UPS 1354 NTSB Investigation - CVR

[mm43]

Here is a graphic that wasn't added to the original thread, as it was effectively "conjecture" at that time - 20130903, but I suspect the flight path indicated will be very close to the fact.



Below is an updated graphic of the approach profile using the integrated accelerometer altitudes from the NTSB Aircraft Performance Study included with the UPS1354 docket. The altitudes fit well with the Google Earth terrain data, which includes the treetop heights as drawn.

[porterhouse]

Maybe somebody with a lot more experience than I have could help me understand how this accident happened. I am brain fried on this one

Certainly not the first and last accident which is hard to understand.
I am an astute student of aviation accidents (both big and small crafts) and one thing I learned a very long time ago there is no shortage of most bizarre screw-ups pilots are capable of pulling off. This one frankly doesn't even make the "10 strangest" list.

[tubby linton]

For a monitored approach the pilot flying the approach has his head in and the pilot landing had his head out looking for the runway There is no mention of either pilot spotting the papi on the cvr or any rad alt call outs.

[Lonewolf_50]

Tubby:
(From a few posts up, your post begins with apologies).
Key point number 1: late let down to approach. Crew is aware that they are in catch up mode for the approach they were cleared for, and we see comments to that effect in the CVR transcript.
Key point number 2: approach not stable by 1000 feet, per various well stated points in other posts.
Key point number 3: go around decision not made.
From your last post:

There is no mention of either pilot spotting the papi on the cvr or any rad alt call outs.

Data points suggesting that the crew was behind the aircraft.
=================
Flyboyike: what the heck is wrong about trying to look your best in a photograph?
Not everyone is photogenic.
It is quite possible (my wife runs into this alot) that you have a lot of photos taken of you that look "not right" and only a very few that "look right."
Maybe she'd run into the problem my wife runs into, and decided to go to a professional to get a picutre that looked nice.
You are being uncharitable, at best.

Not everyone, me included, is as photogenic as Ms Wagstaff.

[flyboyike]

F/Os aren't qualified to fly? That's a new one.

[Sunnyjohn]

As always with accidents, lots of holes in the cheese but, AFAIR, no-one has yet mentioned:

The circadian rhythm makes people’s desire for sleep strongest between midnight and dawn, and to a lesser extent in mid afternoon. In one study, researchers instructed a group of people to try to stay awake for 24 hours. Not surprisingly, many slipped into naps despite their best efforts not to. When the investigators plotted the times when unplanned naps occurred, they found peaks between 2 a.m. and 4 a.m.

The accident happened at 04.47.

[tubby linton]

In the testimony it is mentioned that the F/O fell asleep on every sector of her previous nights work.

[Old Boeing Driver]

@ Tubby

Are you referring to the NA for night ops on the approach plate.

I think this was covered back on the old thread, and the NA was incorrect.

See aterpster's #340 on the thread under tech log.

There were quite a few other posts on this, but I didn't want to look through the 1,000 plus posts.

Cheers,

OBD

Edit: I found your comments, as well as others in the #450-460 post range on the old thread. I understand what you are saying.

[Old Boeing Driver]

Puzzled

One of the things that has puzzled me since the CVR release is why nobody called the PAPI or REILS in sight.

Having landed on that runway at night (a long time ago), it seems that they would have surely seen 4 reds.

Regards,

OBD

[tubby linton]

OBD yes we go back to the N/A box on the chart. And we have discussed it , but the crew do not mention it nor justiify their decision to fly the approach based upon the notam. Was the notam even included in their briefing pack and it was issued months before? I am trying to make the point that if I woke you up in the middle of the night and showed you the chart , would you
A notice the N/A box and B think that there was a notam telling you to disregard it?Jepp are to blame for not changing the chart but we as posters on pprune all missed it. The dispatcher is alleged to have noticed which is why he thought the crew would fly the rnav approach but he failed to bring this to the crew's attention. If there was such an ambiguity should UPS not have issued a notice alerting their crews and requested a correction from Jeppesen. They should have banned operations to the runway using that approach until the matter had been cleared up.In flight operations you should plan that your z list team will always end up trying to land at the most challenging airport despite your best efforts to send them elsewhere and do the best to keep them alive.
What I did not gather from the hearing yesterday was when the crew planned this part of their duty. It was a multi sector night so did they receive the paperwork at the start of the evening, when you would hope that they would have been at their sharpest or on the turnround prior to their last sector? Was it given to them in a well lit office with a set of charts available or on the flight deck. The A300 cockpit isn't particularly well lit and I and my coleagues often struggle to complete a manual loadsheet in the gloom so reading a multi page briefing when you are tired is going to be difficult.

[xcitation]

@OBD

One of the things that has puzzled me since the CVR release is why nobody called the PAPI or REILS in sight.

Having landed on that runway at night (a long time ago), it seems that they would have surely seen 4 reds.

Good point.
Assuming the previous graphic is indication of actual events:
PAPI could have been visible initially and then disappear at around 9000' onwards being obscured by the hill and low altitude,
The entire runway should have gradually disappeared as they drop behind the hill.
The EGPWS gave them less than 5 seconds to pull out/TOGA.
Very tragic.

[Old Boeing Driver]

Good Points

Tubby. Good points. I also don't understand why the dispatcher did not speak directly to the crew at any time.

Back in the "old days" we actually talked things over and agreed to things.

He was fairly new at UPS, but a very experienced dispatcher from Atlas.

Cheers,

OBD

[Old Boeing Driver]

The 1,000 foot call

I believe we agree that the 1,000 foot call is at 1,000 feet above the airport, or about 1,600 MSL in this case.

If you look at mm43's graph, her call was at 1,000 feet on the altimeter. (MSL)

Tubby mentioned this in one of his posts.

They didn't realize where they were vertically.

Has they reacted with a go around at the first "sink rate" announcement, things might be different.

Regards,

OBD

[GBV]

From the CVR transcript, the 1000ft call was done at 4:47:02,9, comparing with the FDR that would be about 1500ft or just above 1000ft on radio altimeter. Also, they call visual at 4:47:27.9 about 150ft radio altitude or just above 900ft. They went below MDA not having the runway in sight, the "hundred above" and "minimum" calls would have saved the day as usual.

Interesting enough there is some talk about the MDA when they are approaching and passing 1200ft:

"Alright ah DA is twelve ah hundred"

"Twelve hundred yeah..."

"two miles"

"it wouldn't happen to be actual [chuckle] (they were passing 1100ft at this point)

Do you guys believe they would have intentionally descended below MDA not being visual?

[Airbubba]

It appears the possible requirement for 1000 feet per minute down at the FAF to catch the profile was part of the canned briefing read by the PF:

04:25:47.0 HOT-1

and in the last note. select profile and verify P descent on an ILS
glideslope out approaches or localizer approaches when the VNAV
path crosses the final approach fix below the FAF minimum altitude.
start a one thousand feet per minute descent at the FAF and
immediately select profile mode to capture the path.

Would this possibly apply in the LOC 18 or does the path automatically cross BASKIN at 2300 as depicted on the chart? Some Boeings have an autothrottle surge that can occur on a computed VNAV path descent due to slight discrepancies between the calculations and charted crossing restrictions.

Does it look like they missed the fact that they could descend down to 2300 from 2500 when established on the localizer inbound to BASKIN? If they were at 3000 eleven miles from Baskin they should have been able to cross BASKIN at 2300 easily since COLIG is on profile 8.1 miles before BASKIN at 3500.

Playing catchup from above in a widebody is no fun on an ILS, harder on a non-precision since you need to be fully configured and on speed prior to the FAF on most aircraft.

It's been quite a while since I've flown an A306, would there be path guidance out to COLIG if the FO had sequenced the approach correctly? Could the crew capture the profile prior to BASKIN? Would there be guidance but no capture with a discontinuity in the box before the approach?

Should the crew have briefed a D-DA of 1250 instead of a DA of 1200? That one may depend on UPS Ops Specs for the A300.

[Showbo]

Airbubba:

Would this possibly apply in the LOC 18 or does the path automatically cross BASKIN at 2300 as depicted on the chart?

No, it doesn't apply to the LOC 18, it only applies to the circumstances in your quote. Based on what they said, I suspect the crew were a little hazy about this.

Does it look like they missed the fact that they could descend down to 2300 from 2500 when established on the localizer inbound to BASKIN?

Yes, I think so. Moreover, they could have stayed at 2500 and captured the path from there provided they made sure they didn't cross BASKN below 2300 but they didn't seem to appreciate that and the cvr chatter seems to indicate mild annoyance that they were fed in 200 ft high and that they thought they would have to drive down to the path from above using Vertical Speed mode to capture it. Note where they say.......

...yeah I'm gonna do vertical speed.
yeah he kept us high.
kept ya high. could never get it over to profile (we didn't) do it like that. uh uh I know

We've all seen it in the sim, especially in initial training, where we've forgotten something, and the "Football" path indicator doesn't appear, and the path doesn't capture and we think, "What the .....??" I rather suspect the captain thought the reason was because of the "high" feed in and that he thought he was well above the path and that is why he continued to drive down in VS. All the while, the real reason the path never appeared was because the box hadn't been sequenced correctly. Quite why all the other subsequent opportunities to abandon the approach were missed remains to be seen.

would there be path guidance out to COLIG if the FO had sequenced the approach correctly?

In theory, yes, although the "football" is not visible unless you are quite close to the path (within 200 ft I think), somewhat like an ILS in full scale deflection. The box actually constructs a glidepath at (in this case) 3.28 degrees backwards from the TCH all the way out to infinity. In theory you can capture it anywhere once Final Approach Mode has been activated and Profile has been armed. In practice, ATC and TERPS restrictions prevent this from happening.

Could the crew capture the profile prior to BASKIN?

Yes.

Would there be guidance but no capture with a discontinuity in the box before the approach?

No..... There would be no guidance and no capture.

Should the crew have briefed a D-DA of 1250 instead of a DA of 1200? That one may depend on UPS Ops Specs for the A300.

A 1200 ft DA was correct.

[BOAC]

I cannot recall if this was in the original thread (now in Tech Log), but the puzzle over the 'no PAPIs' call may be that the PAPIs were never in sight due to cloud/vis such that when they 'saw' the runway it would be the far end and the PAPIs were already below the sight line?

[ManaAdaSystem]

I would like to apologise for comparing this crew with the Korean crew in SFO. I just read the transcript and thought this was a night visual gone bad, when in fact it was a non precision IFR gone bad. A world of difference from a day visual.

The three biggest holes in the Swiss cheeese when it comes to CFIT are:
-Night/darkness
-Non precision
-Terrain

Add what appears to be a weak captain and a FO who did not take proper rest prior to the flight... I'm not impressed, but it is not comparable to Korean.

[Airbubba]

Thanks for your answers to my questions Showbo, I really appreciate your analysis.

The last time I trained in Toulouse, the Concorde sim was still down the hall.

I rather suspect the captain thought the reason was because of the "high" feed in and that he thought he was well above the path and that is why he continued to drive down in VS. All the while, the real reason the path never appeared was because the box hadn't been sequenced correctly.

Would the football be pegged at the bottom in this case or would path guidance not be displayed at all? It seems that the scale and 9999's were on the screen at some point from the panel questioning.

A lot of the questions in the hearing derive from issues raised in the interview summaries in the accident docket.

The 'technique' of putting something lower than the MDA on the approach page of the FMC seems to be mentioned repeatedly in the pilot interviews. If this was done it would explain why the autopilot was still on below MDA in the accident.

[Desert185]

Good Points
Tubby. Good points. I also don't understand why the dispatcher did not speak directly to the crew at any time.

Back in the "old days" we actually talked things over and agreed to things.

He was fairly new at UPS, but a very experienced dispatcher from Atlas.

UPS dispatch provides the flight paperwork one hour before departure for domestic and 1 1/2 hours before departure for international in a briefing area. There is no personal contact with dispatchers. Keep in mind that there is a significant number of flights departing Louisville within just a few hours. It would be impossible for a dispatcher to give a face to face brief to every crew. If anything, the paperwork is overly cluttered with details not directly relevant to the flight, at least it was before I retired.