SWA MDW crash

[SegmentKing]

Boeing Boy, would you have landed at MDW on that day, in these conditions?

 

[BoeingBoy]

IIRC from when I was on the bus (UA1446 was an A320 in TED config) the highest setting for landing was MED. The A320 has three settings LOW MED and MAX. MAX is the boeing equivilant of RTO and MED is the landing maximum setting equal to the Boeing setting of MAX for landing. The inference of the post that MED was used as a less than best setting is incorrect. While labled MED it is the HIGHEST Airbus landing setting and was being utilized correctly for the conditions.

Never having flown the Airbus (and while the NTSB didn't label it "Ted" they did say it was an A320), thanks for the info. I would point out, however, the difference in brake pressure that the autobrake systems applied on the UA flight vs the WN flight (448# in MED vs ~2500# in MAX). Sure doesn't seem "equivalent" to me.

By the way 11 seconds of deceleration is a significant timeframe. Would you happen to have the groundspeed from which the brakes were deactivated on UA 1446?

Not at my fingertips, but it is included in the data traces from the FDR which is available on the NTSB's site.

-----
[Edit to add the following]
Ok, I've looked up the traces. UA1446 deactivated the autobrakes at approximately 75-80 kts (depending on whether you want g/s or not). After disengagement, the brake pressure increased to approximately 800#, from less than 500# that the autobrakes were applying.
-----

Close scrutiny is being paid at some techniques used by the SWA crews . One involves the slamming of the throttles to idle. In this instance there is a chance the throttles bounced off the stops due to the techniques used by the SWA pilots. This small bounce may have played a role in the Capt not being able to pull the reverser levers up. It may very well have been the F/O that finally pulled the to the aft point allowing for selection of reverse thrust.

The fact that not having the forward thrust levers at the idle stops could prevent reverse thrust activation was mentioned in one of the reports - I think the System Group's. As far as being under "close scrutiny" or having to do with technique - that's another question.

Also the SWA non-use of Autothrottles, LNAV/VNAV is a workload issue that could come under review. SWA is setting up crews with long duty days to not have available safety elements to protect the passengers. By not using automation installed on the aircraft the safety chain is being stretched on a daily basis.

Where to start....

Lots of things "could come under review" - as they should if they have any relevance to this investigation. Like the delay in following the airport's policy to close a runway for snow removal when a report of "poor" braking action is received for any part of the runway (they were setting up spacing to do that after the inbound flight behind WN1248). Like the somewhat lax FAA policy concerning runway margin requirements in degraded braking situations (notice the "clarification" they're issueing). Like the policy of "grandfathering" airports so they don't have to install adequate overruns/engineered material stopping devices. The list is long.

WN's scheduling results in few "long duty days" compared to the legacy carriers - does that mean they're "safer" than the rest?

Automation is a tool, not a cure-all. Like any tool, it has pros and cons. For what it's worth, there wasn't any automation installed that this crew didn't use, apparently (including the HUD, although I'd have to go back to be absolutely positive - wonder if Ted used theirs?). By your definition, that means that this crew was operating as safely as they possibly could.

When all is said and done, WN1248 never achieved a deceleration rate greater than ~0.25 gee while applying ~2500# of brake pressure or more even after T/R were deployed - for most of available runway the deceleration rate was pretty stable @ 0.2 gee.

For comparison, UA1446 achieved 0.3 gee deceleration with the autobrakes applying less than 500# of brake pressure and T/R, and peaked at 0.4 gee deceleration after autobrakes were deactivated and brake pressure increased to ~800#.

That is the difference between parking at the gate and running off the end. Given WN1248's deceleration rate after T/R deployment, the result would have been the same if there had been no problem deploying the T/R's. That leads me to the tentative conclusion that runway conditions deteriorated either more than the braking action reports indicated or faster than the reports could keep up with. As the NTSB simulator recreations show, only poor braking action for the full runway length resulted in going off the end - not "fair, poor at the end" or "fair to poor", but "poor" for the entire runway.

Jim

 

[LD3]

Please excuse my ignorance, but at one time were the CVR transcripts non-releasable to the public? I thought that the only the air to ground transcipts were released.

Thank you.

 

[magsau]

Never having flown the Airbus (and while the NTSB didn't label it "Ted" they did say it was an A320), thanks for the info. I would point out, however, the difference in brake pressure that the autobrake systems applied on the UA flight vs the WN flight (448# in MED vs ~2500# in MAX). Sure doesn't seem "equivalent" to me.

The Airbus A/B works from the IRU's g/s to give a deceleration rate. I do remember this from the fleet. That is to say if the decelrate was not adequate it would have provided greater pressure. Again there are only three settings in the bus, LOW MED and MAX. Max is only allowed for RTO and MED is the highest landing setting, equal to using the highest selectable boeing A/B setting for landing. While you are at it do you have the Vref of the two aircraft, TCH Airspeeds and Ldg Gross weights of the two aircraft. Those are also factors in stopping distance. I would not count on my aircraft using the same runway distance as a Gulfstream just because it was a jet.
Not at my fingertips, but it is included in the data traces from the FDR which is available on the NTSB's site.

Needless to say litigation will be heavily involved in this at every angle.

 

[Breath_of_Fresh_USAirways]

When all is said and done, WN1248 never achieved a deceleration rate greater than ~0.25 gee while applying ~2500# of brake pressure or more even after T/R were deployed - for most of available runway the deceleration rate was pretty stable @ 0.2 gee.

For comparison, UA1446 achieved 0.3 gee deceleration with the autobrakes applying less than 500# of brake pressure and T/R, and peaked at 0.4 gee deceleration after autobrakes were deactivated and brake pressure increased to ~800#.

Hi, Jim.

Your insights, experience, and evaluation of this accident are very much appreciated.

I was wondering, to be sure that we're comparing apples to apples, here, would the touch-down speeds of the above 2 descriptions be similar -- if touchdown speed even matters when it comes to gee.

And, besides the snowy/wet runway, didn't SW have a tailwind (or a fairly large tailwind component), also? Would that have had an affect on he decelation achieved, or not?

Also, IIRC, the Airbus has carbon brakes and the 737 has steel brakes -- does that make a difference, too? I mean, a 2000# pressure difference seems quite significant, in my opinion, so I suspect that the systems, too, are quite different. But, I could be wrong -- wouldn't be the 1st time; won't be the last, either. B)

 

[magsau]

This airplane will fly like a golf ball with the dimples in place. Just hope it does not fly like I had hit it, otherwise it will only make right turns.....

IDENTIFICATION
Regis#: SWA1916 Make/Model: B737 Description: 737
Date: 06/27/2006 Time: 2215

Event Type: Incident Highest Injury: None Mid Air: N Missing: N
Damage: Unknown

LOCATION
City: PHOENIX State: AZ Country: US

DESCRIPTION
SOUTHWEST AIRLINES SWA1916, A BOEING B737 ACFT, ON FINAL TO LAND
ENCOUNTERED A HAIL STORM CAUSING DAMAGE TO THE AIRCRAFT, NO INJURIES
REPORTED, PHOENIX, AZ

 

[BoeingBoy]

Please excuse my ignorance, but at one time were the CVR transcripts non-releasable to the public? I thought that the only the air to ground transcipts were released.

Thank you.

Transcripts have always been released (well, as long as I can remember, anyway). It's the CVR tape recording that's not supposed to be released.

The tower tapes are routinely released.

I was wondering, to be sure that we're comparing apples to apples, here, would the touch-down speeds of the above 2 descriptions be similar -- if touchdown speed even matters when it comes to gee.

Well, there's always the possibility. If the extra speed causes hydroplaning, braking action would be effectively nil until speed dropped enough to stop the hydroplaning (of course, the spoilers are supposed stop that too). Other than something like that, speed would affect the stopping distance but not the deceleration rate. The brakes are capable of stopping tire rotation even on a dry runway - it's the antiskid's job to insure that doesn't happen. So the brake system as a whole will extract whatever deceleration is available up to what's commanded - remembering that when used the autobrake setting determines the deceleration rate commanded.

The limiting factor in this case was undoubtedly coefficient of friction - how much deceleration can the brakes generate thru the tire contact with the runway. As magsau explained for the A320-series, the autobrake landing settings on the 737-200/300/400 (and presumably on the 700) attempt to attain a deceleration rate, relying on the anti-skid system to prevent locking the brakes. MAX will attempt to achieve well over what any of the aircraft documented achieved that night, meaning the limiting factor was the tire/runway interface.

And, besides the snowy/wet runway, didn't SW have a tailwind (or a fairly large tailwind component), also? Would that have had an affect on he decelation achieved, or not?

Yes, they had a tailwind - as did the 5 aircraft that landed before them. I didn't notice whether the FDR traces included a wind component up to touchdown, but based on what was reported I think it was about an 8 kt tailwind component.

Also, IIRC, the Airbus has carbon brakes and the 737 has steel brakes -- does that make a difference, too?

Don't know about the Airbus - magsau could answer that. The 737 (presumably including the -700) has steel "rotors" as you said. Assuming that the wheels don't stop turning, I don't know which would provide more ultimate stopping power. As I said above, however, I assume that the Airbus brakes would also be able to "lock the wheels" if it weren't for the antiskid system. So it comes back to how much brake effectiveness can be extracted from the tire/runway contact.

I mean, a 2000# pressure difference seems quite significant, in my opinion, so I suspect that the systems, too, are quite different.

Based on what magsau said about MED being equivalent to MAX on the 737 and the brake pressure applied by the autobrakes on the UA A320 vs WN 1248, I also assume that there's some differences in the system. It could be that the Airbus has some sort of feedback from the antiskid so that only as much brake pressure is applied as the antiskid system says can be used. In other words, if the antiskid is cycling, that may limit the amount of brake pressure applied. Maybe magsau can fill us in. In the 737-200/300/400, that's not the case - the autobrakes will increase pressure up to the limit if necessary to achieve the commanded deceleration rate. There's no connection that allows the antiskid to say "Stop applying more pressure - I've got all I can use". I assume that the 700 is the same, considering the brake pressure applied by the autobrakes that night.

Jim

 

[magsau]

Here is what I found on the company web site manual section for Airbus A/S and autobrakes:

Antiskid: The antiskid system provides maimum braking efficiency by maintaining the wheels at the limit of an impending skid. If a skid is detected by the BSCU (that is airbus for brakes steering control unit..that one I remember!), brake release signals are sent to the normal and alternate servo valves as well as to the ECAM which displays the released brakes. Once the skid is no longer detected, the antiskid system return to normal skid detection and protection. The system is automatically deactivated when:
Groundspeed is below 20 knots.
A/SKD % NW STRNG switch is selected off.

AUTOBRAKES SYSTEM
The autobrake system operates through the normal brake sytem to maintain a selected deceleration rate during landing and to reduce stopping distance during a rejected takeoff

Arming
Autobrakes can be selected and armed when:
Green Hydraulic pressure is available.
Antiskid is powered
No failures exist in braking system
At least one ADIRU is available.

 

[BoeingBoy]

Boeing Boy, would you have landed at MDW on that day, in these conditions?

Sorry, SegmentKing. For some reason the little button for "go to first unread post" skipped by your post till just now.

Sitting at home without all the available paperwork, dispatch access, etc, it's hard to say for sure. All I have available is what's in the NTSB data and the airplane manual.

Based on just that info, most likely.

The weather was above landing minimums, other aircraft had landed and reported at worst "Fair to Poor" braking action, and the chart in the airplane manual for braking action "Poor" says I'd have about 700-1,000 ft of extra runway available (this last depending on whether I was flying a -400 @ 120,000# or -300 at max landing weight of 114,000# - Wn 1248 weighed ~119,000#).

Jim

 

[LD3]

Transcripts have always been released (well, as long as I can remember, anyway). It's the CVR tape recording that's not supposed to be released.

The tower tapes are routinely released.

Thank you Jim....

 

[BoeingBoy]

While you are at it do you have the Vref of the two aircraft, TCH Airspeeds and Ldg Gross weights of the two aircraft. Those are also factors in stopping distance. I would not count on my aircraft using the same runway distance as a Gulfstream just because it was a jet.

Sorry, magsau. I haven't been able to find Vref used by the crew of the UA flight. As I recall, the Vref used by the WN crew was in the CVR transcript.

From the narrative, the UA flight's touchdown airspeed was 123 Kts, gndspd was 128 kts, and gross weight was 114,040#. Maybe you can expain this since I know nothing about the Airbus - there are three FDR readouts for A/G sensors with two of them showing "gnd" ~1 second before the 3rd. The FDR data airspeed showed a negligable difference (119.03 kts vs 119.28 kts) between the two times, but the gndspd showed 128 kts when the 1st two registered "gnd" and ~125 kts when the 3rd registered "gnd" (126 kts 1/2 second before and 124 kts 1/2 second after). Are there A/G sensors on both mains and the nose gear?

WN 1248's narrative said airspeed was 124 kts, gndspd was 131 kts, and gross weight was 118,280# at touchdown. The FDR readout is again slightly different, showing airspeed was 126 kts just over 1 second before and 124.25 kts 0.15 sec after the A/G sensor registered "gnd". Groundspeed, per the readout, was 133 kts 1 seconds before and 131.5 kts 0.15 seconds after the A/G sensor registered "gnd".

Jim

ps - about forgot the G-IV....

You're right - two different aircraft types can definitely have different stopping charasterics. Aside from the obvious differences that would affect stopping distances for one specific aircraft - speed, weight, and wind - there are the factors of brake effectiveness, tire contact area differences, tire width (in contamination especially), reverser effectiveness, etc.

Never having flown the Gulfstreams, I have no idea how it stops. Knowing how the 737-200/300/400s stop, however, I doubt that the G-IV could do much better than the 3000 feet at max landing weight (dry runway, obviously) that the Boeing can achieve.

 

[magsau]

Sorry, magsau. I haven't been able to find Vref used by the crew of the UA flight. As I recall, the Vref used by the WN crew was in the CVR transcript.

No problem.

Are there A/G sensors on both mains and the nose gear?

I believe this is correct. The struts work via a system of computers and an LGCIU (landing gear control...interface unit, I think) that uses each strut position for certain items and the strut on the nosegear is used for various system changes. Too many computers and no control+ALT+Delete switch is the thing I remember the most.

 

[Diversion]

19:08:37.4
HOT-2: if it's [braking action] poor we don't wanna hear it.

19:08:38.8
HOT-1: no.

Wow.

 

[KCFlyer]

19:08:37.4
HOT-2: if it's [braking action] poor we don't wanna hear it.

19:08:38.8
HOT-1: no.

Wow.

I notice you post mostly in the AA forums. There is something about "glass houses" that one needs to bear in mind:

The controller did not receive any further transmissions from flight 1420. About 2349:02, the first officer asked the captain, "want 40 flaps?" The captain indicated that he thought he had already called for the landing flaps, after which the first officer stated, "forty now." About 2349:10, the controller informed the flight crew that the wind was 330 degrees at 28 knots. Two seconds later, the captain stated, "this is a can of worms."

Wow.

 

[Busdrvr]

KC, if it was reported that all SWA pilots smoked crack, you'd be here telling us how it enhances their senses and makes you feel safer... I honestly haven't seen an objective post from you in a couple of years.

For the record, I honestly don't see any flagrant violations of the FARS. Was there some levity in the cockpit below 10? sure. But personally, I think it was more along the lines of "breaking the tension" (and there WAS a LOT of tension about doing this approach). I can't see how the remarks detracted from the safety of the flight.

Two other things I noticed.
1. during cruise, they were discussing the poor weather and the use of Autobrakes (Good), but it appears that the company policy for autobrake use was unclear (bad).
2. Here we have ON TAPE, a SWA capt telling an SWA F/O that it would be perfectly fine with him to lie about the times to stay legal for the next leg. SWA NEEDS ACARS!!!