unusual.
In other words they saw something unexpected that warranted inspections such as incorrect assembly order but not by itself the root cause.
Would a hard over servo runaway result in the reported (by some only) grinding noises?
It is also possible that the SB resulted from an observation of condition of the recovered assembly. By itself does not 'prove' that this is the cause of the accident.
In other words they saw something unexpected that warranted inspections such as incorrect assembly order but not by itself the root cause.
In other words they saw something unexpected that warranted inspections such as incorrect assembly order but not by itself the root cause.
