Very 'close' Qantas jets came within 800m of each other | Hacker News Reader

Very 'close' Qantas jets came within 800m of each other

airlineratings.com

My first coding job was at an avionics subcontractor. One of the senior engineers there told me once that TCAS (Traffic Collision Avoidance System, equipped on all large planes) is theoretically capable of resolving up to 127 aircraft all converging on the same point. As long as all the pilots follow the directions, there won't be any collisions. I have no idea if that's actually true, and could definitely be remembering it wrong, but it always seemed impressive.

Well, TCAS works great in an ideal world.

When you mix civil and military planes, and ATC and pilots who are non-native English speakers, and when your TCAS can be broken, then you get accidents like the German or Brazil collisions.

In the military regime, you must follow orders. In the US civil aviation world, the pilot in command is the final authority and can deviate from any ATC order, with possible paperwork and/or consequences.

Although English is the official language for international flying, local ATC and pilots use their local languages.

There was a scandal in Hawaii when one flight school targeting Japanese students was issuing FAA licenses to pilots who couldn't understand English.

TCAS is not ATC though, no? Also I don't know if TCAS is localized, but it's only two words if I'm not mistaken, "Climb!" or "Descend!"

And if TCAS says "Climb!" and the ATC says "Descend," what do you do?

Think quickly! Several hundred souls are on the line.

For civil pilots, that's easy: Whatever TCAS says.

For military pilots, it's more complicated because they're now deviating from orders. Hopefully it's follow TCAS anyway, but more shades of gray regarding who can deviate.

Climb - TCAS always takes precedence. This was standardized after this incident: https://en.wikipedia.org/wiki/2001_Japan_Airlines_mid-air_in...

Depending on the plane, it's also a digital reading on the radar showing the speed and altitude for surrounding planes relative to yours. You don't have to wait until the collision warning to try and clear things up with ATC.

The pilot in command is the final authority in the military also.

Having been in and out of Sydney a lot, I find this very easy to believe. The airport is straining at the limit for capacity, so they have been quite adventurous in getting planes in and out of there for some time. Particularly with the curfew limiting the hours of operation.*

* Not saying the curfew is a bad thing

Tight curfews are a bad thing, I think. There's an airport just down the road from me, Hanscom (KBED) with an 11P-7A curfew. The shenanigans you see at 10:55P with sometimes three simultaneous landings (two aircraft and one helo to the ramp) are no good for safety.

I've been picking people in the morning and I've more than once seen multiple airplanes circling the field jockeying to be the first one down final at 7:00:00 like it's a sailboat race. Now, would the neighbors rather hear airplanes flying multiple low-level patterns at mid-power from 6:45 to 7:05, or would they be better off if airplanes could come in and land directly [typically at a lower power setting and with only one approach]?

I didn't think there was any race to land as ATC gets full ownership of the pattern and who gets to land when?

That airport is part-time towered, so air traffic control does own the airspace but only when the tower is open. When the tower is closed, it reverts to an uncontrolled (or pilot-controlled) airport with insanely high landing fees (so few airplanes actually land and almost no one would land at 6:55 rather than wait the extra 5 minutes).

So, while it's pilot controlled, we self-announce and negotiate with the other airplanes and plan our pattern so as to be on a 1-2 mile final when the tower opens. (Think of it as the aviation equivalent of a four-way stop sign or rotary which drivers self-navigate every day.)

"Attention all aircraft, Bedford tower is now open; class Delta airspace in effect" "Bedford Tower, Bugsmasher 1234 2 mile final, Runway 29" "Bugsmasher 1234, Runway 29, cleared to land." "Tower, Bugsmasher 97Victor, 3 mile left base for Runway 29" "Bugsmasher 97Victor, #2 following company traffic over the numbers, Runway 29, cleared to land."

Meanwhile, everyone who insisted on the curfew had to hear the bug smashers circling for 5-10 minutes previously.

Thanks for all the detail. I wasn't aware that you could have "closed airports" where you could still land.

Sometimes they even turn their lights off, and you click your mic into a frequency to turn them on:

https://www.aopa.org/news-and-media/all-news/2017/march/flig...

Damn well that's something interesting to learn. Have an upvote.

Out of interest, how high is the insanely high landing fees? I tried Googling but could not find an answer.

For each operation (landing and takeoff charged separately) during curfew: $63 surcharge (plus the normal landing fees) under 12,500# max takeoff weight. $458 surcharge 12,500# or over. After 5 in a calendar year, the 6th and subsequent double.

The sports teams and large jets that are based at the field (and so don't pay an after-hours callout fee to access the ground side of the field) pay it as needed.

There are no races to land before the curfew starts at Sydney airport.

The curfew only affect large planes, and late arriving jets still land normally in sequence, though the airlines pay a steep increase in landing fees.

It is worth noting that the Melbourne-Sydney air route is the second busiest in the world [0] in regards to aircraft movements, and the the flight paths are over some of the most densely populated parts of the city.

The noise was intense and getting worse each day.

Aircraft that arrive early circle well off the coast and do not cause noise issues.

[0] https://www.oag.com/hubfs/Free_Reports/Busiest%20Routes/2019...

> multiple airplanes circling the field jockeying to be the first one down final at 7:00:00 like it's a sailboat race

I don't think it's up to the aircraft who lands first. That's not how airports work.

Depends on the airport. There are plenty of uncontrolled fields out there which definitely rely on inbound aircraft to sort themselves out.

Agree. By the numbers, there are about 500 airports in the US with control towers (some are part-time). There are over 20,000 airports without control towers in the US.

Sydney is one of the airports with runways close enough together that they require special training and procedures. Nothing like SFO though.

Given how much Sydney traffic gets routed over populated areas at relatively low altitudes I can definitely sympathize with wanting a curfew. In fact the noise itself has a cute little term 'Marrickville Pause' because it's so damn loud.

https://www.youtube.com/watch?v=OoYikXvnU6M

For information, it's the "Petersham Pause" or "Marrickville Moment" :) There should also be the "Tempe Tumult" because you can basically reach up and touch one of the bastards at that stage.

Flying around the states I always got the impression that things were a little rickety, that SFO vid is nuts though.

Ah, that's the one. I've lived more or less under an SFO flight path for ages and used to commute down 101 right next to SFO which gets overflown constantly. Sydney was a whole different world though. Just walking sitting on a bench in Enmore Park it felt like you really could reach out and touch one of them.

Oddly enough about the only flight delay out of SFO I've experienced was SFO-SYD, a flight that gets delayed explicitly as a matter of course to avoid the curfew.

Yeah, the fastest I've ever gone in a plane was due to something trying to make it to Sydney for curfew. Can really get nuts sometimes.

Aviation Herald report: http://avherald.com/h?article=4cb85fbe&opt=0

This should really be the linked article, much more informative. The charts at the bottom are particularly helpful for understanding the sequence of events (and seeing the conflict between standard departure and go around routes).

It will be interesting to see the investigation for this, as a layman I don't know if such conflicting routes are standard, and how ATC is supposed to respond to such a situation (presumably not by leaving it to the pilot to avoid, like in this case, however).

It's quite common, because a departure route and the missed approach are both designed around the same terrain/obstacles and often use the same navigation beacons.

The way to solve it is for ATC to provide alternative missed approach instructions. So they would say something like "Quantas 123, go around, turn left heading 250, climb 2000ft" knowing that the departing aircraft is going to make a right turn.

> The Airbus A330 involved in the August 5, 2019, incident was taking off on runway 34 Right as a Boeing 737-800 landing on the same runway was instructed by air traffic control to perform a go-around.

This makes no sense. The seem to be blaming the go around. But what really seems to have happened is they allowed and aircraft on a runway a plane was about to land on. Who authorized this and why? What if the plane didn't go around? Would it just have wiped out the plane on the runway?

And it sounds as if the 330 was both taking off ans flying at the same time.

> The A330 first officer, who was pilot flying, then saw the 737 in close proximity and, in response, reduced the aircraft’s angle of bank to reduce the turn towards the 737.

I don't get it.

I don't think you should read any of that as "blaming the go-around", but as stating facts.

> And it sounds as if the 330 was both taking off ans flying at the same time.

Time passed between the two events.

From how I read it: A330 is ready for take-off, 737 is in landing approach. Tower realizes this is too close, tells 737 to abort and turn right to give the A330 space. But the A330 will take off and also turn right because that's where it is supposed to go. And that happens: A330 takes off, goes into a hard right turn, alarms go off because it gets too close to 737 which is still low and in soft right turn, A330 pilot doesn't turn quite as hard to leave more space.

What exactly the mistake was and how grave it is is for the safety boards to figure out (shouldn't have let the 737 get as far as it did? should have stopped the A330 from takeoff? Should have warned both sides they'd get close to each other? It actually was acceptable in the end because the controller had it in sight?)

I still don't really understand. Unless the 737 was about to crash land right on top of the 330, it must have turned right well before the airport. If the 330 took off, it would have been a few kilometers beyond the airport before it turned right itself. I just can't visualize this 800m thing. If the 737 was so close it's go round intersected a plane that took off, this story really should read "737 narrowly averts crash landing into an airbus that was mistakenly let onto runway". Smells like a coverup to me.

There are standard "missed approach" routes published for each runway. Generally they involve flying more or less directly over the runway before turning to circle around for another attempt.

The 737 flew that route. As they were doing that, the 330 was taking off and turning in the same direction after takeoff (as they were directed), so they both wound up near each other a few km beyond the runway as they were both turning to the right.

Now I am just flabbergasted. How on earth could they let the 330 take off in that case? And it still seems to me if the 737 didn't execute a missed approach, there would have been a crash on the runway. That is just as big a deal as the 800m flyby. This story should be a bigger deal than is being made out.

You don't get the throughput major airports require without vectoring planes to land before other planes have taken off from the same runway. If the plane taking off takes too long you tell the landing plane to go around. This happens daily - I've had planes I'm a passenger on do it twice. It wouldn't have been a big deal at all except for them both having been directed to turn right in conflicting ways.

This is entirely standard procedure.

In Europe (and ICAO standards) the controller will say to the second plane "number 1, traffic departing ahead" to let them know and then only say "cleared to land" after the other plane has left the runway.

In the US this is a bit different, they will say "cleared to land, traffic departing ahead" and are trained to cancel the landing clearance if the other plane isn't off the runway fast enough.

In both cases the second plane is made aware of the first, so they're ready (should always be) for a missed approach.

> And it still seems to me if the 737 didn't execute a missed approach, there would have been a crash on the runway.

When you're approaching an intersection controlled by a traffic light, do you feel the same anxiety? There are cars occupying the same bit of road that you're going to be occupying in a few moments, and if you don't stop at the light, there will be a crash. I'm guessing you don't give it a second thought (you may double check that traffic has stopped before proceeding, and pilots do the same thing when they're landing).

This is basically correct. There are two other contributing factors:

- The controller who was talking to the two aircraft was a trainee (and being supervised by a senior controller).

- The sequence happened after sunset, but before EENT, and the controller cleared the aircraft after judging the distance visually. Most things aviation-related, and in particular visually judging distance, become much harder after the sun sets.

Incident info: http://avherald.com/h?article=4cb85fbe

Departing flight: https://flightaware.com/live/flight/QFA459/history/20190805/...

Sunset table: https://www.timeanddate.com/sun/australia/sydney?month=8&yea...

It is very, very common for one plane to be on approach for a runway while another one is taking off. If the plane taking off is delayed at all, the approaching plane is told to go-around (which is what happened here). Happens all the time.

I haven't read up on this incident, but I suspect the approaching plane wasn't given clearance to land yet, and would have gone around on their own if they hadn't gotten that clearance in time.

Well there are go rounds and there are missed approaches. It appears this was the latter. I can't see how a standard go round involves flying at low altitude directly above a plane that is about to take off or taking off. Obviously planes can approach runways from a distance while others are taking off, it seems here it didn't just approach, but came close to actually landing.

A plane on an instrument approach when told to go around is expected to follow the missed approach procedure. Which 9 out of 10 times includes flying over the runway.

Isn't missed approach procedure to fly runway heading, however not to actually overfly the runway (ie, offset of centerline)?

If you're lined up on the runway, flying "runway heading" will overfly the runway.

Here's the details for the runway in question: https://www.airservicesaustralia.com/aip/current/dap/SSYGL06...

For a visual flight in a small plane that's the way to do it (and then technically runway track instead of heading)

For an instrument flight there is a defined missed approach route that usually includes specific altitudes and turns to get you to some holding position.

> Well there are go rounds and there are missed approaches.

Yes there are, and they both use the same procedure.

> Obviously planes can approach runways from a distance while others are taking off, it seems here it didn't just approach, but came close to actually landing.

It got to point where the controller had to give them final clearance to land, or tell them to go-around. Obviously that point is going to be reasonably close to the runway.

One plane landing shortly after another departs is standard procedure, provided they have sufficient separation [0]. However, the separation in this case became too tight after the previous landing had vacated the runway, so ATC issued a go-around - but the missed approach procedure and standard departure route for that runway coincided, resulting in the close approach [1].

Note that not enough separation in this case (for the landing) is probably about a mile.

[0] http://tfmlearning.faa.gov/Publications/atpubs/ATC/atc0310.h...

[1] http://avherald.com/h?article=4cb85fbe&opt=0

Reply on news.ycombinator.com