Turkish Airlines cargo 747 crashes in Kyrgyzstan

[Icelanta]

747-8driver,

Thank you for the very informative video.

Indeed, it might depend on the ground antenna used.
We are made aware of the danger of pitch-up autopilot behaviour in case of a false G/S capture.

[Bleve]

Originally Posted by Kulverstukas

2) ... GS CAPTURE and a/c begin automatic descend parallel to glideslope ...
3) ... G/S pointer at PFD was at the bottom ...
4) ... A/P engaged ...

This is not an official account, but if true, I don't understand why the A/P flew parallel to the G/S. With G/S capture and the G/S pointer indicating fly down, I would expect the A/P to do exactly that - fly down to get on the G/S. Why didn't it, why did it only fly parallel? Either the account quoted is not complete and/or there is another technical problem that hasn't been revealed yet.

In any event the crew should have detected that the aircraft wasn't on the correct G/S (standard distance/OM height checks), so a tragic crew failure as the last line of defence.

[Machinbird]

Is there a maximum descent rate or angle that the autopilot will use to try to capture a glideslope from above? I expect aircrews would become really uptight if their aircraft made a lawn dart type dive to capture the glideslope, so there probably is some type of limit. Maybe the best it could do was to parallel the false glideslope?

[RichardN]

Once the (mirror) glide slope was locked and followed shouldn't there be a HUGE ALERT to the crew (or at least freeze current ALT) in case the signal is lost again (here after 1 sec)?

[RichardN]

Originally Posted by Kulverstukas

3) ... G/S pointer at PFD was at the bottom all the time but crew was not aware either.

When hitting the false glideslopes (6, 9, 12), I'd expect the needle to move up and down and vice versa.

[skkm]

On the B777, after having become established on and captured the GS (false or otherwise), if the GS signal subsequently becomes invalid (as presumably happened in this case after the short period during which it captured and initiated descent), the AFDS will enter a stabilising mode with an amber line through the G/S FMA (and, later, an AUTOPILOT caution, mind you!), maintaining the vertical speed that existed prior to the loss of valid GS data - thus in this scenario continuing on the observed parallel descent profile above the real glideslope, which remains too far off the scale to re-capture automatically. Does the 744 behave similarly?

[megan]

Aero 21 - Erroneous Glidescope

The incident referred to.

https://www.caa.govt.nz/Occurrences/...rt_00-2518.pdf

JT makes the observation elsewhere with respect to the above incident, "I guess I was lucky in that my apprenticeship was under training captains who encouraged a "don't trust it too much" philosophy ... but there are many, particularly among our younger PacMan brethren who, perhaps, are more susceptible to GIGO problems with computers. I see this report as being a very useful wakeup call to all of us who necessarily have to put our trust in computers over which we may have little direct control."

[Kulverstukas]

Originally Posted by Bleve

In any event the crew should have detected that the aircraft wasn't on the correct G/S (standard distance/OM height checks), so a tragic crew failure as the last line of defence.

Originally Posted by Avherald comments

By mike on Thursday, Feb 9th 2017 01:10Z

A friend who was former military provided this comment:

I was stationed at Manas Air Base (Bishhek) back in 2002. We were flying combat missions with our F/A-18's out of there and into Afghanistan. The ILS was so bad, that we had to bring in USMC controllers and a portable ILS system. At that time, they were using QFE altimeter and a very low transition level. There was some belief that the Russians were deliberately jamming the ILS in an attempt to screw with us. There is a secret Russian communications facility about 15 miles north of Manas.

We almost lost a USMC C-130 because of that issue but the crew knew something was wrong and did a go-around before it was too late. Just a little intel from someone that flew out of there for 6 months and lived there for almost 8 months.

So possible they were aware that ILS is lost but military experience coupled with quote above tricked them into this event. They were ready for GA from 100ft, but they expect it will be concrete rwy under the gears, not ground, trees and concrete fence.

[Airbubba]

So they thought the ILS indication, pegged to the bottom of the scale, was not valid but they still went down to a 100 feet DH?

Hopefully we will get more pieces to the puzzle soon, thanks for your many contributions.

[Chronus]

Never mind ILS, markers, DME and everything else, it seems a case of chuck the anchors out we are going to dock come what may and when the prefabs are spotted coming up fast into view, up anchors. I dare say, a touch late to get the hell out of Dodge in a hurry.

By the way Airbubba I `d guess they must have gone down to 100 feet searching for the glideslope, it had to be there somewhere.

[RichardN]

GS CAPTURED and GS TRACKING seems not to be the same.

Well, I'm not sure if a 9° glide path is actually flyable for the a/p.

I'd like to know what happens when when the 744-AFDS loses or declines the signal after the [mirror] glideslobe was captured:

1. Is the GS mode reset to ARMED?
2. Does the AFDS revert to V/S continuing with the ROD set?
3. Is there any form of alert?

Many thanks from a rookie!

[Icelanta]

The Autopilot on B744 is only certified for an ILS with a glideslope of max. 3.25degrees.

The AFDS nor the airframe (aerodynamically) is able to fly a 9 degree ILS obviously...

[Herod]

Several pages back (post 118) I remarked that I was always taught to INTERCEPT THE G/S FROM BELOW. That way, you don't get a false G/S.

[Airbubba]

Originally Posted by RichardN

I'd like to know what happens when when the 744-AFDS loses or declines the signal after the [mirror] glideslobe was captured:

1. Is the GS mode reset to ARMED?
2. Does the AFDS revert to V/S continuing with the ROD set?
3. Is there any form of alert?

I'd also like to know. I don't think there is any indication in the data we've seen so far that the 747 went down at a final descent path steeper than the glideslope. Instead, it somehow never intercepted the real glide slope and roughly paralleled it to a point past the end of the runway.

I'm really curious what vertical mode and FMA indications would result from the momentary capture of the false glideslope. Was there a 'No Autoland' at 600 feet RA? Whatever the case, something was very wrong and going down to a 100 DH with the glide slope indicator pegged off the bottom seems just as incredible as the OZ 214 crew at SFO letting the airspeed drop to 103 knots with a target speed of 137 .

[RAT 5]

I'm really curious what vertical mode and FMA indications would result from the momentary capture of the false glideslope.

I may have missed it, but what vertical mode were they in? LVLCHG, VNAV or V/S? They can't have been GPTH.

[RichardN]

Originally Posted by Airbubba

I'm really curious what vertical mode and FMA indications would result from the momentary capture of the false glideslope.

I think the AP may have tricked them into believing, the ILS GS was finally followed.
After capturing the false GS the AP commenced the 3 degree descent, but lost/declined the signal instantly. So the a/c was on 3 degree glidepath initiated by the AP/FD without any reason to do so, since the crew's last setting was ALT/HLD 3400 with LOC CAP and GS ARMED.

An alert at this point dispatched by the AFDS software could have saved the lives of 17 children and 24 adults.

This is not meant to be an excuse for the crew, who neither monitored ALT/DME (both VOR and ILS had DME), nor FMA pitch (GS/VS), nor GS pointer, didn't notice level off flight between the markers and initiated TOGA too late.

[Chronus]

Bleve`s question "I don't understand why the A/P flew parallel to the G/S".

Answer from Aviation Herald report :

"On Feb 8th 2017 information sent out to Russian Flight Crew (see the attached graphics of the flight trajectory below) became known stating, that the aircraft was conducting a CATII ILS Approach to Bishkek's runway 26, the crew however did not monitor their altitude, the aircraft passed the final approach fix 650 feet above the glideslope, therefore no glideslope capture occurred and the aircraft continued to descend to 3400 feet at 200 KIAS and levelled off at 3400 feet (Autopilot Modes: LOC CAP and ALT HOLD), the glideslope deviation indicator showed a full scale down deflection. About 0.8nm before the runway threshold the aircraft passed through the false glideslope (9 degrees) resulting in valid glideslope deviation indications for about one second, the autopilot changed to G/S CAP (Capture) mode and the aircraft began to descend, the glideslope deviation indicator returned to a full scale down deflection. The crew did not detect that it was not possible to descend to the runway from 3400 feet (runway elevation at 2055 feet) within 0.8nm, did not cross check their approach trajectory nor did the crew notice the full scale down deflection of the glideslope deviation indicator. The aircraft thus descended parallel to and above the actual glideslope. At 100 feet radio altitude the captain called for a go-around due to lack of visual references, TOGA was activated at 52 feet radio altitude, the aircraft however touched down 900 meters past the runway end and 60 meters to the right of the extended runway center line, broke through a concrete fence, bounced and fell into the village about 1000 meters past the runway end, following the impact at the concrete wall the aircraft began to break up and spilled fuel which resulted in a fire. "

[172_driver]

So the AFDS must have been really confused then. After G/S CAP (of the 9 deg path at 0,8 nm, that lasted 1 sec) it started a descent? What was the AFDS trying to achieve at that point? What was the thrust FMA at that point? I've seen odd AFDS behavior in the past but it's usually been transient and the computer has finally made up its mind.

[Tu.114]

So... no DME check at the presumed GS intercept, no outer marker check, and also the two ADFs (presumably) set to B and BK already pointing to the rear when the GS was captured did not raise a red flag?

If the suspicion raised by Kulverstukas above about the somehow not properly working ILS was true, any of these thecks would have caught the error...

[RichardN]

Originally Posted by Tu.114

So... no DME check at the presumed GS intercept, no outer marker check, and also the two ADFs (presumably) set to B and BK already pointing to the rear when the GS was captured did not raise a red flag?

A crew that doesn't watch its altitude and distance to the primary navaids won't care about the position of marker beacons either.

If the suspicion raised by Kulverstukas above about the somehow not properly working ILS was true, any of these thecks would have caught the error...

Correct. Suspected ILS jamming would have drawn their attention even more to ALT+VOR/DME or NDB origin.

This is so weird.
If there weren't a TOGA attempt, I'd say the were incapacitated.