Indonesian 737 down?

[flyboy2548m]

http://avherald.com/h?article=4e18553c&opt=0

On Jan 19th 2021 the KNKT reported that the FDR contained 330 parameters of a total of 18 flights. The KNKT have already developed a general picture of the accident, however, need further data from the cockpit voice recorder before publishing any information. Inspections of Boeing 737-300, 737-400 and 737-500 in Indonesia have been ordered.

So it was more than just pilot error....

Noted. Both places, thank you.

[tds]

Gabriel - I think you're assuming the two facts (that the FDR data has been downloaded and that inspections have been ordered) are related. Besides appearing in the same paragraph in an Avherald story is there any reason to think that?

Ordering extra inspections / determining compliance on previous inspections, ADs etc, could equally well be "we know nothing but it can't hurt" ass-covering.

[Gabriel]

Gabriel - I think you're assuming the two facts (that the FDR data has been downloaded and that inspections have been ordered) are related. Besides appearing in the same paragraph in an Avherald story is there any reason to think that?

Ordering extra inspections / determining compliance on previous inspections, ADs etc, could equally well be "we know nothing but it can't hurt" ass-covering.

Right, could be... I was just loosely connecting the dots: They have a picture of how the accident happened, they want extra confirmation from the CVR before commenting, they ordered inspections of the specific family that suffered the accident, but not specifically for the airline that suffered the accident. And "inspection of airplane" sounds like going and checking the actual plane physically, not just a documental verification that ADs or such were completed. Also, if you give an order to inspect the whole fleet of a family of planes in a country, the instruction will be more or less specific, not just "go and check them".

It sounds to me that, based on what they learned so far, they suspect something more or less specific was wrong with the plane.
I can be wrong and I hope I am.

[3WE]

Todays WSJ mentions "a problem with the autothrottle on one engine".

I will update my speculation:

Weather MIGHT be contributing as a distraction, but is not a major contributing factor. (Imagine being in some lively turbulence- or "simple" IMC/disorientation after the plane goes to hell.)

The plane, flying automatically, got slowed up and crossed up.

Dude's didn't monitor airspeed enough.

Plane gets to minimum-controllable airspeeds and then stall, spin, crash, burn and did died (with some likelihood of mismanagement).

Total outsider ass-hat pontification that we need a nice gentle, but clear warning- "Dudes, we are slower than we should be". This is in addition to OTHER "Oh shit" speed warnings of "Dude, we're gonna stall real soon". A little heads up so folks can calmly address things, instead of "startle-factor-relentless pull ups" (This is a recurring scenario). (Additional we's not italicized as I refer to actual pilots)

That along with better screening, training, culture improvements, and more oversight, of course.

[tds]

You don't even have to stall for asymmetric thrust (re "autothrottle on one engine") to cause an upset if corrective inputs aren't applied.

[3WE]

You don't even have to stall for asymmetric thrust (re "autothrottle on one engine") to cause an upset if corrective inputs aren't applied.

Concur...several things can happen with dangerously slow speeds (and can cascade into each other)...SSCBDD is just a nice phrase with some accuracy.

[Gabriel]

You don't even have to stall for asymmetric thrust (re "autothrottle on one engine") to cause an upset if corrective inputs aren't applied.

Concur...several things can happen with dangerously slow speeds (and can cascade into each other)...SSCBDD is just a nice phrase with some accuracy.

Concur. Given how fast it went down (less than 1/2 minute from more than 10K ft to some feet BELOW water), i think that something in the line or rolling inverted fits better than stall. Of course that a stall can be followed by rolling inverted.

[Gabriel]

...SSCBDD is just a nice phrase with some accuracy.

Not even all-mighty Google helped me with that.

[3WE]

...SSCBDD is just a nice phrase with some accuracy.

Not even all-mighty Google helped me with that.

Context matters: Stall, Spin, Crash, Burn, Did Dieded.

This place is cutting edge. There and Google are yesterday’s news,

[Gabriel]

...SSCBDD is just a nice phrase with some accuracy.

Not even all-mighty Google helped me with that.

Context matters: Stall, Spin, Crash, Burn, Did Dieded.

This place is cutting edge. There and Google are yesterday’s news,

In a spin you are not falling fast enough to make more than 10K feet in less than 30 seconds.

[3WE]

In a spin you are not falling fast enough to make more than 10K feet in less than 30 seconds.

Quit being pedantic. We will need the final report to know the EXACT sequence and maneuvers. Going out of control leads to all sorts of nasty things and gyrations and who knows if it even stayed in tact. Was there SOME stalling or spinning? Who knows, but it’s POSSIBLE.

As an airplane slows to Vmc(or whatever it technically is) in a climb with an engine “out” you have a pull up and crossed controls and are nicely set up to enter a stall/spin/spiral/structural break up, blah, blah, blah.

[flyboy2548m]

Phugoid.

[3WE]

Gabieee: So it was more than just pilot error....

This AM, a Yay-who article specifically stated: Power imbalance caused a roll and dive, and said it happened around 8000 feet of altitude.

(No real difference from all of our ass-hat speculation, but more of a confirmation that that is what happened.)

Conversely, if a plane starts rolling, shouldn't you grab the yoke and power and make corrective inputs...(Including checking and adjusting airspeed with potential measured nose-down inputs?)...

...and shouldn't 8000 feet be enough room for recovery unless you really blow it?

In other words, rather basic pilot error is the primary cause.

Footonte: Not attacking Gabriel's post, it just provided a good lead in.

[Not_Karl]

AvieHeraldie has a link to the preliminary report.

[Gabriel]

AvieHeraldie has a link to the preliminary report.

And... according to it, passing 8000 ft the autothrottle started to slowly reduce thrust on one side only, and the AP started making increasing compensation aileron inputs, until upon reaching 10900 ft with the AT still reducing thrust further and further on one side only, the AP reached its operation limit, disconnected, and let go on the controls which would have caused the ailerons to return to neutral eliminating all the compensation the AP was doing, the plane quickly rolled inverted and 20 second later it was under water.

It seems that the pilots did a bad job monitoring the flight before the upset (thrust levers and engine instruments had been increasingly split for almost 1 minute and 30 seconds, and either airspeed or rate of climb must have diminished with the reduction of total thrust), but once the AP disconnects... having a simultaneous AP disconnect warning for no apparent reason together with the plane rolling like crazy for no apparent reason, man I feel sympathy for them, very tough to get your shit together, react and recover in such short time with no notice, with everything happening together apparently out of the blue. Especially when apparently upset recovery training was lacking.

[3WE]

...man I feel sympathy for them, very tough to get your shit together...

I dunno...IN THIS CASE, the instruments and manual controls were working normally and it’s not a friggen fighter plane...

The gluteal AI works real good for short time periods...

Beeeeeep, oh shit, we’re rolling left...grab yoke hard right input...check speed and nose up/nose down...correct accordingly...Blue side up....

Three seconds and it should have been going the right way.

At least that’s how it works on my bicycle.

[Gabriel]

At least that’s how it works on my bicycle.

So you never fell?

Let's wait until the final report to see how violent the uncommented upset was. But if it took 20 seconds from fat dumb and happy to impact, there was not a lot of time for startle + getting shit together + proper reaction + attaining recovery (because even proper reaction doesn't guaranty that you will achieve recovery at a positive altitude AGL).

[Gabriel]

At 14:40:05 LT, the FDR data recorded the aircraft altitude was about 10,900 feet, which was the highest altitude recorded in the FDR before the aircraft started its descent. The AP system then disengaged at that point.

At 14:40:10 LT, the FDR data recorded the autothrottle (A/T) system disengaged and the pitch angle was more than 10 degrees nose down. About 20 seconds later the FDR stopped recording.

So it was 25 seconds from FDH to H2O. Even when it was not a fighter.

[3WE]

So it was 25 seconds from FDH to H2O. Even when it was not a fighter.

I get that.

The flip side is that aeroplanies can go wonky and ideally you are correcting it within 3 seconds...

I’m very serious.

I feel sympathy for ruder hard over guys and AA-191 where the airplane is not working right...

This plane had fully functional controls and instruments.

Autopilots occasionally cut out...

“Listening” to your buttocks, locking on the AI, making a correction and checking speed, power, with 10,000 ft of buffer ought to be salvageable...

...hell, there’s training for wake turbulence recovery at MUCH MUCH lower altitudes...

Sure, let’s add ANOTHER warning light that “the autopilot is working pretty hard to stay level”...no one likes startle factor...

PS- need to fire up MSFS and royally screw up and see what recovery looks like...

Here’s something for comparison- I counted 25 seconds for this recovery. I could not read the altimeter. It’s not apples to apples, but the airplane gets inverted.

https://youtu.be/L2CsO-Vu7oc

[Gabriel]

Are you serious?

These are PRFESSIONAL TEST PILOTS testing a STALL IN A SIDESLIP, flying manually, automation off, their hands already in the controls, their eyes already in the instruments, super hyper trained in upset recovery, and mentally expecting that things can go wrong at anytime and mentally ready to react to that.

They react INSTANTLY as the plane starts departing controlled flight (interestingly, turning the yoke to the left in response to an uncommanded and incontrollable roll to the left, which may be counterintuitive but the right thing to do together with right rudder and down elevator, what the pilot does simultaneously).

AND IT STILL TOOK THEM 25 SECONDS AND GOD KNOW HOW MANY THOUSAND FEET.

Now put an average pilot with less than ideal upset recovery training (even for an average pilot standard) flying dumb fat and happy on autopilot while looking at the radar and negotiating headings and altitudes to avoid weather and conflicting traffic when suddenly and out of the blue the autopilot goes off and the plane goes inverted, and do you have ANY expectation that the pilots response will be anything like that video?

[3WE]

... and do you have ANY expectation that the pilots response will be anything like that video?

Yes, period. (And I’ll ignore your absolute statement).

Repeating: I expect (within 3 seconds) to have the attitude indicator analyzed and corrective control inputs...

A 737 takes a few seconds to roll inverted...

They had 10,000 feet.

I THINK there is upset training...this was an upset...

I don’t expect perfect performance, but I do expect gross, but near-full corrective inputs...just like you hit a wake on final approach where you have a lot less altitude to phugoid with.

[Gabriel]

Repeating: I expect (within 3 seconds) to have the attitude indicator analyzed and corrective control inputs...

A 737 takes a few seconds to roll inverted...

Well, it took the 717 / MD95 2 seconds to go from 15 degrees tight bank to 90 degrees left bank. The 737 was already baked and and still rolling to the side it inverted (despite the AP applying maximum input to the other side) when the AP let go and the ailerons got neutralized.

They had 10,000 feet.

They had 25 seconds.
And how many feet they the test pilots lose again? (I counted 18 numbers passing by, for sure these were not hundreds, perhaps thousands or 500 hundreds).

Again, waiting for the final report to see how violent the upset was.

I THINK there is upset training...this was an upset...

On 11 January until 3 February 2021, the Directorate General of Civil Aviation (DGCA) conducted special inspection to all Boeing 737-300/400/500 air in Indonesia. The areas of the inspection were as follow:

- pilot training program implementation, including weather avoidance and upset recovery training program;

On 28 January 2021, the DGCA initiated a discussion with aircraft operators on the implementation of upset prevention and recovery training (UPRT) program.

On 28 January 2021, the Standard, Quality, and Training Division of Sriwijaya Air included the upset recovery training as part of the training syllabus in the next Line Oriented Flight Training (LOFT) – Pilot Proficiency Check (PPC).

There were many items, I only left the ones related to upset training.

For me, it is an indication that upset training was may be not absent but probably lacking.

[elaw]

I'm with Gabe on this one. The DC-717 guys recovered from an upset that they knew was going to happen because they intentionally caused it. Also it was in a sim, so while the "aircraft" went inverted, their physicalness did not - suddenly and unexpectedly becoming upside-down can be a little unnerving for those of us accustomed to having our head more distant from the planet's surface than our buttocks*.

Also, the did_crashed aeroplanie became upset with a gross thrust asymmetry... which very rarely happens with no indication of engine failure or intent of the pilot(s). That of course is going to affect the plane's behavior and the upset-recovery steps. Someone with a clue can feel free to correct me, but I bet sudden autopilot disconnects with the engines at dramatically different power settings in otherwise-normal conditions aren't something often practiced in the simulator.

* For the record that's a political joke, not intended as a dig on present company!

[3WE]

How many seconds are needed for corrective input?

I know it was a terrible oh poop, WTP moment...

And if the plane broke something IN the recovery...

And what is the roll rate of a moderately laden 737? (Which model?)...It didn’t POP onto its back.

Repeating- how many seconds: Oh shit, grab yoke, full corrective rudder, ailerons, and a crisp (but measured nose-down shove (or did they make a hard pull up? (which is not_unheard of) )

How many seconds? Rule 1: Aviate...

Footnote: 25 seconds is too long...so, is 10. Upsets happen...

[elaw]

How many seconds are needed for corrective input?

In this case, apparently a few more than they had available.

I know it was a terrible oh poop, WTP moment...

And if the plane broke something IN the recovery...

And what is the roll rate of a moderately laden 737? (Which model?)...It didn’t POP onto its back.

Repeating- how many seconds: Oh shit, grab yoke, full corrective rudder, ailerons, and a crisp (but measured nose-down shove (or did they make a hard pull up? (which is not_unheard of) )

How many seconds? Rule 1: Aviate...

Footnote: 25 seconds is too long...so, is 10. Upsets happen...

My (totally parlour-talking of course) feeling is there are two factors that interact to a degree. First, the "oh shit moment"... recognizing the autopilot-disconnect horn has sounded and that the a/c has started to roll typically would not take too long. But recognizing a) that this is unintentional and represents a failure (vs. the PF having seen a bird or something and taking deliberate but controlled evasive action) and b) that it's a failure that requires rapid and aggressive correction can take much, much longer.

Keep in mind that one safety philosophy that's (in my perception) being trained a lot lately is that when you first notice a problem you should do nothing. The thinking is that taking a moment to think things through and take appropriate corrective action is better than rapidly and "instinctively" taking an action that may not be appropriate... like pulling up relentlessly or shutting down the good engine when one fails. That probably is a good approach in most cases, but a bad one in this case.