Asiana flight crash at San Francisco

[James7]

Suninmyeyes

Like someone else I wondered if the speedbrake had been pulled out earlier

Very good point, many times this has happened to me, the LG is selected down and speed brakes then armed as part of the sequence, then the PF uses the speedbrake and just stows them without re-arming them. Fortunately this is picked up on the ECAM landing checklist -Speedbrake .... and armed then selected. This usually happens in the Sim following multiple failures and me sitting in the right seat, not my usual seat. It has also happened a few times on the line, sitting in the right seat as PNF. I wonder if this was only picked up at 500' when the landing checks were done. I imagine the ECAM landing checklist is similar to the bus. This would explain the rapid drop in airspeed.

I believe the Autothrottle mode should also be displayed on the ECAM checklist. I know in the Bus the first line of the LDG checklist is Autothrust and the answer is Speed. However this is often missed ie just saying something that is not actually there. Also a common occurrence.

It would be a very simple thing to have -Autothrust ..... Speed. As a line on the ECAM. I am sure this would have saved a few embarrassing moments and in this case a crash.

[rudderrudderrat]

Hi MPN11,

IMO it starts with a badly designed approach plate which is evidently inappropriate, sets the aircraft up at 1800' (i.e. above the notional glideslope), and only works because those in the know ignore 180/5 and play 180/7.

Really?
The Jepp plates for the RNAV RW28L (3 deg slope)
http://yurik.flightgear.ru/KSFO.pdf show crossing the bridge at AXMUL at 1800 ft.
The quiet bridge visual shows crossing the bridge at 1900 ft (recommended).
Are you seriously saying that 100 ft higher is inappropriate?
The time taken to travel 2 nm at 180 kts is 40 secs.
By slowing down from 180kts to 160kts (average 170) takes 42.35 secs.
Are you seriously saying that taking an extra 2 seconds to reach the runway is significant for aircraft spacing?

I don't see why so many posters believe the visual approach / ATC were in some way to blame.
It was all down to the crew.

[Bob Zuruncle]

Cows getting bigger

[QUOTE]An interesting read from a couple of years back.

http://www2.icao.int/en/RunwaySafety...ons.pdf/QUOTE]

Excellent link! Vey pertinent to the current discussion.

[MPN11]

With respect, rudderrudderrat ... The relevant plate is the visual to 28L, not RNAV.

No evidence that I recall that they were doing an RNAV.


Not that I would ever argue with a pilot.

[OK465]

(14:55) OBSERVATIONS FROM FLIGHT CREW (Not yet corroborated by CVR)

Approach asked them to maintain 180 knots until they were about 5 miles out.

In this entire thread, the above is the first reference I can find to '180 at 5' and is from Post #1353 (page 68 currently), an NTSB briefing summary by Knot Apilot.

It is, as is stated, evidently from the Flight Crew comments and is, at that point in time, also as stated, "Not yet corroborated by CVR".

The FCrew comment, as quoted, doesn't reference 5 DME or even just (180 at) 5 miles, but is quoted as an apparently TLAR (?) 'about 5 miles out'.

Shortly after this post, the comment was latched on to by posters and run with as '180 at 5' or variously '180 at 5 DME'. edit: or above just now '180 to 5'.

Not wanting to read every post between then and now, has this been confirmed by the CVR? Has the CVR info even been released?

(additional edit: the referenced poster does use the term 'paraphrased'.)

[junebug172]

Some new pics of Asiana being hauled away. Amazing more people weren't killed.

https://docs.google.com/file/d/0B08e...it?usp=sharing

[Lonewolf_50]

From CGB's nice link. ( good stuff)

VMC stable by 500ft VMC (The aircraft should aim to be stable by 1,000 ft even in VMC conditions. If the crew fail to achieve this then the gate can be reset to 500 ft, but only if the crew anticipate that the aircraft will be stable by 500 ft, if not a missed approach has to be carried out).
Note: The altitude at which the gates are set and by which the approach is
judged to be stable, varies depending on airline SOPs and can range from
1500ft-500ft

Note: you have to be ahead of the aircraft to make that judgment call. Asiana's SOP's will hopefully be included in the NSTB report.

EC Law now states:
• Without Visual Ground Reference:
It is recommended that stabilisation be achieved at the latest when
passing 1,000ft above runway threshold elevation. If ATC procedures require higher speeds and is allowed in the OM, the above gate may not be met, in this case stabilisation should be achieved by 500ft.
• With Visual Ground Reference: Stabilisation should be achieved by 500ft
(however it is still recommended that pilots use the 1000ft gate as
above). If the above gates are not met pilots should consider initiating a
go-around maneuver.

Additionally, note the use of the words “recommended” and “should”.

[JPJP]

rudderrudderrat - The quiet bridge visual shows crossing the bridge at 1900 ft (recommended).
Are you seriously saying that 100 ft higher is inappropriate?
The time taken to travel 2 nm at 180 kts is 40 secs.
By slowing down from 180kts to 160kts (average 170) takes 42.35 secs.
Are you seriously saying that taking an extra 2 seconds to reach the runway is significant for aircraft spacing?

1. They were on the TipToe Visual. The recommended ("should cross") altitude at the bridge is 1900 feet. The bridge is 5nm from the runway (6DME SFO). Therefore: at 1900 feet, the aircraft is 400 feet high, 5 miles from the runway.

2. If they were asked to maintain 180 knots to either a 5 or 7 mile final, they would still be slowing and configuring as they approached the bridge.

The above two factors may have combined to create what was obviously an unstable approach.

http://155.178.201.160/d-tpp/1307/00...TOE_VIS28L.PDF

[Lord Spandex Masher]

They weren't 400 feet high. They were 400 feet above a 3 degree profile. There's a difference.

[olasek]

Therefore: at 1900 feet, the aircraft is 400 feet high

No, it is about 280 ft higher above the 3 deg. slope.

Clearly, whatever the numbers, the overwhelming number of flights have no problem with this approach. Also '1900' is only 'recommended', it is not a mandatory requirement.

[CONF iture]

Originally Posted by suninmyeyes

The low speed warning on the Eicas would still work, even if in FLCH, but probably came at about 150 feet when with the low speed and rate of descent and with the drag of flap 30 and gear down it would have been too late to be of any use.
...
On the central Eicas a low speed warning message appears with an aural warning.

What is exactly that aural message ?
NTSB has not yet mentioned anything like it on the CVR ...
Could it be inhibited due to higher priority aural message or stick shaker ?

[rudderrudderrat]

Hi JPJP,
Thanks for TipToe Visual Chart.

Therefore: at 1900 feet, the aircraft is 400 feet high, 5 miles from the runway.

Even if they were 400 feet higher than the 3 deg slope at 5nm, then their approach path is still only 3.75 degs. (Marseille Provence RW 31R has ILS GP set at 4 degs).
It should not be a problem for crews ahead of their aircraft.

[JPJP]

Lord Spandex Masher - They weren't 400 feet high. They were 400 feet above a 3 degree profile. There's a difference.

You are correct. In the sense that they are not high until they cross their stabilized approach gate. Whatever that is for Asiana.

Can we agree that the goal is to be stabilized on glide path or glideslope, at a specific altitude ? In this case, they must lose 400 feet in addition to continuing the descent towards the runway. That requires a rate of descent that is higher than normally required on a 3 degree glide path. Five miles from the end of the runway.

A normal rate of descent at Vapp is approximately 800 FPM. Traveling at over 2 miles per minute, 5 miles from the runway. If you descend at a safe rate of descent of 1000 FPM, it takes 2 minutes to lose the extra 400 feet. At this point you are now 1 mile from the end of the runway. Hopefully at 500 feet. All doable if well managed, fully configured and on speed when you cross the bridge. Even I can do it, on numerous occasions. It certainly requires greater that 1000 FPM ROD if you are still configuring and slowing.

What was their rate of descent ? Were they still slowing and configuring when they crossed the bridge 400 feet above a 3 degree glide slope ?

[JPJP]

Hi JPJP,
Thanks for TipToe Visual Chart.
Quote:
Therefore: at 1900 feet, the aircraft is 400 feet high, 5 miles from the runway.
Even if they were 400 feet higher than the 3 deg slope at 5nm, then their approach path is still only 3.75 degs. (Marseille Provence RW 31R has ILS GP set at 4 degs).
It should not be a problem for crews ahead of their aircraft.

I completely agree. None of this is an excuse for crashing an aircraft.

[olasek]

Therefore: at 1900 feet, the aircraft is 400 feet high

You keep repeating this 400 feet like a broken record. It was never 400 feet, it is actually around 230 feet when you use rigorous trigonometry, much less dramatic number.

[cwatters]

Some new pics of Asiana being hauled away. Amazing more people weren't killed.

https://docs.google.com/file/d/0B08e...it?usp=sharing

Quite a lot of seats seem to have failed where they attach to the floor, despite the forces being survivable?...or are they designed to break free?

[BOAC]

May I interrupt all this discussion about "3 degrees" to point out that despite what the TipToe chart tells you:-

a) A LOC does not have a (3 degree) glideslope

b) The VGSI are set at 3 degrees, but the RNAV arrival has a 2.85 slope, so if they were trying for that, they would see the odd red.

[Airbubba]

In this entire thread, the above is the first reference I can find to '180 at 5' and is from Post #1353 (page 68 currently), an NTSB briefing summary by Knot Apilot.

You can find the actual verbiage from Debbie Hersman's July 9, 2013 NTSB briefing here at about 15:10 in the clip:

C-SPAN

The 160 knot landing flap limit cited by Ms. Hersman seems low to me, as discussed earlier, perhaps this is not correct.

[Vasco dePilot]

Simultaneously reducing speed and height in the 777.

1. 1900ft at 5 miles is high on any 777's approach. That is a fact!
Whether the 5 mile point is 280ft above a 3deg slope or 400ft above the 300ft/mile formula used by most pilots is a moot point. The NTSB report will probably comment on that.
2. If the airplane is on the glideslope, being 43 knots above Vref at 180 knots at 5 miles might be OK. However, the combination of 280'-400' above the glideslope and 43 knots above Vref is a dilemma. Experienced 777 pilots will have skill and techniques to resolve it, however a trainee may not. When he is being supervised by a new training captain, the cheese holes are getting nicely lined up.
I invite other 777 pilots to tell us how they prefer to deal with this situation.

[nigegilb]

Couple of points about the visual into SFO. Generally speaking 28L is fairly straightforward and often given to heavies. 28R can be much more challenging because airspace restrictions and noise abatement often mean a very high initial approach path. Someone earlier mentioned that ATC always know the ac limits. I beg to differ. With the greatest respect to SFO ATC I often think they presume a Heavy can do the same thing as a 737 we can't and some times the approach to the visual is absolutely on or beyond the limits of the ac performance. Someone mentioned the last crash at SFO being 40 odd years ago. Of much more interest would be the stats on go arounds at SFO. I think they should be looked at as part of this safety review. I have never seen as many ac go around at an international airport.

A final point on speeds. Heathrow ATC are always alert to what speed you have bugged on the MCP because they can see it. Heathrow controllers routinely space ac 3-31/2 miles apart on the approach, anyone tempted to bring the speed back early ( normal to get 160 to 4 DME ) can often get a correcting call by the approach controller because he/she is working so incredibly tightly with following ac. This could well have an effect on crews operating elsewhere. I am much more relaxed taking an "overview" on speed control with US ATC because the margins generally are not as tight, but if you slowed 2nm early at LHR you could well get ATC on your back in the nicest possible way. Not sure how compelled the Asiana crew felt to comply but it is good to get this stuff into the open.