UPS cargo crash near Birmingham AL

[aterpster]

Pope:

V/S or Profile can be used with a LOC appr and LOC* then LOC would be the FMA in either case. The A300 has a very old FMS and unless upgraded by UPS, it has likely been suffering from memory issues with all of the stuff that didn't exist when it was created back in the early 80's. EPGS, complex OPD's, RNAV approaches, huge influx of new enroute waypoints, etc.

Long story short, something has to give and operators sometimes selectively remove approaches and waypoints from the database to fit the space. UPS has a big theatre of ops, so it's easy to see this being an issue. It's possible that there were no database approaches available to the crew, which would X out the use of profile and put them in dive and drive mode with vertical speed. Also the NTSB reported 140kt approach speed is pretty high for an A300, probably pretty close to the landing weight limit for a flaps 30/40 landing

Excellent post!

I doubt we will hear on the CVR from the PNF, "PAPI in sight."

[filejw]

Airbubba..Just repeating what a Birmingham controller wrote on a union blog....

[A Squared]

Originally Posted by aterpster

In the U.S. an unrestricted VGSI must clear all obstacles 1 degree below the commissioned angle, out to 4 n.m, and with a 15 degree splay.

Just out of curiosity, I projected this out to the see how much clearance would be afforded by this. I used the spot which JC identified on Google earth which everyone mocked him for. Turns out that a obstacle clearance plane of 2.2degrees would give a clearance of about 65 feet to an airplane exactly on a 3.2 degree PAPI, 3700 ft from the threshold. That's remarkably close to the clearance from the terrain which JC figured using Google earth. Granted, the first point of impact was quite a bit further out and lower than this, so this is only peripherally related to the accident. However it is interesting to note how little clearance you might have over a half mile from the threshold and perfectly centered on the PAPI.

[Airbubba]

Airbubba..Just repeating what a Birmingham controller wrote on a union blog....

Thanks for sharing it, most of what I know about the mishap outside of the NTSB briefings is similarly anecdotal.

Possibly not having any Rwy 18 NavDB stuff would definitely explain why they opted for the LOC (manually tuned I assume) and not the RNAV.

Perhaps not a player with UPS but some of the A306's I flew years ago had FMS's with very limited memory and internationally some procedures would not be in the database or, even worse, waypoints and constraints would be missing. I flew with Sperry, Smiths and some other kind of FMS, I can't remember. I think UPS only uses the A306's domestically these days so maybe this isn't a problem anymore.

[Capn Bloggs]

Originally Posted by A Squared

Turns out that a obstacle clearance plane of 2.2degrees would give a clearance of about 65 feet to an airplane exactly on a 3.2 degree PAPI, 3700 ft from the threshold. That's remarkably close to the clearance from the terrain which JC figured using Google earth. Granted, the first point of impact was quite a bit further out and lower than this, so this is only peripherally related to the accident. However it is interesting to note how little clearance you might have over a half mile from the threshold and perfectly centered on the PAPI.

This is ridiculous. How could such an approach be allowed, especially at night, with such little terrain clearance? The threshold crossing height is 50ft; 6000ft/1nm further back, you only have another 50ft higher (100ft in total) to give 2W/2R on the PAPI, with wheels hanging probably 30ft below that?

[A Squared]

This is ridiculous. How could such an approach be allowed, especially at night, with such little terrain clearance? The threshold crossing height is 50ft; 6000ft/1nm further back, you only have another 50ft higher (100ft in total) to give 2W/2R on the PAPI, with wheels hanging probably 30ft below that?

Surprising isn't it? There's always the possibility that I misunderstood aterpster, but that's how it pencils out.

[Sorry Dog]

It will be interesting to see when more of the CVR is released. Considering the AP wasn't disengaged (or dis-engaged itself) until 1 second before it, one could wonder if they recognized the terrain issue in time to realize how low they really were. In context of the "black hole" visual anomaly, wouldn't it be rather ironic if lack of lighting and other visual features from the houses removed (for safety reasons) actually ended up contributing to the accident.

I really wonder if this still would have happened if there was some kind of marker lighting on hill where the cockpit currently resides or on the plateau towards the threshold.

I know I would rather have that to help judge the distance.

[mixduptransistor]

wouldn't it be rather ironic if lack of lighting and other visual features from the houses removed (for safety reasons) actually ended up contributing to the accident.

To be clear, the homes were purchased/removed for noise abatement, not safety.

[skysign]

Surprising ....but real....

As I posted before the last 1/2 to 1 mile you litteraly buze the hill all the ways down to the RWY. It is like doing a low flyby over a downhill slope to the rwy.

It has been several years since I have done that approach, so I can not really say how high above the terrain you are. But the first night I did that approach the only think that came to my mind was " Only sh.. that is close !!! and that is being on the Papi.
In my opinion, on a "normal approach" the ground appear that close when you are really close to crossing the threshold and not being 1/2 to 1 mile away from the rwy.

[WillowRun 6-3]

The Careful Enough Statement of a Problem -

....implies the solution.

"This is ridiculous. How could such an approach be allowed, especially at night, with such little terrain clearance? The threshold crossing height is 50ft; 6000ft/1nm further back, you only have another 50ft higher (100ft in total) to give 2W/2R on the PAPI, with wheels hanging probably 30ft below that?"

So I'm assuming that FAA is the authority that decides what specific approach parameters will be published as officially sanctioned (or otherwise designated as safe and effective). And that in the ordinary course of the business of regulation of civil aviation in the US, FAA will await the outcome of the NTSB investigation and report process, prior to changing the approach parameters. Or is this not correct: might FAA change the published approach parameters on an interim basis? Has this happened after other mishaps on final? (And I'm sorry if there is a term more technically appropriate than 'parameters' for the (evidently) altitude and/or glideslope limits.)

Also, is there some periodic review process by which FAA checks on whether approach corridor limits have been shown, through the experience of the aviators who flew them, to need adjustments? It is easy to have the sense that FAA has its hands full just keeping the creaky ship afloat, and things like checking for any needed updates to APP corridors is a luxury the taxpayers have neither demanded nor realized they need to pay for. How does it work?

[ironbutt57]

obstacle clearance from FAF to VDP/MAP is SUPPOSED to be 250'

http://www.skybrary.aero/bookshelf/books/811.pdf

[roulette]

WillowRun 6-3, A-Squared, Capn Bloggs & All those who believe GoogleEarth Elevation/Hgt data:

GoogleEarth provides good approximations of real-earth elevations and is good for relative differences, but it cannot be considered accurate at any single location within confidence levels that you'd want to consider whether you have appropriate clearance or not for approach/dep procedure design. Sometimes it's spot on to the nearest metre, other times not. Depending on where the location is and also often where the location is in relation to nearby terrain and buildings, the elevations can easily be out by up to say ±8 mtrs / ±26.5ft. This is due to the various sources of the data and how the data is modelled. The positional referencing of the image data to WGS84 lat/longs however is far more accurate.
In a sense it's a bit like the FlightAware data, in that it's sampled from actual observations using available data sources. The data is not certified to a particular integrity level though and thus needs to be used and interpreted with care and qualifications, depending on the purpose.

If Capt Bloggs and ASquared calcs are actually more or less correct, then that will come out in the investigation as result of FLT inspection of the PAPIs and resurvey of the terrain and remaining obstacles near the crash site, as well as review of the procedure flown (as selected in the FMS and as flown). In any case, if as it seems to be that the aircraft was simply way too low at that point, the issues you raise may, or may not, in fact be relevant or significant.


ironbutt57:
Yes, you're right: obstacle clearance for an NPA (with a FAF) in the primary area of a final approach segment is 250' (246' actually, or 75m). Take the highest obstacle between the FAF and the MAPt, add the obstacle clearance margin (and any comfort buffer that the designers feel appropriate for whatever reason, including taking into account potential inaccuracy tolerances in the obstacle and terrain data they're using) and round up to the nearest 10' = MDA.
Use of stepdown fixes allow interim descent points in the final segment, obviously.

However, when flying on the VDA (or even dive & drive), once you get to the MDA, that's it: can't continue descent unless RWY in sight and intending to land, and from then on you're in the visual segment and obstacle clearance is up to the pilot.

See also the next bit.

For info: re Obstacle Clearance in the Visual Segment (of Charted APCHs):

Some time back ICAO mandated that all States should assess new and old procedures (when reviewed/amended) for obstacles in the visual segment - ie, between the MDA/DA and the THR (actually, 60m prior usually).
Basically this is done by assessing a pseudo surface at 1.12° less than the promulgated APCH angle. Safety assessments, mitigations (which could include removal or cutting down of obstacle) &/or amendments to the procedure would be required if an obstacle penetrated the visual segment surface. Safety assessment might also indicate additional mitigations, such as additional marking/lighting of obstacles / terrain in the visual segment - eg, like the hill area 1/2-1 mile prior to the runway - ref skysign's post #380).

As I posted before the last 1/2 to 1 mile you litteraly buze the hill all the
ways down to the RWY. It is like doing a low flyby over a downhill slope to the
rwy.

It has been several years since I have done that approach, so I can
not really say how high above the terrain you are. But the first night I did
that approach the only think that came to my mind was " Only sh.. that is close
!!! and that is being on the Papi.
In my opinion, on a "normal approach"
the ground appear that close when you are really close to crossing the threshold
and not being 1/2 to 1 mile away from the rwy.

Aterpster can elaborate on current FAA implementations of this ruling if required (in addition to the analysis that they do of 34:1 surfaces).
In any cases, statements regarding obstacles in the visual segment and implications for pilots are contained in the FAA AIM (relevant extracts are in this thread).
Note also previous posts on the 1° clearance for the PAPIs when installed and calibrated as additional mitigation.


Missed Approach:
To those who have discussed the relevant techniques when reaching the MDA whether flying dive & drive (heavies seriously still doing that?) or CDFA (or even DDA = MDA+50 for CANPA technique):
Unless the operator has a specific approval to do otherwise, the MDA is a legal minimum and busting it is a serious no-no.

With regard to this thread though, they'd gone well past the MDA and info provided by NTSB indicates that they were proceeding to land, so further discussion of handling the missed in this case is irrelevant.


ATerpster:

Thx for trying to correct the various misunderstandings of the different types of APCHs with regard to LNAV, LNAV/VNAV, RNAV GNSS APV, LPV, RNAV RNP AR, etc. Saved me some time. Well done on providing all of the other relevant AFM and FAA material for reference too.
Ian W: If, after reading Aterpster's reply and other material from qualified sites, then you could continue the discussion on the TechLog, but not in this thread as those issues are irrelevant here.

[Speed of Sound]

and UPS normally use r/w 18 if available .

Given that the UPS cargo area is at the south west end of 6/24 I can't see any reason at all why anyone would choose to land a widebody on a shorter, narrower runway with less navaids and a two and a half mile taxi to get where you need to be.

Which brings up another point about the flight planning on UPS1354. Whether or not 6/24 was ever an option. Although 6/24 was NOTAM'd closed until further notice it is being claimed that aircraft were being landed on it less than 15 minutes after the crash.

So when was 6/24 reopened and when was the decision to reopen made?

Were the crew of 1354 made aware of 6/24's availability albeit after a short delay and a few extra track miles? I appreciate that if your preferred runway is NOTAM'd 'CLSD UNTIL UFN' you plan for another approach but if it is onto an unfamiliar runway that puts you down in the wrong part of the airport surely at some point in the sector the question would have been asked "Does anybody know when 6/24 is due to reopen?"

[ironbutt57]

guess what we all don't know yet is whether the airplane had levelled off, then drifted below the appropriate altitude for that segment of the approach, (unlikely since the autopilot was engaged) or the aircraft was on a high rate of descent and continued below acceptable altitudes in an attempt to regain profile..(likely given the "sink rate" callout") wait and see what the DFDR reveals

[de facto]

All that said, if you're continuing down a NPA based on approach lights alone, you are already rolling the dice IMO. Because it's allowed by regulation, my company has decided to give the option to pilots. Whether or not it is good practice in reality is another discussion.

Ahhh now i know why at some airports,there is a note that says,for those pilots: Do NOT mistake the highway lights for the approach lights.

[Cows getting bigger]

Maybe 06/24 was reopened because there was a pile of smouldering wreckage and a few fire engines in the undershoot of 18.

[Ian W]

Capn Bloggs
Quote:
Originally Posted by A Squared
Turns out that a obstacle clearance plane of 2.2degrees would give a clearance of about 65 feet to an airplane exactly on a 3.2 degree PAPI, 3700 ft from the threshold. That's remarkably close to the clearance from the terrain which JC figured using Google earth. Granted, the first point of impact was quite a bit further out and lower than this, so this is only peripherally related to the accident. However it is interesting to note how little clearance you might have over a half mile from the threshold and perfectly centered on the PAPI.

This is ridiculous. How could such an approach be allowed, especially at night, with such little terrain clearance? The threshold crossing height is 50ft; 6000ft/1nm further back, you only have another 50ft higher (100ft in total) to give 2W/2R on the PAPI, with wheels hanging probably 30ft below that?

And as Airbubba raised a lot earlier in the thread, the small vertical error margin does make altimeter setting something to return to - the difference between 29.92 and BHM QNH 29.97 just happens to be 50ft low - "with wheels hanging 30ft below that".

[Capn Bloggs]

if you're continuing down a NPA based on approach lights alone, you are already rolling the dice IMO.

Here, there are very few NPAs that have a true Visual Segment, where the visibility minimum is less than the distance to the runway. On any NPA, I would hazard a guess that if following a 3°-ish profile, you will always be able to see the threshold at the MDA if the vis is on the minimums or greater.

Originally Posted by Ian W

the small vertical error margin does make altimeter setting something to return to

True, however at this point the crew would be operating visually on the PAPI and well below the MDA, and the altimeter value would be not a consideration.

[legomaniac]

I wonder if the crew was fighting sleep. Micronaps are not uncommon flying overnight.

[captjns]

The FAA as well as EASA need to establish and approve GPS/LPV approaches at all airports in lieu on the traditional N/P/As.

It will help to keep pilots out of the dirt a mile short of the runway, with the advances in the glass cockpit for both GA and Transport Category Aircraft. TAWS technology could be improved too.