Approaching the glideslope from above! Fraught approaches.

[TheGreenAnger]

A350 sank far below glideslope as shortened approach stressed pilots. January 2020 incident saw widebody descend to less than 700ft while still nearly 6.5nm from Frankfurt runway threshold

German investigators believe a shortened approach route given to a Thai Airways Airbus A350-900 crew generated increased time stress, resulting in a botched high-speed descent to Frankfurt that took the twinjet far below the glideslope. The aircraft, arriving at night, was just 668ft above ground, while still 6.43nm (11.9km) from the threshold of runway 07R, before it climbed away during a go-around.

German investigation authority BFU states that four pilots – the captain, first officer, and two cruise pilots – were in the cockpit. The cruise pilots were supposed to act as safety observers, and intervene if necessary, but neither pointed out that the approach was too low.

The crew had been expecting to follow a preceding aircraft to 07R, and the approach route had been entered into the flight-management system.

But a passenger medical situation had developed earlier in the flight and, as a result, the controller feeding traffic into the ILS path gave it priority and issued the crew with instructions that shortened the approach, putting the A350 ahead of the preceding aircraft.

The inquiry believes the medical event was not an emergency and did not justify the shortened route.

“This instruction exposed the flight crew to time stress,” says BFU, and the crew’s actions were subsequently “uncoordinated” and resulted in “loss of situational aware ness”. It adds that the orders of the first officer, who was flying, were “no longer clear”.

While the approach path had to be adjusted in the flight-management system, to allow the vertical guidance to provide the correct path indication, the inquiry believes the crew programmed the system, with a remaining distance “significantly longer” than the actual one. “It is very likely that the indication on the [primary flight display] of the calculated vertical flightpath did not correspond with the mental image of the pilots,” it states.

Open descent BFU believes the crew thought the aircraft was far too high above the required approach path, and the first officer flew an open descent, extending the landing-gear, flaps and speedbrakes to lose altitude.

At about 3,000ft above ground the A350 passed through the glideslope from above, with a descent rate of about 3,000ft/min. It pitched increasingly nose-down, from 3° to 8°, and its airspeed rose.

Having previously overshot the localiser, with the approach mode not activated, the aircraft subsequently captured it at 1,820ft above ground with the descent rate in creasing to more than 4,000ft/min. Sink-rate and glideslope warnings were triggered at 1,340ft – the aircraft was far below the glide slope and still descending at nearly 3,400ft/min – but a go-around was not initiated until the aircraft was just 936ft above ground. It sank a further 184ft and was 688ft above terrain, 4.1 dots below the glide slope on the cockpit indicator, before it started gaining altitude. The aircraft conducted a second approach to 07R and landed without further incident. Cockpit-voice recorder information shows no discussion of the aborted first approach took place before the second attempt to land.

None of the 306 passengers and 18 crew was injured and the jet (HS-THF) was undamaged during the 1 January 2020 occurrence. All four cockpit crew members were experienced, and the captain and first officer respectively had 400h and 1,500h on the A350. BFU says that, during interviews, the two pilots “did not give a statement” after being asked to explain the low approach. The cruise pilots were also interviewed, the inquiry adds: “They were asked whether they had realised that the aircraft had been too low during the approach, and had they given guidance to the two pilots. Both questions were answered in the negative.” Investigators point out that, along with the incorrect programming of the flight-management system, the first officer changed modes on the flight-control unit several times, resulting in operating errors.

BFU also indicates that the captain had difficulty understanding and communicating in English, with comprehending the Frankfurt controller’s queries about the medical situation, and with following the conversation during investigators’ interviews.

From Flight International. September 2022.

[TheGreenAnger]

The incident noted above made me apt to think of Turkish Airlines Flight 1951

[Boac]

The risk in 'capturing from above' is the capture of a false glidepath. Several separate glidepaths are radiated by a GP transmitter. This Airbus training aid may help understand https://safetyfirst.airbus.com/lining-u ... ide-slope/.

Capturing one of the steeper glideslopes will take the aircraft to the designated landing point, but at a much higher descent rate with associated dangers.

All can be avoided if situational awareness is maintained, which it was not in your 2 examples. There is nothing wrong with 'capturing from above' as long as the right glidepath is captured!

Even the word 'capturing' now-a-days implies the use of autopilot rather than the 'old-fashioned' fly the needles techniques and it is so easy to be suckered into thinking the a/p knows what it is doing without double checking. Several times Amsterdam ATC would vector aircraft so tight that one would finish up above the GP.

It is/(was?) tradition to confirm the distance from touchdown at which the GP was actually intercepted thus enabling detection of a false GP to be made. I really hope that is still done!

The Turkish accident was seriously compounded by a malfunctioning autothrottle caused by a radalt failure and then lack of monitoring by the crew.

[TheGreenAnger]

The risk in 'capturing from above' is the capture of a false glidepath. Several separate glidepaths are radiated by a GP transmitter. This Airbus training aid may help understand https://safetyfirst.airbus.com/lining-u ... ide-slope/.

Capturing one of the steeper glideslopes will take the aircraft to the designated landing point, but at a much higher descent rate with associated dangers.

All can be avoided if situational awareness is maintained, which it was not in your 2 examples. There is nothing wrong with 'capturing from above' as long as the right glidepath is captured!

Even the word 'capturing' now-a-days implies the use of autopilot rather than the 'old-fashioned' fly the needles techniques and it is so easy to be suckered into thinking the a/p knows what it is doing without double checking. Several times Amsterdam ATC would vector aircraft so tight that one would finish up above the GP.

It is/(was?) tradition to confirm the distance from touchdown at which the GP was actually intercepted thus enabling detection of a false GP to be made. I really hope that is still done!

The Turkish accident was seriously compounded by a malfunctioning autothrottle caused by a radalt failure and then lack of monitoring by the crew.

Yes, as you say, the approach from above is sub-optimal but not an issue in of itself, in an approach that can still be completed safely.

It was in that sub-optimal spirit that I mentioned being above the glideslope, almost as a metaphor for things going rapidly downhill in the approach thereafter, rather than anything else. The Air Turkey channel A radio altimeter failure and the long litany of further automation and piloting failures, culminating in all 3 pilots failing to monitor the speed was pretty staggering really. Interestingly there were monitoring pilots up front in the first case above as well. Are 3 to 4 pilots more of a hinderance than a help in such situations?

As for ILS false glide paths, back beams, etc. etc. I guess we should start another thread to to canvass that very interesting subject.

[Boac]

Are 3 to 4 pilots more of a hinderance than a help in such situations?

How long is a piece of string? In theory, yes, but often in practice............. The basic concept of 'Pilot Flying' and 'Pilot Monitoring' works if used correctly. Now bring in 'cultural' influences, seniority, over-bearing personalities etc etc and we have a murky pool.

Single pilot - well, yes - better be on the ball!

[CharlieOneSix]

Yet an unnoticed descent to a 737 on an ATC requested go-around warrants no comment?.....At least it would have just missed the city and crashed in farmland if the descent had continued for another 30 seconds or so: viewtopic.php?p=342605#p342605

During the manually flown go-around, which was initiated at 2,250 ft amsl, the aircraft initially climbed, but just before it reached the cleared altitude of 3,000 ft amsl it began to descend. It descended to 1,780 ft amsl (1,565 ft agl) with a peak rate of descent of 3,100 fpm, and accelerated to an airspeed of 286 kt (the selected airspeed was 200 kt) before the crew corrected the flightpath. The aircraft descended for a total of 57 seconds before the climb was re-established.

[Boac]

Cannot really comment - as I said "if used correctly".

[Fox3WheresMyBanana]

These guys aren't pilots. They are autopilot inputters.

[TheGreenAnger]

These guys aren't pilots. They are autopilot inputters.

They probably hadn't actually flown anything but a simulator for a couple of months before this "Incident". As they say "use it or lose it!"

My neighbour, a training Captain on the 737 for an airline operating locally from Stansted, a good stick and rudder jockey, who came to his position the hard way, via hours building, instructing, flying Sheds, etc. etc. said that getting back into flow of things after so many months "off" was a strain, even for him. What hope do lesser mortals have?

[TheGreenAnger]

Yet an unnoticed descent to a 737 on an ATC requested go-around warrants no comment?.....At least it would have just missed the city and crashed in farmland if the descent had continued for another 30 seconds or so: viewtopic.php?p=342605#p342605

During the manually flown go-around, which was initiated at 2,250 ft amsl, the aircraft initially climbed, but just before it reached the cleared altitude of 3,000 ft amsl it began to descend. It descended to 1,780 ft amsl (1,565 ft agl) with a peak rate of descent of 3,100 fpm, and accelerated to an airspeed of 286 kt (the selected airspeed was 200 kt) before the crew corrected the flightpath. The aircraft descended for a total of 57 seconds before the climb was re-established.

ZF.jpg

C16, sorry I did read the report you posted and should have appended my posts after yours (mea culpa).

In fact I believe there is another thread lurking around here somewhere, might have even been started by that Goblin slacker that was focused on the Aberdeen fiasco as well.

TUI ups and downs near Aberdeen! - viewtopic.php?p=315614#p315614

[FD2]

How much of the base checks for these guys are flown with no autopilot? Only a small part of our six-monthly IBCs was flown without an autopilot - it was actual handling instead of button pushing. At the time we also had some machines which only had a basic stabilisation system fitted anyway so all the procedural approaches were hands on apart from a radar to ILS at the end which usually happened to coincide with an engine failure or other 'emergency'.

The matter of crews being unable to recover from unusual attitudes manually has often been raised after accidents but what has been done to teach the boys and girls how to fly an aircraft with their hands and feet again? The loss of altitude/height (don't know what pressure settings they were using) after the go around was initiated is alarming in this Tui aircraft.

[TheGreenAnger]

How much of the base checks for these guys are flown with no autopilot? Only a small part of our six-monthly IBCs was flown without an autopilot - it was actual handling instead of button pushing. At the time we also had some machines which only had a basic stabilisation system fitted anyway so all the procedural approaches were hands on apart from a radar to ILS at the end which usually happened to coincide with an engine failure or other 'emergency'.

The matter of crews being unable to recover from unusual attitudes manually has often been raised after accidents but what has been done to teach the boys and girls how to fly an aircraft with their hands and feet again? The loss of altitude/height (don't know what pressure settings they were using) after the go around was initiated is alarming in this Tui aircraft.

Truth is, automation adds its own ineffable complexity.

[G~Man]

Truth is, automation adds its own ineffable complexity.

Exactly...Look at the Asianna Airlines crash with Captain Sum Tingwong, FO Wi Tu Low, etc..... And YES, this is a REAL newscast:

[TheGreenAnger]

Truth is, automation adds its own ineffable complexity.

Exactly...Look at the Asianna Airlines crash with Captain Sum Tingwong, FO Wi Tu Low, etc..... And YES, this is a REAL newscast:

Holy **** man...

[Fox3WheresMyBanana]

The backstory is that the TV news team were handed the joke names by some wag at the airport.
They did then phone the FAA for confirmation. The FAA weekend guy on the phones was a student intern, who thought it would also be a good laugh to confirm those names.
The intern wasn't an intern a week later, but it was probably worth it

[FD2]

Truth is, automation adds its own ineffable complexity.

The funny thing was that we used to fly manual ILS approaches whenever we could in order to 'keep our hands in' but it was also a recency requirement. The Sperry system could fly you down to 50ft and up the runway if you wanted it to but there were still those who didn't trust it and they used to fly manually to practise for when they 'really needed to fly it in an emergency'. The reason we flew them manually was in case of a system failure but if there had been a catastrophically wrong weather forecast and we didn't have diversion fuel I would have trusted Mr Sperry's system implicitly, to get us on the runway. CAA reports can be written later. The other alternative was to do a 180 and fly back offshore where again the autopilot system would reliably get us back on the helideck and the paperwork was less. I'd agree with you about complexity though TGA - we also needed to practise coupled manoeuvres frequently because it could get quite challenging working out what to do in, for instance, a fully coupled backbeam go around!

[Alisoncc]

I suppose the days when one flew the Outer, Middle and Inner markers are long gone.

[FD2]

Long gone in the age of simulators and autopilots I think Alison. I was being a little tongue in cheek about some of the aids we were told and shown could be coupled for various manoeuvres, but the point was made by TGA that:

'Truth is, automation adds its own ineffable complexity.'

[421dog]

On the contrary, the local tower/approach is a training facility, and I often get vectors that put me over the FAF at altitudes and speeds that vex my autopilot and multiple magenta-line-generation devices leading to things not doing what they’re supposed to and resulting in a whole bunch of vociferous “effs” on my part (as well as a punch of the red button and a hand flown approach….)
Fortunately, I’m of a vintage where steam was the normal rather than a backup.