Atlas Air 767 down/Texas

[Machinbird]

The pinger is normally removed prior to shipment according to NTSB CVR guidelines
Thanks Airbubba. That was a useful reference.
Guess I've seen pictures of too many CVRs being recovered from other countrys' waters.:(

[Chu Chu]

Presumably, the more rigidly the pinger is attached, the more likely it is to sustain damage -- or to damage the crash-protected part of the recorder itself. A detached pinger would still help locate the wreckage in general (not that they seem to have been too effective in open-ocean crashes recently . . .). So I guess there's a trade off.

[GlueBall]

Conceivably, the ULB (pinger) is removed to preclude activation during shipment of the recovered FDR/CVR.

[DaveReidUK]

Presumably, the more rigidly the pinger is attached, the more likely it is to sustain damage -- or to damage the crash-protected part of the recorder itself. A detached pinger would still help locate the wreckage in general (not that they seem to have been too effective in open-ocean crashes recently . . .). So I guess there's a trade off.

No trade-off required.

It's perfectly possible for a pinger to be both securely attached, so that it won't separate from the recorder in an impact, and also readily removable once its job is done and the unit has been recovered (note the 4 Allen screws):

https://cimg9.ibsrv.net/gimg/pprune.org-vbulletin/840x470/flight_data_recorder_evolution_where_next_1cbc8c43929e3a75e3 ae1ecc8bfd1967a50bf509.jpg
The pinger is, of course, externally mounted so that its batteries can be changed without having to open up the recorder itself.

[Airbubba]

Remains of all three crewmembers have now been officially identified.

From the CCSO Facebook page:

Sheriff Brian Hawthorne reports that at 6:05 p.m. today he received notification confirming the identity of Captain Ricky Blakely, 60, recovered in Trinity Bay from the crash of flight 3591.

Chambers County Sheriff’s Office extends their heartfelt condolences to his family, friends and coworkers.

[Machinbird]

The pinger is, of course, externally mounted so that its batteries can be changed without having to open up the recorder itself.
That is one of the advantages of course.
Ultimately, whatever mount used has to allow the pinger/ULD to couple acoustic energy effectively into the water.

I suppose that the CVR/FDR recorder manufacturers have tested the mounting technique under dynamic conditions, however the use of what visually appears to be 1/4 inch or 5-6 mm Allen head screws seems weak should the mount receives a blow in the direction of the pinger axis.
I have sheared plenty of 1/4 inch hardware in my life, and going up at least one hardware size would significantly improve strength of the mounting.

One of the AF447 pingers was knocked off its mounts. Neither was detected. That was not a very high energy impact as crash energies go.
Lessons learned from the AF447 search (https://www.hydro-international.com/content/article/lessons-learned-from-the-af447-search)
What we discovered is that you are dealing with a lot of uncertain information in such a search operation. You cannot count on the pingers: after a crash it is always unclear if they are going to work. They towed the pinger locators right over the wreck shortly after the crash, and they didn’t hear anything. When the black boxes were found, only one pinger was still attached. The other one was never found.

[tubby linton]

FDR now found according to NTSB.

[Airbubba]

From @NTSB_Newsroom:


https://cimg7.ibsrv.net/gimg/pprune.org-vbulletin/1435x950/atlas_fdr_1_704f5dfcd28b09227dd324a087024cbf116c2a8c.jpg
https://cimg6.ibsrv.net/gimg/pprune.org-vbulletin/1432x946/atlas_fdr_2_c445756431796066f9f920e2d90536f9c1e13d26.jpg

[Machinbird]

Fresh from the accident scene, muddy, without a pinger attached.
Somehow, I don't think they were keeping an Allen wrench handy to immediately remove the pinger as soon as the unit was discovered.

[DaveReidUK]

Fresh from the accident scene, muddy, without a pinger attached.
Somehow, I don't think they were keeping an Allen wrench handy to immediately remove the pinger as soon as the unit was discovered.

I don't see any reason why a team looking specifically for the FDR shouldn't have the (simple) means to remove the pinger, given that that's SOP.

Maybe they didn't need it, maybe they did.

[lomapaseo]

I don't see any reason why a team looking specifically for the FDR shouldn't have the (simple) means to remove the pinger, given that that's SOP.

Maybe they didn't need it, maybe they did.

Does the pinger keep on ticking after they retrieve the CVR, maybe that's why the remove it to save their ears when they pack it in water

[Machinbird]

If they show the end of the unit that had the pinger attached, you can look to see if the sheared off tips of the bolts are still there.
Considering the relatively high pinger frequency, I doubt that people in the vicinity will find it very disturbing, particularly if they have been around jet engines any amount of time.

[Airbubba]

Does the pinger keep on ticking after they retrieve the CVR, maybe that's why the remove it to save their ears when they pack it in water

At 37.5 kHz most of us wouldn't have to worry about hearing the pinger. But, it does have a lithium battery which may be damaged.

From the NTSB FDR Handbook:

3.5. If there is any visible damage to the underwater locator beacon (ULB or “pinger”), it should be removed prior to shipping. Use caution when handling damaged beacons.

https://www.ntsb.gov/investigations/process/Documents/FDR_Handbook.pdf

[jugofpropwash]

At 37.5 kHz most of us wouldn't have to worry about hearing the pinger. But, it does have a lithium battery which may be damaged.



Wouldn't it helpful to have the pinger at a frequency where it could be easily heard? (Or quite possibly at both a high and low frequency.) Wouldn't an audible ping help divers/searchers to quickly locate the box under certain conditions?

[FIRESYSOK]

Wouldn't it helpful to have the pinger at a frequency where it could be easily heard? (Or quite possibly at both a high and low frequency.) Wouldn't an audible ping help divers/searchers to quickly locate the box under certain conditions?

Hydroacoustics are best left to dolphins and submarines. Humans can hear sound under water but determining direction and distance is not possible without help. Additionally, the higher the frequency, the more ‘directional’ it becomes.

[Airbubba]

The NTSB is urging the FAA to require 25-hour CVR recorders after incidents like the one with Air Canada at SFO in 2017 where pertinent CVR data was overwritten.

The National Transportation Safety Board (NTSB) is providing the following information to urge the Federal Aviation Administration (FAA) to take action on the safety recommendations issued in this report. These recommendations address the need to install cockpit voice recorders (CVR) with a minimum 25-hour recording capability on all newly manufactured airplanes required to have a CVR and retrofit these CVRs on existing aircraft required to have flight recorders. These recommendations are derived from the NTSB’s experiences with investigations that lacked access to relevant CVR data. Information supporting these recommendations is discussed below.

https://www.ntsb.gov/investigations/AccidentReports/Reports/ASR1804.pdf

[NSEU]

Does the pinger keep on ticking after they retrieve the CVR, maybe that's why the remove it to save their ears when they pack it in water

It stops after it has been pulled from the water. Even when I was young, I couldn't hear the ultrasonic pinging. We (maintenance engineers) had a special test box which converted the sound to audible clicks.

[gums]

Salute!
@Bubba!! PLZ keep those tweets coming as I refuse to have a tweet or face account.

Gums whines...

[Ripper3785]

I don't see any reason why a team looking specifically for the FDR shouldn't have the (simple) means to remove the pinger, given that that's SOP.

Maybe they didn't need it, maybe they did.

Getting pedantic here- As someone else mentioned, you can see the tabs are damaged, almost identically with bent and sheared tabs(for a moment I thought they were the same recorder but there is some difference in the damage to the labels) If you think it likely they removed the pinger from a bent tab(assuming it stayed connected to the one that is still attached to the recorder), well, then, ok.

Here's a picture of an undamaged one that looks more similar to what was recovered - photovault.com/data/comps/IAC/IACV01P05_17.jpg

[fdr]

Wouldn't it helpful to have the pinger at a frequency where it could be easily heard? (Or quite possibly at both a high and low frequency.) Wouldn't an audible ping help divers/searchers to quickly locate the box under certain conditions?

Nope.




...



low frequencies will travel further, however they take more energy to develop than the higher frequencies. Low frequencies are also in the main background frequencies of the ocean, below the snapping shrimp and whale song freqs​. Shipping noise has primary frequencies in the very low levels, with blade rates and shaft rates that are down in the bottom of the spectrum, (still audible to human ears though), but they have harmonics above that which can be strong, particularly when a multiple of the number of blades.Higher frequencies will attenuate faster than the lower frequencies, but they are also usually easier to analyse with fewer artifacts from the sampling rate that is used for the transform. ​​​​​​

[Airbubba]

Salute!
@Bubba!! PLZ keep those tweets coming as I refuse to have a tweet or face account.

Here's a picture of the International Brotherhood of Teamsters pilot investigators with the Chambers County Sheriff. Since the operating pilots were represented by the IBT, the union is a party to the NTSB investigation. Traditionally a union pilot typed in the accident aircraft will be part of the CVR Group that auditions the CVR recording for the official transcript.

Sheriff Hawthorne respects the professional relationship that Captain John Jester IBT Teamsters (Pilots Union) Chief Accident Investigator from Hoofddorp, Netherlands and Captain Howard Lentz Accident Investigator from Cameron Park, CA, have with the National Transportation Safety Board. The Sheriff has reached out to these gentlemen for expert advice regarding equipment and systems of the aircraft which has been instrumental in assisting the Chambers County Sheriff’s Office with our role and responsibilities in the recovery of flight 3591.


https://cimg5.ibsrv.net/gimg/pprune.org-vbulletin/1830x1899/53196633_2103529986401851_9096452523442569216_o_cafddca30c34 d16124185b1ed7ec91fee7cf4ace.jpg

The Teamsters have an unfortunate long history of close ties to what is euphemistically called 'The Mob'. I believe James Hoffa, the current IBT president, is the first head of the Teamsters since 1952 not to be indicted on felony federal charges. Not all of the predecessors went to prison, Jimmy Hoffa (James' father) disappeared, Jackie Presser died of a heart attack before reporting to Club Fed and Ron Carey was acquitted.

[Airbubba]

Update on the CVR and FDR analysis from the NTSB:

NTSB Laboratory Completes Initial Review of Cockpit Voice Recorder, Recovers Flight Data Recorder

3/5/2019​Engineers at the National Transportation Safety Board’s Office of Research and Engineering Vehicle Recorder Division (https://www.ntsb.gov/about/organization/RE/Pages/office_re.aspx) completed the initial review of the Atlas Air Flight 3591 cockpit voice recorder (https://www.ntsb.gov/news/pages/cvr_fdr.aspx) Saturday evening and recovered the airplane’s flight data recorder (https://www.ntsb.gov/news/pages/cvr_fdr.aspx) Sunday.




Three people (the two pilots for the flight and a non-revenue jump-seat pilot) died when Atlas Air Flight 3591, a Boeing 767-300 cargo jet, crashed in the muddy marshland of Trinity Bay Feb. 23, 2019, about 40 miles from Houston’s George Bush Intercontinental Airport. The airplane was destroyed. The airplane was carrying cargo for Amazon.com Inc., and the US Postal Service from Miami to Houston.The condition of the accident site made locating the recorders challenging. Directors from the Office of Research and Engineering and the Office of Aviation Safety conducted an audition of the CVR as part of the NTSB’s ongoing investigation of the accident. The audition revealed the following information, which is preliminary and subject to change as the investigation continues:The length of the recording is approximately two hours and was obtained from a download of a solid-state type cockpit voice recorder.

The recording included the final portion of the flight; however, the quality of the audio is poor.

There are times during the recording when the content of crew discussion is difficult to determine, at other times the content can be determined using advanced audio filtering.

The crew was in communication with air traffic control and were being provided radar vectors for the runway 26L approach into George Bush Intercontinental Airport.

Crew communications consistent with a loss control of the aircraft began approximately 18 seconds prior to the end of the recording.

The flight data recorder arrived at the NTSB’s Recorder Lab Sunday at 11:45 p.m. The memory module was disassembled, cleaned and dried, and download of the data was achieved Monday afternoon. Initial review of the data revealed:The accident flight was captured, and the FDR contained a total of about 54 hours of data from 17 flights.There were approximately 350 parameters recorded by the FDR detailing the motion of the aircraft and operation of its engines, flight controls and other systems.


NTSB recorder investigators are currently verifying and validating the FDR data, and the NTSB plans to provide a summary in an investigative update in a few days. Technical experts in the CVR group will convene in the coming week to review the entire recording and produce a transcript of the accident recording. It will be a time-consuming process to complete the transcript. The CVR group is one of the seven investigative groups established by the Investigator-in-Charge for the accident investigation.






In this photo, taken Saturday in the NTSB laboratory in Washington, an NTSB engineer from the Office of Research and Engineering’s Vehicle Recorder Division inspects memory boards from the cockpit voice recorder of Atlas Air Flight 3591 for signs of damage and water intrusion. Atlas Air Flight 3591 crashed Feb. 23, 2019, about 40 miles from Houston’s George Bush Intercontinental Airport, and the NTSB recovered the airplane’s CVR March 1, 2019. NTSB photo.

https://cimg3.ibsrv.net/gimg/pprune.org-vbulletin/2000x1504/33406167018_1dea71586f_k_b1b0d20f7925f0ece7024e4d6e6b198d31b b6f86.jpg

[Airbubba]

B-roll video from the NTSB of the FDR arrival at the Vehicle Recorder Lab.

https://www.youtube.com/watch?v=LC79o2_C1Gc

[Educated Airman]

@ FDR Thanks for your response.

Here is the "proposed AD

" ....in 2000"On January 31, 2000, there was an accident involving a McDonnell Douglas Model DC-9-83 (MD-83) airplane. The National Transportation Safety Board (NTSB) determined that the probable cause of this accident was a loss of airplane pitch control resulting from the in-flight failure of the acme nut threads of the jackscrew assembly of the horizontal stabilizer trim system.
The NTSB concluded that the thread failure was caused by excessive wear, resulting from insufficient lubrication of the jackscrew assembly.Start Printed Page 58621The drive mechanism of the horizontal stabilizer on McDonnell Douglas Model DC-9-83 (MD-83) airplanes has a jackscrew assembly with an acme screw.
The drive mechanism of the horizontal stabilizer on Boeing Model 767 airplanes uses a ballscrew. Acme screws and ballscrews have some differences in design, but perform similar functions and have the same airplane level effect following failure. The manufacturer's safety analysis of the 767 drive mechanism found no safety problems with the configuration of the drive mechanism, but showed that changes to the maintenance procedures and maintenance intervals are required to keep the drive mechanism properly maintained and operating as designed.
We have received a report indicating that the ballscrew in the drive mechanism of the horizontal stabilizer on a Boeing Model 757 series airplane showed extensive corrosion, which could lead to excessive wear. The ballscrew on Boeing Model 757 airplanes is similar to that on Boeing Model 767 airplanes that are the subject of this proposed AD. Therefore, both of these airplane models could have the same unsafe condition.
We are considering separate action for the Boeing Model 757 series airplanes and other similar Boeing airplanes.Extensive corrosion of the ballscrew in the drive mechanism of the horizontal stabilizer, if not corrected, could cause an undetected failure of the primary load path for the ballscrew and subsequent wear and failure of the secondary load path, which could lead to loss of control of the horizontal stabilizer and consequent loss of control of the airplane."
I am a commercial pilot and A&P and was an inspector at a major repair station with a B.S. degree in Aeronautical Science. We did freighter conversions on Boeing 727s that had been pulled out of storage. Many of the 727s had pitting corrosion on the jack screws and had to be replaced. It would be interesting to know where the Atlas aircraft had been stored as the environment of storage to determine the probability and speed with which corrosion could occur. As an inspector, I had rejected the lubrication of jackscrew work numerous times. Proper lubrication involves pumping grease into the ball nut until clean lubricant comes out, cleaning the screw, moving the jackscrew, recleaning the screw, applying oil to the screw, and re-greasing the nut. This is a timely and very messy operation if performed correctly. After the Alaska Air disaster our MRO initiated the procedure that only specific individual mechanics would perform lubrication procedures and that all lubrication work cards would be inspected by an inspector. Your theory of a jack screw failure seems the most likely scenario to me as I could see the aircraft rolling in a vertical dive on the video. As a pilot, I think that if I experienced a sever unintentional dive without elevator response, I would try to manipulate any other controls available to avoid imact. It would be interesting to know if the pilots of the Alaska Air crash migt have used roll control to avoid their demise.

[Airbubba]

Pictures of the FDR processing from the NTSB. My workbench also has a vise with a bent handle and a can of WD-40.


https://cimg5.ibsrv.net/gimg/pprune.org-vbulletin/1440x1080/47291601331_5647d1c01c_o_large__25b0d37d1bf32ea8960c38ab7a15 a3e99f7a4cf2.jpg
https://cimg6.ibsrv.net/gimg/pprune.org-vbulletin/1440x1080/47291601471_348239503f_o_large__17a202f2aab39294e02380fdd9b7 87b7794cd3c0.jpg
https://cimg4.ibsrv.net/gimg/pprune.org-vbulletin/1440x1080/33415991658_c137e3ef02_o_large__4dd9e52d1c7b22fc21e4d49dc831 d489c738c8f6.jpg

[PastTense]

Crew communications consistent with a loss control of the aircraft began approximately 18 seconds prior to the end of the recording.

Does this support or rule out any of the possible causes which have been discussed?

[Ranger One]

Does this support or rule out any of the possible causes which have been discussed?

That is an extremely neutral form of words from the NTSB, as it should be at this stage; it simply confirms that the crew were conscious and had situational awareness. I wouldn't read any more into it than that.

[lomapaseo]

Does this support or rule out any of the possible causes which have been discussed?

Since no comment about an overt sound of a warning horn was mentioned it might limit some theories

[filejw]

Since no comment about an overt sound of a warning horn was mentioned it might limit some theories

Don’t expect a comment on anything specific untill they do a full read with all teams in the investigation.

[Chu Chu]

An ultimate loss of control may not disprove the speculation about a deliberate act, but "crew communications consistent with a loss of control" seems to come pretty close.

[PJ2]

Don’t expect a comment on anything specific untill they do a full read with all teams in the investigation.

Agree, unless analysis of the data shows that there are reasons to communicate preventative safety actions on a timely basis.

[lomapaseo]

Don’t expect a comment on anything specific untill they do a full read with all teams in the investigation.

agree, but I would have expected the team members to already be there and that no comment would have been given by the NTSB unless the team members agreed.

[filejw]

agree, but I would have expected the team members to already be there and that no comment would have been given by the NTSB unless the team members agreed.

Well the press release said the the team will assemble in the coming weeks so looks like it will be awhile.

[tdracer]

The exception would be if they quickly suspect there was some sort of failure that could exist on other 767s - in which case I'd expect something out very quickly to alert the rest of the fleet of a potential safety risk. Otherwise expect the sounds of silence for a while as they work through the voice recordings and data.
Satisfying public curiosity is not high on the NTSB agenda.

[PJ2]

The exception would be if they quickly suspect there was some sort of failure that could exist on other 767s - in which case I'd expect something out very quickly to alert the rest of the fleet of a potential safety risk. Otherwise expect the sounds of silence for a while as they work through the voice recordings and data.
Satisfying public curiosity is not high on the NTSB agenda.

Precisely.

[Airbubba]

We should get a summary of the FDR data in the next few days according to the NTSB:

NTSB recorder investigators are currently verifying and validating the FDR data, and the NTSB plans to provide a summary in an investigative update in a few days.

[Matvey]

It would be interesting to know where the Atlas aircraft had been stored as the environment of storage to determine the probability and speed with which corrosion could occur

Goodyear and Marana for about 2 years. Did the freighter conversion in Singapore.

[voyageur9]

Satisfying public curiosity is not high on the NTSB agenda.

Still the NTSB provides far more information, including detailed factual sequences, far faster and with far more transparency than other similar accident investigation bodies. Compare, for instance, the foot-dragging and secrecy endemic in Canada's TSB.

[SteinarN]

I did some calculation.

Starting at 6000 feet/2000 meter, the time in free fall to ground is 20,2 sec. This aircraft took about 18 sec from 6000 feet to ground. This means the G throughout the dive, from start to end, would have to be close to zero or slightly negative.

This seems to imply that the elevator/h.stabilizer must have been at average at approximately zero degrees angle of attack thus producing zero force.

I do wonder, if the jack screw for the stabilizer breaks loose, in which position will the aerodynamic forces put the stabilizer into? Would it "free float" into zero degrees AoA?

[DaveReidUK]

I do wonder, if the jack screw for the stabilizer breaks loose, in which position will the aerodynamic forces put the stabilizer into? Would it "free float" into zero degrees AoA?

While that might be true, more or less, for a control surface that's hinged at its leading edge (say an aileron), I don't think you can make that assumption for a THS where the pivot point certainly isn't at the L/E..

[Sailvi767]

I did some calculation.

Starting at 6000 feet/2000 meter, the time in free fall to ground is 20,2 sec. This aircraft took about 18 sec from 6000 feet to ground. This means the G throughout the dive, from start to end, would have to be close to zero or slightly negative.

This seems to imply that the elevator/h.stabilizer must have been at average at approximately zero degrees angle of attack thus producing zero force.

I do wonder, if the jack screw for the stabilizer breaks loose, in which position will the aerodynamic forces put the stabilizer into? Would it "free float" into zero degrees AoA?

I don’t think it would free float. It’s not a hinged device. Having said that if the aircraft was loaded with a forward CG and the stabilizer system suffered a catastrophic failure it would have thousands of lbs of force trying to move it to a neutral airload position like the Alaska MD80. That kind of force could drive it to that position and I doubt there is enough elevator force available to override it. To prevent that Boeing has both a main and secondary brake system on the stabilizer if I recall correctly. It would be interesting to know the CG of the aircraft as loaded that day.
Another possibility is that as the aircraft slowed the stabilizer was frozen and not moving. The autopilot would compensate with up elevator until at some point the autopilot desengaged which would produce a instant negative G nose down pitch mode.
It might be very difficult under negative G to apply much if any nose up elevator via the yoke depending on seat position and how tight the lap belts were plus shoulder harness usage. In fact I think it would be nearly impossible. The catch to this theory is the system has a stab out of trim caution to alert to this type of failure and the yoke itself would have been visually aft of its normally position.
If the aircraft had recent work in stabilizer area it would be interesting to know what was done.

[SteinarN]

While that might be true, more or less, for a control surface that's hinged at its leading edge (say an aileron), I don't think you can make that assumption for a THS where the pivot point certainly isn't at the L/E..

The question then would be if the pivot point is in front of, at, or behind the center of lift? If the pivot point is in front of the center of lift then the stabilizer should self adjust to zero degrees AoA. If the pivot point is very near or at the center of lift then the stabilizer position would be unpredictable. If the pivot point is behind the center of lift then the stabilizer would deflect to one or the other extreme position, either fully up or fully down.

[B2N2]

I did some calculation.

Starting at 6000 feet/2000 meter, the time in free fall to ground is 20,2 sec. This aircraft took about 18 sec from 6000 feet to ground. This means the G throughout the dive, from start to end, would have to be close to zero or slightly negative. Incorrect

This seems to imply that the elevator/h.stabilizer must have been at average at approximately zero degrees angle of attack thus producing zero force. Incorrect

I do wonder, if the jack screw for the stabilizer breaks loose, in which position will the aerodynamic forces put the stabilizer into? Would it "free float" into zero degrees AoA?


There is some flawed thinking here.
An object released to freefall will experience a 1G (gravitational) acceleration until atmospheric friction/drag will stabilize its speed at terminal velocity.
Depending on surface area and drag coefficient this terminal velocity may be significantly less then a dive while still under engine power.
https://www.grc.nasa.gov/www/k-12/airplane/termv.html

Recovery attempt from a steep dive can cause very high G-forces on the airframe


https://cimg9.ibsrv.net/gimg/pprune.org-vbulletin/220x185/220px_g_force_svg_a11c717f0f2ca6b648bb8da73aa9bc9d062698d4.p ng
An example of this is here where they descended under power at appr. 12000fpm (!)
https://en.wikipedia.org/wiki/China_Airlines_Flight_006


The fact that the time of freefall more or less equals the time to descent is nothing but coincidence.

[FLCH]

On a less technical aspect, the jumpseat rider on flight 3591 Sean Archuleta had a class date with United Airlines.
The class was scheduled for Mar 12 a vacant chair will be in that particular class for Sean in his honor.
Sean's wife who flew up from Colombia on United was presented with Sean's wings and epaulets out of respect for him.

[B2N2]

SteinarN

Here is another example.
In this case an airshow crash of an F-18.
At 1:20 you see a high rate of descent/ high G combination

https://www.youtube.com/watch?v=keCnK4kh65U

[SteinarN]

There is some flawed thinking here.
An object released to freefall will experience a 1G acceleration until atmospheric friction/drag will stabilize its speed at terminal velocity.
Depending on surface area and drag coefficient this terminal velocity may be significantly less then a dive while still under engine power.
https://www.grc.nasa.gov/www/k-12/airplane/termv.html

Recovery attempt from a steep dive can cause very high G-forces on the airframe


https://cimg9.ibsrv.net/gimg/pprune.org-vbulletin/220x185/220px_g_force_svg_a11c717f0f2ca6b648bb8da73aa9bc9d062698d4.p ng
An example of this is here where they descended under power at appr. 12000fpm (!)
https://en.wikipedia.org/wiki/China_Airlines_Flight_006


The fact that the time of freefall more or less equals the time to descent is nothing but coincidence.

No, you are not correct here.

As I am sitting here completely at stand still on my chair in front of my computer I am experiencing exactly 1G (in the vertical direction), the same 1G as the unfortunate pilots in this flight experienced while they briefly was semi stable at 6000 feet. If a 6000 feet hole suddenly appeared under my chair I would instantly start free falling at zero G. If I was to reach the bottom of the hole in 20 sec I would need to fall at an average zero (vertical) G. The acceleration you are describing does not count as a positive 1G. An object which is free falling is by definition experiencing zero G, this is according to laws set by Albert Einstein.

[B2N2]

gravitational[/i] acceleration until atmospheric friction/drag will stabilize its speed at terminal velocity. ]

Corrected for semantics.

[pattern_is_full]

Actually, with Alaska 261, once the jackscrew cut loose from the acme nut, the THS was driven fully leading-edge up (nose down) by the aerodynamic forces. And beyond - it actually tilted leading-edge up so far and so hard that the leading edge ripped right through the motor fairing (the bulge on the tip of the MD vertical stabilizer), to 25° or more (normal limit was 2.1°). Resulting in unrecoverable nose-down aircraft pitch. In fact that aircraft performed a half-outside-loop (English Bunt) and was flying inverted on the reciprocal heading when it hit the water.

(There's a little NTSB video presentation on the failure progression here: https://en.wikipedia.org/wiki/Alaska_Airlines_Flight_261 )

767 may, of course, have different structural geometry - tdracer would probably know.

I'd rate a failed THS jackscrew mechanism in this incident as conceivable and possible, but, thus far, of unknown probability. Effects of a possible turbulence encounter on a weakened jackscrew unit, or switching from auto to manual control at that phase of flight, are suggestive, however.

[Contact Approach]

There is so much bollocks written on this forum as soon as any accident occurs.

[Sailvi767]

SteinarN

Here is another example.
In this case an airshow crash of an F-18.
At 1:20 you see a high rate of descent/ high G combination

https://www.youtube.com/watch?v=keCnK4kh65U

He actually did not have enough airspeed to generate a high G condition. He was however at 35 degrees AOA at impact.

[SteinarN]

gravitational acceleration until atmospheric friction/drag will stabilize its speed at terminal velocity. ]

Corrected for semantics.

I think we in reality agrees, but we might have a slightly different meaning regarding acceleration.

According to Einstein a sattelite in orbit is free falling around the earth in its orbit, experiencing zero G. Standing still on the face of the earth we are experiencing 1G.

What you was describing is what most people would count as an acceleration, the (vertical) speed starts at zero, then the speed increases as the object is free falling towards the earth. And thus the object must be experiencing a G of some magnitude since it is falling faster and faster. But this acceleration is only perceived and is dependent on the observer standing still and experiencing 1G. An observer free falling together with the object would see no acceleration of the object and would therefore conclude the object was not under any acceleration and thus experiencing zero G.

[B2N2]

Yes thank you.
The rest of your assumptions remain flawed though.

[Airbubba]

There is so much bollocks written on this forum as soon as any accident occurs.

According to Einstein a sattelite in orbit is free falling around the earth in its orbit, experiencing zero G.

Einstein said that? Who knew? I always thought it was Fig Newton... :)

[mustangsally]

I understand some of the pprune group needs to vent all the ideas from the comfort of there arm chairs. But, all this discussion is not helping answer the question of "what really happened." Let us all wait for the professional to produce the result.

We lost three fellow airman, may they rest in peace.

[Hotel Tango]

mustanfsally, there's nothing wrong with venting ideas as long as it comes from those sufficiently qualified to make comment (though admittedly that's not always the case). That's what PPRuNe is all about didn't you know? If it upsets you, I guess it's best just to ignore this thread.

[SteinarN]

Einstein said that? Who knew? I always thought it was Fig Newton... :)

Well, they both did. You have Newtonian gravity, and you have Einsteinian gravity. In most cases the equations yeld practically identical results, however, when the values are extreme, as in an object in close orbit around another extremely dense object, like a white dwarf, a neutron star or a black hole, the equations yeld completely different answers.

Here on earth we need to resort to Einsteinian gravity and his theory of relativity in order to get the gps system to work, among other things. If we didnt apply Einsteins equations to the gps system we would see the position of a non moving gps receiver on the surface of the earth drift a couple km each hour. This is due to the fact that the time, and hence the clocks, on the gps satelites is running ever so slightly faster than the time and clocks on earth.

[CONSO]

RE 767 ballscrew AD

https://www.federalregister.gov/documents/2005/10/07/05-20267/airworthiness-directives-boeing-model-767-airplanes

[SteinarN]

Yes thank you.
The rest of your assumptions remain flawed though.

Just a honest question.
What do you calculate the average vertical G to be in order to start at zero vertical speed at 6000 feet and then reach the ground in 18sec?

[Herod]

this is according to laws set by Albert Einstein.

No. The law was DISCOVERED by Einstein. It was always there.

[FCeng84]

Response of the stabilizer to a failure of the jack screw or its nut and braking system would be greatly influenced by the position of the attached elevators. The horizontal tail normally carries as downward load (less so at aft CG, more so at forward CG). With the stabilizer hinge line forward of its center of pressure, the load on the jack screw is usually in the direction associated with rotating the stabilizer in the leading edge up (i.e., airplane nose down) direction. If the linkage controlling stabilizer position were to let loose I would expect the stabilizer to rotate in that direction causing the airplane to pitch nose down. In response (either by way of the autopilot or the crew) the elevator would be commanded in the airplane nose up direction (elevator surface trailing edge up) that would put further moment on the stabilizer to drive it in the airplane nose down direction.

We also need to remember that the lift generated by the tail is only a small portion of the total lift of the airplane. For the overall airplane to enter an essentially ballistic free-fall the wing angle of attack would have to decrease to that for zero lift. A horizontal tail mechanism failure that allows the stabilizer to float will result in nose down pitching moment that would tend to decrease wing AOA, but it is not correct to assume that a floating stabilizer leads to the wing experiencing AOA corresponding to zero lift. It is very possible that such a failure would result in sufficient nose down pitching moment to drive the wing to a significantly negative AOA to cause negative g.

[PJ2]

CONSO, this indicates a "proposed" rule-making. Was the AD actually published and is it in force today? Tx.

[B2N2]

Just a honest question.
What do you calculate the average vertical G to be in order to start at zero vertical speed at 6000 feet and then reach the ground in 18sec?

I don’t know as this would depend on where terminal velocity is reached and this is dependent on a set of variables.

Unless I’m recalling my rusty theory incorrect G-forces are accelerating forces and indicate a rate of change rather then velocity. Same as a VSI shows rate of change.
Standing in an elevator I feel acceleration and deceleration forces.
If the vehicle that I’m in reaches a constant speed and I’m at the same speed I won’t feel ( measure) anything.
If the speed of the vehicle changes and I’m firmly attached to this vehicle I will sense (measure) a change.
Starting at altitude I can have an average rate of descent which is based on altitude change/time but it can be continuously accelerating or deceleration after initial acceleration.

This is where Vg diagrams come in and the definitions of the various speeds such as Va.

[RatherBeFlying]

Just a honest question.What do you calculate the average vertical G to be in order to start at zero vertical speed at 6000 feet and then reach the ground in 18sec?Drag, thrust and lift work in three dimensions. Only gravity works in a single dimension.

Etkin's Dynamics of Atmospheric Flight devotes 75 pages of serious math to General Equations of Unsteady Motion

[gums]

Salute!

@ the wait-and-see folks ----- we should be fairly confident that our discussion of the MCAS on the 610 accident was worthwhikle and may have prevented another accident, ya think? And all that good info was before much of a formal report was released from the relevant safety agency. As tdracer , et al have stated, an obvious system or hardware failure that might exist, or be possible on other aircraft of this type would likely be announced early on by the NTSB.
If I were flying this type today I would be glommed onto this thread like flies on sierra.

Gums sends...





.

[DaveReidUK]

CONSO, this indicates a "proposed" rule-making. Was the AD actually published and is it in force today? Tx.

AD 2008-06-06. Still in force.

[DaveReidUK]

I don’t know as this would depend on where terminal velocity is reached and this is dependent on a set of variables.

Correct me if I'm wrong, but average vertical acceleration would simply be a function of vertical distance travelled and time (if initial vertical velocity is zero).

6000' in 18 seconds gives a = 37 ft/s² if my maths is correct.

[SteinarN]

Response of the stabilizer to a failure of the jack screw or its nut and braking system would be greatly influenced by the position of the attached elevators. The horizontal tail normally carries as downward load (less so at aft CG, more so at forward CG). With the stabilizer hinge line forward of its center of pressure, the load on the jack screw is usually in the direction associated with rotating the stabilizer in the leading edge up (i.e., airplane nose down) direction. If the linkage controlling stabilizer position were to let loose I would expect the stabilizer to rotate in that direction causing the airplane to pitch nose down. In response (either by way of the autopilot or the crew) the elevator would be commanded in the airplane nose up direction (elevator surface trailing edge up) that would put further moment on the stabilizer to drive it in the airplane nose down direction.

We also need to remember that the lift generated by the tail is only a small portion of the total lift of the airplane. For the overall airplane to enter an essentially ballistic free-fall the wing angle of attack would have to decrease to that for zero lift. A horizontal tail mechanism failure that allows the stabilizer to float will result in nose down pitching moment that would tend to decrease wing AOA, but it is not correct to assume that a floating stabilizer leads to the wing experiencing AOA corresponding to zero lift. It is very possible that such a failure would result in sufficient nose down pitching moment to drive the wing to a significantly negative AOA to cause negative g.

I was giving some thought to what the wings AoA would be after the stabilizer had reached zero AoA. As the aircraft should be more or less nose heavy the nose should drop down decreasing wings AoA. My thinking is that when the wings reach zero AoA then the complete aircraft is free falling (when talking about vertical movement) and the aircraft is no longer nose heavy, as there is zero G on every part of the aircraft. However there would still be some momentum left, causing the tipping down of the nose to continue past the zero AoA of the wings, and then the wings should start to produce negative lift as you speculate. When the wings produce negative lift and the nose is still heavy (as in center of mass beeing forward of wings center of lift) then the aircraft should go more nose up again. So there might be some ocillations. My thinking is then that as the vertical speed increases the aircraft must point more and more nose down in order to keep the wings at pluss/minus zero AoA.

So, I would speculate that a free floating HT where the hinge point is forward of the center of lift line will cause the aircraft to quickly reach approximately zero G and thereafter increase vertical speed at approximately 10m/s2 (1G) which would take an object from 6000 feet to the ground in 20 sec.

[PJ2]

. . . It is very possible that such a failure would result in sufficient nose down pitching moment to drive the wing to a significantly negative AOA to cause negative g.Though the entire empennage is lost, the nose-down effect is the same - note the extreme downward bend of the wings just prior to impact. This came up briefly during the MCAS discussion: A26 video. (https://duckduckgo.com/?q=bouncing+bomb+crash&t=ffsb&atb=v119-6__&ia=videos&iax=videos&iai=kgSiIojF7fE)

[SteinarN]

Correct me if I'm wrong, but average vertical acceleration would simply be a function of vertical distance travelled and time (if initial vertical velocity is zero).

6000' in 18 seconds gives a = 37 ft/s² if my maths is correct.

My idea with the "average" value is the same as yours, namely what value of constant G or acceleration would cause the object to travel the required vertical distance in the required time.

6000 feet in 20 sec gives zero G, free fall or approximately 10m/s2, 6000 feet in 18 sec gives your slightly larger value which is a slightly negative G.

[DaveReidUK]

My idea with the "average" value is the same as yours, namely what value of constant G or acceleration would cause the object to travel the required vertical distance in the required time.

6000 feet in 20 sec gives zero G, free fall or approximately 10m/s2, 6000 feet in 18 sec gives your slightly larger value which is a slightly negative G.

Yes, though that doesn't seem to be relevant to what actually happened.

A "zero G" descent (or indeed a descent at any constant vertical acceleration value) follows a parabolic profile (ask any trainee astronaut). The actual descent profile, as far as it's possible to ascertain from FR24, was at an almost constant flightpath angle.

[SteinarN]

Yes, though that doesn't seem to be relevant to what actually happened.

A "zero G" descent (or indeed a descent at any constant vertical acceleration value) follows a parabolic profile (ask any trainee astronaut). The actual descent profile, as far as it's possible to ascertain from FR24, was at an almost constant flightpath angle.

Well, I am not arguing this or that beeing the cause of the crash, I am only speculating as to the effect of an free floating HS.

With that said I really thought the vertical speed increased significantly all the way down to the ground?
If it was at an almost constant flight path angle, then the transition from more or less level flight about 20 sec prior to the aircraft hitting the ground, to the constant nose down angle must have happened extremely quickly and hence required a very high value of negative G for a few sec

[FCeng84]

Taking a step back it appears to me that something quite dramatic occurred to rapidly transition this airplane from relatively stable, level flight to a very steep dive that took it from 6000' to the water in less than half a minute. We can postulate all sorts of root causes - hopefully the combination of CVR, FDR, and examination of key portions of the wreckage will shed light on what actually went wrong. This event happened so quickly that I don't see how it could be associated with stabilizer motion at rates that the jack screw is capable of generating. It is possible that elevator hard over (both sides) could have led to this result. I would hope that the time histories of the tail surface positions, pitch attitude, roll attitude, and normal acceleration are made public sooner than later.

For now I think that it would be prudent for current 767 operators to make sure that they are up to date on all maintenance and inspections associated with horizontal stabilizer and elevator control. The 767 has a very extensive and solid history of safe operation so I would be very surprised to see design listed as a cause of this loss.

[tdracer]

767 may, of course, have different structural geometry - tdracer would probably know.

Sorry, but my expertise is in propulsion/engines - I'm not very knowledgeable about the structural aspects or details of how the flight controls work.

Pretty much all I can add in that respect is that I was on the Boeing Safety Review Board (SRB) for many years - in the aftermath of the Alaska MD-80 jackscrew crash Boeing did a study to determine if any of the Boeing models had a similar susceptibility and determined that the Boeing designs were still safe. But I don't recall any meaningful details.

BTW voyageur9 - I have worked with the NTSB on several occasions and have the highest respect for the people I worked with. Yes, they do tend to get information and reports out quickly - especially when things are reasonably clear cut. Just saying that the NTSB releasing information to the public is not a priority unless they believe there is some safety aspect that should be quickly addressed.

[filejw]

Taking a step back it appears to me that something quite dramatic occurred to rapidly transition this airplane from relatively stable, level flight to a very steep dive that took it from 6000' to the water in less than half a minute. We can postulate all sorts of root causes - hopefully the combination of CVR, FDR, and examination of key portions of the wreckage will shed light on what actually went wrong. This event happened so quickly that I don't see how it could be associated with stabilizer motion at rates that the jack screw is capable of generating. It is possible that elevator hard over (both sides) could have led to this result. I would hope that the time histories of the tail surface positions, pitch attitude, roll attitude, and normal acceleration are made public sooner than later.

For now I think that it would be prudent for current 767 operators to make sure that they are up to date on all maintenance and inspections associated with horizontal stabilizer and elevator control. The 767 has a very extensive and solid history of safe operation so I would be very surprised to see design listed as a cause of this loss.

Take a look at post 97, some type of roll cloud or wind rotation could be the trigger to an aircraft structural failure or improper response by either the autoflight system or crew.

[DaveReidUK]

This event happened so quickly that I don't see how it could be associated with stabilizer motion at rates that the jack screw is capable of generating.

I don't think anyone will disagree with that. A screwjack/ballnut failure, on the other hand, could be implicated.

[Airbubba]

We lost three fellow airman, may they rest in peace.

Not in any way trying to detract from the technical discussion, here is a blog post about a moving tribute to the jumpseat rider.

United Airlines Presents Flight 3591 Widow with Wings and Epaulets, Leaves Indoc Seat Vacantthepilotwifelife/ 6 comments (https://thepilotwifelife.com/united-airlines-presents-flight-3591-widow-with-wings-and-epaulets-leaves-indoc-seat-vacant/#comments)

On February 23, 2019, tragedy shocked the aviation world when Atlas Flight 3591 went down in Trinity Bay just outside of Anahauc, TX just before 12:45 pm CST. The flight was a cargo transport from MIA to IAH for Prime Air using a twin-engine Boeing 767 and had three people aboard – Captain Ricky Blakely, FO Conrad Aska, and Mesa jumpseater Sean Archuleta. There were no survivors.

When tragedy strikes the aviation world, it affects all of us. We are a tight-knit family, and we are all only a few degrees from the incident. If we did not know those involved, we know those who did. We all pray, hope, and believe that these types of events will never rock our world, but when they do we stand together as one family, regardless of type rating, airline affiliation, or even country of origin. Even though the media did not feel like it was a ‘big story’ and it quickly fizzled out because the jet was cargo and there apparently weren’t enough people aboard to warrant wide coverage, to us it is our family. It affects us all to the core and leaves us with an empty hole in our hearts and in our extended family.

Captain Archuleta was, by all accounts, a standup guy who was easy to be around and who cared immensely for others. “I was friends with and flew with Sean,” states Captain Cole Goldenberg. “He would’ve been the first person there for everyone else in a tragedy like this. He is a very special guy.”https://i0.wp.com/thepilotwifelife.com/wp-content/uploads/2019/03/img_3478.jpg?w=960&ssl=1Captain Archuleta (R) as remembered by his friend and fellow pilot Cole Goldenberg (L). Photo with permission.

Just before the tragic accident that took his life, Captain Archuleta had been hired by United Airlines (UAL) and had already been assigned his class date – he was on his way to the majors! This is a celebratory event that we all look forward to as we move through the life of aviation, and it was not different for the Archuleta family. However, fate took away their moment and left a PW sister behind with her young children to figure out how to face another day without her Captain.

In a gesture of kindness, respect, and utmost class, United Airlines is doing something incredible. During Sean’s Indoc class on March 12th, they will leave his seat vacant in memory of the great loss that rocked the aviation world and the hole that is left in all of our hearts. What an incredibly honorable and classy gesture!

Furthermore, they presented his widow with UAL wings and epaulets. As told by UAL Captain Gunn:

“An update regarding Captain Sean Archuleta. He was the jumpseat rider on the prime Atlas/Prime Air crash who had a class date to begin at United, this week I believe. I operated flight 1009 to Bogota on Friday March 1st. Prior to heading to the airport, I was contacted by the IAH Chief Pilot and informed that his widow would be on our flight returning to Colombia where she currently lives. He asked if I could deliver a package to her from United. Of course I agreed to help in any way that I could. United had her booked in First Class and escorted her to the Polaris Club before the flight with plans for a personal escort to the airplane for departure. I coordinated to have her escorted to the plane and be in her seat about 5 minutes prior to general boarding. His wife Titania speaks only Spanish and although I speak some Spanish it’s definitely not conversational. I offered my condolences on behalf of United Airlines and all United pilots. The circumstances were difficult especially considering that I had never met her. I delivered a stack of condolence cards form both United and Mesa Airlines (his current employer) as well as a set of United wings and Epaulets. She was a very lovely lady but she broke down when I gave her the wings/epaulets. It was clearly very emotional for her to receive the wings. However, I could tell that she was very moved and it meant a great deal to her. Through the interpretation of the flight attendant, she told me that she would save these for her children to see and so that they would know that he was a United pilot. I simply said that it would have been a pleasure and and honor to fly with her husband. She thanked me and told me that United had been very good to her and that she was grateful for their support. It appears that United stepped up and treated her with respect and helped an already tragic situation.”

Hats off for an awesome gesture of kindness, respect, and utmost class. I think the fact that they did not feel the need to publicize their act of kindness as even more upstanding and honorable. There is enough negativity in the world. There is enough hatred and stereotyping to fill up volumes. The media often clings to those things that are certain to up readership and line pockets. However, THIS is the stuff we should make viral, friends!

Well played UAL, well played.



https://thepilotwifelife.com/united-airlines-presents-flight-3591-widow-with-wings-and-epaulets-leaves-indoc-seat-vacant/


https://cimg1.ibsrv.net/gimg/pprune.org-vbulletin/959x963/27625118_10156152438859886_7570115695178979446_o_5449d3aae1c f662ffccef1ab91a035450b36590a.jpg

[CONSO]

CONSO, this indicates a "proposed" rule-making. Was the AD actually published and is it in force today? Tx.

I'm sure those more familiar with proposed versus final AD, and how to trace can find or give a definitive answer- especially on this forum. I posted it for a link to related information on 767 horiz stab jackscrew and the mention of ballscrew versus MD " NUT" issues.

I have found a bit further related information at

https://www.federalregister.gov/documents/2007/08/21/E7-16424/airworthiness-directives-boeing-model-767-airplanes

and I'll let the experts- and those most familiar with either maintainance records of a given airframe follow up with links if possible to factual data on that 767.
FWIW-around 1982 I had a small task to quickly make a manufacturing ' tool ' to drive the 767 horiz stabilizer during factory checkout before installation of any electrical or hydraulics. I was able to use a piece of 1/4 inch airhose, two screw band type clamps , one on hose over drive spline and one on hose over a 1/8 socket attached to a a low speed 1/4 hp air nutrunner. That kluge was used for a few months until a nicely fitting female match to the spline drive was made attached to a low speed nut runner to move the stabilizer thru its full motions after mounting and basic installation. Of course with no air-loads it was VERY easy and required very little torque. Sort of like your car steering wheel with front wheels off ground- which also uses a ball screw- jack type mechanism.

[wingview]

Correct me if I'm wrong, but we only know it was uncontrollably the last 18 seconds but not before that. Was there a problem already or not? So we can do the math with G forces or not, nobody knows at this time.
One thing I think we can scratch is if it was deliberate because than it would have been controlled.

[PJ2]

Thank you, DaveReidUK & CONSO for info on the AD.

[mustangsally]

mustanfsally, there's nothing wrong with venting ideas as long as it comes from those sufficiently qualified to make comment (though admittedly that's not always the case). That's what PPRuNe is all about didn't you know? If it upsets you, I guess it's best just to ignore this thread.

HT, you miss read me. But, that is OK, my wife does the same. But frequently the comments posted come from someone with little knowledge of 767 systems and how they relate to this accident. I personally have over thirty five years as a crew member but have zero 767 experience. So I don't speculate on what happened or the cause and effect. If we could weed out the rift raft and get those with good understanding of the specific aircraft systems then I would agree with you. But that just does not happen in this pprune world. Just go back and reread the discussion on elevator trim and pivot points of such, jack screws and nuts et all. Maybe that is why I only look at this web sight for entertainment with a couple of pints.

Cheers

[CAP A330]

If the aircraft was in free fall, it would be accelerating at 1 g, and experiencing 0 g. Two different things.

[DaveReidUK]

If the aircraft was in free fall, it would be accelerating at 1 g, and experiencing 0 g. Two different things.

If by "two different things" you mean two different frames of reference, then yes. File under "interesting, but irrelevant".

[Old Boeing Driver]

Another video of the crash.

https://www.chron.com/news/houston-texas/houston/article/New-video-shows-Atlas-Air-plane-crash-into-13668603.php#next

[Rhino power]

In the video clip posted above by, Old Boeing Driver, in the zoomed in part of the video, 0:15s onwards, the rate/angle of descent moments before impact seems to decrease slightly, perhaps control was regained and a pull out of the dive was being attempted? Or maybe it's just an anomaly in the video footage?

-RP

[McGinty]

In the second video it does look to me as well as if the plane begins to pull up a little. At that speed of descent such a manoeuvre would likely tear the wings off if the ground had not intervened first.

[Viscount Way]

Without access to the Alert Service Bulletins mentioned in the AD it's difficult to determine how much of the content is applicable and the compliance times. The Bulletins would also indicate the means of terminating inspections and any other closing actions. Given that Wear Measurement and Lubrication are involved it's quite likely
that the repetitive actions have been included in the aircraft's Maintenance Program by now.

[B2N2]

At that speed of descent such a manoeuvre would likely tear the wings off if the ground had not intervened first.

Utter rubbish...again.
The rate of change determines the forces on the airplane. Do you even know what the V-speeds mean?
You can easily recover from a Vne dive if you do not exceed load limitations.
So nothing “tearing off”.

[pattern_is_full]

Two halves of the same thing:

A pull-out with careful attention to avoiding load limitations won't pull the wings off - it may very well also not be possible in the altitude/time available (BAD)
A pull-out with a hard enough pull to avoid the ground in the altitude/time available may very well exceed the load limitations (also BAD). Or produce an accelerated stall (see Hawker Hunter air show crash - also BAD)

Takes fine judgement and ingrained knowledge of an aircraft's capabilities to thread that needle - in the middle of a "startle" event.

[B2N2]

Two halves of the same thing:

A pull-out with careful attention to avoiding load limitations won't pull the wings off - it may very well also not be possible in the altitude/time available (BAD)
A pull-out with a hard enough pull to avoid the ground in the altitude/time available may very well exceed the load limitations (also BAD). Or produce an accelerated stall (see Hawker Hunter air show crash - also BAD)

Takes fine judgement and ingrained knowledge of an aircraft's capabilities to thread that needle - in the middle of a "startle" event.


This ^^^
People please read it again and again.

[SteinarN]

This ^^^
People please read it again and again.

Yes. And you can start your self reading it.

When I was reading reply #336 by McGinty I was reading and understanding that reply exactly like pattern_is_full explaines it in reply #339; " A pull-out with a hard enough pull to avoid the ground in the altitude/time available may very well exceed the load limitations (also BAD)."
And i was reading it like that even with my very limited cognitive ability. Some here with claimed better cognitive ability seemingly was not able to read reply #336 that way...

[DaveReidUK]

Despite being misinterpreted (wilfully?), McGinty's post was perfectly clear.

He (or she) did not assert that the wings would fall off in a dive. :ugh:

[MLHeliwrench]

Say the horizontal stab somehow became able to float up and down the screw shaft on its own - would the elevators act as a servo surface like how some aircraft have a smaller flight control surface move opposite the desired deflection of the main surface?

In this hypothetical scenario, pulling back the yolk would cause the horizontal stabilizer to deflect leading edge up. This would cause a dive.

Can anyone attest to the capability of full yolk back (under normal travel limits) to overcome a horizontal stab trimmed all the way leading edge up?

[B2N2]

Despite being misinterpreted (wilfully?), McGinty's post was perfectly clear.

He (or she) did not assert that the wings would fall off in a dive. :ugh:

Insinuated the wings would rip off.
Which they didn’t “before the ground interfered”.

[DingerX]

That situation is colloquially known as "over easy".

Through the noise: the relevance of the free fall calculation would have been to get a back of the envelope estimation of what an in-flight break up would look like. The videos pretty much make that exercise pointless. They also make a compelling case for a primary or secondary flight control issue, whether the yolk stayed where it was, got scrambled with the white, or ended up in a nose-down omelette.

[FCeng84]

Say the horizontal stab somehow became able to float up and down the screw shaft on its own - would the elevators act as a servo surface like how some aircraft have a smaller flight control surface move opposite the desired deflection of the main surface?

In this hypothetical scenario, pulling back the yolk would cause the horizontal stabilizer to deflect leading edge up. This would cause a dive.

Can anyone attest to the capability of full yolk back (under normal travel limits) to overcome a horizontal stab trimmed all the way leading edge up?

I concur with your suggestion that a floating stabilizer would be impacted by the elevators in much that same manner as a tab-driven trailing edge control surface. Airplane nose up elevator deflection puts loads on the stabilizer in the opposite direction and thus would drive a floating stabilizer in the airplane nose down direction. For small motions, elevator has about half as much pitch authority as stabilizer on a degree for degree basis. The story is not quite that simple, however. Below are a few factors to consider:

First is the question of where the stabilizer is positioned for trim. If the stabilizer were near its airplane nose down limit when trimmed and then went to that limit there should be enough elevator to counter. If, however, the stabilizer were a long way from its airplane nose down limit when in trim, having it go all the way to its airplane nose down limit would probably generate more nose down pitching moment than the elevator could counter.

Second is the question of elevator blow-down due to limited hinge moment capability. At approach speed the elevator control actuators are able to generate enough force to drive the elevators to their travel limits. At cruise (and faster) speeds elevator travel will be hinge moment limited as the actuators cannot generate enough force to reach elevator travel limits. This can become a serious issue in a dive as the higher the airspeed gets, the less elevator deflection the system can generate to command the needed pull-up to recover.

Third is the question of elevator effectiveness with the stabilizer at its full airplane nose down limit. With the stabilizer leading edge up (i.e., airplane nose down) it may disturb flow over the elevator in such a way as to reduce elevator effectiveness. This would be particularly true if elevator is at first able to generate nose up pitching moment to generate significant wing AOA needed to pull out of a dive. In that case the AOA seen by the tail will be that of the airplane (i.e., the wing) plus the angle of the stabilizer. If this is high enough, flow over the stabilizer will begin to separate severely reducing elevator effectiveness - particularly in the airplane nose up direction.

The bottom line is that having the stabilizer significantly out of trim is a bad situation. Herein lies the crux of why control systems that require detection and arrest of uncommanded stabilizer motion have requirements to do so within just a couple of degrees of errant stabilizer motion.

[DaveReidUK]

Say the horizontal stab somehow became able to float up and down the screw shaft on its own - would the elevators act as a servo surface like how some aircraft have a smaller flight control surface move opposite the desired deflection of the main surface?

Bear in mind that when the Alaskan MD-80's jackscrew let go, it pulled through the mechanical stop and just kept going. The elevators didn't make a blind bit of difference.

[Lonewolf_50]

Maybe that is why I only look at this web sight for entertainment with a couple of pints. I thought that was a requirement. :E

@Old Boeing Driver: had there been another aircraft impacting the accident aircraft, I am pretty sure that lost or damaged plane would be accounted for/reported, by now. (So I think I agree with you)

Unless it's a drone at 6000'? (That is a piece of raw guesswork to answer the hypothesis of "something hit the 767" which hypothesis so far neither ATC nor the NTSB has alluded to as a possible cause factor)

PS: thanks for the video, but it also made my guts clench up.

[B2N2]

A lot of things are possible but fewer and fewer are plausible

[Old Boeing Driver]

I thought that was a requirement. :E

@Old Boeing Driver: had there been another aircraft impacting the accident aircraft, I am pretty sure that lost or damaged plane would be accounted for/reported, by now. (So I think I agree with you)

Unless it's a drone at 6000'? (That is a piece of raw guesswork to answer the hypothesis of "something hit the 767" which hypothesis so far neither ATC nor the NTSB has alluded to as a possible cause factor)

PS: thanks for the video, but it also made my guts clench up.
The video was gruesome to me as well. By the looks of the speed coming down on the video, I still lean to a failure in the tail somewhere.

Have a great evening.

[Gipsy Queen]

An object which is free falling is by definition experiencing zero G, this is according to laws set by Albert Einstein.

I suspect that a fellow called Newton thought of this first.

[BluSdUp]

So they did 240kts at ca 6000feet.
ATC lost radio and Radar contact.
So problems with Electrics and or Electronics.
Maybe due to turbulence and cargo shift, maybe not?

But as it have been showed in quite a few accidents it only takes a few seconds with the wrong elevator inputs and it is all over.
FlyDubai is case in point, and that Russian 737.300 was the same.
But the Bombardier RJ in Sweden as described in an earlier post was the worst.

Lets for argument sake say they lost a generator or a bus, so that the aircraft was dumped in the hands of PF without any warning!
All it takes is a few seconds of zero or negative G and it becomes critical, THEN depending if they have all displays , You have to be quick and correct in the recovery.
This would also explain why the Transponder stopped working.

There was nothing wrong with the Stabilizer, is my wild guess, mainly because no Boeing to date has had this problem.
Minor OR major problem in Display, AP, Electronics and Electrics, with a small or big human factor problem I am afraid is most likely.
Regardless , a gruesome way to end.
As terrible as this is for friends and family, It could be worse.
It could be a regular passenger flight.
So lets learn.

Peace
Cpt B

[a330pilotcanada]

Good Evening All:

I have been out of the "industry" for about nine years now but I found this link for an Airworthiness Directive for all models of the B-767.

For those who are more up to date than I am please comment.


https://www.federalregister.gov/documents/2008/03/12/E8-4677/airworthiness-directives-boeing-model-767-airplanes

[Skywagon1915]

Any other way to view the video ? I can't open it at all.

[sablatnic]

Any other way to view the video ? I can't open it at all.

It is here on youtube:
https://m.youtube.com/watch?v=IHZa8OpGDOw

[lpvapproach]

Those videos appear to show the airplane intact and flying, descending down to the ground

[DaveReidUK]

Good Evening All:

I have been out of the "industry" for about nine years now but I found this link for an Airworthiness Directive for all models of the B-767.

For those who are more up to date than I am please comment.

https://www.federalregister.gov/documents/2008/03/12/E8-4677/airworthiness-directives-boeing-model-767-airplanes

See post #316 from two days ago.

AD 2008-06-06. Still in force.

[JH_CAMO]

I’m sure people will understand that for professional reasons I can’t post maintenance records online, but the AD line of thought is a bit of a red herring.

Reference my earlier post, this aircraft was in full compliance with all applicable ADs.

Generically I can say that the SB mandated by the AD requires three separate actions - a detailed inspection, lubrication, and a free play check. All of these have been performed on time since the AD was effective.

[ManaAdaSystem]

So they did 240kts at ca 6000feet.
ATC lost radio and Radar contact.
So problems with Electrics and or Electronics.
Maybe due to turbulence and cargo shift, maybe not?

Lets for argument sake say they lost a generator or a bus, so that the aircraft was dumped in the hands of PF without any warning!
All it takes is a few seconds of zero or negative G and it becomes critical, THEN depending if they have all displays , You have to be quick and correct in the recovery.
This would also explain why the Transponder stopped working.



Did they loose the transponder?
Where does the transponder get power from on the 767?
On the NG the transponder will work even when you are down to battery power. Is the 767 different?

[DaveReidUK]

Did they loose the transponder?Where does the transponder get power from on the 767?
On the NG the transponder will work even when you are down to battery power. Is the 767 different?
ATC stopped receiving the transponder signal shortly before impact. That doesn't mean it had stopped transmitting at that point.

SSR data captured by FR24 continued up to around 2-3 seconds before the aircraft hit the ground. Allowing for the granularity of the captured data, it's perfectly possible that the transponder was transmitting right up to impact.

[ManaAdaSystem]

Thanks.
And to clarify, it’s transponder #1 that is wired to the battery. If transponder #2 is operating, you need to switch it to #1 if you loose AC power.
Easy to forget in a high stress situation.

[Orestes]

The video linked below gives a view of some wreckage that has been moved to a warehouse. In it there can be seen the remains of at least one engine. For what it's worth, there is a good view of the aft end of the engine showing a turbine disk. From the way the blades are broken and bent, it appears that the turbine disk was rotating at impact.

https://www.youtube.com/watch?v=qOa4V4JF4t4

[lomapaseo]

From the way the blades are broken and bent, it appears that the turbine disk was rotating at impact.

From a frontal impact which should stop the fan dead?

Lots more evidence is needed

[Airbubba]

The video linked below gives a view of some wreckage that has been moved to a warehouse.

Excellent video quality and professional reporting. I'm sure some trained eyes here will recognize many more of the pieces than the reporter or me.

Here is a follow up report on the KHOU evening news.

https://youtu.be/4ImIZVFhnJM

[tubby linton]

At 4:31 in the video in post 362 is the item in the background the jack screw for the stab?

[SteinarN]

At 4:31 in the video in post 362 is the item in the background the jack screw for the stab?


https://cimg3.ibsrv.net/gimg/pprune.org-vbulletin/1024x768/120_2004_img_116b096ca856bbf58f8d0d3f48f74f05cd3437bd.jpg


This should be a pic of the jack screw on a 737. I dont know how the 767 one looks, but the screw in this picture and the triangular bracket going to the tail plane is very similar to the screw and bracket in the video at about 4.30. If so, then the jack screw in the video seems to be at about 2/3 nose up position.

But, can this be possible? If the jack screw is almost full nose up, how come the airplane plunged like it did? Even if we suppose the trim was running towards nose up at the time of impact, how many seconds would it need to run to this apparent position from lets say a neutral position? Would it be possible at all in the available 18 to 20 sec to first get the airplane to start plunging where the jack screw obviously cant have been in almost full nose up position, and then at some time before impact start running in the nose up direction and have time to reach the position it seems to have had at impact? And another question, considering the very high indicated airspeed the aircraft had at impact, would an almost full nose up trim position be compatible with wings AoA below a value where generated lift wouldnt tear the wings off of the aircraft?

What would the AoA be for the airspeed the aircraft is assumed to have had at impact and at the same time generating say 3G? I believe somewhere between say 5 to 9 degrees AoA at say 400 knots at low altitude would produce so much lift that the wings would tear off?

So, could it be possible that the tail plane was not in the postion that the screw jack is indicating, ie that the screw jack had broken loose in one end before impact?

[PJ2]

Not sure...I don't see threading, but the resolution at this enlargement size isn't clear enough.


https://cimg1.ibsrv.net/gimg/pprune.org-vbulletin/1224x960/2019_03_08_092247_stabilizer_jack_x2_010f795c2f852778bdb39ea 39b267888888ba2b2.jpg

[SteinarN]

Maybe this is not the tail plane jack screw.... It seems small? Could it be a gear actuator?

Hope a 767 mechanic can chime in on this.

[TheHardWay]

I concur with your suggestion that a floating stabilizer would be impacted by the elevators in much that same manner as a tab-driven trailing edge control surface. Airplane nose up elevator deflection puts loads on the stabilizer in the opposite direction and thus would drive a floating stabilizer in the airplane nose down direction. For small motions, elevator has about half as much pitch authority as stabilizer on a degree for degree basis. The story is not quite that simple, however. Below are a few factors to consider:The elevator is hinged at the leading edge, so the servo tab mounted at the trailing edge has the mechanical advantage or leverage to force the elevator to move up or down.
The stabilizer is hinged at the trailing edge with the actuator at the leading edge, so the position of the elevator mounted at the trailing edge of the stabilizer is going to have little or no effect on the position of a free floating stabilizer.

[Meester proach]

Is that not in really bad taste, inviting a TV reporter to look at the debris ?
cannot see the AAIB doing this.

[PJ2]

Image of a B737NG screw-jack - certainly looks similar.

The photograph is taken looking down and slightly rearward from the starboard side of the fuselage. The position of the ball-screw on the jack-screw, (nearer to the base & motor), I believe would be "leading-edge-down", for "nose-up" trim.

https://cimg1.ibsrv.net/gimg/pprune.org-vbulletin/805x768/2019_03_08_094608_b737ng_screwjack_xl_a1b0d54ce51fcf688842f2 85e88bb7779e5df352.jpg

[B2N2]

Is that not in really bad taste, inviting a TV reporter to look at the debris ?
cannot see the AAIB doing this.

Freedom of Information Act.
Unless it’s been deemed classified or not suitable to be released.
In this case the wreckage is “news” and it’s a local TV station.

[SeenItAll]

Is that not in really bad taste, inviting a TV reporter to look at the debris ?
cannot see the AAIB doing this.

Not too much different from the photos that were displayed post the PanAm 103 crash over Lockerbie. And in that one 270 people died. See https://www.gettyimages.com/photos/pan-am-flight-103?sort=mostpopular&mediatype=photography&phrase=pan%20am%20flight%20103

[Airbubba]

Is that not in really bad taste, inviting a TV reporter to look at the debris ?
cannot see the AAIB doing this.

In this widely published iconic picture of the Lockerbie crash the bodies of former colleagues Captain Jim MacQuarrie and First Officer Ray Wagner are under blankets with their flight kits embedded in the soil. Few here seem to call this bad taste or question the media's right to publish photos of the wreckage even years later on the 30th anniversary of the crash. Wasn't the AAIB involved in this investigation?


https://cimg6.ibsrv.net/gimg/pprune.org-vbulletin/1024x566/ap960719110574_fd79113fe83fff812a2267e707c6b8c9ed15cc44.jpg

[Airbubba]

Freedom of Information Act.Unless it’s been deemed classified or not suitable to be released.
In this case the wreckage is “news” and it’s a local TV station.


Actually, information about an ongoing investigation is specifically exempted from FOIA requests under NTSB regulations:

FOIA Exemptions

The National Transportation Safety Board embraces administration policy and applies a presumption of openness in responding to FOIA requests. The NTSB has also taken steps over the past two years to make discretionary releases of record as evidenced by the number of dockets posted to the public website.

However, certain records retained by the NTSB may fall under exemptions of the FOIA. The NTSB will deny requests, either in whole or in part, if the request seeks information that falls within one of the exemptions of the FOIA. The NTSB may deny a request based on the following exemptions:

...Exemption 7 : Information from an investigation that is ongoing.

https://www.ntsb.gov/about/foia/Pages/foia_exemptions.aspx

[derjodel]

Sorry, but the people who state the airplane hits at full speed nose down are plain wrong.

1. Yes, camera lenses have distortions. However, distortion is about the same in equal radius from the centre. Meaning that an object in an upper screen would be about equally distorted as an object at the bottom of the screen for the same distance from the image centre.
2. That said, this particular lens seems a normal lens (not wide, not tele). One can draw a fairly straight line over horizon. On top of that, all the visible objects (e.g. huts) don't show any major visible distortions.
3: Given points 2 and 3, lens distortion can be ignored as it plays no role in assessing the flying path of this video.
4. As noted in point 3, the video is at an angle. An observer might get a better perception by rotating the screen to the left.
5. The airplane enters the video at about 6''.
6. Just before 7'' we can clearly see the airplane between the tree branches. It's at an angle we can't determine, but on my screen the fuselage is 26px long (only relative size matters). Wings at the root are 4 pixels long. Engines and other features are easily seen and recognisable.
7. The airplane reaches the centre of the area between upper bound of the screen and the horizon (approx. where the line crosses the central trees) at 8''. Thus, it needed 2'' to travel that distance (in 3d space, and we can't determine how fast it's moving away from us).
8. The airplane hits the ground at 11''. Seconds are not detailed enough to determine the right numbers, frame by frame analysis for which I don't have time would be needed.
9. However, it is my estimation that the plane needs about 2'' to reach the relative centre and it needs 3'' to hit the ground.
10. If the airplane was moving in constant speed on the down axis and away from us, the speed should be constant.
11: Therefore I believe the downwards speed in the lower segment is lower and the speed of moving away from camera is higher, in other words, the airplane seem to be changing the trajectory in its final moments.
12. In the final moments there's no clear view of the airplane, as it's exactly following the path the branch is hiding.
13. The airplane in this segment is much harder to see. Engines are not visible. Wings, which are clearly visible at the beginning are now just blips of interpolations. They measure 2 px.
14: Airplane is creating smaller image on the sensor either because it moved a considerable distance away from camera, or because it's pitch got higher, or combination of both.
15. At this point I would invite the reader to watch the recent video "aircraft lands at Asuncion Airport, Paraguay when staff is performing runway works". Watch closely the airplane which is approaching the camera at the landing speed. For the first 5 seconds of the video it stays virtually the same size. In fact, due to perspective humans often underestimate how fast are fast distant objects approaching.
16. The airplane would need to be moving away from the camera very fast to become that much smaller.
17: Hence, the only conclusion is, the pilots were trying to pull up, but unfortunately they did not have enough height.


Any my non factual speculation if I may: downdraft/windshear caused loss of speed while in the clouds, either pilots or systems pushed the stick forward to regain speed, lost awareness as no visual contact with the ground, once they were out of the clouds it was too late to safely pull up.

[warbirdfinder]

Amazing what one can get from a few seconds of fuzzy video, you should call the NTSB and tell you have the probable cause worked out!

[B2N2]

Any my non factual speculation if I may: downdraft/windshear caused loss of speed while in the clouds, either pilots or systems pushed the stick forward to regain speed, lost awareness as no visual contact with the ground, once they were out of the clouds it was too late to safely pull up

I can only shake my head.
Pixels and inches moved on a screen and THAT conclusion?
Its eyewatering to read.

[B2N2]

Actually, information about an ongoing investigation is specifically exempted from FOIA requests under NTSB regulations:



https://www.ntsb.gov/about/foia/Pages/foia_exemptions.aspx

Well...mute point as they did give access to the hangar.

[DaveReidUK]

Pixels and inches moved on a screen and THAT conclusion?

The references to 6", etc, are seconds into the video, not inches.

[Airbubba]

Well...mute point as they did give access to the hangar.

Mute point? Here, here!

[Herod]

Full marks to the FAA for the video. There is a faint chance that someone who knows what they are talking about may spot a point overlooked by the investigators. Note I did say someone who knows what they are talking about.

[Ranger One]

Is that not in really bad taste, inviting a TV reporter to look at the debris ?
cannot see the AAIB doing this.

From the fact that it was the Sheriff being interviewed in the second clip, I think it's fair to conclude that it was the Sheriff that invited the reporter in, not the NTSB; it seems the NTSB do not have 'exclusive custody' of the wreckage.

[NSEU]

Where does the transponder get power from on the 767?

It depends on the individual aircraft. On some of our old 767's, we used to have the Left ATC powered only by the 115Vac Left Main Bus and on the rest of the 767 fleet, it was powered by the 115Vac Left Main Bus with backup from one of the HMGs (Hydraulic Motor Generators). You would need to look at the wiring diagrams/schematics for that particular aircraft.

[Airbubba]

Full marks to the FAA for the video.

I don't think the FAA had much to do with it, the NTSB is an independent agency. But I agree, nicely done for those of us who are interested in the investigation.

[Airbubba]

From the fact that it was the Sheriff being interviewed in the second clip, I think it's fair to conclude that it was the Sheriff that invited the reporter in, not the NTSB; it seems the NTSB do not have 'exclusive custody' of the wreckage.

I disagree. I think the Sheriff comes in only with permission of the NTSB. 49 CFR § 830.10 is normally cited as giving the NTSB this authority over the wreckage until it is released with a Form 6120.15.

If there is evidence of a crime, the DOJ may also claim jurisdiction. Normally this is not a problem but there have been disputes, or so it is claimed. One famous example is the TWA 800 crash in 1996.

[NSEU]

Maybe this is not the tail plane jack screw.... It seems small? Could it be a gear actuator?

Hope a 767 mechanic can chime in on this.

I haven't spent much time in the forward stabilizer compartment, but your image seems to match the maintenance manual.


https://cimg3.ibsrv.net/gimg/pprune.org-vbulletin/660x493/767hstjackscrewa_3195308af62e12c2e3dae34871b5c3dd1bc3227b.jp g
There is no mention of total jackscrew length or diameter in the manual, so it will be difficult to create scale diagrams of the assembly. The trim only moves the assembly 21.51 inches during control column trimming (less if the flaps are less than 5 units). i.e. from 0.25 units (full nose down) to 12.8 units (full nose up) with column switches. Maximum range of the stabilizer is 0~14.2 units. 2 units is neutral.

At full nose down (0.25 units), the top of the moving part (ballnut) is within an inch of the upper stop.

[Airbubba]

I haven't spent much time in the forward stabilizer compartment, but your image seems to match the maintenance manual.

...At full nose down (0.25 units), the top of the moving part (ballnut) is within an inch of the upper stop.

Wow... :eek:

[pattern_is_full]

Good find.

Positioning in the video looks "close to" neutral. Not close to either stop. Unless it jammed there (and neutral was not the trim needed) or the whole unit broke loose from its mounts, that seems to minimize the odds of a THS jackscrew problem. (But you just never know).

On another note - probably having no significance to the accident - I have tentatively ID'd the "biggest part found so far" (shown at 5:10 in the reporter's first walk-around video, post 362) as the bottom half of the left winglet, and its junction with the wing tip.

[dash34]

This is an analysis of the video using the Tracker video analysis program. I used the wingspan to calibrate distance, set the x-axis along the aircraft's track, and used the starboard engine to track the motion. I had to interpolate some frames when the aircraft was hidden behind the tree so take the middle of the graph with a grain of salt. There were also repeated frames in the video for some reason. FWIW, this is the result. Analysis shows that the ROD appears to change from over 200 m/s to just over 100 m/s towards the end of the sequence. Another source of error is using the wingspan to calibrate because the aircraft is banked relative to the camera.
https://cimg1.ibsrv.net/gimg/pprune.org-vbulletin/561x463/capture_4e2bdc07ab4d2ceb08f674f11836e2a9f5770639.jpg

[pattern_is_full]

Dash34 - I take it your graph is inverted (i.e. the top right corner of the graph shows the bottom of the dive)?

[DaveReidUK]

Analysis shows that the ROD appears to change from over 200 m/s to just over 100 m/s towards the end of the sequence.

Given that we know the flightpath angle was approaching -50°, a ROD of 200 m/s would resolve to a velocity of around 500 kts.

Hmmm.

[dash34]

Dash34 - I take it your graph is inverted (i.e. the top right corner of the graph shows the bottom of the dive)?

Correct. The zero reference point was set at the tail of the aircraft in the first frame it was visible. The slope of this graph is ROD. Time is on the x-axis, displacement from the reference point on the y-axis.

[SteinarN]

dash34,
The video is grainy, and the branch is in the way too, but I think I am seing that the wings towards the end of the video is deflected significantly upwards, like in a very hard pull up at very high speed. Such an observation is in line with your suggestion that the ROD was decreasing significantly at the end of the video.

[SteinarN]

I haven't spent much time in the forward stabilizer compartment, but your image seems to match the maintenance manual.


https://cimg3.ibsrv.net/gimg/pprune.org-vbulletin/660x493/767hstjackscrewa_3195308af62e12c2e3dae34871b5c3dd1bc3227b.jp g
There is no mention of total jackscrew length or diameter in the manual, so it will be difficult to create scale diagrams of the assembly. The trim only moves the assembly 21.51 inches during control column trimming (less if the flaps are less than 5 units). i.e. from 0.25 units (full nose down) to 12.8 units (full nose up) with column switches. Maximum range of the stabilizer is 0~14.2 units. 2 units is neutral.

At full nose down (0.25 units), the top of the moving part (ballnut) is within an inch of the upper stop.

Thanks!
This is obviously the HS jack screw. It seems like it is at about position 2.5 or 3. So, close to neutral or slightly nose up.

[DaveReidUK]

Positioning in the video looks "close to" neutral. Not close to either stop. Unless it jammed there (and neutral was not the trim needed) or the whole unit broke loose from its mounts, that seems to minimize the odds of a THS jackscrew problem. (But you just never know).

In the photo, the jackscrew is resting on the actuator end and the broken-off stabilizer bracket. I don't think you can necessarily read too much into the position of the ballnut - if its threads had stripped and it was free to move up and down the screw then the photo could simply be showing how the assembly has been positioned on the floor.

On the other hand, we could be looking at a perfectly serviceable jackscrew. There just isn't enough definition in the photo to tell.

[PerPurumTonantes]

https://cimg3.ibsrv.net/gimg/pprune.org-vbulletin/660x493/767hstjackscrewa_3195308af62e12c2e3dae34871b5c3dd1bc3227b.jp g
At full nose down (0.25 units), the top of the moving part (ballnut) is within an inch of the upper stop.
Based on that, looks to me like recovered part is in full nose down position. Am I missing something?

[donotdespisethesnake]

Based on that, looks to me like recovered part is in full nose down position. Am I missing something?

Appears the same way to me. Another case of runaway trim?

[NSEU]

"Based on that, looks to me like recovered part is in full nose down position. Am I missing something? "

Are you overestimating the amount of travel the ballnut moves? I see about 6~8 inches of chrome above the ballnut (wild guess, not knowing the diameter of the screw). Subtract 1 inch to get the start point of 0.25 units. There is 21.51 inches of travel from here (to reach 12.8 units). That is 1.7 inches per unit. So, 5~7 inches (say 6) is 3.5 units from 0.25 units = 3.75 units. Neutral is 2 units, so 1.75 units nose up. I need a fellow engineer to check my numbers.

Pilot input required here. If you're flying straight and level with the flap setting required at 6000 feet (?), what's the average trim setting? (aka... how long is a piece of string).

[Ivor_Bigunn]

If the Statement:

"At full nose down (0.25 units), the top of the moving part (ballnut) is within an inch of the upper stop."

is correct, then the photographed screwjack is 80 to 90% Nose Down.

(look at the length of screwjack extending below the ballnut to the lower stop).

Whether this was how it was recovered, or if all the threads were stripped, would reduce the significance of the photo.

[Raffles S.A.]

Flaps would be 0 and the trim for 240KIAS about 1 or 2 units nose up

[henra]

Based on that, looks to me like recovered part is in full nose down position. Am I missing something?
???
I believe I see ~ 2-3 inches of spindle above the Ball nut. The silver part above the hinge appears to have the same color and diameter as the lower part. With 1,7" per unit and 0,25 units = max ND, I end up with ~2,5 -3 units. Seems like pretty much spot on for that speed and configuration.
That observation does not rally point into the direction of spindle/jackscrew failure nor runaway trim.
On top of that in the footage shown it appears indeed as if there were some kind of recovery attempt going on.
The speed being somewhere around 240kts this becomes more and more strange.
So at the moment it looks like:
- Flight at correct trim (assumption based on position of Jackscrew)
- At 240kts at 6000ft. (assumption based on FR24 data). (Probably Some 60-70kts above stall speed).
- Flight in IMC (Statement early in the process)
- Coming out of the clouds at high speed (>>250kts)
- Flightpath angle in the part visible in the footage initially ~40 - 50°ND, apparently possibly reducing to ~30 - 40°ND.
- Initial altitude in the video difficult to judge but in a wild guess somewhere in the range 2000 - 3000ft.

What plausible scenarios does that leave?
=> Stall as initiating event rather unlikely (too much speed margin, flight attitude pretty stable and straight, albeit hefty ND with rather lowish AoA in the footage)
=> Out of Trim situation somewhat unlikely if the position of the jackscrew represents the state it was in during the terminal phase of the flight.
=> failed jackscrew/ball nut can't be completely ruled out but is not really support by the picture of the jackscrew nor the apparent started recovery => rather unlikely.

At the moment I'm somewhat at a loss what could have reasonably caused that loss of Control.

[lomapaseo]

If there was an apparent recovery with a jackscrew failure, might the latter step have been from commanding max thrust vs spool up time?

[SteinarN]

???
What plausible scenarios does that leave?
=> Stall as initiating event rather unlikely (too much speed margin, flight attitude pretty stable and straight, albeit hefty ND with rather lowish AoA in the footage)
=> Out of Trim situation somewhat unlikely if the position of the jackscrew represents the state it was in during the terminal phase of the flight.
=> failed jackscrew/ball nut can't be completely ruled out but is not really support by the picture of the jackscrew nor the apparent started recovery => rather unlikely.

At the moment I'm somewhat at a loss what could have reasonably caused that loss of Control.

Some instrument failure causing false and highly unexpected aircraft attitude data beeing displayed for the pilots making them push hard on the yoke, aka CRJ200 crash in Sweden....
Failure of the elevator control/servo system putting the aircraft in a steep dive, aka 737 rudder hardover......
Mechanical breakage/failure of the HS screw jack system including its attachment structure making the HS freefloating, somewhat similar to Alaska MD-80......
........

[Raffles S.A.]

Remember that the jackscrew moves the stab, not the elevator.

[B2N2]

Some instrument failure causing false and highly unexpected aircraft attitude data beeing displayed for the pilots making them push hard on the yoke, aka CRJ200 crash in Sweden....
........

Crossing the waypoints at 7000' or 6000' depending on the transition they were in VMC conditions.

KIAH 232053Z 31010KT 10SM FEW045 BKN210 OVC250 22/08 A2989 RMK AO2 SLP121 T02220083 57010=
KIAH 231953Z 31011KT 10SM SCT039 BKN090 BKN250 23/13 A2990 RMK AO2 SLP123 T02330128=
KIAH 231853Z 32011G19KT 10SM SCT035 BKN080 BKN250 21/12 A2991 RMK AO2 SLP129 T02110117=
KIAH 231802Z 32015G24KT 10SM FEW035 SCT060 BKN080 BKN250 22/12 A2992 RMK AO2 T02220117=

[PJ2]

Concur with your observations, henra.

B2N2, they were still ~35nm away from HOU, so just keeping in mind, drawing no conclusions, Tim Vasquez's (https://www.pprune.org/rumours-news/618723-atlas-air-767-down-texas-5.html#post10399550) work near the start of the thread regarding cloud & turbulence in the aircraft's path and altitude approaching the Trinity Bay.

HOU ATC was apprising them of weather and possible detours. I'm sure the NTSB is examining what route all other aircraft in that area took to find out what their experience was. Same thing was done during the AF447 investigation.

[SteinarN]

Crossing the waypoints at 7000' or 6000' depending on the transition they were in VMC conditions.

Thanks.
In that case, less possible causes for pilot error at least.

[evansb]

Although not a factor in this crash, always remember ceiling heights reported from surface based weather observations, i.e. METARs, are AGL, (Above Ground Level), NOT above MSL (mean sea level). Forecast cloud heights in Area Weather Forecasts are ASL, (above sea level), unless noted.

Also remember wind direction in METARs are "True", NOT magnetic. Winds on the ATIS and winds reported by Tower are, or course , magnetic. There are exceptions to the reporting standard, usually reserved at sites located in extreme latitudes.

Curiously, the prevailing visibility as reported in most North American METARs are in STATUTE MILES. (a unit of linear measure equal to 5,280 feet, or 1,760 yards (approximately 1.609 kilometers, or 0 .86 of a nautical mile).

[Old Boeing Driver]

Failure of the elevator control/servo system putting the aircraft in a steep dive, aka 737 rudder hardover......
Mechanical breakage/failure of the HS screw jack system including its attachment structure making the HS freefloating, somewhat similar to Alaska MD-80......
........

These would be my 2 thoughts regarding the cause of this accident.

[positiverate20]

Apologies if this has already been answered, was the aircraft normal or inverted? I can't tell from the videos I've viewed.

Loss of control (for whatever reason)... followed by inversion, outside loop etc., it might give a different dimension/explanation as to why control surfaces were in unlikely positions.

[evansb]

Do you further speculate the failure of elevator control was due to turbulence?

[Old Boeing Driver]

Do you further speculate the failure of elevator control was due to turbulence?

Don't know if the question was for me, but I would not rule out the turbulence possibility. Early posts showed a possibility of some severe turbulence.

However, I would lean more to a basic structural failure by metal fatigue or corrosion.

No scientific reasons, just my thoughts.

[FCeng84]

The elevator is hinged at the leading edge, so the servo tab mounted at the trailing edge has the mechanical advantage or leverage to force the elevator to move up or down.
The stabilizer is hinged at the trailing edge with the actuator at the leading edge, so the position of the elevator mounted at the trailing edge of the stabilizer is going to have little or no effect on the position of a free floating stabilizer.
The stabilizer hinge line is slightly forward of the point where its trailing edge meets the fuselage. Due to stabilizer sweep angle, its hinge line is close to the center of lift such that horizontal tail angle-of-attack does not have a significant impact on the load applied to the jack screw. Due to the stabilizer sweep angle and the location of the elevators, deflection of the elevators generates a rotational torque on the stabilizer in the opposite direction. Note that all for the elevators are located aft of the horizontal stabilizer hinge line. The outboard end of the elevator is significantly aft of the stabilizer hinge line.

[golfyankeesierra]

Flaps would be 0 and the trim for 240KIAS about 1 or 2 units nose up

You sound very convinced and I never flew 76, but 1 or 2 units nose up sounds not much to me and more like nose down trim.
While low speed and flaps up, in any aircraft, would require quite some nose up trim.

[Educated Airman]

The stabilizer hinge line is slightly forward of the point where its trailing edge meets the fuselage. Due to stabilizer sweep angle, its hinge line is close to the center of lift such that horizontal tail angle-of-attack does not have a significant impact on the load applied to the jack screw. Due to the stabilizer sweep angle and the location of the elevators, deflection of the elevators generates a rotational torque on the stabilizer in the opposite direction. Note that all for the elevators are located aft of the horizontal stabilizer hinge line. The outboard end of the elevator is significantly aft of the stabilizer hinge line.

I used to fly skydivers in an old Cessna 182 that used stab trim. The only "locking" device for the jackscrew was the ball and spring detents on the trim wheel. The pivot point was at the aft spat well behind the center of lift. When I would youthfully speed up my decent by applying down elevator without trimming to achieve yellow line, the rotational force exerted by the elevators was enough to overcome the ball detent and I would get a sudden roll back of the trim wheel resulting in an unexpected 2.5-3 g pull up. With the 767 stab hinge closer to the center of lift, I would also expect control reversal caused by the elevators acting as servo tabs if the stab trim was mechanically disconnected.

[FIRESYSOK]

This ‘thread’ officially derailed pages ago. You all will be unceremoniously stunned when the findings are far, far simpler than a stripped stabilizer jackscrew.

[TheHardWay]

The stabilizer hinge line is slightly forward of the point where its trailing edge meets the fuselage. Due to stabilizer sweep angle, its hinge line is close to the center of lift such that horizontal tail angle-of-attack does not have a significant impact on the load applied to the jack screw. Due to the stabilizer sweep angle and the location of the elevators, deflection of the elevators generates a rotational torque on the stabilizer in the opposite direction. Note that all for the elevators are located aft of the horizontal stabilizer hinge line. The outboard end of the elevator is significantly aft of the stabilizer hinge line.

Thanks for taking the time to explain, FCeng84. What you say does make sense.

[Winemaker]

Forecast cloud heights in Area Weather Forecasts are ASL

The altitude of Houston is like 100± feet ASL.

[NSEU]

If the Statement:

"At full nose down (0.25 units), the top of the moving part (ballnut) is within an inch of the upper stop."

is correct, then the photographed screwjack is 80 to 90% Nose Down.

(look at the length of screwjack extending below the ballnut to the lower stop).

Whether this was how it was recovered, or if all the threads were stripped, would reduce the significance of the photo.

Let me add a little more information from the manual here:

0.00 units: ballnut is 0.5 inches from the top = stab leading edge up +2 degrees (full nose down trim with Alternate Trim?)
0.25 units: ballnut is 0.92 inches from the top = stab leading edge up +1.75 degrees. This is the nose down electric trim limit for the column switches with flaps not retracted
0.50 units: ballnut is 3 inches (?) from the top. This is the nose down electric trim limit for the column switches with the flaps retracted.
2.00 units (neutral): ballnut is 3.84 inches from the top.
12.8 units: ballnut is 22.43 inches from the top. This is the electric nose up limit for column switches. i..e 12.8 units = stab leading edge down minus 11 degrees
14.2 units: ballnut 24.96 inches from the top = stab leading edge down minus 12.5 degrees (full nose up trim for Alternate Trim)

I can't explain the irregularity in the manual with regards to inches versus units. Using up/down column switch limit values, 0.25 units is 0.92 inches and for every unit thereafter, there is an additional (approx) 1.79 inches. However, the 0.5 unit/3 inches value seems to be an aberration. It should be more like 1.3 inches. This 3 inch value is repeated throughout the manual.

For info only.

[PJ2]

Great information, thanks NSEU :ok:

[Machinbird]

About the only way to lose the nut engagement with a ball shaft is to lose the balls out of the nut.
These balls recirculate as the shaft turns, and anything that damages the recirculation pathway has the potential to cause disengagement of the nut from the ball drive shaft.
How could a recirculation pathway be damaged? There is always the potential for FOD in an aircraft. The right foreign object in the correct place and damage can happen despite the best efforts of the design engineers and maintainance engineers.

If the balls are missing from the previously pictured stabilizer actuator nut, the nut would be free to slide, and any position we see it presently would be coincidental.
Should the NTSB held FDR data shows a step jump in stabilizer position, they will be looking closely at this unit to identify the cause. But for now, it needs to sit untouched so as to preserve any potential clues as to what might have happened.
For those who have poor understanding of how a ballscrew system operates, here is an illustration:

https://youtu.be/zjXnv5Uzarw

[c_coder]

Seems needlessly complex. Why not use a simple nut? Sort of a sleeve with thread on the inside.

[sablatnic]

Seems needlessly complex. Why not use a simple nut? Sort of a sleeve with thread on the inside.
Friction and wear.
It is not something new, it was used in cars for many years.

[PerPurumTonantes]

Seems needlessly complex. Why not use a simple nut? Sort of a sleeve with thread on the inside.
Low friction
https://en.m.wikipedia.org/wiki/Ball_screw

[The Bartender]

Seems needlessly complex. Why not use a simple nut? Sort of a sleeve with thread on the inside.

https://en.m.wikipedia.org/wiki/Ball_screw

A simple nut would probably require frequent maintenance and regular replacement.

[bucoops]

Not on aircraft but I have specified ball screws on robotic production line equipment in a previous job for quite a harsh chemical environment and they were ridiculously reliable. Hundreds of thousands of cycles before noticeable wear without any maintenance at all.

[c_coder]

Okay so its like a worm drive, but with ball bearings.

Thanks for the info.

[henra]

This ‘thread’ officially derailed pages ago. You all will be unceremoniously stunned when the findings are far, far simpler than a stripped stabilizer jackscrew.
Looking at the Jackscrew assembly and the video I'm afraid you might be right….

[lpvapproach]

This ‘thread’ officially derailed pages ago. You all will be unceremoniously stunned when the findings are far, far simpler than a stripped stabilizer jackscrew.

If you are saying you know the cause of this crash - do please enlighten us?

[GlueBall]

???
I believe I see ~ 2-3 inches of spindle above the Ball nut. The silver part above the hinge appears to have the same color and diameter as the lower part. With 1,7" per unit and 0,25 units = max ND, I end up with ~2,5 -3 units. Seems like pretty much spot on for that speed and configuration.
That observation does not rally point into the direction of spindle/jackscrew failure nor runaway trim.
On top of that in the footage shown it appears indeed as if there were some kind of recovery attempt going on.
The speed being somewhere around 240kts this becomes more and more strange.
So at the moment it looks like:
- Flight at correct trim (assumption based on position of Jackscrew)
- At 240kts at 6000ft. (assumption based on FR24 data). (Probably Some 60-70kts above stall speed).
- Flight in IMC (Statement early in the process)
- Coming out of the clouds at high speed (>>250kts)
- Flightpath angle in the part visible in the footage initially ~40 - 50°ND, apparently possibly reducing to ~30 - 40°ND.
- Initial altitude in the video difficult to judge but in a wild guess somewhere in the range 2000 - 3000ft.

What plausible scenarios does that leave?
=> Stall as initiating event rather unlikely (too much speed margin, flight attitude pretty stable and straight, albeit hefty ND with rather lowish AoA in the footage)
=> Out of Trim situation somewhat unlikely if the position of the jackscrew represents the state it was in during the terminal phase of the flight.
=> failed jackscrew/ball nut can't be completely ruled out but is not really support by the picture of the jackscrew nor the apparent started recovery => rather unlikely.

At the moment I'm somewhat at a loss what could have reasonably caused that loss of Control.

...Excluding "unlikely" scenarios such as Germanwings 9525, Egyptair 990, Silk Air 185, etc . . . ?

[DaveReidUK]

Seems needlessly complex. Why not use a simple nut? Sort of a sleeve with thread on the inside.

Some aircraft do, the MD-80 for example. Google Acme screw.

[aircarver]

Seems needlessly complex. Why not use a simple nut? Sort of a sleeve with thread on the inside.

Alaska 261
https://cimg2.ibsrv.net/gimg/pprune.org-vbulletin/220x293/220px_screwshavings2_sm_233f47e3be8fe10c7b070236b98ba5b33ff4 f440.png
Stripped nut threads
https://cimg7.ibsrv.net/gimg/pprune.org-vbulletin/600x467/alaska_airlines_261_acme_nut_ntsb_aar_02_01_figure_16_61945e c9265a5fa1ceb3761e006cb31506c3198a.png
Stripped nut

[Lord Farringdon]

If the stabiliser jackscrew is such a prime suspect for this accident, as evidence by this thread, then why is it just sitting on the hangar floor with all the other wreckage? Surely, the NTSB would have been looking for it, retrieved it and sent it directly to the metallurgists for analysis? I can only assume it doesn't figure as highly in their minds as it does in the minds of PPruners? Or am I missing something?

[Raffles S.A.]

We, the Prunes, want to solve the mystery faster than it takes the NTSB.

[henra]

If the stabiliser jackscrew is such a prime suspect for this accident, as evidence by this thread, then why is it just sitting on the hangar floor with all the other wreckage?
I guess it is because that is the (honorable but also somewhat desperate) hope of many here that it was purely mechanical. If it wasn't the Jackscrew (and I do see not much indication that it would have been the Jackscrew, given the picture of the assembly and the footage) the chances of human factors playing a major role drastically increase.

[T28B]

If it wasn't the Jackscrew (and I do see not much indication that it would have been the Jackscrew, given the picture of the assembly and the footage) the chances of human factors playing a major role drastically increase.
/not as a mod
Or, the chances that something the PPRuNe Investigative Board have failed to consider increase.
We, the Prunes, want to solve the mystery faster than it takes the NTSB. Cheeky buggers, the PPRuNe Investigative Board.

[pattern_is_full]

Lord Farringdon - I actually asked myself that same question. And then gave it further thought.

And the conclusion I came to is: that the NTSB takes a methodical, organized and scientific approach to investigations. They aren't going to go haring off after any particular piece of evidence until they have collected all the evidence (and it appears that, at the time of the video, half the aircraft more or less was still out in the mud.)

It's my impression (with no direct experience) that the NTSB first and foremost sets out to determine exactly what happened - relying on evidence such as the videos and track data and coms we've also seen and heard, and also things we have not seen and heard here (FDR and CVR recordings if available; actual radar returns; impact pattern(s), witness descriptions, clock times, and so on). And only once they are sure they know what happened in detail, move on to investigating how and why it happened (which is where the physical jackscrew will - or won't - come in). At which point they will go over to the hangar where the parts are stored, and start testing any and all theories that have developed, against the physical evidence of aircraft parts recovered. As well as sifting through maintenance and crew records and other specifics.

Of course, at that point the physical evidence may suggest additional ideas, and the investigators may go back to the tapes, to see if there was an anomaly (sound, crew comment, control movement or disparity) that went unnoticed on the first reading.

In short, at this point, nothing "figures highly in their minds" - or doesn't "figure highly in their minds," one way or the other. The jackscrew isn't going anywhere - it can wait.

cf: "confirmation bias" - and read your Thomas Huxley: "Sit down before fact as a little child, be prepared to give up every conceived notion, follow humbly wherever and whatever abysses nature leads, or you will learn nothing."

[cappt]

I would estimate at least 3" inches between the nut and upper stop.

https://cimg7.ibsrv.net/gimg/pprune.org-vbulletin/458x309/opera_snapshot_2019_03_10_201020_www_youtube_com_2__faac8adc 69646125715c1841a6b1105ff824199c.png

[FlexibleResponse]

The jackscrew is attached to the stabilizer to change the datum trim point of the stabilizer.

In the wreckage photos, the jackscrew is no longer attached.

The question that will need to be answered is whether this failure of attachment happened pre or post-impact.

That is, did impact forces sever the attachment, or did the attachment structurally fail prior to impact and let the stabilizer float aerodynamically free?

[DaveReidUK]

The jackscrew is attached to the stabilizer to change the datum trim point of the stabilizer.

Yes, the THS/jackscrew combination will be familiar, of course, to anyone who has flown or worked on types such as the 737 or DC-9 over the past 50+ years.

In the wreckage photos, the jackscrew is no longer attached.

The question that will need to be answered is whether this failure of attachment happened pre or post-impact.

That is, did impact forces sever the attachment, or did the attachment structurally fail prior to impact and let the stabilizer float aerodynamically free?

I doubt that particular question will take very long to answer.

It would be very strange had the jackscrew remained attached to what's left of the stab following an impact like the one that occurred. Occam's Razor applies.

[Lord Farringdon]

Lord Farringdon - I actually asked myself that same question. And then gave it further thought.

And the conclusion I came to is: that the NTSB takes a methodical, organized and scientific approach to investigations. They aren't going to go haring off after any particular piece of evidence until they have collected all the evidence (and it appears that, at the time of the video, half the aircraft more or less was still out in the mud.)

It's my impression (with no direct experience) that the NTSB first and foremost sets out to determine exactly what happened - relying on evidence such as the videos and track data and coms we've also seen and heard, and also things we have not seen and heard here (FDR and CVR recordings if available; actual radar returns; impact pattern(s), witness descriptions, clock times, and so on). And only once they are sure they know what happened in detail, move on to investigating how and why it happened (which is where the physical jackscrew will - or won't - come in). At which point they will go over to the hangar where the parts are stored, and start testing any and all theories that have developed, against the physical evidence of aircraft parts recovered. As well as sifting through maintenance and crew records and other specifics.

Of course, at that point the physical evidence may suggest additional ideas, and the investigators may go back to the tapes, to see if there was an anomaly (sound, crew comment, control movement or disparity) that went unnoticed on the first reading.

In short, at this point, nothing "figures highly in their minds" - or doesn't "figure highly in their minds," one way or the other. The jackscrew isn't going anywhere - it can wait.

cf: "confirmation bias" - and read your Thomas Huxley: "Sit down before fact as a little child, be prepared to give up every conceived notion, follow humbly wherever and whatever abysses nature leads, or you will learn nothing."


I don't disagree. My only concern would be if this was a ticking time bomb for the aircraft flying right now. Wouldn't you want to defuse that pretty quick? But then again, it strongly suggests the NTSB are not concerned that it is. They will take it step by step. And so as you say, there the jackscrew will lie until they get to it.

[mostlylurking]

While the Jackscrew could be at fault, I'm not convinced it is.
I have a fair bit of experience of them in the manufacturing industry. They work in continuous high speed, high load applications for a decade or more without issue.
When they do fail it usually begins by developing endplay. If that symptom is ignored, they fail by sizing up solid (no amount of force will get them to move) because the balls become wedged in the return tube.
They do not strip and there is no warning,.

[Educated Airman]

I noticed the jackscrew attachment lug appears to be intact without the bolt in it. Attachment hardware is usually designed to be stronger than the structures they are installed in. I would have expected that if the jackscrew was separated due to impact, that the bolt, nut and fuselage mounted attachment lugs would be present. It would be interesting to see the bolt and nut and determine if the cotter pin had been installed. It is unlikely but not unheard of for this to occur during maintenance. I had a brake torque ling bolt fall out on a 707 that really made a mess of things upon landing due to a forgotten cotter pin.

[hawkerjet]

This accident has got me thinking about something else, such as putting the speed brakes out to slow down and get down quickly and disconnecting the auto throttle. At level off, forgetting to retract the speed brakes and re engaing auto throttles.. The aircraft could have entered a deep stall and given similar nose down profile as captured by the CCTV. I know the guys at the NTSB have their hands full and wish them all the best in finding the cause of this accident so we can all learn from it.

[Espada III]

Given the rapidity with which the AAIB and BEA gave advice regarding the tail rotor of the helicopter model involved in the Leicester City crash, I would assume that the jack screw will get some publicity if the NTSB think it appropriate.

[oceancrosser]

This accident has got me thinking about something else, such as putting the speed brakes out to slow down and get down quickly and disconnecting the auto throttle. At level off, forgetting to retract the speed brakes and re engaing auto throttles.. The aircraft could have entered a deep stall and given similar nose down profile as captured by the CCTV.

Why would you disconnect the Auto Thottle? That’s not how it is done.

[Airbubba]

Why would you disconnect the Auto Thottle? That’s not how it is done.


Agreed, you might get THR HOLD though.

As I posted earlier, it appears that they were late slowing to 250 below 10 so there may have been some speedbrake use just before the upset.

Still, the B-763 has been using autothrottles and speedbrake for three decades with few problems that I am aware of.

[threestable]

Agreed, you might get THR HOLD though.

As I posted earlier, it appears that they were late slowing to 250 below 10 so there may have been some speedbrake use just before the upset.
Throttle hold in descent.... If coming from a dual function mode (FLCH VNAV) autothrottle will go SPD automatically at level off. If speed brakes still out more thrust applied to maintain MCP speed. No need to disconnect autothrottle to get it down - would be back at the stops anyway.

[gums]

Salute!

I take it that inadvertant reverser on one or both engines is not a player? Not sure how to get to reverse thrust, but seems someone here can answer real quick.

Gums asks...

[Sailvi767]

This accident has got me thinking about something else, such as putting the speed brakes out to slow down and get down quickly and disconnecting the auto throttle. At level off, forgetting to retract the speed brakes and re engaing auto throttles.. The aircraft could have entered a deep stall and given similar nose down profile as captured by the CCTV. I know the guys at the NTSB have their hands full and wish them all the best in finding the cause of this accident so we can all learn from it.

No one does that on the 767. If you want to go down fast you simple select flight change and the throttles go to idle. If you want to help you can pull them to idle however the auto throttle system is still connected and will add power to maintain speed at the level off. I never saw the technique you mentioned used even once in 15,000 hours on the aircraft.

[lomapaseo]

Salute!

I take it that inadvertant reverser on one or both engines is not a player? Not sure how to get to reverse thrust, but seems someone here can answer real quick.

Gums asks...

In the air it's tough unless you short a couple of wires and use a screwdriver outside on the wing.

But just from examining the wreckage the investigators could tell if the reverser(s) were deployed, also the DFDR traces

[gums]

Salute. looms!

Thx. Kinda what I thot, but don’t have the documents like my friends here linked to me in the 447 thread.

Reason I asked was it seemed improbable to push over and reach the extreme dive angle without bouncing off the top of the cockpit. So I was considering a uncommaned reverser, then yaw, then roll inverted. Of course, the stab disconnect is also on my mind.

Gums sends....

P.S. Also in Florida in Niceville and watch two pure FBW jets every day, one of which even powers the actuators with electrons. If you look up their demos, they both do things that are aerodynamically impossible, he he.

[Duchess_Driver]

I take it that inadvertant reverser on one or both engines is not a player? Not sure how to get to reverse thrust, but seems someone here can answer real quick.

Not familiar with the 767 but if this were the case here I would suspect reverser unlock and deployment would have caused significant yaw as per the Lauda 767 out of Thailand some years back. IIRC in that instance the result was in flight break up. You have very little time to catch it. The video suggests no noticeable roll, pitch or yaw changes.

[Airbubba]

And, the Lauda 767 that had the inflight reverser deployment had Pratt engines, the Atlas plane was powered by GE's.

[NSEU]

I would estimate at least 3" inches between the nut and upper stop.

https://cimg7.ibsrv.net/gimg/pprune.org-vbulletin/458x309/opera_snapshot_2019_03_10_201020_www_youtube_com_2__faac8adc 69646125715c1841a6b1105ff824199c.png

Seems a little low. I've revised my earlier estimate downwards:

From 767-300 Maintenance Manual scale diagrams and the known maximum of dimension "A" (26.41 inches), ref AMM 27-41-18, the jackscrew is at least 3 inches in diameter. I see about 1.5 diameters of exposed thread. i.e. 4.5", which is about 0.5 degrees nose up (stabiliser position, not aircraft nose or AOA).

[hawkerjet]

OK, guys, I was just looking at different ways the aircraft could have been operated and stalled and that came to mind. My lack of knowledge on the B767 auto throttle system has glaringly showed. I guess thats why we use this open forum, to learn more. Thanks for all the positive comments.

[Roger_Murdock]

NTSB reporting column input full forward and engines to full. No conclusion yet, but this certainly narrows the possible causes.

[9gmax]

On Mar 12th 2019 the NTSB provided an update stating:

The wreckage was situated in a shallow muddy swamp area, and the main debris field was oriented east to west and about 350 yards long by about 200 yards wide (figure 1). One engine and some landing gear components were found beyond the main debris field to the west. Less dense components and a large portion of the cargo floated southward and were recovered up to 20 miles away.

The NTSB reported:

Air traffic control communications and radar data indicated the flight was normal from Miami to the Houston terminal area. About 12:30 pm the pilots contacted the Houston terminal radar approach control (TRACON) arrival controller and reported descending for runway 26L; the airplane was at 17,800 ft with a ground speed 320 knots.

At 12:34, the airplane was descending through 13,800 ft, and the controller advised of an area of light to heavy precipitation along the flight route and that they could expect vectors around the weather.

About 12:35, the flight was transferred to the Houston TRACON final controller, and the pilot reported they had received the Houston Automatic Terminal Information System weather broadcast. The controller told the pilots to expect vectors to runway 26L and asked if they wanted to go to the west or north of the weather.

Radar data indicated the airplane continued the descent through 12,000 ft with a ground speed of 290 knots, consistent with the arrival procedure. The pilots responded that they wanted to go to the west of the area of precipitation. The controller advised that to do so, they would need to descend to 3,000 ft expeditiously.

About 12:37, the controller instructed the pilots to turn to a heading of 270°. Radar data indicated the airplane turned, and the automatic dependent surveillance-broadcast (ADS-B) data indicated a selected heading of 270°. The airplane was descending through 8,500 ft at this time.

About 12:38, the controller informed the pilots that they would be past the area of weather in about 18 miles, that they could expect a turn to the north for a base leg to the approach to runway 26L, and that weather was clear west of the precipitation area. The pilots responded, “sounds good” and “ok.” At this time, radar and ADS-B returns indicated the airplane levelled briefly at 6,200 ft and then began a slight climb to 6,300 ft.

Also, about this time, the FDR data indicated that some small vertical accelerations consistent with the airplane entering turbulence. Shortly after, when the airplane’s indicated airspeed was steady about 230 knots, the engines increased to maximum thrust, and the airplane pitch increased to about 4° nose up and then rapidly pitched nose down to about 49° in response to column input. The stall warning (stick shaker) did not activate.

FDR, radar, and ADS-B data indicated that the airplane entered a rapid descent on a heading of 270°, reaching an airspeed of about 430 knots. A security camera video captured the airplane in a steep, generally wings-level attitude until impact with the swamp. FDR data indicated that the airplane gradually pitched up to about 20 degrees nose down during the descent.

[Airbubba]

Wow. :eek:

The only reasons I can think of that you would 'normally' firewall the throttles down low without a stall warning would be windshear or an EGWPS terrain escape maneuver.

[gums]

Salute!

Yeah, Bubba, looks like the CVR will provide more than the gruesome data plots.
I mentioned something to 'bird about a possible crew change of position nd then a bit of turbulence that moved bodies about. I am not ruling that out except for the power change.
gums...

[atakacs]

That is indeed a headscratcher...

​​​​​​Too bad nothing from the CVR

[Airbubba]

https://cimg7.ibsrv.net/gimg/pprune.org-vbulletin/1280x720/46636091114_2874fb88e9_o_e8035d731e72f124d7263e7645ae0c6e2e9 e3de5.jpg
WASHINGTON (March 12, 2019) — The descent of Atlas Air flight 3591 and the communication between air traffic control and the aircraft pilots on Feb. 23, 2019, is depicted in this graphic. Atlas Air flight 3591, carrying cargo for Amazon and the U.S. Postal Service, crashed about 40 miles from Houston’s George Bush Intercontinental Airport. Three people died in the crash. (NTSB Graphic)

This is in agreement with DaveReidUK's ADS-B data analysis posted here the day after the mishap:

The FR24 data, as usual, contains numerous artifacts and synchronisation issues.

That said, once cleaned up it appears to show a slight but unmistakeable climb interrupting the descent just before reaching 6000', starting about 10 seconds before the beginning of the final dive.

It's exaggerated, obviously, in this foreshortened view (apologies for the skewed verticals):

https://cimg5.ibsrv.net/gimg/pprune.org-vbulletin/464x592/5y3591a_a42ad9ae66c288d28615ee0d9c5146acd45a5f99.jpg

[Lonewolf_50]

Also, about this time, the FDR data indicated that some small vertical accelerations consistent with the airplane entering turbulence. Shortly after, when the airplane’s indicated airspeed was steady about 230 knots, the engines increased to maximum thrust, and the airplane pitch increased to about 4° nose up and then rapidly pitched nose down to about 49° in response to column input. The stall warning (stick shaker) did not activate.
FDR, radar, and ADS-B data indicated that the airplane entered a rapid descent on a heading of 270°, reaching an airspeed of about 430 knots. A security camera video captured the airplane in a steep, generally wings-level attitude until impact with the swamp. FDR data indicated that the airplane gradually pitched up to about 20 degrees nose down during the descent.
Question: if pilot slumps over due to incapacitation / heart attack, would the body movement and eventual pitch fit the above sequence of events?
Yeah, I am reaching, but that seems really weird to me.

[Airbubba]

There aren't many explanations that would push the throttles and yoke forward at the same time...

[pattern_is_full]

This is not even speculation, just a "scenario" that might fit that reported throttle and control-input reading.

Massive incapacitation of PF (bird through the windscreen, medical, other), body pitches forward onto both column and (hand) throttle levers. Nose-over forces make it impossible for the other pilot and jumpseater to clear the controls until too late (that possible last-second attempt to get the nose up).

EDIT: I see I'm not alone - but it would certainly require a "Black Swan" event.

[DaveReidUK]

There aren't many explanations that would push the throttles and yoke forward at the same time...

That might also account for why the NTSB aren't saying what's on the CVR.

[LookingForAJob]

Has it been stated that the CVR cannot provide additional information? As I read the release from the NTSB issued on 7th March it suggests that with filtering it may be possible to determine more elements of FD communications and environmental noise. Today's information offered by 9gmax, although I can't see it on the NTSB website, seems only to include ATC communications which are likely to have come from ground recordings. The earlier release talks about crew communications being consistent with a loss control of the aircraft....but this is not really consistent with FD actions. As with the ET accident, I find the wording of the information that is being issued by key agencies to be interesting - the text from NTSB refers to 'engines increased to maximum thrust', but does not mention TL movement. But, perhaps I am reading too much into this.

Hopefully there is enough information available or ultimately accessible to enable some of the present questions to be answered.

[Airbubba]

Hopefully there is enough information available or ultimately accessible to enable some of the present questions to be answered.

I strongly suspect that the answers to some of these questions are already known.

[derjodel]

FDR data indicated that the airplane gradually pitched up to about 20 degrees nose down during the descent.

Well, so much about ad-hominem attacks on my primitive video analysis in which I suggested they were indeed pulling up.

The excerpt of what NTSB is sharing is interesting and basically is saying „not Beoing‘s fault“ - important due to ET crash. However, it offers no explanation to what initiated the maneuver, which could be very unfair to the crew. Fly Dubai was also not a „black swan”.

[RoyHudd]

Sadly this looks likely to be a sudden and deliberate act.

[TowerDog]

Holy Cr@p..
Highjack, suicide, mental illness on the flight deck, Egypt 990 Allah Akbar, or what.
Hopefully none of the above..:sad:

[SteinarN]

An unexpected and sad turn in this accident....

[henra]

Sadly this looks likely to be a sudden and deliberate act.
Sadly there aren't too many other realistic scenarios. Still the Questions is: Germanwings or hefty spatial disorientation. The fact that they were IMC, entering turbulence and have stopped the descent very shortly before (which will create somatogravic illusions of climbing) plus the fact that they started to try and recover once visual with the ground makes me tend more to the latter scenario. Still a very sobering scenario.

[cappt]

There's still a lot of information to come out. This guy is a respected former investigator whom most likely is talking with former colleagues.

Gregory Allen "Greg" Feith is an American former Senior Air Safety Investigator with the National Transportation Safety Board (NTSB). Today on Denver News Channel 9, while commenting on the 737 Max crashes, Greg Feith said this, verbatim:

“...and a lot of carriers overseas, they are so automation dependent that they don’t know, based on their training, when to intervene, and if there is a problem they continue to try to use the automation. We’ve seen that now in three accidents. Lion Air, Ethiopian, and in fact, Atlas Air, the one that crashed in Houston. The automation was still coupled, the pilots didn’t hand fly the airplane when they lost control and even through the recovery they were fighting the automation.”

[Photonic]

Sadly this looks likely to be a sudden and deliberate act.

Final descending approach to an airport at the end of a flight, is an odd time to choose suicide. Either pilot incapacitation or some loss of control upset in turbulence seems far more likely to me.

[9gmax]

Has it been stated that the CVR cannot provide additional information? As I read the release from the NTSB issued on 7th March it suggests that with filtering it may be possible to determine more elements of FD communications and environmental noise. Today's information offered by 9gmax, although I can't see it on the NTSB website, seems only to include ATC communications which are likely to have come from ground recordings. The earlier release talks about crew communications being consistent with a loss control of the aircraft....but this is not really consistent with FD actions. As with the ET accident, I find the wording of the information that is being issued by key agencies to be interesting - the text from NTSB refers to 'engines increased to maximum thrust', but does not mention TL movement. But, perhaps I am reading too much into this.

Hopefully there is enough information available or ultimately accessible to enable some of the present questions to be answered.

link : https://www.ntsb.gov/investigations/pages/DCA19MA086.aspx

[LookingForAJob]

Ahh, thanks, I was looking for a news/media release.

[Machinbird]

Atlas Air, the one that crashed in Houston. The automation was still coupled, the pilots didn’t hand fly the airplane when they lost control and even through the recovery they were fighting the automation.”

Imagine that! The AP has a dive bomb mode.:suspect:

[SamYeager]

link : https://www.ntsb.gov/investigations/pages/DCA19MA086.aspx

The phraseology in part of that link now reads

The airplane then pitched nose down over the next 18 seconds to about 49° in response to nose-down elevator deflection. The stall warning (stick shaker) did not activate.

Previously the bolded part said "a control input".

[Airbubba]

The phraseology in part of that link now reads

The airplane then pitched nose down over the next 18 seconds to about 49° in response to nose-down elevator deflection. The stall warning (stick shaker) did not activate.

Previously the bolded part said "a control input".

Actually it said:

then rapidly pitched nose down to about 49° in response to column input.

Curiouser and curiouser… :confused:

[SteinarN]

So, it seems that the gist of this now is that the AP got haywire thinking the aircraft was a Stuka bomber attacking some ground forces. And the crew was not able to or didnt succed in disconnection the AP.

[Airbubba]

So, it seems that the gist of this now is that the AP got haywire thinking the aircraft was a Stuka bomber attacking some ground forces. And the crew was not able to or didnt succed in disconnection the AP.

Even then, what pushed up the autothrottles to max? Both autopilot and autothrottles can normally easily be disconnected and/or overpowered.

[wiedehopf]

I wouldn't make too much of that wording change.

Maybe it was a reaction to people saying they were blaming the pilots.

I mean they should have the column forces on the FDR readout.
Would surprise me if the person writing "in response to column input" didn't check the column force before writing it.

[LookingForAJob]

On the contrary, I’m inclined to think that if they went to the trouble of changing the wording it’d because it is in some way significant. Curiouser indeed.

[DaveReidUK]

I mean they should have the column forces on the FDR readout.

AFAIK, control column deflection and control force aren't mandatory FDR parameters per Annex 6.

So it may be that it was inferred from the elevator deflection (which is of course a mandatory parameter) and the inference subsequently corrected.

[ASRAAMTOO]

Does full power on the 767 generate much pitch up. In other words if you had run out of ideas to raise the nose might you try it? Clutching at straws here......

[Feathers McGraw]

A question for 767 knowledgeable crew/engineers. If one of the (two?) autopilots is engaged, and the control columns are pulled fairly/very hard, does the AP disengage at some level of force or does it require positive selection of a button or switch?

I'm thinking of the Eastern Tristar that crashed in the Everglades, in that case gentle pressure on one column was enough to disengage altitude hold but I don't know if that caused the autopilot to release all control surfaces or if it was in one axis. I recall reading that there was a quiet chime that wasn't noticed when this happened whereas I thought that an AP disengaging is indicated in a louder and more attention-getting manner.

I saw up thread that an investigator commented that the Atlas crew "fought the automation" but logic suggests that significant control movement is clearly different from wanting the AP to keep doing what it was previously commanded to do.

[SMT Member]

Descending In IMC, receives sudden W/S warning. Throttles goes to the stops, aircraft pitches up. Spatially disorientated pilot notices airspeed not building, pushes control column forward and holds it until things 'start feeling right'. Realises what's going on, hauls back on column but too late.

Don't know, just an idea.

[Homebrew1]

Does full power on the 767 generate much pitch up. In other words if you had run out of ideas to raise the nose might you try it? Clutching at straws here......

Absolutely and certainly an option if you've run out of nose up pitch authority.

[Homebrew1]

A question for 767 knowledgeable crew/engineers. If one of the (two?) autopilots is engaged, and the control columns are pulled fairly/very hard, does the AP disengage at some level of force or does it require positive selection of a button or switch?

I'm thinking of the Eastern Tristar that crashed in the Everglades, in that case gentle pressure on one column was enough to disengage altitude hold but I don't know if that caused the autopilot to release all control surfaces or if it was in one axis. I recall reading that there was a quiet chime that wasn't noticed when this happened whereas I thought that an AP disengaging is indicated in a louder and more attention-getting manner.

I saw up thread that an investigator commented that the Atlas crew "fought the automation" but logic suggests that significant control movement is clearly different from wanting the AP to keep doing what it was previously commanded to do.

Yes the AP's will disengage. Something to remember if you inadvertently ground engage the AP's on a 777 and wonder why it won't rotate at VR ;)

[wrench1]

On the contrary, I’m inclined to think that if they went to the trouble of changing the wording it’d because it is in some way significant. Curiouser indeed.
Not really. Shortly after the NTSB posted the original update with the "control column input" reference, several individuals on their twitter feed started pointing to "an intentional act." Subsequently, the update got updated again with the "column input" deleted. Just another example of social media leading the narrative vs the people who actually know.

[DaveReidUK]

Not really. Shortly after the NTSB posted the original update with the "control column input" reference, several individuals on their twitter feed started pointing to "an intentional act." Subsequently, the update got updated again with the "column input" deleted. Just another example of social media leading the narrative vs the people who actually know.

It's more likely that nobody knows. The NTSB didn't say that there was an absence of control column input. That suggests that they can't actually tell control column deflection from the FDR.

[tdracer]

This is not even speculation, just a "scenario" that might fit that reported throttle and control-input reading.

Massive incapacitation of PF (bird through the windscreen, medical, other), body pitches forward onto both column and (hand) throttle levers. Nose-over forces make it impossible for the other pilot and jumpseater to clear the controls until too late (that possible last-second attempt to get the nose up).

EDIT: I see I'm not alone - but it would certainly require a "Black Swan" event.

Pattern, I've been thinking sudden incapacitation for a couple of weeks (early on I postulated a large bird through the forward bulkhead as a possibility if you go back enough pages) - or something acute such as a massive heart attack could pitch the PF forward into the controls (granted, also moving the throttles makes it somewhat less likely - but SOP is to have a hand on the throttles during approach so not impossible). 18 seconds is not much time to overcome the startle factor, recognize and react to something completely unexpected - especially with a 200 lb. pilot laying on the yoke. In short, it wouldn't need to be deliberate, just sudden and unexpected.

AFAIK, control column deflection and control force aren't mandatory FDR parameters per Annex 6.
This was an early 1990s build aircraft - and as digital memory was getting cheaper the number of non-mandatory items on the DFDR really exploded in that time frame. Both a benefit and a curse when you were reviewing incident data - with so much data available it could make it a real challenge to sift the wheat from the chaff...

[DaveReidUK]

This was an early 1990s build aircraft - and as digital memory was getting cheaper the number of non-mandatory items on the DFDR really exploded in that time frame. Both a benefit and a curse when you were reviewing incident data - with so much data available it could make it a real challenge to sift the wheat from the chaff...

Are you saying that control column deflection was in fact an FDR parameter for the accident flight?

[flyingchanges]

Descending In IMC, receives sudden W/S warning. Throttles goes to the stops, aircraft pitches up. Spatially disorientated pilot notices airspeed not building, pushes control column forward and holds it until things 'start feeling right'. Realises what's going on, hauls back on column but too late.

Don't know, just an idea.

Pretty sure all wind shear warnings are inhibited at that altitude.

[PJ2]

Are you saying that control column deflection was in fact an FDR parameter for the accident flight?

PMFJI, it could be a parameter, along with the force in pounds depending upon the data frame. The B737-400's -5 data frame has control column and control wheel position. This aircraft's data frame will likely have those parameters, and possibly also control column/wheel force in pounds.

[Airbubba]

Pretty sure all wind shear warnings are inhibited at that altitude.

I think they are inhibited above 1500 feet AGL on the classic 767. Also, you wouldn't get an EGWPS alert at 6000 feet over East Texas. But like everybody else, I'm pulling at straws trying to think of why you would go to max thrust before the pitch over.

I suspect that the NTSB has a good guess between the CVR and FDR but need to be cagey in their public release of information until they have confirmation.

From today's NTSB update:

A cockpit voice recorder (CVR) group was convened and will complete a transcript of the entire event. The CVR transcript will be released when the public docket is opened.

Unfortunately, it is usually months before the public docket is opened in a major investigation. :uhoh:

[Chu Chu]

Having worked in a bureaucracy, the most likely explanation is that someone accidentally posted a version of the press release that didn't include all the edits from the review process (and then corrected their mistake). It seems hard to believe that NTSB is both nimble enough to revise its press release in response to speculation on Twitter, and at the same time stupid enough to think that the revision would make things better.

It also seems hard to believe that those words would have made it into any version unless the FDR data showed column deflection and/or force. Maybe NTSB decided it should formally eliminate any other possible explanation for those readings before publishing the conclusion that there was actual control column input. Or it just didn't want to feed too much speculation ahead of the report (but accidentally did so anyway).

[The Dominican]

Column, elevator... Whatever the intended edited version went out, I think the Human Behavior team is very busy at this time.

[BFSGrad]

From the EgyptAir 990 NTSB accident brief:

p. 25: “The FDR installed on the accident airplane was a Sundstrand Data Corporation (now named Honeywell Aerospace Electronic Systems) Universal Flight Data Recorder…”

“Flight performance parameters recorded by the FDR included the following: pressure altitude; airspeed (computed); engine rpm; pitch; roll; heading; angle of attack; normal (vertical), longitudinal, and lateral acceleration (load factors); left and right elevator positions; left and right inboard and outboard aileron positions; left and right trailing edge flap positions; rudder position; and horizontal stabilizer position.”

“The FDR was not required to and did not record control wheel, control column, or spoiler positions nor did it record control wheel and column forces.”

“Although control column position was not recorded by the FDR, the Safety Board’s testing and evaluation of the 767 elevator system showed that any movement occurring at the control columns would have resulted in concurrent, identifiable movements of the elevators, which would have been recorded on the FDR.”