Controlled Pod Into Terrain episode 10: TACA 110

that time a 737 drank too much and fetched up spark out on someone's lawn outside new orleans - too relatable tbh

featuring, as always, J, the engineering one,

Ariadne, the business and logistics one,

and Kyra, the one who is also known as Admiral Cloudberg

Episode Transcript

(00:00):
Hello and welcome to Control Pod Interterrain. We are a multimedia podcast about air and

(00:24):
space
My name is Kira Dempsey, better known as the aviation writer Admiral Cloudbird.

(00:48):
My pronouns are she and her.
Next slide.
Today we are going to dive into the story of TACA Flight 110, a plane that drank too
much and ended up stranded outside New Orleans.
And honestly, relatable.
But first, let's do some sort of news thing.
It's been a long time.
We got some news to get through.

(01:09):
Yeah.
So we could take any number of news stories, given that it's been checks and notes three
and a half months since our last episode.
We decided that since this is a podcast mostly about air disasters, we're going to go over
what we know about three fatal accidents involving commercial aircraft that happened over the
summer.
So...

(01:29):
Okay.
So let's start with a Sukhoi Superjet crash when Russia invented their own MCAS.
Right.
So on July 12th, a Sukhoi Superjet 100 crashed during a post-maintenance ferry flight between
two Moscow airports, killing all three crew members.
So what is an SSJ-100 anyway?
Okay, remember when we talked about the TU-204 as a version of the 757 that sucked in every

(01:54):
way?
Well, this is that, but for the A220, it's ostensibly a small regional jet with advanced
tech.
That sucks in every way.
It's probably honestly not terrible.
It's just that why would you want one of these when you could get an Airbus A220, which is
the same thing but better?

(02:14):
Right.
So basically, it's an approximately 100-passenger Russian-belt regional jet.
And I say Russian-belt with my tongue in my cheek because most of the systems on it come
from Western countries, but that's a whole other issue.
It's also kind of a widowmaker.
They've only delivered 127 of them, and this is their third fatal crash.

(02:36):
It's not a particularly great record for an aircraft that only entered service in 2012.
I mean, most aircraft these days do a lot better than that, but also it's really hard
to blame Sukhoi exactly for this.
It's mainly for piloting, which again, Russians.

(02:59):
So anyway, this particular aircraft is operated by Gazprom Avia, which is a really weird airline
in some ways because it's wholly owned by the Russian state oil company Gazprom, and
it performs a lot of services for Gazprom, like supporting oil prospecting, air rescue,
and so on.
But you can also just go and buy a ticket on Gazprom Avia to fly from Moscow to someplace

(03:20):
like Barnaul, and for that, it's basically a regular airline.
So this is like if Shell just sold tickets to Amsterdam to tourists.
Yeah, more or less.
In the future, you'll probably be able to do this when the merger between Texaco's
Buc-E's and Southwest Airlines goes through.
So this Gazprom Avia SSJ-100 was in for maintenance at the Luhovitsy aircraft plant outside Moscow,

(03:49):
and it was supposed to be ferried to Zhukovsky Airport, where I assume it was going to reenter
service.
So before the aircraft went in for maintenance during its last takeoff roll, both primary
angle of attack sensors jumped to values around 5 to 7 degrees as the aircraft speed came
alive, which is patently ridiculous because the angle of attack should be zero during

(04:10):
the takeoff roll and tell rotation.
So the AOA sensors were giving erroneously high values for quite a while before this.
What was actually wrong here?
We don't know.
What we do know is that whatever it was, they didn't fix it.
Well, yeah, at this point, average Russian maintenance procedures.

(04:32):
We don't know if the inability to source the necessary Western parts was a contributor
to this crash, and we may never know.
But for the indefinite future, that asterisk is going to have to be applied to any crash
that happened in Russia.
And I think it's a real possibility.
This is certainly a crash where that's the Russian angle to it and not poor piloting

(04:56):
because as far as I can tell in this incident, the piloting was more or less fine.
I think most pilots probably would have reacted this way.
So anyway, what actually happened?
So when the aircraft left maintenance on July 12th with just three crew on board, the angle
of attack sensors pulled the stunt again and both started indicating an AOA that was way
above actual.

(05:18):
So the SSJ-100 being a clone of the Airbus has three angle of attack sensors, but the
third one for some reason wasn't recorded.
It might just not have been working at all, or maybe their Russian clone of the ADIRU
was ignoring it because the other two were reading the same incorrect value.

(05:42):
Yeah.
Well, it wasn't recorded by design, but we don't know whether it was giving the same
wrong AOA value as the other two or if it was working but was rejected.
But anyway, so the pilots got the airplane airborne despite this, but during initial
climb they got an unreliable airspeed alert, which they think is weird because both of
the airspeed indicators are showing the same value, which is also an appropriate value

(06:05):
for the configuration and flight path.
So they just kind of ignore that even though the warning keeps going off intermittently,
which in hindsight was because the flight computers were using incorrect angle of attack
data and thought that the airspeed didn't comport with what it thought the airplane
was doing.
But again, the pilots basically applied the unreliable airspeed procedure here and everything

(06:28):
checked out as fine, so they kept going.
That warning was the airplane showing them a photo of its wife and kids and telling you
how excited it was to go see them.
Yeah.
That's a reference to one of our bonus episodes for anyone who doesn't subscribe to that.
So now, Lukowice Tzizukowski is a very short flight, so the pilots pretty quickly just

(06:48):
left this issue alone to talk with ATC and they were told to level off at 5,000 feet
and then shortly after to climb to 10,000 feet.
So the pilots entered a higher altitude into the autopilot panel and the plane doesn't
climb.
What's going on here is that the SSJ is a fly-by-wire aircraft very similar to an Airbus
and if you remember back in episode 7, we went over how that works, including the basic

(07:13):
flight envelope protection systems.
So the SSJ-100 has an auto trim system and a high angle of attack protection system.
And by this point, the erroneously high angle of attack had reached the high angle of attack
protection threshold, which caused the auto trim to start trimming the stabilizer nose
down in order to prevent the angle of attack from increasing.

(07:35):
So by the time the pilots ordered the autopilot to climb, the trim was below neutral, which
would make it almost impossible to ascend, but the auto trim couldn't trim nose up
because it thought the aircraft would exceed the angle of attack limit if it did, so actually
the airplane started to enter a shallow descent instead of climbing.
In response to this, the first officer who was flying disconnected the autopilot and

(07:57):
pulled back on its side stick.
The captain took control.
And if you've ever read any of my Airbus stall articles, you might remember that when
the pilot pulls back on the stick when the AOA is at the high AOA protection threshold,
which is called Alpha PROT, the fly-by-wire system will allow the AOA to increase only
so far as Alpha MAX, which is just below the stall AOA, but no farther than that.

(08:23):
And that's what happened here.
They pulled up against the nose down trim and they managed to level the plane off, even
entering a slight climb, but the false AOA increased to Alpha MAX and the high AOA protections
prevented any further nose up elevator deflection.
So, a quick aside about one of our favorite planes, which is the Airbus A400M.

(08:43):
It has the same base of GNC and fly-by-wire systems as commercial aircraft, but it's
also equipped with an override switch on the side sticks that instantly disables the flight
envelope protection systems because sometimes you might be in a nose down trim situation
and for some reason, for whatever reason, you might need to pull the nose up.

(09:03):
Like if you're getting shot at, presumably.
I mean, yeah.
So if you've just gotten an IR alert for a man-pads launch and you need to bring the
nose up right now, right fucking now, you can disable the Alpha PROT and force the
plane into I must go, my planet needs me climb mode.
It also helps that the A400M has a thrust to weight ratio of LOL and carbon fiber wing

(09:26):
spars that are unreasonably strong.
So you'll pulp the crew long before you fly its wings off.
Yeah, the A400M actually has more in common with the Super Hornet in this aspect, which
will similarly let you override the G protections because it can pull up a lot farther than
the water balloon in the cockpit can handle.
So back to a plane that is not clever like that.

(09:46):
So our crew have now found that they're in a very shallow climb with climb thrust set
in the autothrottle and their airspeed is rapidly increasing and they can't climb at
a steeper angle to bring the airspeed down.
So the captain tried to fix this by disconnecting the autothrottle, but the airplane quickly
exceeded its maximum operating speed, which triggered an overspeed warning, and it was

(10:08):
chaos in the cockpit by this point.
The controller was calling asking why they weren't climbing, the FO had his radio on
the wrong frequency, the captain was hauling back on his side stick and getting barely
any response, it was madness.
And then this is the clincher.
The SSJ's high airspeed protections kicked in and automatically deployed the god damn
speed brakes.

(10:29):
So the speed brakes reduced speed by increasing drag, but they also decreased lift, so this
immediately put the plane into a descent, which actually caused the airspeed to increase
even more.
And the pilots couldn't pull up to stop the descent because they had already reached
the pitch up limit, the aircraft still thought the AOA was at alpha max.

(10:49):
The captain then reduced engine power to idle in an attempt to slow the descent, but it
was too late, the aircraft just descended in a steadily increasing dive until it hit
the ground.
So basically the aircraft MCAS'd itself.
It turns out that any fly-by-wire plane can have MCAS if it sucks hard enough.
Is the opposite of bad MCAS the F-35B that slowly hovers away even after you eject?

(11:15):
So this actually is kinda similar to Ethiopian Airlines Flight 302, the second 737 MAX crash
in some respects, only in that crash they could have regained control if they had reduced
engine power, which would have reduced the control forces.
Here I'm not sure that would have helped, again, because that was a very analog control
system problem and solution.

(11:36):
In this case it looks like recovery should have been possible if they had just switched
the fly-by-wire system to alternate law, which would turn off the high AOA protections, but
the pilots only realized that their issue was a false AOA reading about 30 seconds before
impact, which was way too late to do any troubleshooting, all around a pretty fucked situation.
Alright, next slide.

(11:57):
Okay, so up next, 12 days after that, a CRJ crashed in Nepal.
This one is a little less complicated than the Russian crash because we actually have
very little idea what happened.
But in short, a Sarya Airlines CRJ-200 crashed right after takeoff from Kathmandu on a repositioning
flight to Pohara with 19 people on board, all of them airline employees.

(12:22):
Only the captain survived.
One of the victims was the airline's director of safety.
I guess he put his money where his mouth is or was.
CRJ-200s crashing on repositioning flights is a thing, apparently.
Between this and episode 8, I'm learning that being an airline exec on your flight doesn't

(12:44):
mean that you're any less likely to eat it.
So from video footage and the preliminary report, what we know is that the airplane
pitched up very rapidly during takeoff at almost three times the prescribed pitch up
rate, then climbed steeply and immediately stalled.
The plane banked about 90 degrees to the right and then fell back to earth next to the runway
where it exploded.

(13:05):
The reason for this hasn't been determined, but the preliminary report has suspiciously
large amounts to say about weight and balance, and all of this is consistent with an improper
center of gravity or a load shift during rotation, and I have seen unconfirmed rumors that a
large maintenance kit wearing several hundred kilos that they were ferrying to Pohara may

(13:27):
have been improperly secured.
If you really want to unlock a nightmare, go watch the haunting dashcam video that documents
the crash of National 102 out of Bagram Air Base.
In that situation, an improperly secured Oshkosh MATE armored vehicle broke free of its mountings,
shifted backwards, and damaged the plane carrying it in such a way as to become instantly unrecoverable.

(13:50):
LARIE Yeah, I heard that there's also someone did
a really good write-up of that accident.
SONIA Yeah, I have no idea who.
So load shift incidents often become unrecoverable really fast, even without damage to the flight
controls.
And also, these guys apparently used the wrong takeoff speeds, but we don't know if that
contributed.

(14:10):
So overall, this is just Nepal trying to keep up its streak of having a fatal commercial
plane crash every single year, so not the best place to fly.
Next slide.
LARIE Alright, now last but definitely not least,
let's talk about the most recent, which was the Brazilian ATR crash.
SONIA Right, so on August 9th, Voipass Airlines Flight
2283 and ATR 72 crashed just before top of descent into Sao Paulo, Brazil, killing all

(14:36):
62 passengers and crew.
And most of you probably saw the bystander videos of the plane spiraling down like a
leaf in a flat spin.
So from day one when I saw that the crash occurred in an area where there was a warning
for severe icing, I had a hunch about what might have happened here.
LARIE ATRs and icing, name a more iconic duo.

(14:57):
SONIA Truly.
I just recently published a 28,000 word article about all the different ways the ATR can go
wrong in icing conditions and why most of them are actually completely preventable by
following basic procedures like this one, but not all of them.
LARIE Seriously, if you haven't read that article
yet, pause this episode and then go read it before you continue this segment.

(15:18):
SONIA Or maybe don't, unless you have like three
hours to kill.
LARIE I mean, we're a YouTube video, so we'll
wait.
SONIA Yeah, so anyway, the reason I say this is
because the preliminary report revealed some absolutely wild information about this crash.
So according to the prelim, the pilots encountered icing conditions during the latter part of
the initial climb and turned on the de-icing boots, which remove ice from the wing leading

(15:42):
edges.
Again, these are the primary defense against airframe icing, but after a few seconds they
got an airframe de-icing fault light.
So uh oh, it's not working.
LARIE Let me guess, the procedure for an airframe
de-icing fault in icing conditions is four letters long, and those letters are G, T,
F, and O, right?
SONIA Yeah, it's exactly that.

(16:03):
It's just to leave icing conditions immediately, which they didn't do, they just turned the
de-icing boots off and kept going.
LARIE If it's cold, hit the bricks, you can simply
leave.
SONIA Yeah, well they didn't.
They climbed to 17,000 feet, which they were actually
limited to because of an air conditioning pack fault that had been there for quite a
while, that's a whole other issue.

(16:25):
But that put them right in the middle of the altitude range where severe icing was forecast
over Sao Paulo, and if you didn't read my article on the ATR icing crashes, the definition
of severe icing is that it's outside what any aircraft is certified to withstand, and
the procedure if you encounter it is also to GTFO, even if your de-icing equipment is
working.
LARIE It actually shouldn't have been legal to

(16:47):
dispatch an aircraft into severe known icing conditions.
We don't know why the airline did it, or whether the pilots knew about it.
That will probably come out in the final report.
SONIA Yeah, but anyway, the flight continued to
encounter intermittent icing conditions throughout the cruise phase, which caused the ice detector
light to go on and off a whole bunch of times, but the pilots didn't really respond to

(17:09):
this, and the aircraft performance remained normal.
However, as they neared Sao Paulo, the ice really started to pick up.
Drag increased rapidly as the ice adhered to the wings, and the plane started to lose
speed.
The pilots responded by turning the de-icing boots back on, even though they had previously
gotten an alert that the system wasn't working.

(17:30):
Why did they do this?
Fucked if I know, don't try this at home, kids.
SONIA Just say no to icing.
LARIE Now, the ATR-72 has a thing called the Performance
Monitoring System that gives the pilots a series of increasingly dire warnings if the
system detects that aircraft performance is being affected by ice.
So as the plane starts to decelerate from where it was before, which was close to 200

(17:51):
knots, it flicks on a cruise speed low warning light, and the pilots don't react.
Then it hits them with degraded performance, but they don't notice that either.
Then it goes to the most dire warning light of them all, which is increased speed.
I think in one of his videos on this crash, Maegnar Nordahl said something like, you should
never see increased speed.

(18:11):
Which is true, because these guys sure as hell didn't see it.
SONIA Turns out you can add a stop this bullshit
light, and it doesn't help if your crew simply ignores it.
LARIE So what were they doing the entire time?
Usually they were talking to a flight attendant about logistical stuff, and they were trying
to get ATC clearance for their initial descent unsuccessfully.
The FO did comment that they had a lot of ice, but they didn't seem to notice that

(18:35):
they had lost like 40 knots of airspeed and it was still dropping.
SONIA I guess because these guys were above 10,000,
this wasn't actually technically a violation of sterile cockpit rules, but it is a pretty
clear example of something that they just shouldn't have been doing.
LARIE Yeah, I mean, yeah again, I get how they

(18:57):
got into that position where they were both focusing on things other than the performance
of the aircraft, but they needed somebody focusing on the performance of the aircraft.
They're in icing conditions, you can't just not do that.
Just as the flight started to turn inbound to Sao Paulo, the aircraft stalled, and this
is because there was so much ice on the wings that air stopped flowing smoothly over them
and they lost lift.

(19:19):
This happened at about 169 knots, which is quite fast.
Normally an ATR 72 is going to stall at least 55 knots below that, I think.
That's how bad this ice was.
But this was totally avoidable if the pilots had seen the performance warnings, or if they
had made sure to keep their airspeed at least 15 knots above the icing minimum speed.

(19:40):
So in this case, that comes out to 180 knots.
So after the stall, the airplane banked sharply left, then rolled over 90 degrees to the right
and entered a spin.
And a spin is when you stall an aircraft with a rotational component, so the airplane is
rotating about its center point as it's falling.
In a light aircraft, you can recover from a spin by applying rudder opposite to the

(20:02):
spin, reducing engine power, and then holding the nose down.
But in an airliner with engines out in the wings, the momentum of the heavy engines spinning
around and around can overcome even maximum opposite rudder, so recovery from a spin can
be close to impossible.
And if you don't apply the spin recovery procedure, and you increase engine power and pull the
nose up, then you'll still be stalled, but now your nose is close to level with the horizon,

(20:25):
and that's what we call a flat spin.
That's what we saw in the videos of this crash, and it's not possible to recover from a flat
spin in an airliner, period.
And to all of that, I want to add that when ice is accumulating that fast, you can declare
an emergency, it's within any pilot's rights to do that, and if they had descended to below
the freezing level at around 12,000 feet, going down as fast as possible, keeping the

(20:47):
speed above 180 knots, they'd probably have been fine.
But this flight never should have even gotten that far, they should have diverted as soon
as the de-icing system failed.
So the TLDR is that this airline and this crew fucked up, and the family should still
go pass into a glidion?
Yeah, I would say that's fair.
Okay, so the airline that we're talking about today actually flew 111 200s, and we are going

(21:11):
to make you listen to us talk about it, even though it has absolutely nothing to do with
the crash.
We apologize for nothing.
You want us to talk about something else besides this weird little British plane?
Well, you'll write 15,000 words on a plane crash and its background and web of systemic
failures.
So we're researching TACA, and the first thing we do when we find one of these, figure

(21:34):
out the airplane's ownership structure.
While we're doing that, I found out it had 111s.
One thing we noticed when we found out how familiar this livery looked.
That's when we realized that Airbus stole this airline's livery and barely changed it.
Airbus copying their homework and promising that they're going to change it so it doesn't

(21:56):
look the same, and then just turning this shit into the Professor.
They do fly the hell out of these 111s because they do so well with high, heavy and high
operations.
They're ideal for this sort of mountainous, central and South American operation.
Of all the ones that we found, one of the ones they had lasted through five hours until

(22:19):
it ended up sitting engineless in Sierra Leone.
I think we should go get it.
The first officer on this flight that we're talking about today was actually qualified
on the 111.
It was the first plane that he flew through the airline, which is good for him.
Let's move on.
Next slide.
All right, so from our last episode on Swiss Air, guess what plane did have fire detection

(22:42):
systems in the attic?
Was it the 111?
Almost.
It was the 1011.
While I was doing a research on the Saudia 163 crash from Mentor Pilot's video essay,
I discovered that the Lockheed L-1011 had a thing called a duct overheat detection system
that basically monitored the entire air conditioning duct network inside the hidden spaces of the

(23:03):
aircraft and would sound an alarm with an indication of the location if an overheat
is detected.
This system was for detecting pneumatic system leaks, not fires, but if Swiss Air 111 had
one of these systems, it probably would have generated a warning well before a fire actually
became visible to the pilots.
What's it mean when you get a duct overheat warning and the pneumatic system appears to

(23:25):
be functioning entirely correctly and also you smelled smoke earlier?
I don't know, you tell me.
So firstly, I'm going to go out there on a limb and say that if they'd actually bought
L-1011s instead of MD-11s, then the accident probably wouldn't have happened because the
L-1011 was awesome.

(23:46):
So true.
Secondly, it's not clear to me what the crew is actually supposed to do with the information
that there's a fire in a space that they can't get at, short of going, well, at least we
know why we're going to die in 11 minutes time.
Presumably it's to give you another couple of minutes warning so you can try and get
it on the ground.

(24:08):
Yeah.
Again, we went over in that episode why, even if they had, realized immediately that there
was a fire that it probably wouldn't have saved them.
But anyway, I still thought I would dunk on the MD-11 while standing the L-1011, the best
airplane ever made.

(24:29):
Sorry, BAC.
So that's why that's here.
CPIT is a Predator hands meme and each of us has a different one.
The BAC-111, the General Dynamics F-111, Glock E-L-1011.
Yeah.
All right, next slide.
All right, time to plug some capitalism.

(24:51):
All right, we always say our Discord is CPIT, it's the writer's room, this podcast is just
the output file.
We also dropped us a bonus recently where we put out a movie commentary for a classic
film with some of the worst spoken and sung Russian ever captured on screen.

(25:11):
And we made our co-host with a master's in Slavic studies listen to it and you get to
enjoy her pain.
We earn your money at CPIT and we respect our audience even if we don't respect each
other.
Next slide.
All right, so what is today's episode even about?

(25:33):
We're here to talk about TACA.
Next slide.
No, that's a taco.
No, that's sticky tack.
That's weird owls tacky.
That's a TACAN beacon.
That's a thumb tack.
We're getting further away.
Oh, there you have it.

(25:54):
That's TACA.
Yep, it stands for Transcuentes Aero de Continente Americano.
It was airline based in El Salvador.
It's actually still around.
It's one of the oldest airlines in the world, but it was bought by Columbia's Avianca and
is now called Avianca El Salvador.
All right, so let's talk about the airplane of today.

(26:18):
That is a 737.
It's an airplane.
That's the most important characteristic.
It is an airplane.
For listeners and those that don't have slides, just picture an airplane in your head right
now.
Bam, you have probably pictured a 737.
Also, this particular 737 is painted in the Airbus livery.

(26:41):
Yeah.
Well, I think what you mean to say, Jay, is that Airbus livery is painted in this aircraft's
livery.
Yes.
As of this episode's recording, it is still the best selling passenger of all time, but
the Airbus 320 is only a few hundred units behind in terms of deliveries, and it should
pass it very soon despite its construction not starting for two decades after the 737.

(27:07):
As with last episode, this plane sucks, but not in a way that's relevant to the accident
sequence today.
This particular model, November 75356, was a 737-300 and was delivered brand new to TAC
Airlines two weeks before the incident we're going to talk about today.
It was delivered through a dry lease, which isn't relevant or interesting, but I don't

(27:28):
get a lot to do this episode, so sorry.
At least you have to write some jokes.
We don't have to feel bad about laughing at...
Next slide.
What is an El Salvador?
Okay, so we're in this country in the 1980s.
El Salvador is a country you used to think about so little you don't realize that's

(27:48):
not El Salvador, that's Costa Rica.
Are we stealing jokes from last week tonight now?
Unabashedly.
This is El Salvador.
El Salvador is and has been...
What's the best way to say this?
Rough.
It has been the murder capital of the world for some time, just edging out Cabot's Cove

(28:10):
May.
The cause of this, from a literal perspective, is a longstanding economic and social inequalities
and a series of authoritarian regimes suppressing dissent, including their current president,
Nayib Bukele, who is a full-blown dictator that also somehow has overwhelming popularity
among his people.
All right, because he sent...

(28:30):
He just shoved all of the gang members or anyone he thought was possibly a gang member
into a hole.
Yeah, he's got some issues, but we don't have time to go into that.
No, we don't.
All right.
So as with every episode of CPET, the first question we have to ask is, how is this Reagan's
fault?
So El Salvador has massive income inequality.

(28:54):
It is exacerbated by the US, and a group of people calling themselves diferente, Pablo
Roldo, Marti Palt, La Libreación Nacional, or FMLN, is trying to exercise political control
in this country.
But these Spanish names are really hard.
Ronnie Fuckhead and his merry band of fascists decided that since the possible new government

(29:16):
of El Salvador at the time was planning to be nice to its people, which was and is not
acceptable, it was time for some war crimes.
They went something, something communism, and they started sending significant financial
and military support to the fascist Salvadoran government.
The CIA started operations to support government forces and undermine the FMLN, which was this

(29:37):
left-wing group that was just seeking to have a foothold in the government.
Naturally, the introduction of massive amounts of American cash and heavy weaponry escalated
this conflict and dragged it out for more than a decade, rendering it one of the bloodiest
and most civil wars of recent history.
The ramifications of this civil war are still being felt today, and El Salvador remains
fucked because of it.
And as if that wasn't bad enough, we allowed El Salvadoran gangs like the MS-13 to develop

(30:01):
in the US and then exported them back to El Salvador like some kind of utterly fucked
supply chain.
Hang on a minute.
Is Henry Kissinger involved in this?
This seems like a Henry Kissinger thing.
Uh, I think, yeah, I think he is actually.
The captain of the plane we're going to be talking about today, in fact, still feels

(30:23):
the ramifications of this civil war, but he apparently doesn't let it get him down.
All right, next slide.
This brings us to the airline we're going to be talking about today, CACA.
They are the flag carrier of El Salvador.
The airline was started as a state-owned one.
It had its roots going back to 1931, which makes it the second oldest airline in Central
American and Caribbean area after Cubana de Aviación, which is the now very, very state-owned

(30:47):
flag carrier of Cuba.
Eventually, it becomes privately held, shares are floated.
They joined the jet age in 1966, which are with the previously discussed BAC 111, which
they flew for 22 years.
However, thanks to the aforementioned fuckery, it was no longer safe or sensible to have
the airline actually in El Salvador, nor was it realistic to ask the El Salvadoran government

(31:11):
to oversee an airline industry on the account of there being an El Salvadoran government
during much of the late 20th century.
At some point, the El Salvadoran business Roberto Created bought a minority stake, which
he eventually turned into an everything stake.
So crucially for the purposes of today, as part of his efforts, he incorporated the airline
in New Orleans, Louisiana.

(31:32):
Nolens.
So I have it on good authority from a friend who is from there that nobody actually says
Nolens.
Are you both sure it's not Nouvelle Orleans?
No, Nolens.
I think you'll find I'm correct.
I live there.
And I speak French.
And it's not my problem.
All right.
So now we have an airline with an excellent paying job and a mixed approach to purchasing

(31:53):
airliners sending a plane home to NOLA.
All right.
Let's go ahead and fly to the scene of the crash.
Next slide.
The date is May 24th, 1988.
And November 7, 5356 with call sign TACA 110 is flying from Camopola International in
San Salvador to Philip Goodson International in Belize and then onward to New Orleans International.

(32:13):
Slide.
All right.
First up, we have our captain Carlos Charlie Dardano, who at the time of this accident
was only 29 and was the third pilot in his family.
However, he had already surmassed a sizable 13,400 hours with 11,000 of those in command.
And that's actually insane.

(32:35):
I mean, I'm sure the only way that's possible to get that many hours by age 29 is if your
company laughs at the concept of duty time limits.
He actually got his private pilot's license at 16 while he was still in high school and
then went to help his father at work.
It took him five months to go from a PPL to a commercial license.

(32:56):
He'd been a commercial pilot for 14 years at the time of the accident, which means that
he was flying a little under a thousand hours a year, which is a ninth of literally all
the time.
And that is a lot.
Like, even at my height, I never managed to ride more than a couple thousand hours a year
and there's no duty time restrictions on do dumb shit on bike.

(33:20):
Although judging by my medical records, there probably ought to be.
He was also monocular, having lost an eye in small arms fire during the aforementioned
R-word induced civil war.
It happened while flying.
As he tells it, he was caught in ground fire while on the threshold for takeoff and caught

(33:40):
a stray round that destroyed his eye.
But rather than simply sit down and feel sorry for himself like we probably would have, the
dude completed the takeoff roll and then flew another 20 minutes to his scheduled destination,
upon which he finally landed and decided to get this bullet wound thing sh**ed out.
To be fair, I wouldn't want to stay at that airport either given that there are people
shooting at me there.

(34:00):
So I mean, if this dude has a career worth of hours by the time he's 30, he's probably
been flying pretty much nonstop since he could reach the rudder pedals, you know, mostly
in single engine planes doing air taxi and delivery services, vital in a country with
poor infrastructure and lots of rebel checkpoints.
So I don't know, it seems he was probably, I mean, he probably didn't know how to stop

(34:22):
flying.
Seriously, we looked into this guy, because obviously he's still around.
The more we looked into him, the more we liked him.
He basically just felt like your cool uncle who took you to a rated R movie when you were
12.
For anyone who's wondering about, you know, his ability to fly a plane, Lila has only
one eye and she's a perfectly cromulent pilot.

(34:44):
Really?
Nothing?
Look, this is me trying, okay?
The first officer, 44 year old Dionisio Lopez, was also experienced with 12,000 hours.
There was a line captain, Arturo Soli, we don't have his age.
He was riding in the jump seat of request of airline management.
He signed off on the new 737's behavior and performance, because this was the first 300

(35:09):
model bought by Taca.
And just for those that don't know, the 300 was the first model of the 737 that had high
bypass engines.
And we'll get to those.
Also, we do know Soli's age.
I just didn't put it in the screen.
So the flight was not particularly full.
There were only 38 passengers and seven crew on board.

(35:29):
And even though the plane was brand new, it had just been delivered by Boeing.
The day before the incident, they decided to replace the battery because it wasn't
providing enough power to get the right hand engine going.
So Taca put in a new battery just before this flight, which is great quality control, Boeing.
Though this would end up having nothing to do with the accident, I just wanted to point

(35:53):
out that Boeing has been unable to do batteries properly for decades, apparently.
They just don't seem to understand that batteries are a solved problem for literally everybody
else.
And you know, I guess it's better to have a nice fresh battery than to not have one,
right?
So before we start, we're going to have to learn about a few things, starting with a

(36:14):
quick overview of jets and their parts that burn things.
This is opposed to a before burner, which is burning jet parts.
Oh my god, you guys, that was one time, okay?
Actually, let's start with what is a CFM56-3B1.
So it's made by a company called CFM International, which doesn't stand for anything, which is

(36:37):
annoying.
They're composed of GE and Snecma, which is now Safran, and their name comes from the
CF6 engine series that GE made and the M56 engine series that Snecma made.
See?
It's stupid.
This just doesn't mean anything.

(36:58):
Anyway, these engines are used on freaking everything.
They've built more than 32,000 of these things.
It's a turbofan, specifically a two spool medium bypass annular combustor turbofan.
Okay, so that was a lot.
Every single part of that is going to be important.

(37:18):
So Jay and I are going to take it piece by piece.
Okay, Jay, what is a turbofan engine?
A turbofan engine is a turbojet engine with a big fan on the front.
Nailed it.
Okay, and how does a turbojet engine work?
So a compressor at the front sucks in and compresses air, which makes it hot.
Then fuel is added very sparingly in a combustor where the fuel combusts, believe it or not,

(37:44):
in the combustor you combust the fuel, and that makes hot gas.
And seriously, we do mean sparingly.
Turbines run unbelievably lean.
It means they can run with a much higher oxygen to fuel ratio than a piston engine does.
And the higher the plane goes, the leaner it can run.
This is one of the reasons planes are more efficient the higher they climb.

(38:05):
It also means, however, that turbines run on the absolute razor's edge of being flooded
and unable to ignite.
As we've seen back in episode two, adding too much fuel can cause it to melt or it can
extinguish the igniter.
Go ahead and post it that.
This is foreshadow.
So the hot gas generated by this combustion in the combustor turns a turbine which drives

(38:27):
the compressor.
The combustor is sometimes called a gas generator.
So round and round this energy goes, generating hot high velocity gas at an exhaust nozzle.
All internal combustion engines operate on this process of suck, squeeze, bang, and blow.
Four stroke piston engines do each one in order.

(38:48):
Two stroke engines sort of overlap two steps in each stroke.
And then in a turbine, all four are happening continuously and constantly, which is why
Rolls Royce names their turbine engines after rivers because, you know, back in the 50s
when they started designing these things, they thought that the continuous flow of gas

(39:10):
through the engine was like water down a river.
This is also why jet turbines are very, very smooth, very, very efficient, because in theory,
they should be perfectly balanced.
As all things should be.
Indeed.
And everything should be happening at exactly the right spot so that this process can happen

(39:31):
continuously and smoothly.
And what does two spool mean?
So it means that there are two concentric shafts.
The outer one is attached to the high pressure compressor and has a high pressure, high temperature
turbine on it.
On the slide, if you can see that, it's shown in purple.
The inner one runs the big fan at the front, a low pressure compressor that compresses

(39:56):
the air so that the high pressure compressor doesn't have to do so much work.
And it's driven by the low pressure turbine that's right at the back of the engine core.
What does medium bypass mean?
So medium bypass means that there's more bypass than a low bypass and less bypass than a high
bypass.
No, really, that's what it means.

(40:18):
Yes, it is.
It means that the big fan at the front pushes about five times as much air through the bypass
duct around the outside of the engine core as passes through the engine core itself.
For comparison, the low bypass engines that we were talking about in episode two, the

(40:39):
Rolls-Royce Spey, had a bypass ratio of about 2, 1.4 to 2.
And high bypass engine like a Rolls-Royce Trent might have a bypass ratio of about 10.
And an ultra high bypass ratio engine like the Ultra fan or any of the geared turbo fans

(41:02):
might have a bypass ratio of 14 or 15.
And this air that goes through the bypass duct is where most of the thrust of the engine
comes from.
You know, a low bypass engine is very good if you're trying to go supersonic because
it has a smaller frontal area so it produces less drag.

(41:23):
But the universe demands a trade.
So it means they're very inefficient.
And because there's only very high speed gas coming out of the back of this engine, it
means they're extremely hot, hot and loud as hell, which, you know, is a kind of a problem
if you live near an airport.
A high bypass engine like the Trent that we all love is even more efficient.

(41:48):
But high bypass engines don't fit on a 737.
Do they, Boeing?
Do they, Boeing?
And what's an annular combustor?
So an annulus is a ring-shaped region of space.
The annular combustor is a ring-shaped space around the center of this engine between the

(42:12):
compressor and the turbine, which has several fuel swirl nozzles and flame holders spaced
around it.
They're sometimes called combustor cans.
There are eight in the CFM56-3B1, but, you know, bigger engines have more and smaller
engines might have fewer.
This particular plane had a brand new pair of CFM56-3B1s.

(42:36):
They only had 81 hours on them.
They were the 20,000 pound thrust version.
And they have the second most EGT margin or range of exhaust gas temperatures they're
designed to tolerate.
So EGT, for those of us that aren't us or Fox, is exhaust gas temperature.
And that's one of the ways pilots can determine what the fuck the engine is doing.

(42:58):
In general, high exhaust gas temperature means that the engine is working harder.
Low EGT is the opposite.
But high EGT only means high thrust if something called N1 is also high.
And N1 is fan rotation speed.
So if your EGT is high, but your fan isn't rotating very fast, that probably just means
your engine is overheating.

(43:19):
It's not actually generating very much thrust.
But if they're both high, that's great.
And we'll come back to this concept.
So CFM56s, even the smaller lower thrust versions, have a surprisingly small EGT margin.
For the 20k pound 3B1, it's only on the order of 92 degrees C with a brand new engine.

(43:42):
And this margin reduces as the lifespan limited parts in the engine age and the profiles of
these turbines gets worn and less optimal.
So what we mean by EGT margin is the difference between how hot the engine normally is versus
how hot it would be to be considered overheating.

(44:03):
And a 92 degrees C difference between those two levels is apparently not a lot.
Yeah, I mean, we're talking about 92 degrees in an exhaust gas temperature, which might
be 800 or 900 degrees.
So since the engine runs hotter if the air coming into the inlet is also hotter, the
compressor has a maximum outside air temperature that it can support.

(44:26):
This is called the sea level outside air temperature limit, which I'm going to insist is an acronym
to be pronounced SLO-TAL.
They think SLO-TAL is fun to say and that's like half the reason we even brought that
up.
I think it's the whole reason.
Yeah, actually it is.
So why is this important?

(44:46):
In the middle of the engine, you're adding fuel and you're burning it with all of this
compressed air to make hot gas.
The turbines are then responsible for extracting energy from this gas, which cools the gas
down and also turns the compressor and the accessory drive and, you know, makes your
plane work, right?

(45:06):
As the turbine parts age or are damaged, they get less effective at extracting energy from
the gas.
So they're less efficient and there's more energy left in the exhaust gas and the EGT
rises.
And this is what sets the service interval, by the way, because obviously if the EGT gets
too high, things in the engine will start to melt.
The point is that this brand new engine should operate at absolutely optimal efficiency and

(45:32):
should therefore have a low EGT.
If the EGT goes very high on a brand new engine, the obvious diagnosis is that the turbine
is damaged.
This is going to be important later.
We'll come back to these engines, but all you need to take away from this section right
now is that these engines have a relatively narrow band of happiness.

(45:53):
The hotter they get outside of their happy range, the worse it is for everyone involved.
Next slide.
So after completing the first leg from San Salvador to Belize, the flight leaves again
for New Orleans at 10.55 local time.
And for the first hour and a half, everything is smooth with no issues.

(46:16):
And at 12.35, Houston's center clears flight 110 to begin descending from 35,000 feet to
11,000 feet.
So the pilots plugged 11,000 feet into the autopilot and the autothrottle rolled back
engine power to idle for the descent.
And by this point, the pilots had already spotted some thunderstorms ahead on their

(46:36):
weather radar.
Their onboard weather radar showed areas of light to moderate precipitation with some isolated
red cells indicative of heavy precipitation on both sides of their intended flight path,
which were sort of embedded in a large area of milder weather.
So it's more or less instrument conditions with surprise thunderstorms hidden inside.

(46:58):
Yeah, the kinder egg of weather, only instead of a toy you get murdered.
These thunderstorms were actually strong enough that the National Weather Service had issued
a SIGMET, or Significant Meteorological Report, warning of severe thunderstorms in the areas
south and west of New Orleans with cloud tops reaching 52,000 feet.
But the SIGMET wasn't in the pilot's dispatch paperwork, and the Houston Area Control Center

(47:23):
had broadcast the SIGMET over all frequencies at one point, but it was shortly before TACO
110 contacted Houston Center, so they never heard it.
As far as they knew, and they being the pilots, these were your typical New Orleans afternoon
thunderstorms that they had seen however many times before.
Now you still don't want to fly into a thunderstorm, no matter whether it has a SIGMET or not,

(47:46):
because there can be extreme turbulence, hail, microbursts, and things like that.
So the pilots were thinking about how they were going to get past this, and the Houston
controller suggested that they could get around the worst of the weather by deviating 5 degrees
right of course during the descent, which they did.
And the idea was that they could make sort of an end run around the worst part of the

(48:09):
weather, and then cut back through the area of instrument meteorological conditions toward
New Orleans, avoiding the embedded thunderstorms.
But I want to add though that Air Traffic Control had no idea what the conditions were
like out there.
No one else had flown through this area, and actually TACO was the only airline that regularly
approached New Orleans from that direction at all.

(48:31):
So as they were descending through 30,000 feet, they entered the overcast clouds at
quite a high altitude, and the temperature was quite low, so they recognized that they
were potentially in icing conditions.
Now this was an experienced and capable crew, so they followed the procedure for icing conditions.

(48:51):
They activated continuous ignition.
This is a setting that keeps the igniters back of the engine firing so that combustion
continues even if combustion chamber temperature drops too low to be self-sustaining.
They accelerated to the turbulence penetration speed of 280 knots.
And they turned on the engine anti-ice.
This was not automatic on the 737 in 1988.

(49:11):
Yeah, and it's still not automatic today, is it?
Yeah, why would you even ask?
The fact that it's not is one of the reasons the MAX 10 hasn't been certified by the FAA.
Obviously, the other reason is that Boeing keeps screwing up so they didn't get the
waiver.
Yeah, so anyway, they're taking these steps in order to protect the engines from supercooled

(49:34):
rain that can cause icing, and from precipitation that can lead to overwatering.
And all of this is done as a precautionary measure to avoid a flame out.
If you know what's coming, you're probably already giggling.
Slide please.
This is for purely illustrative purposes.
This is not what they actually saw that day.
But anyway, they were now descending through the clouds, aiming for a gap between cells

(49:57):
on the eastern end of the thunderstorm band somewhere just south of the Mississippi Delta.
Because they were in instrument meteorological conditions and they couldn't see jack shit.
They were navigating with their weather radar, much like you see on screen.
Now, do you remember back earlier this year when we covered Air France 447 and discussed
radar?

(50:18):
Yeah, that's the one where we learned that you can terrify and frighten first officers
by fucking around with the radar and telling them it's for the internet.
Yes, but it's also where we learned that the weather radar kind of sucks at the best
of times, and it sucked even more in 1988.
So there are two main ways that navigating around thunderstorms using weather radar can
go horribly wrong, and one of these happened here, but we don't know actually which one

(50:42):
it was.
One thing that can go wrong is radar attenuation, which is where the area looks free of precipitation
because there's just too much shit between it and you, and the microwaves from your radar
can't get back there to give you a signal.
Also ice doesn't absorb or reflect microwaves as well as liquid water does, because water's

(51:04):
physics are weird.
This is also why defrosting stuff in the microwave oven kind of sucks.
And the kind of radars you get on planes for weather purposes are not the most sophisticated
things ever built, because they have to fit in limited space, which is the radome at the
front of the plane, which actually on the 737 is quite small.

(51:26):
Yeah, and if you remember what Fox told us in the Air France 447 episode, the flight
crew doesn't really get much training on how to use the radar and get the most out of it,
so there's that as well.
Yeah, podcasts hadn't been invented yet, so pilots didn't have a reason to just play with
things like...
Yeah, another way you can get screwed by radar is that it presents this snapshot of what

(51:47):
the route ahead looks like right now this minute, and not what it will look like when
you actually get there.
So you can be aiming for a gap that is entirely real, right?
It 100% exists, but it then suddenly closes up on you and now you're screwed.
That's why the advice is to stay at least five miles away from any thunderstorm cell,

(52:09):
because they're very high energy weather systems, or in other words, don't penetrate a gap that's
less than 10 miles across.
That said, this is still done all the time, because otherwise in some parts of the world
like New Orleans, you wouldn't fly if you avoided weather like this.
So long story short, the gap that Dardano and Lopez were aiming for was either not real,

(52:35):
or it was real, but it closed up right as they got there because the entire squall line
was hauling ass eastwards at 38 knots, which is pretty freaking fast for big weather systems
to be moving.
And so shortly after they start their course correction back towards New Orleans, bam,
the gap is gone and they're in the shit without warning.

(52:56):
Pretty much the instant they hit the storm, they started really getting beaten up by it.
Yeah, basically as soon as they penetrate the storm wall, they start flying into a car wash
that is also an ice maker.
Turbulence, heavy torrential rain, a cow flying past.
Another cow.
I think that was actually the same cow twice.
The turbulence was reportedly bad enough the crew had difficulty reading the ins...

(53:19):
And there was also hail, lots and lots of hail, big hail.
Thunderstorms don't always have this much hail, but you know, here's the damn season.
What is a hail?
A hail is a raindrop that gets frozen over and over again.
It's not terribly complicated.
Updrafts caused by reasons that are beyond the scope of this episode carry this frozen

(53:39):
ice back up into the cloud where it gets another coat of ice.
Then it falls, it gets picked up again over and over and over again.
The longer that the hail is stuffing these updrafts, the bigger it gets.
These were big hail.
We're not going to go into a lot of detail about how and why is hail, when is hail right
now.
Are you not paying attention?
And then where is hail?

(54:00):
It's in the cloud because that's where the planes live, Alephs.
So they're flying through the storm, lots of waters running through the engines, lots
of hail too.
Now this shouldn't be a problem, especially for a brand new engine like the ones these
guys have.
They are tested to flying to storms.
But in this case, it was.
To find out why, we've got to talk about enthalpy real quick.

(54:23):
Oh boy, next slide.
So enthalpy is the sum total of all heat energy that's contained in something.
In systems, we use it to describe how much work we can make that heat do.
But it's also the amount of heat needed to change the state of a system from one state
to another.
The graph that you can see here is how the heat and pressure of the gas flowing through

(54:46):
the core of a jet engine changes as it goes through the compressor, the combustor, the
turbine, and the exhaust.
As you can see, the turbine extracts a lot of energy.
That energy is needed to keep the compressor running, and even more of that energy is needed
to spin the big fan at the front.
In a CFM56, about six times as much air is getting pushed around by that big fan, so

(55:10):
it needs a lot of power just to keep it spinning.
The entire premise of a jet engine is predicated on this delicate balance of pulling out exactly
the right amount of heat, exactly in the right spot at the right time.
The engine's optimum efficiency and its best ability to deal with bullshit is when they're

(55:33):
running at high power, climb or cruise power.
Cruise power is actually relatively low, but climb power, absolutely.
Next slide.
These guys were descending into the ice machine fire hose with the engines at flight idle.
Descent with engines at flight idle is a particularly vulnerable time for the engines for a couple
of reasons.
The first, the engines run with very little fuel in the combustor because they're trying

(55:56):
to lose energy so they can slow down and descend at the same time.
Obviously the big fan doesn't need to spin so fast, so the engine only needs enough power
to keep the compressor running as we discussed.
It's running cool and slow, so it doesn't have spare enthalpy to deal with any bullshit.
The second thing that applies in this situation is something called the scoop effect.
So a jet engine is designed to scoop up as much air as possible, and if it is a bypass

(56:20):
turbofan even more so.
But remember that a jet engine can't control the amount of air, just the amount of fuel.
This means the lower your throttle setting, the less of the air getting rammed into the
engine is being heated and used for work, but the same amount of air is being forced
in by the ram effect because you're in less the same size.

(56:40):
So this causes an area of high pressure in front of the inlet that forces some of the
oncoming air to flow around the engine instead, which is what you see on this diagram, where
it says air spillage, and whereas at high power, this air would be sucked in and put
to work.
Now, if the oncoming air has water droplets or hail in it, these would have a much higher

(57:01):
mass than air and therefore more inertia.
The way the excess air gets shoved around the outside of the engine, the rain or hail
is going to want to move straight through.
This means that for a given concentration of precipitation in the air, the amount of
water ingested by the engine is always proportional to the size of the inlet, regardless of how
much air is flowing through the engine.

(57:21):
So when the engine is at low power, it will ingest less air, but the same amount of water.
In this case, that air had a lot of water in it, because these guys aren't just an
irregular air, they're in something that these two are unfamiliar with, something that
I know very well, which is a big fuck you southern thunderstorm.
And because of that, you have all this liquid and frozen water getting shoved into the

(57:43):
engine.
So these are, as Forrest Gump would say, big old fat brains that have a lot of inertia.
Okay, but that's not all.
Not only does the scoop factor make water ingestion worse at low power, the effect is
additive with something else called the centrifuge effect.
When water droplets hit the spinning fan blades, they get centrifuged radially outwards towards

(58:05):
the edges of the engine, where they get sucked through with the bypass air, which is fine.
You know, and the bypass duct is just a duct.
You can put water through the bypass ducts all day.
It doesn't matter.
But some percentage of the water droplets are going to slip in between the fan blades,

(58:26):
like some kind of fucked up game of Frogger.
And if this water is in line with the engine cord, that's where it'll go.
It won't get centrifuged out to the bypass duct.
If the engine is moving at low power, then the fan is spinning slower, which means that
more droplets proportionally will win the Frogger game, and an even higher concentration

(58:48):
of water goes into the core.
And this is on top of the scoop effect.
So let's talk about why this is bad, shall we?
Hale is not liquid water.
It's frozen water.
You need to melt it before you can deal with it.
This, as you can imagine, takes a lot of energy.
And that energy comes from the combustion process, which means less of it is able to
go into the keeping the engine ignited process.

(59:11):
Thanks to our listener Beck, we actually know how much energy is required to flash a hailstone
for a game.
And for a typical 2 to 3 centimeter hailstone, you're looking at 75 kilojoules.
That's about the amount of energy you'd need to take a cup of water from room temperature
to boiling.
Now, that doesn't sound like much.
You have to multiply that by lots of hailstones, and they just keep coming.

(59:31):
Oh, and also, we can't forget about the absolute fire hose of rain coming into the engines.
So if the engines are running at high power, taking away a little bit of heat to melt all
of the ice and evaporate the rain isn't a big deal.
A CFM56-3B1, I actually calculated this, makes 5.6 megawatts of output power at full toga

(59:55):
thrust, which is about enough to power 5,000 San Francisco homes or about 1,000 homes in
suburb Georgia if they have air conditioning running.
But if the engine is at low power, then ingesting large amounts of water can take so much heat

(01:00:17):
away from the combustion process that the fuel will stop burning.
This is called a flame out because the flame goes out.
Yeah.
Engine engineers are really clever.
Yeah, very, very clever people.
So the turbine spins the compressor.
So energy is being extracted from the exhaust and sent forward in the engine to operate

(01:00:39):
the compressor.
In order for the engine power to increase, you need to make more hot gas, which means
you need to burn more fuel and therefore you need enough air to burn that fuel.
Usually the combustion is lean enough that there's air to spare and you can just add
some more fuel.
The air coming out of the compressor is normally about 900 degrees Fahrenheit just through

(01:01:03):
the adiabatic compression that the compressor applies to it.
And this is necessary because believe it or not, jet fuel just isn't very flammable.
You have to get it good and hot before it'll burn very well.
Of course, if you suddenly add a shitload of water to the front end of this process,
then things are going to become problematic in short order since if this recycling of

(01:01:27):
energy is disrupted, this Brayton cycle, the fuel won't be hot enough to burn properly.
It won't make as much hot gas as it should.
The hot gas won't have enough energy to turn this turbine fast enough.
The compressor will then slow down because it's being driven by the turbine and the engine
will slow down.

(01:01:48):
It won't be able to burn even more of the fuel and eventually it will flame out.
It's a vicious circle where there's not enough air being forced into the combustor by the
compressor to support adding more fuel.
So you can't speed up, you can only slow down.
This was something the aviation industry has known about for a long time since well before

(01:02:09):
this incident.
After a crash in 1977 in Georgia outside Atlanta, by the way, shout out to Ari, the FAA put
out a bulletin advising pilots to increase engine power in heavy precipitation, but none
of these guys had read it.
What was found after the accident is that you want to have at least 45% N1, to remind

(01:02:32):
you N1 is fan rotation speed.
With the power settings required to spin the fan at 45% of redline speed, there's enough
combustion heat to handle basically any conceivable amount of precipitation.
And increasing N1 above 45% doesn't really add any value to that.
You've already reached maximum precipitation evaporation efficiency at that point.

(01:02:54):
So it's worth noting here that on the original 737-200, which was also operated by TACA,
the high idle setting with the engine anti-ice on was 55% N1, which was what Captain Dardenneau
was used to.
He had only transitioned to the 737-300 a few weeks ago.

(01:03:17):
But this meant that on the dash 200, it was basically impossible for precipitation to
flame out the engine if anti-ice was turned on.
And that's because if anti-ice was off, then the N1 would be lower and it would be possible.
But when flying in clouds, the anti-ice is generally on.
And by the way, that's because engine anti-ice uses bleed air siphoned from the hot engine

(01:03:40):
core so the engine is programmed to use a higher idle setting with anti-ice active in
order to compensate for that.
However, on the dash 300, the flight idle N1 was only about 35% even with anti-ice on,
which is actually a feature of all medium and high bypass turbofans.
They have a lower idle N1.
The dash 200 on the other hand had low bypass for turbofans, which is why it had a high

(01:04:03):
idle.
And by the way, low bypass turbofans also produced comparatively less power, so it was
a lot rarer to operate them at flight idle in the first place.
Point is, none of the pilots had any idea that the new CFM56 engines on the 737-300
couldn't handle intense precipitation at flight idle.
And even if they had known this, they didn't have a lot of time to react, because they

(01:04:27):
really only had about 30 seconds after entering the heavy preset before all hell broke loose.
Because at 1243 and 42 seconds, at an altitude of 16,500 feet, the engines flooded with too
much water and flamed out.
Next slide.
This means our accident sequence has begun.
Takes us to… pull up!

(01:04:50):
So immediately after the dual engine flameout, the entire electrical and hydraulic systems
went offline, air traffic control lost contact with flight 110, and they dropped off radar
when the Mode C transponder stopped getting power.
They were at 16,500 feet and descending rapidly.
Dardenneau announces it's his aircraft, he tasks First Officer Lopez with working the

(01:05:11):
radios and getting the auxiliary power unit started, and then he tasks the observer captain
Soleil in the jump seat with getting the engines restarted.
Now, these days, proper CRM is for the captain to hand-slide these FOs so they can work on
problem-solving and diagnostics.
But in this case, the diagnosis was that they had no power, and the problem they needed

(01:05:31):
to solve was get power back.
So Dardenneau is flying with the battery-powered attitude indicator, altimeter, and compass.
His first priority is he turns northeast to try and get the hell out of the storm.
We would be remiss not to point out that the 737 doesn't have a backup for hydraulics
in the event of total power loss, because unlike any Airbus or even other Boeing planes,

(01:05:55):
it doesn't have a rat.
There's no ram air turbine, which is the little windmill that automatically drops out into
the airstream of the plane so that it can power emergency hydraulics and electrical
systems.
Now, the backup procedure in the 737 is a business card for Gold's Gym, and on the back, somebody's
just written, hope you've been lifting, bro.

(01:06:17):
Yeah, so more or less.
So Dardenneau was muscling through the storm on manual backup with only basic instruments
against heavy turbulence, rain, and hail, with only a limited idea of where he was.
So three minutes and ten seconds after the power failure, the plane descended through
10,500 feet, and it was at that point that Lopez was able to get the auxiliary power

(01:06:42):
unit started, which restored electrical power and hydraulics.
The APU is a little turbine generator in the tail if you're new to this kind of stuff.
Yeah.
So with the radio now functioning, Lopez immediately broadcast a Mayday, Mayday, Mayday call.
He tells ATC they've lost both engines and are running on the APU alone.

(01:07:02):
Yeah, we talked about this action extensively in the host chat.
After a lot of discussion, tears, and self-reflection among the three of us, we all came to the
conclusion that immediately calling Mayday and telling ATC they've lost both engines
is the proper action to take as opposed to lying to ATC.
Yeah, that's not me glaring at Jesse and Peter right now.

(01:07:22):
Go watch episode five if you don't remember what we're talking about.
So while F.O. Lopez was getting the APU started, the observer captain, Soleil, attempted a
couple of windmill restart attempts using the airflow from the plane's descent to
get the engines up to speed.
Oh, you mean instead of pulling back on the stick as hard as you can, overriding the stick

(01:07:44):
pusher and core locking the engines.
Yeah, so a windmill restart just means using the airflow over the dead engine to spin the
core and then when you ignite the fuel, the process will hopefully become self-sustaining
and the engine will power up.
This is opposed to a normal start where core rotation is initiated by pumping air into
the engines from the starter system or ground start equipment.

(01:08:08):
So there's a minimum airspeed for a windmill restart, which they had met, no problem.
But the engines wouldn't light off because they were still ingesting so much rain, the
combustion chamber temperature was too low to sustain it to ignite the fuel.
So after Lopez got the APU going, they decided to attempt a more traditional engine restart
process using the APU.

(01:08:29):
So with Dardano still hand flying, finally with instruments working now, Lopez worked
with Soleil to start running the standard restart procedure using the main engine starters
powered by the APU.
And he first tries to start one engine then the other.
Presumably this is aided by all of them sitting in the correct seats and not, for example,

(01:08:51):
looking for Pepsi.
You might say these guys were the pinnacle of skill.
Yes, we will continue to dunk on Beavis and Butthead as long as we are making this show.
So at this point the engines light off, but the engines just flatly refuse to spool up
to produce meaningful thrust when Captain Dardano attempts to apply power.

(01:09:11):
The reason for this is kind of complicated and we don't have definitive evidence for
what actually happened, but we have a theory that we're going to explain.
Yeah, so the way an engine startup process normally works is the APU powered starter
system starts spinning the core while the fuel valves open and the igniters spark to

(01:09:32):
initiate combustion.
The combustion then causes the core rotation to accelerate until it becomes self-sustaining
and then the starter shuts off.
Core rotation speed is called N2.
It's expressed in terms of a percent of redline speed, just like N1 is fan rotation speed.
The lowest ground idle N2 is somewhere north of 50%, so keep that in mind.

(01:09:57):
Anyway, in this case when the observer Captain Soleil turned on the starter with the ignition
switches on, combustion did happen, but it was only partial.
I should note that this part is, as I say, what we've inferred from the engine behaviour
described by the pilots and the results of the general electric engine tear down.

(01:10:18):
It wasn't explicitly stated in any documents, but it's what we think happened.
The pattern of damage they found in engine number two very strongly suggests that this
is what happened.
Partial combustion means only some of the combustors lit off.
Again, we talked about how there are eight combustor cans in the CFM56-3 and it looks

(01:10:38):
like maybe three of them successfully lit off, the other five didn't.
And this was because the engine was still full of water.
So only some of the combustors were dry enough for the fuel to ignite.
So with only three of the combustors actually working, there was insufficient combustion
to sustain core rotation without the assistance of the starter.

(01:10:59):
So N2 never went above 26%, which is way below ground idle.
But what the pilots would have seen was that ignition had successfully taken place and
the engines were working, and this is what Soleil reported to Captain Dardenneau.
There wasn't some kind of warning that says, hey, only a third of your combustor cans are
lit.

(01:11:19):
So Captain Dardenneau went, great, I have working engines now, and he pushed both thrust
levers forward to stop the descent and power didn't increase.
Next slide, please.
Instead they started getting overtemp warnings.
The exhaust gas temperatures were out of limits.
So what was the deal here?
So advancing the thrust levers just adds more fuel to the engine, which means usually more

(01:11:45):
burning, which means more energy, which means more hot gas, which means fan goes spinny
faster.
But remember, these engines were too wet to sustain the amount of combustion required
to increase N2 beyond 26%.
So adding more fuel didn't cause N2 to increase.
Most of the extra fuel didn't even ignite.

(01:12:06):
Instead it appears to have flowed back from the combustion chamber into the turbine section
where it mixed with the hot gases produced by the fuel that did ignite in the combustion
chamber.
This then auto ignited inside the low pressure turbine section, which is bad.
It didn't spin the turbine faster.
It just caused the exhaust gas temperature to increase until it was above limits.

(01:12:29):
It seems to have ignited, as I say, in the low pressure turbine based on when the overtemperature
damage was found, which drives the fan at the low pressure compressor.
So even if it did increase N1 rotation speed a little bit, it wouldn't have had enough
of an effect on N2 to get the engine properly started.
And this overheating kept occurring whenever the throttles were advanced, which Captain

(01:12:54):
Dardano immediately noticed.
He says, what do I do about without power in English?
And then repeatedly says, I don't feel power in this thing.
This thing is not starting in Spanish.
Remember how earlier in this episode we said that if a pilot sees a high EGT on a brand
new engine, their first thought is going to be, this engine is damaged?

(01:13:19):
So yeah, Dardano concluded that the engines wouldn't produce power because there was
something wrong with them.
They were damaged.
And even if they weren't damaged, they would be very shortly because the high EGTs would
melt the turbines, start fires, you know, that kind of thing.
So he took the only option available, which was to shut down both of them.
And actually, by the time he did so, parts of the right engine's low pressure turbine

(01:13:43):
had actually started to melt.
The left engine wasn't damaged, though, and interestingly, it could have been restarted
if there was time, which there wasn't.
Also, both engines could have been restarted on the first attempt if the pilots had shut
them down completely, let the starter run for a bit to clear the water out, and then
reintroduced fuel flow.

(01:14:03):
But there was no procedure for that, and this was only discovered in hindsight.
So let's now pause for a moment and talk about what was happening with air traffic
control while all of this was going on with the engines.
Initially, ATC couldn't see the plane on radar or contact the crew for three minutes during
the electrical power loss until the APU was started and First Officer Lopez made his Mayday

(01:14:24):
call.
After the crew reported that they lost both engines, the controller initially attempted
to direct the crew to turn left toward Navy calendar field, but there was a great big
red thunderstorm sail off to their left, and not wanting to go through that stupid bullshit
again, the crew refused.
The controller provided the distance to that airfield twice, but the track of the aircraft

(01:14:46):
continued moving north, away from calendar.
Then after the crew believed they got their engine back, they requested vectors to New
Orleans, and the air traffic controllers in New Orleans suggested a closer airport at
Lakefront, but then when the engines refused to spool up, the pilots realized they weren't
going to make it to Lakefront either.

(01:15:08):
So these guys are rapidly running out of kinetic energy, and they're pretty sure they're
not going to make it to a runway.
I would also like to point out that this is approximately the point at which the CVR starts
because part of it got taped over, unfortunately.
But do you know what would have helped them at this moment?
I may have a guess.

(01:15:29):
Yes, you know it.
You love it.
Jato Bottles would have given them enough energy to make it back to Nouvelle-Aulillard.
No, we talked about this.
Tell us.
So they had popped out of the clouds at 5,000 feet around the time the engines didn't spool
up, and they had about two minutes to impact at this point.

(01:15:49):
The video was okay-ish with rain showers, and they were somewhere over Lake Bourne.
The crew scouted the area and assessed crash landing options, maybe in some of the swamps,
but Dardenneau decided that their best option is probably to circle once and ditch on Lake
Bourne, where they had a relatively flat area of open water.
The air traffic controller, on the other hand, gives them the position of Interstate 10 and

(01:16:12):
suggests landing there, but it was obvious to the crew that that was also too far away.
Even as we learned in episode 2 of this podcast, landing on the Interstate is a horrible idea.
At this point, the pilots spot the Gulf Intracoastal Waterway, which Dardenneau prefers over Lake
Bourne because it's right alongside some really big industrial facilities, and they'll
be rescued faster.

(01:16:32):
So Dardenneau lines up with the Intracoastal Waterway with the flaps and gear retracted,
fully ready to put this plane in the water.
He's riding right on the edge of energy preservation because if they stall at this
altitude it's game over.
First officer, but also, he needs to be close to stall or else they'll drop too fast and
hit the water too hard.

(01:16:53):
So first officer Lopez is working to finish the ditching checklist, they're preparing
for a water landing, but then, next slide, right then, Lopez spots a grass levee off
to the right of the canal and he goes, wait a minute, that looks very flat and very long,
maybe we should land there instead.

(01:17:14):
So Lopez says, look, look at that one over there.
Dardenneau is letting the other two pilots look for a spot to land.
He's concentrating on stretching this glide and flying the plane, and this is textbook
what he does.
Yeah, he is aviating, letting them do the navigate and the communicate.

(01:17:35):
Also for non-American listeners, a levee is an artificial dirt berm that's used for flood
and water control.
Sometimes it's also called a dike, an embankment, a flood bank.
These will sound very familiar to our Dutch listeners.
Once they listen to this episode, when they're done compulsively building water control systems
out of sticks, leaves, and their fur oils, these levees are all over New Orleans.

(01:17:58):
They're built by the US Army Corps of Engineers because New Orleans is below sea level.
It's also an affront to God, much like the Dutch.
Yes, so the crew do an incredibly quick assessment of the situation, I'm talking like five seconds,
and they decide to take a chance on the levee.
This is by no means a sure thing though.
These guys have no idea what the consistency of the ground is, whether they'll sink and

(01:18:22):
rip the wheels off.
Dardenneau says, do we go in there on the grass?
He's still concentrating on flying the plane.
Lopez says, see you home brave.
Besides the weather is so shit, landing in deep mud and getting stuck is very much a
risk.
However, they don't know this yet, but this particular levee is special because of who

(01:18:44):
built it and where it is, so just keep that in mind.
So Dardenneau agrees to use the levee at the last moment.
They put the flaps and gear down, but this is not going to be easy because first off
they're slightly too high, and second there's a raised area just before the proposed touchdown
zone which they'll hit if the approach angle is too shallow.
Soleil says, wait a moment, wait a moment, we've got to keep on trying, and he even

(01:19:07):
suggests that landing in the water is better.
But Dardenneau's got this.
So what Dardenneau does here, because he was an absolute chad with balls of steel, is he
extends the speed brakes and then he puts the plane into a slip in order to lose speed
and altitude simultaneously, allowing him to come in steeply enough to dodge the raised
area without landing so fast that they can't stop after they're down.

(01:19:32):
And all of this is done while aiming for a levee that's only about 120 feet wide, which
is like 30 or 40% narrower than a major airport runway with an embankment running down the
left side and water on the right side.
Dardenneau remembers that he needs to tell the passengers to get ready for what might
be a rough landing.
He says, prepare the cabin, come on, come on.

(01:19:53):
But Lopez is enthusiastic.
Vamanos, Ali Vamanos, let's go.
We feel we are remiss if we do not remind you that the pilot that is pulling this off
is 29 and has one eye.
And then they fucking nail it, they touch down as softly as they can get away with,
and Dardenneau is ginger on the brakes because he doesn't want to dig that nose wheel in.

(01:20:14):
Up to this point, it's been tense in the cockpit.
But as soon as Dardenneau puts this thing down in what has to go down in history as
one of the greatest landings of all time, the relief is palpable.
Everyone's laughing before they even stop rolling.
But the ground holds and the brakes work and after 2,500 feet, they come to a stop as smooth

(01:20:36):
as butter.
Next slide, please.
And that's where they are.
And that's it.
Our accident sequence has ended with the plane on its wheels, intact, and a cabin full of
very relieved passengers.
At this point, the crew discuss whether to evacuate and Captain Dardenneau says, and
this is a direct quote from the translation of the CVR, no, get everybody out because

(01:21:01):
otherwise these gringos are going to start saying things.
Yeah, so they ordered an evacuation and everyone disembarked via the slides without a single
injury except for a girl who had had her appendix removed two days before and popped a stitch.
Yeah, and she was okay.
So for the record, this is a landing that it's appropriate to clap for.

(01:21:24):
Right, save your claps for this, not when your Delta Connection CRJ successfully lands
in Indianapolis with unlimited visibility and no wind.
Oh, another fun thing is air traffic control had no idea these guys were safely down.
So the probably quietly panicking controller sent another aircraft was in the area to look

(01:21:45):
for a missing 737.
And then the pilot reported, hey, I see a 737 down there.
The slides are out.
There's a bunch of vehicles streaming toward it.
And the controller asked if he was on the highway and the pilot reported he was on one
of these embankments and it was intact.
The controller then asked, is there any smoke or fire?
Like, okay, I get it.
This guy's nervous.

(01:22:07):
And the pilot just responded, negative.
Everything looks okay.
Looks like he did a pretty good job on it.
And that was that.
The flight crew also decided after they double checked that everybody in the aircraft had
evacuated safely and that rescue crews were on the way, that you know what, it's raining
and they're going to go ahead and stay inside the plane where it's dry.

(01:22:27):
Thanks.
And you know what?
There, y'all earned it.
Next slide.
All right, talk about the aftermath.
Next slide.
All right, so first off, fun fact.
They totally at random picked probably the best off airport landing site imaginable.
Yeah, they were landing at a facility owned by NASA and operated at the time by Martin

(01:22:52):
Marietta.
It's part of Lockheed Martin now.
It's the Michoud Assembly Plant, one of the largest buildings in the world.
It's where they built the Saturn V's first stage and where at this time in 1988, they
were getting ready to restart the production line for the shuttle's external tank.

(01:23:14):
Yeah, the purpose of the levee in question, which is actually visible in this photo, it
was a place they could load the shuttle's external tanks onto barges so that they could
be sailed through the Gulf and around to the Cape to be launched.
So it was compacted, engineered soil.
There's a nice gentle ramp at the end so they could get a heavy crawler up it.

(01:23:39):
How heavy?
Well, the standard weight tank weighed 77,000 pounds dry and the crawler that carries it
weighs almost a third of a million pounds.
By contrast, a 737-300 weighs 72,490 pounds with no gas or asses.
Of course, the factory had also made the S1C, which weighed 303,000 pounds dry.

(01:24:05):
This 737 obviously still had a bit of fuel left and a medium load of people in bags,
but it was still like rolling a Honda Civic onto a road designed for an 18-wheeler.
It was fine.
Yeah, the levee was in pretty good shape for handling some pretty crazy Salvadorian landing
a narrow-body jet on it.
Michoud also had heavy equipment to help them with the post-crash activities.

(01:24:26):
They towed the plane off the levee into the parking lot alongside, which you can also
see in this photo in the background.
They removed both of the engines and sent them to GE, fitted new engines, and then used
Saturn Boulevard, which is the road that you can see just behind the large cube-shaped

(01:24:47):
part of the building, as a runway to fly it to Nouvelle-Orléans.
I think they only actually replaced the right engine.
I'm pretty sure the left engine was original when they took off because it wasn't damaged.
I'm pretty sure they removed both of them.

(01:25:07):
I seem to remember reading that they put the left one back on because it was fine.
I mean, same difference, right?
Yeah, I guess.
Yeah.
So the road they were on, they used as runway, Saturn Boulevard, it was built for moving
the S1 stage between the factory and the levee, so it was both very wide and very straight.

(01:25:27):
You know, unlike any of us.
So apart from the one engine that was launched and some hail damage to the nose and the leading
edges, there was no damage to the plane and no damage to the plane from the landing at
all, which is pretty amazing for an off-field landing, really.

(01:25:47):
History does not record how much the Martin Marietta guys laughed at the Boeing guys for
needing their help.
But this does prove how easy it is to fly a stranded plane away, which makes Ural Airlines
all the more cowardly for refusing to do it from a field in Siberia.
Next slide.
I have to imagine this was the easiest crash investigation of that particular NTSB team.

(01:26:08):
The plane is just sitting there, the pilots are all there.
Hell, they probably had investigators on site before that storm was even over.
Picturing an NTSB guy putting rain into one of those evidence bags.
Another one carefully putting a yellow number of placard next to a hail stone.

(01:26:32):
This particular plane, N75356, ended up being fine, as we discussed.
It re-entered service a little bit later with TACA.
It came with them into Avianca when they merged.
It got sold through a few other airlines.
It finally ended up at Southwest, where it flew until 2016 when it was stored at Pinal
Airport.

(01:26:53):
This thing flew for Southwest for 21 years, which is almost an entire era.
Hey, I turned 27 real soon, okay?
Alright, let's talk about the investigation and lessons learned.
What the fuck did we learn?
Next slide.

(01:27:13):
So one interesting thing is that Boeing didn't have a twin-engine out procedure until three
weeks after this happened.
Yep, these guys did everything with no applicable emergency procedure.
Just in case everything else they did wasn't impressive enough.
Another thing Boeing did after the accident was they changed the first step on the in-flight
engine restart checklist to shut down the engine, because this basically hard resets

(01:27:37):
the process and prevents an unbalanced condition like what happened on flight 110.
More importantly, though, they did a bunch of testing on the CFM56-3 engine and found
that it was way more vulnerable to precipitation than anyone had guessed.
The engine had only been in service for a couple years, and this wasn't even the first
time this had happened.
In 1987, an Air Europa 737-300 was descending into Thessaloniki in Greece with its engines

(01:28:05):
at idle when it ran into a hail column so dense the pilots later called it a quote wall
of ice.
Both engines flamed out, but the pilot successfully windmill restarted them and made a safe landing.
The investigation found that the concentration of water in the storm cell they encountered
was well above the certification limit that the engines were designed for.

(01:28:25):
That's right.
Engines are certified to ingest large amounts of water and ice and keep running, and actually
the amount of water that they had to ingest at that time was pretty similar to the estimated
water concentration encountered by TACO 110.
But according to the CFM56-3 certification tests, they should have been able to keep

(01:28:46):
operating at atmospheric water concentrations 400% higher than this.
So what went wrong?
Well, the answer is twofold.
First off, the certification tests were conducted at high engine power and low forward airspeed,
which is the least adverse condition for precipitation ingestion.
We already went over why high engine power is better, but low airspeed also means less

(01:29:10):
water is ingested per unit of time as well.
On the other hand, when engine power is low and forward airspeed is high, like during
a high altitude descent, for example, the amount of water ingested into the core can
be over three times greater at the same atmospheric water concentration.
The second issue is that certification tests were focused on liquid water, not hail.

(01:29:34):
It was assumed that when ingested, hail basically behaves the same way as rain, and hail encounters
were mostly considered in terms of possible impact damage, but as it turns out, this assumption
is untrue.
For an engine, ingesting hail is actually quite a bit worse than ingesting rain for
several reasons.
As we discussed earlier, first off, because the mass of the hail stones is absolutely

(01:29:56):
larger than that of a raindrop, the inertia increases the severity of the scoop factor
and decreases the effectiveness of the centrifuge effect.
So when the engine ingests that hail, more water ends up inside the engine core than
it would be the case if the atmospheric water concentration was exactly the same, but the
water was liquid rain.
And the second, because hail is frozen, more energy is required to vaporize and expel it

(01:30:19):
with the exhaust, which increases the risks of engine flame out.
A third possibility is that large enough hail stones can actually damage blades of the engine
and limit its ability even further to create power.
Because of this, the flame out on TACA 110 happened at a precipitation level that's
really quite common inside large thunderstorms.

(01:30:40):
It wasn't even the heaviest precipitation Dardano had encountered in his career.
And yet it flamed out anyway because of this.
None of this was discovered after the Aer Europa incident in 1987 because the calculated atmospheric
water concentration in that incident was above the certification limits, so they didn't
really bother to study whether the presence of hail made things worse or not.

(01:31:02):
But after TACA 110, they found that not only was the entire certification basis flawed,
but that the CFM56 in particular did not meet the certification requirements if engine power
was low, not by a long shot, regardless of whether it was ingesting rain or hail.
Yeah, and CFM International ultimately identified two main factors that made the CFM56-3 more

(01:31:25):
vulnerable.
The first issue was that the spinner, you know, the cone thingy in the middle of the
engine was too conical.
It didn't do much to stop hail stones from sailing straight into the core.
To fix this issue, CFM simply made the spinner more dome-shaped so that when hail impacts,
the hail stones tend to deflect radially outward and into the bypass ducts rather than into

(01:31:49):
the core.
The second issue was that the splitter was too close to the fan rotor.
So what's the splitter anyway?
The splitter is just a static structural element that divides the airflow into the core stream
and the bypass stream.
On the CFM56, the splitter was closer to the fan rotor in front of it than on any other

(01:32:10):
similar engine, and this reduced the effectiveness of the centrifuge effect because water droplets
had to travel farther along the fan blades before sloughing off or else they would be
stopped by the splitter and redirected into the core instead of being flung all the way
radially outward and then into the bypass duct.
CFM fixed this issue by moving the splitter about 1.75cm farther away from the fan rotor,

(01:32:37):
leaving more room for droplets to slough off the blades and into the bypass duct.
And in addition to all of these fixes, CFM also increased the number of variable bleed
valve doors in the engine core airflow path, which provide a path back into the bypass
ducts after objects have already entered the core.

(01:32:57):
The idea is that when the airflow zigs, hail and rain will zag because they have more momentum,
they're denser.
So they'll get flung out through the variable bleed valve doors while the airflow continues
on its merry way towards the combustion chamber.
More of these doors means that this works better.

(01:33:19):
In later versions of the CFM56, the zig in the core airflow path was made more zaggy,
amplifying this effect even further and unfortunately providing an opportunity for James Dyson to
sell vacuum cleaners.
The spinner redesign, the splitter modification and the additional bleed valve doors were

(01:33:42):
all made mandatory thanks to an FAA airworthiness directive.
Another airworthiness directive later mandated an increase to the minimum N1 idle on the
CM56.
Airworthiness directives, they work.
And no CFM56 engine ever again flamed out due to…
Oh sorry, I'm getting a phone call.

(01:34:03):
Hello?
Huh.
It's the NTSB, they say it did.
Next slide.
So in 2002, a Garuda Indonesia Boeing 737-300 with CFM56 engines encountered a record breaking
rainfall inside a thunderstorm over Java which exceeded the certification limits and both

(01:34:27):
engines flamed out.
Actually if I recall correctly, the precipitation encountered by that flight was, at least at
that time, the heaviest rain ever known to have been encountered by an aircraft in flight.
In that case, the pilots made several engine restart attempts before starting the APU and
it turned out that the plane's main battery was faulty so after two restart attempts there

(01:34:49):
was no longer enough juice to start the APU either and they were shit out of luck.
Remember how he said the 737 doesn't have a ram air turbine?
Yeah.
Oof.
What is it with Boeing and batteries?
Batteries are not that hard, guys.
As the pilots ended up ditching the plane on the Solo River just outside Surakarta, one

(01:35:11):
flight attendant was killed after the tail struck the river bottom and ripped out the
floor where she was sitting, but the other 59 passengers and crews survived.
The investigation into that accident didn't find anything wrong with the CFM56 engines
specifically though they just flew into a rainstorm that was way out on the thin trailing
end of the bell curve.

(01:35:32):
Anyway, after TACA 110, the FAA updated the certification criteria for water and hail
ingestion and operating procedures for jets with medium and high bypass turbofans were
changed to encourage pilots to increase N1 to at least 45% immediately when they encounter
heavy precipitation.
Anything else anyone wants to add?
Yeah, just a quick note.

(01:35:54):
One of the two engines was left running a bit too long before shutdown and there was
actually some debris found in its tailpipe on TACA 110.
So they did a borescope inspection and they observed that there was damage to the second
stage of the low pressure turbine where the gas is supposed to have had energy and heat

(01:36:14):
taken out of it by the high pressure turbine already.
The debris they found in number two's tailpipe was most likely Rhean 77, which is a high
temperature alloy that's used in turbine blades.
Surprising nobody.
But also it had been exposed to temperatures of at least 2450 degrees Fahrenheit, which

(01:36:35):
is way too hot for the second stage of the low pressure turbine.
That's about the temperature the gas is supposed to be at the inlet face of the high pressure
turbine, not the low pressure turbine, which all supports the theory that there was fire
happening in the wrong place in these engines.
And as we've seen, you really only want fire to be in clearly defined places in an aircraft

(01:37:01):
and not anywhere else as several of our earlier episodes can attest.
This doesn't really add much to our understanding of the accident.
I just thought it was interesting.
Oh yeah, and another random thing, they never did toxicology testing on the pilots.
So you're telling me they could have been sniffing glue this whole time.

(01:37:21):
It's that performance enhancing glue?
Yeah.
Does the epoxy they used to bind the layers of the 111 frame count as performance enhancing
glue?
Yes, absolutely.
One note I will say is we did have a discussion whether the flooding that caused the flameout
was fuel or water.
By the time they performed boroscope inspection of the engine after it had landed, the hail

(01:37:45):
had melted.
The water had evaporated away, so no body no crime I guess.
And we'll never really know the true cause of the flameout.
Yeah, you can flame out an engine by flooding it with fuel just as easily as you can with
water.
Yeah, and we just don't know.
Also, Charlie Dardano is a badass and we all really appreciate him.

(01:38:07):
He kept flying airliners until 2023 by the way.
When he retired, it made the news.
Next slide.
So, we wanted to do this incident because our last few have all been extremely brutal
ones to cover.
We've had refusal to reevaluate situational awareness leading to seafit, a deep misunderstanding

(01:38:29):
of the limits of fly-by-wire leading to a stall, colonialism and crimes leading to a
botched takeoff, and then capitalism and crimes leading to an in-flight fire.
So we want to do a crash that had as happy an ending as possible.
Our listeners and readers of Cura's articles are probably very familiar with the Swiss
cheese model, but just in case, it's the notion that an accident happens when all the

(01:38:50):
holes of the Swiss cheese line up in such a way as to allow the accident.
Well, in this case, they got Swiss cheese in two directions.
Yes, it was extremely unlucky for them to get fucked by this hellstorm and fly through
it in such a way as to unlock this edge case where both engines died and couldn't be
restarted.
But, then they got extremely lucky that they happened to be very close to shore, saw this

(01:39:13):
landing spot, happened to have three extremely experienced guys on board, and oh yeah, the
landing site they could reach with their onboard kinetic energy just happened to be usable
as a runway and next to a major aircraft and aerospace facility.
We should also note that these sections of the CVR that we quoted are from the last few
minutes of flight when these guys knew they were going to make it.

(01:39:34):
And I should note that the last few minutes was all we had because some FAA chucklefucks
accidentally taped over most of the CVR when they powered the plane back on for some checks
a few days later.
And right after they actually get it on the ground is when they start really laughing
and joking around.
They are deadly serious right up to that point when they're working through the issue and

(01:39:56):
yeah, it's understandable.
When it breaks, the relief is absolutely palpable.
This crash was nice for us to break down in the host chat because most of the crashes
that we cover and that Kira covers don't end with so little damage to the plane that
a plane built in 1988 goes on to fly until halfway through phase one of the MCU.
Though we thought this would be a fun one to celebrate with you.

(01:40:18):
Absolutely.
I hope you enjoyed this fun story as much as we did and shout out to Captain Charlie.
He's a real one.
Godspeed.
I just had a thought which is that these people had bought tickets to New Orleans and this
emergency landing was like 25 minutes drive outside of the city.

(01:40:38):
It's close enough.
Right.
It's close enough that they could just take cabs to get where they were going.
All right.
Thanks everyone for joining us.
Our next episode will be on Malaysia Air 370.
Have a good night.
Thank you.
Bye.
Bye

(01:41:54):
You