Go Back  PPRuNe Forums > Flight Deck Forums > Tech Log
Reload this Page >

Maintenance Lapse Identified as Initial Problem Leading to Lion Air Crash

Tech Log The very best in practical technical discussion on the web

Maintenance Lapse Identified as Initial Problem Leading to Lion Air Crash

Old 2nd Jan 2019, 13:45
  #61 (permalink)  
 
Join Date: Jan 2008
Location: Reading, UK
Posts: 10,507
Originally Posted by infrequentflyer789 View Post
KNKT also shows the four sectors from the maint log (in the prelim report and posted in full further up the thread):

26 October 2018 :: Tianjin Binhai to Manado
27 October 2018 :: Denpasar to Manado
27 October 2018 :: Manado to Denpasar
28 October 2018 :: Denpasar to Jakarta

Methinks that Occam's razor comes down on the side of "four consecutive" being a translation error, as opposed to teleportation of a 737 from Manado to Denpasar - I could be wrong though...
I'm quoting from the statement made by the KNKT on 5th November, where they identified problems encountered on four consecutive flights (including the accident flight):

(Denpasar)-Manado
Manado-Denpasar
Denpasar-Jakarta
Jakarta-Pangkal Pinang

KNKT: Lion Air Airspeed Indicator Damaged since 3 Flights Before
DaveReidUK is online now  
Old 2nd Jan 2019, 14:03
  #62 (permalink)  
 
Join Date: Jun 2009
Location: florida
Age: 76
Posts: 1,059
Salute!
and Happy New Year!

@ 'bird, et al..... re: digitizing the analogs. Even way back in the early 90's most of our analog signals were digitized by single purpose chips. A very few really "special" systems would have the A2D conversion circuits as part of a much more complex chip/ckt board versus separate components. In a very few cases dealing with analog sensors that had no A2D data used by more than one subsystem ( think guidance seekers, we had to "capture" the signals before they got to the "user" box in order to properly troubleshoot.

The previous gripes seem to all concern airspeed disagreement and such, and the AoA faults only showed up when maintenance codes were examined. As I understand it, some of the systems use AoA to "condition" the airspeed indications and such. Someplace between those values used for displays/STS and the AoA source should bear the brunt of investigation besides a "simple" mounting error. Considering that the speed gripes were there before the AoA vane change, I tend to rule out the sensor itself and its internal electronics.

Is it the sensor? Is it a BB rolling around on the ckt board? Is it a dip switch setting that lets the A2D processing account for left versus right sensor? And so forth.

So I like looking at the SMYD for the problem. Hell. just yank the thing out and replace! Not possible now, but we may see a similar sequence in another plane and we could keep this one in mind if a gripe shows up two times in a row.

Gums sends..

Last edited by gums; 2nd Jan 2019 at 14:16. Reason: revised wording, typos
gums is offline  
Old 2nd Jan 2019, 16:29
  #63 (permalink)  
 
Join Date: Feb 2009
Location: Seattle
Posts: 379
AOA vane angle is converted to body angle

Originally Posted by Machinbird View Post
Conso, either find a picture of the nose of a NG/MAX 737 and count the screw holes or go to the airport and make your own picture. I doubt if it has more than 6.
They can limit the interchangeability of probes by playing with the screw hole spacing, the mounting plate size, the electrical connectors, etc. There is an indexing pin involved as well, and that can control proper installation too.
With 9 holes you you would have a 40 degree error. That doesn't fit the scenario.
When considering the potential for AOA sensor installation to have played a role in this event, it is important to recognize the difference between AOA vane angle and AOA body angle. The local flow that the vane measures is distorted by the fuselage. As a result, the vane angle changes much more than one degree for each degree change in airplane AOA. Within the AOA signal processing logic is conversion of the vane angle back to the corresponding body angle. This conversion has a gain factor of just about 0.5 to get from vane to body.

The data that we have seen for the Lion Air event shows an AOA bias of approximately 20 degrees body between left and right signals. That could result from the difference in vane angles being approximately 40 degrees.
FCeng84 is offline  
Old 2nd Jan 2019, 18:24
  #64 (permalink)  
 
Join Date: Mar 2014
Location: WA STATE
Age: 73
Posts: 1
RE FCeng84 # 63

The data that we have seen for the Lion Air event shows an AOA bias of approximately 20 degrees body between left and right signals. That could result from the difference in vane angles being approximately 40 degrees.
Interesting - a pic of " typical " AOA Sensor shows 8 holes in post # 58

If 9 holes then 360/9 = 40 degrees but if 8 holes then 360/8 = 45degrees
assuming a plus or minus travel gives either 20 degrees or 22.5 degrees. which IMHO gets close to the constant 20 degree offset noted due to fit it or force it miss- installation ( clocking) .
CONSO is offline  
Old 2nd Jan 2019, 19:20
  #65 (permalink)  
 
Join Date: Jan 2008
Location: Reading, UK
Posts: 10,507
If only Boeing had thought to Murphy-proof the AoA sensor, say by incorporating a locating pin that ensured it could only be fitted in one orientation ...

I guess we've just been lucky that before this event nobody has fitted one incorrectly in the 50-year history of the 737.
DaveReidUK is online now  
Old 2nd Jan 2019, 19:56
  #66 (permalink)  
 
Join Date: Jun 2009
Location: florida
Age: 76
Posts: 1,059
Salute!

according to FCeng

When considering the potential for AOA sensor installation to have played a role in this event, it is important to recognize the difference between AOA vane angle and AOA body angle. The local flow that the vane measures is distorted by the fuselage. As a result, the vane angle changes much more than one degree for each degree change in airplane AOA. Within the AOA signal processing logic is conversion of the vane angle back to the corresponding body angle. This conversion has a gain factor of just about 0.5 to get from vane to body.
Well, FCeng says that the vane deflection requires a "0.5 gain factor" someplace in the sftwe or A2D chip/device, wherever that is.. Is the "gain" two degrees of wing chord to free stream flow? Is it sensed AoA times two or divided by two? Hard to believe the AoA difference between the sensor and the wing chord remains constant when airflow about the nose doesn't seem to be at a constant angle when the plane changes the actual AoA, so a linear "correction" seems too easy. Who does that signal processing? And most important....and according to latest diagram we have of the "system"....

Our diagram shows analog AoA is sent from the SMYD and thence to the stick shaker. But it has been stated that AoA is also used to "correct" airspeed and maybe that value is used for the STS versus raw pitot dynamic pressure. Where do the ADIRU's , cockpit displays and FCC boxes and the MCAS get AoA? From the SMYD? . There must be plenty of room for mischief here. Ya think?

Gums sends...
gums is offline  
Old 2nd Jan 2019, 19:57
  #67 (permalink)  
 
Join Date: Jan 2006
Location: Sydney Australia
Posts: 2,068
Originally Posted by DaveReidUK View Post
If only Boeing had thought to Murphy-proof the AoA sensor, say by incorporating a locating pin that ensured it could only be fitted in one orientation ...

I guess we've just been lucky that before this event nobody has fitted one incorrectly in the 50-year history of the 737.
More to the point: If only Boeing had thought not to Murphy-enable the MAX by introducing a system, whereby the aircraft will attempt to destroy itself (along with everyone on board) if the right set of conditions existed.

As far as I know, this particular outcome is unique to the 737 MAX. So, the statement regarding being “lucky” over the last 50 years really doesn’t apply. I’ll even go further and suggest that over the past 50 years it is probable, or even certain that a similar maintenance FUBAR would have occurred, given the the enormous amount of hours flown by the type.

I know this is ”Tech Log”, but the vast discussions surrounding the almost mind bending set of variables, only serves (IMHO) to muddy the waters as to the ultimate cause of this tragedy. An elephant in the room, that many on this, or any other forum, seem unable, or unwilling, to come to grips with?


KRUSTY 34 is offline  
Old 2nd Jan 2019, 20:47
  #68 (permalink)  
 
Join Date: Jun 2009
Location: florida
Age: 76
Posts: 1,059
Salute Krusty!

so....
I know this is ”Tech Log”, but the vast discussions surrounding the almost mind bending set of variables, only serves (IMHO) to muddy the waters as to the ultimate cause of this tragedy. An elephant in the room, that many on this, or any other forum, seem unable, or unwilling, to come to grips with?
And what would that elephant be, Krusty?

More to the point: If only Boeing had thought not to Murphy-enable the MAX by introducing a system, whereby the aircraft will attempt to destroy itself (along with everyone on board) if the right set of conditions existed.
There are libel and slander laws that come into play, although I feel many of us here could make assertions and render opinions without facing the law, but sometimes we need to leave our escape door open. And BTW, Krusty, I am more in your camp than nit-picking maintenance procedures, simple mis-alignment, BB's on the ckt boards, etc. Nevertheless, I am exercising caution by looking for something that can be corrected easier than the management decisions of a major aerospace company.

Gums sends...
gums is offline  
Old 2nd Jan 2019, 21:11
  #69 (permalink)  
 
Join Date: Jan 2008
Location: Reading, UK
Posts: 10,507
Originally Posted by KRUSTY 34 View Post
I know this is ”Tech Log”, but the vast discussions surrounding the almost mind bending set of variables, only serves (IMHO) to muddy the waters as to the ultimate cause of this tragedy.
The waters were already muddied before we started this discussion . We can take it as read that this isn't one of those very rare accidents that turn out to have a single, straightforward "ultimate cause".

I would expect, when the investigation report finally appears, that there will be both multiple factors listed under "probable cause" and several other "contributory factors" (that's NTSB-speak, but most AIBs make a similar distinction).

One finding I don't expect to see is that the engineer replacing the AoA probe either sawed off the locating pin or hammered it through the skin of the 737.
DaveReidUK is online now  
Old 2nd Jan 2019, 21:31
  #70 (permalink)  
 
Join Date: Dec 2002
Location: UK
Posts: 1,832
CONSO, et al; if you are looking for a number, then reconsider the two charts “40% shift in SIN value” and “a mix of positive offset SIN and negative offset to COS” in https://www.satcom.guru/2018/12/angl...ure-modes.html
”…about a 20 degree shifted output greater than normal… offsets to SIN and COS can cause shifts in the resolver output. These shifts either maintain nearly a constant value across the angular range ….”
safetypee is offline  
Old 2nd Jan 2019, 21:38
  #71 (permalink)  
 
Join Date: Jul 2013
Location: Everett, WA
Age: 64
Posts: 2,332
Originally Posted by DaveReidUK View Post
If only Boeing had thought to Murphy-proof the AoA sensor, say by incorporating a locating pin that ensured it could only be fitted in one orientation ...
What makes you so sure they didn't? Murphy can be extraordinarily creative. In the early days of the PW4000, it used a motor driven actuator for the high pressure fuel shutoff in the fuel control - and for some reason it wasn't the same as the one they used on the JT9D - basically RUN and CUTOFF were reversed. To prevent use of the JT9D actuator, they changed the locating pin location. Turned out that due to the vibration environment on the PW4000, the actuators failed quickly and often (quickly replaced by a solenoid setup that they should have used in the first place).
We got more than one JT9D motor actuator back - with complaints that it didn't work on the PW4000 - with the locating pins neatly machined off so it would fit.
We saw similar issues with throttle resolvers - the locating mechanism to prevent left/right swap being neatly machined off. I think it was Wernher von Braun who said 'It's really hard to make something idiot proof, because the idiots are so creative.'

I can't help but wonder if for some reason they changed the AOA sensor for the MAX (with appropriate Murphy Proofing). When the Lion Air mechanics determined that the NG sensor wouldn't fit the MAX, they "made it fit".
tdracer is online now  
Old 2nd Jan 2019, 21:44
  #72 (permalink)  
 
Join Date: Jan 2006
Location: Sydney Australia
Posts: 2,068
Red face

[QUOTE=gums;10349998]Salute Krusty!

so....


And what would that elephant be, Krusty?



I’m really not sure how to answer that gums!?
KRUSTY 34 is offline  
Old 2nd Jan 2019, 21:57
  #73 (permalink)  
 
Join Date: Jan 2008
Location: Reading, UK
Posts: 10,507
Originally Posted by tdracer View Post
Originally Posted by DaveReidUK View Post
If only Boeing had thought to Murphy-proof the AoA sensor, say by incorporating a locating pin that ensured it could only be fitted in one orientation ....
What makes you so sure they didn't?
I was being sarcastic - of course the AoA sensor is Murphy-proofed (as several posters have previously confirmed).

I can't help but wonder if for some reason they changed the AOA sensor for the MAX (with appropriate Murphy Proofing). When the Lion Air mechanics determined that the NG sensor wouldn't fit the MAX, they "made it fit".
While I don't know for a fact, I would be extremely surprised if the NG and MAX sensors were different. Besides, if they were, given that they are serialised rotable components, the investigation would undoubtedly have been able to ascertain by now if the wrong part number had been fitted.

And, while I may be biased from my years spent in airline engineering, I simply don't believe that any LAE, anywhere in the world, would "make a part fit" if it clearly didn't.
DaveReidUK is online now  
Old 2nd Jan 2019, 22:02
  #74 (permalink)  
 
Join Date: Jan 2006
Location: Sydney Australia
Posts: 2,068
Originally Posted by DaveReidUK View Post
The waters were already muddied before we started this discussion . We can take it as read that this isn't one of those very rare accidents that turn out to have a single, straightforward "ultimate cause".

I would expect, when the investigation report finally appears, that there will be both multiple factors listed under "probable cause" and several other "contributory factors" (that's NTSB-speak, but most AIBs make a similar distinction).

One finding I don't expect to see is that the engineer replacing the AoA probe either sawed off the locating pin or hammered it through the skin of the 737.
Agree with you 100% on that score Dave. Quite often the prevention of the last hole lining up though, is the action(s) of the crew. I doubt that anyone here with at least a reasonable amount of experience hasn’t seen this in one form or another. I know I have.

There can can be no doubt that the process leading up to this disaster is complex, perhaps even complicated! I have always believed though that when the automation goes haywire, disconnect and fly attitude + power to = performance. If that hapless crew did that on the day, and they still all died, then WTF!

KRUSTY 34 is offline  
Old 2nd Jan 2019, 22:33
  #75 (permalink)  
 
Join Date: Mar 2014
Location: WA STATE
Age: 73
Posts: 1
seems to me if one carefully reads the entire link
https://www.satcom.guru/2018/12/angl...ure-modes.html

it comes down to Electrical issue unlikely, mechanical issue probable. OK. mis-clocking when attached is a possibility- but perhaps unlikely ? . Seems to leave a clocking between vane and resolver shaft as a possible scenario. A) BUT the probability of two units being factory assembled wrong being used on the same side of the same airplane IMO has way to many zeros behind the decimal point. (but NOT impossible ). B) So how might the shaft be attached to the vane ? C) Or is it one solid piece and the mounting of the internal- attached coil possible to be incorrect - but that gets us back to A) .

Or is occams razor double sided ? Looks like recovery of the AOA sensor and how attached to left side may be the only way to resolve ....
CONSO is offline  
Old 2nd Jan 2019, 22:44
  #76 (permalink)  
 
Join Date: Feb 2009
Location: Seattle
Posts: 379
Originally Posted by gums View Post
Salute!

according to FCeng



Well, FCeng says that the vane deflection requires a "0.5 gain factor" someplace in the sftwe or A2D chip/device, wherever that is.. Is the "gain" two degrees of wing chord to free stream flow? Is it sensed AoA times two or divided by two? Hard to believe the AoA difference between the sensor and the wing chord remains constant when airflow about the nose doesn't seem to be at a constant angle when the plane changes the actual AoA, so a linear "correction" seems too easy. Who does that signal processing? And most important....and according to latest diagram we have of the "system"....

Our diagram shows analog AoA is sent from the SMYD and thence to the stick shaker. But it has been stated that AoA is also used to "correct" airspeed and maybe that value is used for the STS versus raw pitot dynamic pressure. Where do the ADIRU's , cockpit displays and FCC boxes and the MCAS get AoA? From the SMYD? . There must be plenty of room for mischief here. Ya think?

Gums sends...
Gums - 737 Max conversion from AOA vane to AOA body is an approximation that uses a scale factor of slightly more than 0.5 plus a small second order term and a constant offset. This is clearly a simplified fit to a non-linear relationship that was designed to be sufficiently accurate over the range of interest. A vane movement of 2 degrees equates to a free stream AOA change of about 1 degree. The sensor base plate has eight holes and a pair of alignment pins that look from pictures I have seen to be located roughly fore / aft along the horizontal line of symmetry of the circular base plate. I do not know the details about whether or not installation is critical with regard to which pin is forward and which pin is aft. I suspect that the same sensor is designed to work on either side of the plane so there must be something in the wiring or signal processing down stream that differentiates between left and right and sorts out whether clockwise motion is positive or negative AOA change.
FCeng84 is offline  
Old 3rd Jan 2019, 00:02
  #77 (permalink)  
 
Join Date: Jun 2009
Location: florida
Age: 76
Posts: 1,059
Salute FC eng

Thanks for the clarification.
Seems the folks on this thread are looking downstream of the actual sensor/vane to advocate an electrical or signal processing factor versus a poorly installed AoA vane.
as you said:
I suspect that the same sensor is designed to work on either side of the plane so there must be something in the wiring or signal processing down stream that differentiates between left and right and sorts out whether clockwise motion is positive or negative AOA change.
Thnks again,

Gums...
gums is offline  
Old 3rd Jan 2019, 01:30
  #78 (permalink)  
 
Join Date: Dec 2018
Location: 22L
Posts: 3
Originally Posted by FCeng84 View Post
Gums - 737 Max conversion from AOA vane to AOA body is an approximation that uses a scale factor of slightly more than 0.5 plus a small second order term and a constant offset. This is clearly a simplified fit to a non-linear relationship that was designed to be sufficiently accurate over the range of interest. A vane movement of 2 degrees equates to a free stream AOA change of about 1 degree. The sensor base plate has eight holes and a pair of alignment pins that look from pictures I have seen to be located roughly fore / aft along the horizontal line of symmetry of the circular base plate. I do not know the details about whether or not installation is critical with regard to which pin is forward and which pin is aft. I suspect that the same sensor is designed to work on either side of the plane so there must be something in the wiring or signal processing down stream that differentiates between left and right and sorts out whether clockwise motion is positive or negative AOA change.
I'm confused by this. When installing an AOA a reference check is done against the aoa angle and smyd. The quick check uses three points. The upper aoa stop is 100deg, middle is 0 deg and lower stop is -100 deg. These correspond to actual angles as measured at the aoa and as seen on the smyd display. I'm confused as to why a conversion would be needed.

The aoa cannot be installed backwards as it only has 200 degrees of movement. Symd 1 and 2 contain different functionality so I would think the processing of aoa angles is done there with regards to the aoa being in the left or right position.
Bleedtrip is offline  
Old 3rd Jan 2019, 02:49
  #79 (permalink)  
 
Join Date: Jul 2013
Location: Everett, WA
Age: 64
Posts: 2,332
Originally Posted by DaveReidUK View Post
I was being sarcastic - of course the AoA sensor is Murphy-proofed (as several posters have previously confirmed).

And, while I may be biased from my years spent in airline engineering, I simply don't believe that any LAE, anywhere in the world, would "make a part fit" if it clearly didn't.
Sadly, during the course of my career I've seen several examples where they did exactly that - a couple of which I outlined in my previous post (admittedly none of the cases I'm familiar with had bad outcomes, there was a certain amount of luck involved).
I lived in Indonesia for the better part of a year in the mid 1990s - working with what was then IPTN on new aircraft development. I saw things going on there in the manufacturing arena that were down right scary - and had to go to the chief engineer to ground an airplane one time because the lightning protection was so flawed as to be blatantly unsafe (to his credit, once I explained the risk he followed my recommendations).
tdracer is online now  
Old 3rd Jan 2019, 02:51
  #80 (permalink)  
 
Join Date: Mar 2014
Location: WA STATE
Age: 73
Posts: 1
Question

Originally Posted by gums View Post
Salute FC eng
Thanks for the clarification.
Seems the folks on this thread are looking downstream of the actual sensor/vane to advocate an electrical or signal processing factor versus a poorly installed AoA vane.
as you said:
Thnks again,
Gums...
Seems to this SLF that if it was a simple clockwise versus counterclockwise type change, for the issue a hand, one vane would show 20 degrees pitch up and the other would show pitch down- BUT only one showed a bias in the ' wrong' direction - ???
CONSO is offline  

Thread Tools
Search this Thread

Contact Us Archive Advertising Cookie Policy Privacy Statement Terms of Service

Copyright © 2018 MH Sub I, LLC dba Internet Brands. All rights reserved. Use of this site indicates your consent to the Terms of Use.