Go Back  PPRuNe Forums > Flight Deck Forums > Rumours & News
Reload this Page >

B-738 Crash in Russia Rostov-on-Don

Rumours & News Reporting Points that may affect our jobs or lives as professional pilots. Also, items that may be of interest to professional pilots.

B-738 Crash in Russia Rostov-on-Don

Old 26th Mar 2016, 02:46
  #721 (permalink)  
 
Join Date: Apr 2010
Location: St. Petersburg
Posts: 272
The media message, in essence, is:
"We do not know nothing and feel obliged to warn you about it. However, we managed to obtain a textual representation of the voices' record, what was heard - we were let to know about the last seconds of the record. There were 3 phrases there followed by a shout. The phrases were this, and that, translated into Russian, the shout we won't let you know of for ethic considerations.
Based on that, we deduced that "large atmospheric scissors" happened to the plane, together with mis-play of the "button of the joystick / button on the joystick / button of the button/ no buttons, joystick. - anyway, what the aviators call "knuppel" - either in the result of the accidental touch of it by the pilots or it was the "automation" itself. At 01:40:40, by Greenwich, sharp."

Last edited by Alice025; 26th Mar 2016 at 03:04.
Alice025 is offline  
Old 26th Mar 2016, 04:27
  #722 (permalink)  
 
Join Date: Feb 2009
Location: N/A
Posts: 4
Russian TV at it's best :-/

CVR translated to Russian with some noise added so it sounds like the real recording, with blame speculations to follow...

Anyway, the word "Knuppel" sounds in a sentence "Stabiliser is activated by a button that pilots call Knuppel"
The word "button" goes at the moment when CGI'ed pilot's left thumb flicks the switch on the CGI'ed yoke.
SergeyKa is offline  
Old 26th Mar 2016, 04:40
  #723 (permalink)  
 
Join Date: May 2012
Location: Virginia
Posts: 9
In what way does the FR24 data contraindicate a significant L or R roll? Honest question as all I see is vert. velocity and ground speed. It will help me better understand the data as it is.
mommaklee is offline  
Old 26th Mar 2016, 04:45
  #724 (permalink)  
 
Join Date: Aug 2008
Location: Betwixt and between
Posts: 666
01:40:50GMT ‘Pull! Pull! Pull!
01:40:54GMT ‘Aaaaaa’
What is this? The crash happened at 00:42Z (03:42L). (Edited: this can only mean 01:40 mins into the CVR recording).

Also, the notion that the pilot accidentally held the trim button down to such an extreme degree is something i find very very hard to believe. It is actually quite ridiculous. It is much more likely that, in a dark noisy 737 cockpit, the continuous and uncommanded motion of the trim wheel could have easily not be noticed,especially in rough air and considering that many crew wear ANC headsets to protect their hearing which do attenuate all cockpit noises to varying degrees.

Last edited by Sciolistes; 26th Mar 2016 at 04:59.
Sciolistes is offline  
Old 26th Mar 2016, 05:11
  #725 (permalink)  
 
Join Date: Apr 2009
Location: Oakland, CA
Age: 68
Posts: 427
I agree about this accidental pressing on the switch but I also find this uncommanded motion of trim very hard to swallow. We better have a theory why their vertical speed varied widely from +6000 fpm down to +2000 fpm and then up again to +4000 fpm (all within about 75 sec) to finally plummet to negative values. I don't see how uncommanded trim motion in the same direction would help explaining that.
olasek is offline  
Old 26th Mar 2016, 05:19
  #726 (permalink)  
 
Join Date: May 2012
Location: Virginia
Posts: 9
Quick Question

Also really quick: I am wondering what happened about 30 sec prior to the last data point, and about 32 sec prior to impact

I plotted the comparison between first and second attempts relative to when each attempt both descended through 1500' GPS altitude. (Point in question is near 65 sec past in second attempt.)

Images via Flickr:

GPS altitude and estimated pitch: https://flic.kr/p/FEWW2p
Est Vert Accel and est pitch: https://flic.kr/p/EKvmjG

If any are interested in the calcs I did (hopefully correctly), I'm happy to share the Excel file.
mommaklee is offline  
Old 26th Mar 2016, 05:27
  #727 (permalink)  
 
Join Date: Jan 2011
Location: island
Posts: 1
Wink

Originally Posted by KKN_ View Post
Whatever stab problems one can think of (technical, inputs), this alone seems not to explain the notable bank angle inferred from CCTV recordings.
Originally Posted by DaveReidUK View Post
The FR24 data does not support the hypothesis that the GA involved any significant bank angle.
Originally Posted by enola-gay View Post
But the video footage does. The video at post #298 shows 2 landing lights in view of the ground camera situated at right angle to the descent path. It looks like a high bank angle as the a/c impacted

This question is equally puzzling to me:

How could the plane have gone from 0 bank angle to 45-90 bank angle
WITHOUT lateral course deviation ?



FZ981 was well lined up with the runway on second/final approach, hence ~0 bank angle:




Even AFTER "start of dive" FZ981 continues on its straight ground track without almost
any lateral deviation. Here better seen from above:




But during the fatal dive, FZ981 had a bank angle somewhere between 45-90 (probably
closer to 90), as can be seen in the various videos posted here:

Originally Posted by Wrist Watch View Post

http://youtu.be/9aBq8saXy2M

Pause this video in a precise moment between 0:06 and 0:07 so that aircraft in fall (or rather its lights) can be seen shortly becore impact. Observe the position of landing lights. Conclude the bank angle being close to 90 degrees left, with a steeply negative pitch attitude.



Nevertheless during [final approach + GA + fatal dive] FZ981 doesn't sensibly deviate
from its straight ground track, which can not only be seen in image #2 above but is also
corroborated by the plane crashing onto the runway (hence no lateral deviation):




avherald says: "Radar data suggest the aircraft on final approach was to the left of the localizer and just to the
left of the left runway edge and corrected to the right while over the runway bringing the aircraft just within the
runway edges moments before ground contact."





So how does the plane go from 0 to 45-90 bank angle without lateral deviation ?
Forward motion provided, a passenger jet makes a turn when it banks, isn't it ?
  • the plane banked (see image #3 above)
  • the plane was in forward motion: 200 kts when starting to bank and it still kept some forward motion during the final dive (see images below)
  • yet the plane didn't make a turn (image #2) but crashed onto the runway (image #5)
how come ?


images showing forward motion (and banking) throughout the final dive:








(Google Earth Images produced by IG member Victor Ianello)
lovelove is offline  
Old 26th Mar 2016, 05:43
  #728 (permalink)  
 
Join Date: Aug 2008
Location: Betwixt and between
Posts: 666
A trim runaway of jammed elevator does seem unusual. But as they are QRH procedures then one assumes they are conceivable. It would also explain the disengagement of the A/P - again probably uncommanded. If the rate of climb was undesirable, the initial reaction by the crew would have probably been to select Vertical Speed mode and not disconnect (unless it was extreme - which it does not appear to be).

Regarding the vertical speed values and ground speed values. I don't see too much wrong with this. We know they were in rough air, rough enough for them to not want to continue the approach. Perhaps we can expect bumps and spikes in speeds and altitudes over that time. The general trends look consistent with an initial climb, acceleration to commanded speed and then once that speed was reached another steeper climb.

Does anybody know how the Mode S system derives or calculates these values? Obviously vertical speed is a function of altitude change over time and ground speed of position change over time. If the Mode S system calculates these values rather than extracting directly from the Air Data Computer then perhaps we can expect the data to be somewhat erratic depending on the sampling rate (frequency) of altitude and position data by the Mode S system. It would seem sensible for these values to be calculated from a systems integration unit interoperability point of view. If I was the designer of the Mode S system, I think I would calculate as much as possible within the module.
Sciolistes is offline  
Old 26th Mar 2016, 06:46
  #729 (permalink)  
 
Join Date: Oct 2008
Location: UK
Posts: 361
From my limited understanding of the transcript, it reads as though the aircraft encountered wind shear during the go around, whether annunciated or crew detected. It appears that the autopilot was then disconnected, presumably manually, to fly through the shear. This is probably an automatic response to training, as I suspect many operators do not routinely teach the autopilot coupled escape manoeuvre.

If the aircraft is for some reason grossly out of trim, this will have been masked by the autopilot, and the out of trim condition only apparent on disconnection. Assuming that the trim is for nose down, it would explain the pitch over. This will almost certainly have come as a surprise to the crew, but the "Don't worry, don't worry" sounds like it could be the captain, and indicates to me he has recognised a problem and at this moment feels it is something he can control. The subsequent CVR transcript of "Don't do this" I'm pretty sure is directed at the aircraft and would suggest that the aircraft is not responding as expected.

Why they couldn't recover, only further investigation will tell.
Otto Throttle is offline  
Old 26th Mar 2016, 06:46
  #730 (permalink)  
 
Join Date: Aug 2012
Posts: 1
knyupel thingy

Just found this on "vesti" website.
http://www.vesti.ru/doc.html?id=2735707

Почему включился в режим пике стабилизатор — вопрос. Он приводится в действие кнопкой, которую пилоты называют "кнюпель".
The phrase itself, if translated to English, would suggest that the knyupel thing is responsible for re-posistioning of a horiz.stab ...

Direct translation is:
It is a question why the stabilizer was activated(started trimming, I presume) nose-down (пике literally means - negative pitch situation).
It (stab) is activated by a button that pilots call knyupel.

And I have to warn that this is a news organization and normally, when they post something about aviation or any other technical field, their wording in those articles usually sounds extremely weird and cumbersome.

Last edited by Sunamer; 26th Mar 2016 at 07:01. Reason: added source
Sunamer is offline  
Old 26th Mar 2016, 06:51
  #731 (permalink)  
 
Join Date: Nov 2009
Location: Poland
Age: 33
Posts: 11
Correct me if I am wrong, but is it only a dual channel ILS approach (autoland) that allows for autopilot go-around? Otherwise pushing TOGA disconnects the A/P.
ipilot86 is offline  
Old 26th Mar 2016, 07:02
  #732 (permalink)  
 
Join Date: Oct 2008
Location: UK
Posts: 361
Ipilot, correct.

The news link re the CVR transcript appears to suggest that the autopilot had been re-engaged during the go around itself.
Otto Throttle is offline  
Old 26th Mar 2016, 07:04
  #733 (permalink)  
 
Join Date: Mar 2005
Location: N/A
Posts: 2,895
It is a question why the stabilizer was activated(started trimming, I presume) nose-down (пике literally means - negative pitch situation)
They could possibly mean the stab itself was trimming nose down, meaning an aircraft pitch up reaction.
megan is offline  
Old 26th Mar 2016, 07:08
  #734 (permalink)  
 
Join Date: May 2008
Location: Hang on, let me check the FMS...
Posts: 195
Hypothetically, if during the GA at light weights with full thrust and flaps down, the PF trimmed almost full nose down at high trim speed (the 737 trim is faster when flaps are down) to keep the nose up attitude under control.

Then when levelling off the speed increases rapidly, so PF makes a large thrust reduction while PM is retracting flaps correctly according to the speed schedule.

The aircraft will then have a large pitch down motion due to reduction of thrust, coupled with the nose down trim position and then with the flaps now up, a slower trim wheel or maybe no trim adjustment at all due to a failure or pilot error/fatigue, would the PF have enough elevator authority to correct the aircraft attitude?

The CVR transcripts says the autopilot disconnected, the PF could try and engage the autopilot but if grossly out of trim, would immediately disengage and the autopilot disconnect horn would sound in the flight deck.
FlyingTinCans is offline  
Old 26th Mar 2016, 07:09
  #735 (permalink)  
 
Join Date: Dec 2013
Location: Paso Robles
Posts: 261
It appears that the autopilot was then disconnected,
I don't think autopilot was ever engaged during this go-around, it was regular CAT I approach, perhaps even localizer only approach (this is not clear) and execution of go-around disconnects autopilot.
porterhouse is offline  
Old 26th Mar 2016, 07:16
  #736 (permalink)  
 
Join Date: Nov 2009
Location: Poland
Age: 33
Posts: 11
I believe re-engaging AP during a flight director go-around is a standard procedure? It can get messy with the AP modes (LVL CHG), full thrust, flap retraction, trimming.
ipilot86 is offline  
Old 26th Mar 2016, 07:21
  #737 (permalink)  
 
Join Date: Dec 2013
Location: Paso Robles
Posts: 261
I believe re-engaging AP during a flight director go-around is a standard procedure?
Yes, at some point however looking at their weird vertical speed gyrations it doesn't look like autopilot would be commanding something like that.
porterhouse is offline  
Old 26th Mar 2016, 07:34
  #738 (permalink)  
 
Join Date: Jul 2005
Location: btw SAMAR and TOSPA
Posts: 565
Originally Posted by Sciolistes View Post
A trim runaway of jammed elevator does seem unusual. But as they are QRH procedures then one assumes they are conceivable. It would also explain the disengagement of the A/P - again probably uncommanded. If the rate of climb was undesirable, the initial reaction by the crew would have probably been to select Vertical Speed mode and not disconnect (unless it was extreme - which it does not appear to be).

Regarding the vertical speed values and ground speed values. I don't see too much wrong with this. We know they were in rough air, rough enough for them to not want to continue the approach. Perhaps we can expect bumps and spikes in speeds and altitudes over that time. The general trends look consistent with an initial climb, acceleration to commanded speed and then once that speed was reached another steeper climb.

Does anybody know how the Mode S system derives or calculates these values? Obviously vertical speed is a function of altitude change over time and ground speed of position change over time. If the Mode S system calculates these values rather than extracting directly from the Air Data Computer then perhaps we can expect the data to be somewhat erratic depending on the sampling rate (frequency) of altitude and position data by the Mode S system. It would seem sensible for these values to be calculated from a systems integration unit interoperability point of view. If I was the designer of the Mode S system, I think I would calculate as much as possible within the module.
Data are from Air Data Computer. No data are calculated outside the airframe. Most likely recorded FDR data will appear 100% the same, albeit at a higher sampling rate.
MSL altitiude is from GPS (height over ellipsoid) and transmitted as an offset to QNE baro altitude.
Open remains the source of VS. There are two options: GPS or Baro. Depends on the supplier of the avionics.
threemiles is offline  
Old 26th Mar 2016, 07:43
  #739 (permalink)  
 
Join Date: Oct 2008
Location: UK
Posts: 361
The weird vertical speed gyrations aren't unheard of when the autopilot is controlling the aircraft in low level turbulence. It flies like a piece of [email protected] It wouldn't be unusual for a pilot to disconnect and fly an average pitch attitude through it and accept a small amount of speed deviation.
Otto Throttle is offline  
Old 26th Mar 2016, 07:44
  #740 (permalink)  
 
Join Date: Aug 2000
Location: Monrovia / Liberia
Age: 58
Posts: 734
On the B737, during a 'dual channel' approach (i.e. an approach being made with both autopilots engaged), a press of a TOGA button will cause the aircraft to go-around and the autopilot / auto-throttle will remain engaged (i.e. the autopilot / auto-throttle will fly the go-around, all the pilots then have to do is raise the landing gear, and move the flap lever in order to accelerate the aircraft).

During a single channel approach (i.e. an approach made with just one autopilot engaged), a press of the TOGA button will cause the selected autopilot to disengage (along with an associated aural warning of such), it's then up to the pilots to action the go-around manoeuvre and all subsequent associated actions.

In flydubai 'dual-channel' approaches are not approved, wherein Cat II / Cat III (low visibility approaches & landings) are entirely hand-flown via use of the HGS / HUD... and accordingly it would be true to say that ALL landings in flydubai are hand-flown (i.e. use of 'auto-land' is not permitted) and therein at some point during an approach (regardless of it being either Cat I/II/III) the autopilot must be disconnected, and following which the remainder of the approach & landing (or go-around) is flown manually.

During a go-around, the normal procedure is to hand-fly the aircraft until the flaps are fully up and then to engage the autopilot... that said, some might (for workload reasons) engage an autopilot & auto-throttle once the aircraft is above the autopilot's 'minimum use height'.
Old King Coal 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.