Changelog

The changes a controller would notice, newest first. Entries appear here on the day the change went live.

Work you could not see from the scope is recorded in the repository and is not listed here.

October 2026

Emergencies have their own section in the documentation

10 October 2026 · Added · Documentation

The documentation now has its own section on emergencies. An overview explains how a trainer starts, steers and ends one. The radio failure page, which explains each step the aircraft takes and why, has moved there from the command pages.

The section is at https://docs.aurora.schwingenschloegl.online/emergencies

Trainers can start a radio failure on an aircraft

9 October 2026 · Added · ATC Commands

In an aircraft's private chat, a trainer can now start one of three radio failures:

• emergency comms starts a complete failure. The aircraft stops answering and no longer follows instructions on the frequency. It keeps flying its last clearance, and you can still steer it from the private chat.
• emergency comms transmitter fails only the transmitter. The aircraft still receives, so it follows instructions on the frequency and idents when asked, but it never answers.
• emergency comms unaware starts a complete failure that the crew has not noticed. The aircraft stops answering but keeps its squawk code.

In the first two, the aircraft squawks 7600 between 30 seconds and two minutes after the failure starts.

To start the failure later, add in and a number of minutes from 1 to 60, for example emergency comms in 10. To start it when the aircraft passes a fix, add at and the fix, for example emergency comms at ROKIL. The aircraft tells you privately what it will do next.

emergency now skips the wait, emergency pause and emergency resume hold and restart it, and emergency status shows where it stands. emergency cancel ends the failure, and the aircraft's next answer on the frequency explains what happened. emergency list in a private chat with SERVER shows every running emergency.

The radio failure page in the documentation explains each step and why the aircraft takes it: https://docs.aurora.schwingenschloegl.online/emergencies/radio-failure

Trainers can steer aircraft privately and make them speak

8 October 2026 · Added · ATC Commands

Trainers now have functions of their own on the server. Everyone with a staff position in IVAO's Training department is a trainer here, without being added by hand. When you log in as a trainer, or become one while connected, the server tells you so and links to what you can do.

As a trainer you can:

• write an instruction to an aircraft in a private chat. The aircraft flies it and reads it back there, so the trainee sees only what it does on the scope.
• type say followed by a text in that chat. The aircraft then says the text on the frequency of the controller working it.

Instructions on the frequency work for every controller as before. A private chat with an aircraft now only works for trainers.

Arrivals that are high on the approach now lose the extra height faster

7 October 2026 · Changed · Flight Model

An arrival that is still high and fast when you clear it for the approach now gets down to its profile within a few miles. It puts the speedbrake out and descends noticeably steeper, about 2 300 instead of 1 700 feet per minute on a 737 and about 4 000 on an A320, then stows the boards again. Before, the speedbrake hardly changed the descent, and the same aircraft kept it out for about thirty miles.

This applies to the 737, 757, 767, 777, 787 and A320. Other types, including the 747 and the rest of the A320 family, still descend at their usual rate with the speedbrake out.

Arrivals whose route begins with a direct leg appear on their route

7 October 2026 · Fixed · Sessions

Some session arrivals appeared beside their route instead of on it, flying towards their entry point from the wrong side. This happened when the filed route began with DCT, which is about one flight in thirty.

These arrivals now appear on the last leg of their filed route, like every other flight.

Arrivals enter where the October agreements hand them over

7 October 2026 · Fixed · Sessions

Session arrivals now appear at the handover points and levels that IVAO's letters of agreement set for AIRAC 2610.

At Vienna, arrivals enter shortly before MASUR, BARUG or NIGSI. Before, they appeared at the start of their arrival route, up to 130 NM out.

At Frankfurt, arrivals over TUTOV are part of random sessions. Before, a session left them out and said TUTOV was not a gate.

Random traffic at eight airports, built from real flights

6 October 2026 · Added · Sessions

Random sessions now start at Frankfurt, Munich, Berlin, Düsseldorf, Hamburg, Cologne, Vienna and Friedrichshafen. Every arrival is a real flight that landed there in early October. Its strip carries that flight's callsign, aircraft type, registration, operator, origin, filed route and cruising level, and a departure time that fits the moment it appears.

Start a session in the runway direction you want to work. You get only the flights that actually landed in that direction, with the route they filed on such a day, spread over the entry fixes the way they really came in. The arrival route on each strip matches that direction.

If an airport saw no arrivals in a direction, the start says so and does not begin. If a gate has too few different flights for the rate you ask for, the start tells you that as well.

The recorded Munich sessions are no longer offered.

Updates and connection drops are handled more carefully

4 October 2026 · Changed · Sessions

The whole server is now tested the way you connect to it, and new behaviour can be checked live in Aurora before it ships.

When the server is updated, it now shuts down cleanly and signs off the controllers who were online instead of keeping them registered. A controller who drops during a brief outage inside the server is signed off as well. One faulty part of the server can also no longer stop it from clearing up once everyone has left.

IVAO training staff can connect to the server

4 October 2026 · Added · Sessions

Everyone who holds a post in IVAO's Training department can now log in to the server without being added by hand.

If your access is withdrawn while you are connected, you get a message saying so before the connection closes.

When the server turns a login away or rejects a command, Aurora now shows the whole message with its reason. Until now it showed only "Connection refused" or "Command rejected".

A320 and 737 descend at full rate above FL100 with a steady vertical speed

4 October 2026 · Fixed · Flight Model

Between FL100 and FL200 at their normal descent speed, the A320, 737-800 and 737-900 now come down at the full rate idle thrust allows. Before, they descended up to about an eighth shallower there.

An arrival slowing towards 250 knots above FL100 no longer has its vertical speed jump by about 200 feet per minute every few seconds. The vertical speed now changes only at the start of the descent and just before FL100, where the aircraft really is faster than planned.

On these three types the readback adds "Energy budget tight" to a combined descent and speed reduction more often than before.

A session carries on after a short server interruption

4 October 2026 · Fixed · Sessions

If part of the server restarts while you are working a session, the session now carries on. You stay connected. Your commands go unanswered for about ten seconds, and then everything works again. Your aircraft and your tracks are all still there.

Until now the server stopped answering after such a restart. It stayed that way until somebody restarted it by hand.

A controller whose Aurora connection drops is signed off within a minute

4 October 2026 · Fixed · Wire Protocol

If your connection drops or Aurora freezes, the server signs you off after about 45 seconds of silence. Until now it could go on counting you as online for a long time.

Sessions and unassumed traffic therefore end on time once a controller is gone. If you come back within five minutes, your tracks are yours again.

Your tracks and label values come back when you reconnect

3 October 2026 · Fixed · Wire Protocol

If your connection drops and you are back within five minutes, the aircraft you had assumed are yours again. The values you had set in their labels return as well, such as a cleared level or a scratchpad note.

Until now they came back unassumed and with empty labels, and you had to take every aircraft again by hand. They now return a moment after you log in.

Descending aircraft slow down at the rate the arrival planning expects

3 October 2026 · Fixed · Flight Model

An aircraft told to slow down while it is descending now loses speed at the rate the arrival planning assumes. Higher up it used to lose speed at about half that rate or less, so it reached the new speed later than planned. An arrival still faster than its normal descent speed as it nears FL100 now has its speed down to 250 knots before it gets there.

On an approach the aircraft starts slowing down slightly earlier than before.

A320 and 737 descend closer to their manufacturers' published rates

2 October 2026 · Changed · Flight Model

An Airbus A320 cleared down at 250 knots below FL100 now comes down at about 1600 feet per minute, close to Airbus's own figures. The Boeing 737-800 and 737-900 now descend close to Boeing's published idle rates. All three used to come down noticeably shallower and took longer to reach the level you cleared them to.

Between FL100 and FL200 the A320 is still a little shallower than the real aircraft. Other types descend as before.

Server commands and problem reports now work in a private chat

1 October 2026 · Added · ATC Commands

Server commands now also work in a private chat with SERVER. In Aurora, type .chat SERVER to open it and type the command as you would on the frequency. The answer appears in that chat and nobody else reads it, so a problem report sent with feedback stays between you and the server.

Instructions to an aircraft stay on the frequency and are refused in this chat. pause and resume are still announced on your frequency, because they affect everyone. A session started from the chat calls its pilots on your primary frequency.

When the server refuses a line you typed on a frequency or in the private chat with SERVER, the reason now appears there rather than in the message tab.

September 2026

Arrivals now appear where their own route is handed over to you

25 September 2026 · Changed · Sessions

Session arrivals now appear shortly before the point where the area controller hands them over to you, and that point is read from each flight's own flight plan. Where an arrival flow can be handed over at more than one point, each flight appears at the one its route actually uses, at the level agreed for that point, instead of every flight of the flow appearing at the same place. Arrival routes whose handover lies further along the route now bring their traffic in close to that handover, rather than far out in the area controller's airspace.

How far before the handover point an arrival appears follows how early the frequency change comes, which is at the latest two minutes or ten miles before the handover, whichever comes later, unless an airport publishes its own figure. An arrival calls you as soon as it appears. At fields whose arrival routes start far out, that is sooner than before, when arrivals waited until they were fifty miles from the airport.

Two arrivals never appear on top of each other any more. Where several arrival routes are handed over at the same point, the second waits until the first has moved on.

Arrivals now enter at the handover levels set in the IVAO LoAs

24 September 2026 · Changed · Sessions

At 23 airports in Germany and Austria, session arrivals now appear at the points and levels that the letters of agreement (LoAs) of the IVAO divisions set for handing traffic from the area controller to approach, including levels that differ by runway direction. Before, only Munich worked this way and every other field used the level on its arrival chart, or FL110 where the chart gave none.

In recorded sessions, a flight whose arrival route begins where no LoA applies now appears about forty miles out along its route. Most of these are long routes that are handed over further along, such as Frankfurt's, so the flight appears close to where the real handover happens instead of far out in the area controller's airspace.

At fields without an LoA for their arrivals, traffic with no published level now arrives at FL130 instead of FL110, which is closer on average to the known LoAs.

Squawk and ident in one instruction, and a code typed after ident is refused

23 September 2026 · Fixed · ATC Commands

A squawk code typed after the ident, as in si 2000 or sq ident 2000, used to be read as a number pad instruction. Depending on its first digit the aircraft descended, turned left or turned right, all under an ordinary readback. Such a line is now refused whole, and the reply shows the spelling that works.

Squawk and ident together is written the way it is said on frequency, code first: sq 2000 ident, or sq2000 si in the short form. The aircraft reads it back as "Squawk 2000 and ident."

A clearance to a level of zero or below is refused as well, whatever the spelling, so 2000 typed on its own no longer sends an aircraft down towards the ground.

A session ends five minutes after the last controller for its airport leaves

23 September 2026 · Changed · Sessions

A session now belongs to the controllers who work its airport. When the last of them disconnects, the session has five minutes. If one of them connects again in that time, nothing changes. Otherwise its pending arrivals are cancelled, its traffic is removed, and the end is announced on the frequency the session was started from. Until now a session ran on as long as anybody at all was connected, so a controller working another airport kept an abandoned session flying, and the next controller at that airport inherited traffic nobody had been working.

A controller counts for every airport their position works, whether it is named after the airport or covers it. An aircraft whose track a controller who is still online has assumed stays when the session ends, and from then on belongs to no session, like one spawned by hand. A session started while nobody who works its airport is connected starts on the same five minutes, and the reply to the start says so.

Randomised sessions now start at airports beyond Munich

23 September 2026 · Added · Sessions

Any airport with a traffic pool can now run a randomised session, not only Munich. Without a setup made for the airport, arrivals enter where its published arrival routes begin, at the level on the chart or at FL110 where the chart gives none. Friedrichshafen is the first field this opens, and the start message says which setup is in use.

Ask an aircraft for its speed, heading, level or distance

23 September 2026 · Added · ATC Commands

Aircraft now answer questions on the frequency. report speed (rs), report heading (rh), report level (rl) and report distance MAH (rd MAH) are answered from what the aircraft is actually doing, in standard phraseology. Above FL250 the speed comes as a Mach number, the unit you assign speeds in up there. confirm 100 (rl100) checks a level, and the aircraft either affirms it or says where it really is. Requests can share a line with a clearance, d080 rs for instance.

Turn readbacks now use the magnetic heading you gave. t 270 reads back 270 degrees, not the true heading the aircraft flies.

The problem report is now feedback <what happened>, because report belongs to the aircraft. It works with an aircraft selected. A report typed on its own files nothing and says what to type instead.

Aircraft outside every controller's range now disappear after five minutes

23 September 2026 · Added · Sessions

When no controller has assumed an aircraft and it stays outside every connected controller's radar range for five minutes, the server now removes it. Previously, an arrival that nobody worked would keep its heading and fly on indefinitely, and so would a track released into empty airspace.

Aircraft you have assumed are never removed this way as long as you stay connected. Traffic that nobody has seen yet is also left alone, so an arrival that appears beyond your range will still reach you.

The range is the one set in your radar client. In Aurora, it is measured from the centre of your radar screen. If you move the screen far away or choose a small range, unassumed traffic may disappear. Assume the aircraft you are working, and they will stay.

A session start at one airport more than the server runs at once is refused

21 September 2026 · Added · Sessions

A server now runs traffic at a set number of airports at once. Two is the default, and whoever runs the server can set another. The number is a setting rather than a measured limit. The simulation has room for more, and two is where the first setup with several airports on one server runs, so the number is raised as that setup proves itself. Starting a session at one airport more than the server is set for is refused. The reply names the limit and the sessions that hold it, so you can end or delete one before starting yours. Starting a session at an airport that already has one still replaces it, and a session that has ended keeps its airport until its remaining aircraft have been worked down.

Descents no longer run steeper than idle thrust can hold the speed on

21 September 2026 · Fixed · Flight Model

An aircraft descending to a cleared altitude now descends no faster than it can while holding its speed with the engines at idle. On the A320 family, the 737 and about sixty other types the descent at their usual descent speeds, below FL100 at 250 knots and on some of them higher up as well, is shallower than before, and a jet cleared fast and high reaches its altitude later instead of arriving on speed for free. The speedbrake still does less than it should on those descents; that half is open on the same item.

Direct to Stuttgart no longer points at Lake Constance

20 September 2026 · Fixed · Navdata

About three percent of airports sat in the wrong place in the navigation data, Stuttgart among them: dct EDDS turned the aircraft towards a heliport near Allensbach, 59 nm from the airport. A direct or a route to any of those airports now goes where the airport is. Approaches and procedures were never affected.

Route fixes are now joined with commas in every spelling

20 September 2026 · Changed · ATC Commands

route MIQ,ROKIL,ANELA is the route command now, in the long and the short rt form alike, the way the number-pad form *MIQ,ROKIL,ANELA was already written. A list with spaces in it comes back as unrecognised.

Some fix names look like commands. D2 is a fix near Ajaccio and also a descent, HAAKE is a fix and also a hold over AAKE, and a route with spaces in it ended wherever such a name happened to sit: route MIQ HAAKE ROKIL was refused for a hold nobody typed, and route MIQ ROKIL D2 flew two fixes and a descent to 200 feet. With the commas the list is one word and every fix name is safe inside it.

Commands still follow the list after a space, so route MIQ,HAAKE,ROKIL d100 is three fixes and a descent. A speed glued onto the number-pad list without that space now reads as one more fix and is refused; *MIQ,ROKIL ,220 is the spelling.

Procedures can be cleared in words, and the reference explains the number pad

20 September 2026 · Added · ATC Commands

clr ALG2E clears the departure, clr ROKI1B the arrival and clr ROK26 rwy 26L the transition: the clearance that until now existed only as cALG2E can be written out like every other command. The short form is unchanged.

The command reference now shows the third spelling as what it is: the number pad, with its signs (*, /, ,, +, -) in one column and its digits (8, 2, 4, 6, 3) in the next, instead of a Symbol and a Numpad column that read as two different things. Six spellings that never sat on the pad (i26L, l26L, hMIQ, a bare h, s2000 and #26L) are now listed beside the short form they belong with. Nothing you type today changes.

A paused simulation now says so, and refuses what it cannot carry out

20 September 2026 · Changed · ATC Commands

While the simulation is paused, an aircraft instruction, a session start or a spawn is now refused with a reply that names who paused it and since when, and tells you to type resume first. Until now the instruction was accepted in silence and flown whenever somebody typed resume, possibly minutes later, and a session started into a pause answered with the usual started line and then spawned nothing.

Everything that describes or reduces the picture keeps working while paused. Session status now ends with the pause and its start time, delete and session end still work, and the confirmation of a problem report tells you the simulation is paused when it is. The report itself records the pause for the operator.

A pause no longer outlives the last controller

20 September 2026 · Fixed · Sessions

When the last controller disconnects, the five minutes that clear all traffic now lift a pause as well. Until now the pause stayed set after everybody had gone, so the next controller to connect found a frozen scope, paused by somebody who had left an hour earlier, and had to type resume before anything would move. That happened whether or not aircraft were left behind: with none, nothing was scheduled at all and the pause stood until somebody noticed.

A controller who returns within those five minutes finds the simulation exactly as they left it, paused and with their traffic in place, the same as before.

Clearances are read at the airports your position works, not only at Munich

20 September 2026 · Changed · ATC Commands

The server no longer assumes one airport for everybody. A departure, arrival, transition, approach or expected runway is looked up at the airports the position you logged in as normally works, so a Frankfurt or Hamburg controller clears their own procedures without the server reaching for Munich's. Where a runway number exists at more than one of those airports, the aircraft's filed destination decides, then the airport your position is named after; an aircraft filed to an airport outside your coverage is cleared there, and a position the server has no entry for still works through the flight plan. A clearance nothing can settle is refused with the candidates named.

Spawning traffic at a fix that exists more than once now picks the one nearest your airports. Magnetic headings convert with the magnetic variation at the aircraft itself rather than one value for the session, which matters once one server serves airports far apart. The runways in landing use, which the server uses to judge how far an arrival still has to fly, now come from the session you start rather than from a setting; nothing needs configuring for that.

One more thing the review of this change turned up is fixed with it: the approach clearance readback now quotes the localiser course printed on the chart, 261 for Munich's 26L, where it used to say 259.

Instructions typed one after the other are now carried out in that order

19 September 2026 · Fixed · ATC Commands

Two lines sent within a second or two of each other could be carried out the wrong way round, because each one waited its own random moment before the pilot acted on it. The case that showed it was a heading followed straight away by an ILS clearance: about half the time the clearance was taken first and the heading then cancelled it, so the aircraft flew through the localiser while you held a readback saying the approach was armed. Each instruction still takes its own moment, but a later one now never overtakes an earlier one. A go-around stays immediate, as before.

A heading no longer gets dropped when the aircraft reaches its cleared altitude mid-turn

19 September 2026 · Fixed · ATC Commands

An aircraft given a heading and then, on a later line, an altitude used to stop turning the moment it levelled off, if the level came before the turn was complete: it rolled out on whatever heading it happened to be passing and held that. Typically this showed on a long base turn with a descent to platform altitude given right after it, and the aircraft then flew off at an angle to the final approach. The heading now stays in force until the next instruction, whatever the altitude does in the meantime.

An arrival cleared a long way out no longer slows down in the cruise

19 September 2026 · Fixed · Flight Model

An aircraft cleared for an approach 40 nm or more out used to be told to slow down at once and to take flap a long way from the runway. Its plan assumed it had to lose the speed on the straight line down to the runway, and a clean jet cannot slow down on a three degree descent, so the plan called it late from the moment it was cleared. It now plans on the shallow segment a real aircraft slows down on, keeps its own speed until somewhere around 16 to 20 nm, and slows from there; flown against Munich, the flap sequence that used to start 35 nm out now starts at 15 nm. On the final approach itself nothing changes: from the point where the procedure's platform altitude meets the glidepath the aircraft is established and the plan stops assuming it can level off. That point comes from the procedure rather than from a fixed distance, so it sits about 11 nm out at Munich and nearer 6 at most other airports. One thing is unchanged on purpose: a jet cleared far out at cruise speed still uses its speedbrake for a while, because whether it can make the approach is judged on the straight line to the runway. That judgement is the next step.

A heading given after an approach clearance no longer cancels it

19 September 2026 · Changed · ATC Commands

Until now a heading, a direct or an orbit given after an approach clearance cancelled it, so a small correction on the intercept meant clearing the approach again. The clearance now stays until you end it with cancel app, go around, a new approach clearance or a new route such as a STAR, a transition, a route or a hold. A new route replaces the clearance and the readback says so.

A cleared aircraft flies your heading and joins the localiser when it meets it at a usable angle. That also holds for a heading across the final, so say cancel app first when you want to take a cleared aircraft across. An aircraft that is already established leaves the localiser on your heading, keeps its altitude until you give it one, and the readback tells you it has left. Your next heading arms it again and it joins on the next intercept without a new clearance. An aircraft that loses the localiser in wind or on a bad intercept keeps its clearance as well and waits on its present heading for a new intercept vector.

An aircraft cleared while it is already inside the localiser beam but pointing slightly away from it now turns onto the course instead of drifting off and losing it.

Controllers now read each other, not only the aircraft

11 September 2026 · Fixed · Wire Protocol

Type on a frequency and the other controllers working it read you, the same way an aircraft's answer reaches them. Until now only the answer travelled: a frequency carried the readback and not the clearance it answered, so half of every exchange was the machine's. A private message between two control positions arrived nowhere at all, and now reaches the station it names.

This matters most at an observer position, which exists to watch somebody work and until now saw every readback and no instruction.

Present heading and squawk ident are now written the way every other verb is

9 September 2026 · Changed · ATC Commands

ph becomes present heading with ph as its short form, and id becomes sq ident with si as its short form. * still flies present heading and si still squawks ident, so the fastest way to type either is unchanged.

The two were the only verbs whose short form sat in the full-spelling column, which left one of them shortening to a bare p and the other with no short form at all. Every other verb that takes no argument is already a two-word phrase shortened to its initials: cancel speed and cs, go around and ga. These two now match.

p and id are no longer accepted. Type ph and si, or present heading and sq ident.

The command reference now names the short way to clear a procedure

8 September 2026 · Fixed · Documentation

The three spelled-out clearances for a departure, an arrival and an approach transition showed "none" in the Short column, so a controller scanning the table for a quicker spelling found none and had to read the prose underneath to learn that cALG2E and cROKI1B exist. Each of the three rows now names c<NAME> directly.

It stays a pointer rather than a second spelling in the cell, because the two are not always the same clearance: clr star ROKI1B rwy 26L answers one runway where cROKI1B answers both parallels, since the cycle files that arrival for each.

Arriving aircraft now fly the approach at the weight they land with

8 September 2026 · Changed · Flight Model

An arriving aircraft is flown at the weight it lands at rather than the weight it departed at. It has burned the sector's fuel by then, so it is lighter and sheds speed more readily on the way in.

Most types reached the gate at their approach speed either way and look no different. Four do not, and on those a controller can watch it: the ATR 72 that used to arrive at a thousand feet 55 kt above its approach speed now arrives 37 kt above it, and a Hawker business jet gains 52 kt of the same. Aircraft configure a little earlier and slow a little harder across the board.

The types that look unchanged are the ones whose landing weight has not been researched yet, and they will follow as it is.

Altitude limits can now be typed in the symbol spelling

8 September 2026 · Added · ATC Commands

/100+ is a floor and /100- a ceiling, the same two characters the speed already used in ,210+ and ,250-. Until now the limits stopped at c100+, and the symbol column said none.

The slash carries no direction, so a limit on it cannot contradict a verb: /100- at an aircraft below FL100 is a ceiling it already meets, where c100- has to be read as a descent under a climb verb.

Six commands can now be typed on the number pad

3 September 2026 · Added · ATC Commands

Climb, descend, both turns and the two approach clearances gained a spelling made of digits alone, listed in the new Numpad column of the command reference. The first digit is the verb and where the key sits is what it means: 8 climbs, 2 descends, 4 turns left, 6 turns right. So 8090 is c 090, 4270 is tl 270, and 6090 2070 ,220 is one clearance mixed with the other spellings.

3 clears the ILS onto the runway you noted with expect and 33 the localiser under it. Neither takes a runway of its own, so 326L is refused, and so is 3 with nothing noted. There is no keypad speed and no keypad limit. Unlike /, the keypad vertical still watches the direction: 2090 at an aircraft below FL090 comes back refused.

Pick which entry fixes a randomised session feeds traffic from

2 September 2026 · Added · Spawn

A randomised session no longer has to use every entry fix. Name the ones you want and give each its own rate: session start eddm random 26 rokil 20 betos 10 gives you twenty an hour over ROKIL, ten over BETOS, and nothing over LANDU or NAPSA. The rhythm mode still goes on the end, and leaving the gates out works exactly as before.

Ask a gate for more than it can pass and you are told its limit instead of quietly getting fewer aircraft than you typed. Naming a gate the direction does not use, or one with no flights behind it, is refused the same way.

Session commands say what is wrong instead of going quiet

2 September 2026 · Fixed · Spawn

Some session commands used to answer with nothing at all. No error, no refusal, just silence, which looks exactly like a command that never arrived.

Two cases are fixed. A rate with far too many digits, as in session start eddm random 26 99999999999, now tells you the number is unusable. And when the server cannot read one of its own session files, it says so and names the file, instead of leaving every session command at that airport unanswered.

August 2026

Controllers can turn an aircraft through a full circle

31 August 2026 · Added · ATC Commands

An aircraft can now be told to orbit. Type "orbit left" or "orbit right" and it flies one full circle in the direction you name, then carries on with whatever it was doing before: a heading, a route or an arrival, picked up where it left off. There is nothing to type to end one. The short forms are "ol" and "or", and the symbol form doubles the turn sign, so "++" is right and "--" is left.

How long the circle takes follows the aircraft rather than a fixed rate. Down in the approach an A320 at 180 knots comes round in just over two and a half minutes and the same aircraft at 250 knots in about four, while a light aircraft turns no faster than three degrees per second and runs a little over two minutes. Higher up every turn is gentler, so an orbit in the cruise takes considerably longer than one in the descent.

An orbit cancels an approach clearance, in the same way a heading does, and the readback says so. Clear the approach again when you want the aircraft back on it. Clearing it on the same line as the orbit does the opposite and is the tidier way round: the clearance stands, the aircraft flies its circle without joining the localiser, and it takes the beam once the circle is behind it.

Four commands now accept only the spellings the reference lists

30 August 2026 · Changed · ATC Commands

The relative turn, the at-or-above and at-or-below limits, the minimum and maximum speeds and the hold each accepted spellings that appear in no reference table. tlby30, tl b 30, abv100, smin180, s 180+ and holdMIQ all worked, next to the spellings that are actually written down. Nineteen of them are gone, so these four commands now read the way the rest of the set already did: one spelling per form, and anything else is answered with a refusal instead of being guessed at. Type tl by 30 or tlb30, abv 100 or c100+, smin 180 or s180+, and hold MIQ, h MIQ or hMIQ.

One spelling that worked properly has gone with them, and it is worth knowing if you hold at a fix whose name begins with OLD. Those fixes can no longer be entered in the one-character hold form, because hOLDOD and the word hold run together with OD are the same thing typed. The server was reading every such line the second way: hold OD typed without its space became a hold over OLDOD and was read back as understood. Write hold OLDOD or h OLDOD instead, which both still work.

Approach transitions work on Frankfurt 07L and 25R again

30 August 2026 · Fixed · ATC Commands

Clearing an approach transition for Frankfurt's 07L or 25R answered that no approach exists for that runway, although six are published for each. The go-around was affected the same way. Both work now.

The cause was how an approach is named in the navigation data: where two approaches of the same kind serve one runway, a letter is appended to tell them apart, and the server read that letter as part of the runway. Frankfurt's parallel runways carry it on every approach, so nothing matched. Berlin and Hannover had the same problem on individual approaches rather than on whole runways.

Arrivals fly the deceleration profile instead of reaching approach speed early

30 August 2026 · Fixed · Flight Model

An arrival that is planning its own approach now holds each speed until the track it has left makes the next reduction necessary, and reaches its approach speed at the stabilisation point rather than before it. It used to go for the landing speed the moment it started slowing: on a full transition into Munich that meant all four flap settings inside two and a half miles, approach speed with eleven miles still to run, and eleven miles flown in landing configuration. The same aircraft now takes flap 1 at 15 miles and 230 knots, flap 2 at 11 miles and 200, and finishes configuring inside 7 miles.

Aircraft also no longer configure for landing when an approach is cleared from far outside the range the plan covers. One cleared 150 miles out used to end up in landing flap at cruise altitude.

Clear any published procedure by naming it

29 August 2026 · Added · ATC Commands

Typing c followed by a procedure name clears it, whatever kind it is. cALG2E clears that departure, cROKI1B that arrival and cROK26,26L that approach transition: the name says which of the three it is, so you no longer type sid, star or trans.

The runway usually comes from the name as well. Most procedure names serve exactly one runway, so cALG2E is the whole clearance. Where a name serves several, the reply lists them and you name one.

Both spellings of a name work. Navigation data shortens a name that does not fit its field, so the arrival published as ROKIL 1B is stored as ROKI1B: type either and you get the same arrival, and the reply always names the procedure the way the data has it.

The spelled-out clearances are unchanged: clr sid ALG2E rwy 08R still works and still writes everything out. What went away are the half-written mixtures in between, such as sid ALG2E 08R, which made one clearance readable four different ways.

Tell an aircraft which runway to expect, then clear the approach without naming it

29 August 2026 · Added · ATC Commands

Typing EXPECT 26L notes the runway an aircraft is going to. It is not a clearance: nothing is armed, the aircraft keeps the heading and altitude you gave it, and it will not turn onto the approach. What it saves you is the clearance afterwards, which no longer needs the runway: CLR ILS, or a bare I, clears onto the noted one, and L does the same for the localiser alone. The short spellings of the note are E26L and #26L.

To change the runway, name the new one; a second note replaces the first. There is no separate command to take a note back, matching the phraseology, which has none either: clearing an approach uses the note up and a go-around drops it.

One spelling changed with it. A bare I or L used to cancel an approach and now clears one onto the noted runway, because one letter meaning two opposite things was worth removing: I26L armed an approach and I cancelled one, told apart only by whether a number followed. Cancelling an approach is CA, in every spelling.

Turboprops climb at the speed their handbooks give, not a jet's

18 August 2026 · Fixed · Flight Model

This finishes what the August 15 note left open. Turboprops were flying a climb speed schedule that speeds up with altitude, which is what a jet does. Above ten thousand feet they had to accelerate on the way up, and the energy that went into going faster did not go into climbing, so they arrived at their cruising level later than they should have.

They now hold the speed their own handbook publishes. Across the propeller fleet the effect is large: an ATR 72 used to be told to accelerate from 170 to 210 knots during the climb and now holds 170 throughout, and a Beech 1900 climbs at 134 knots where it used to be pushed to 180.

Every one of the 102 propeller types in the fleet changed. Where a handbook could be read, the figure comes from it: 63 types now carry a speed traced to a manufacturer handbook or a certification document, with the page and revision recorded. For the remaining 39 no handbook could be obtained at any price, because the manuals are sold rather than published or are issued only to operators. Those types hold a single speed for the whole climb instead, which is what the certification rules describe and what most handbooks that could be read actually publish. They are marked as derived rather than researched, so nobody later mistakes one for a figure somebody read.

Jets, helicopters and piston aircraft are untouched.

Replies no longer arrive with question marks and missing characters

17 August 2026 · Fixed · ATC Commands

A route clearance read back as "Route via MIQ ? ROKIL", an approach clearance quoted its course with the degree sign missing, and a spawn confirmation did the same with the heading. The characters could not survive the connection to the radar client, and three of them failed in three different ways. Replies now stay within the character set every client reads the same way: a route reads "Route via MIQ > ROKIL", an angle reads "course 259deg M", and the long dashes are gone in favour of ordinary punctuation. The wording is otherwise unchanged. If a command does contain a character the radar client cannot carry, the reply now names it instead of quoting back a blank.

A crashed radar client comes back to the traffic it was working

16 August 2026 · Fixed · Wire Protocol

Two things that went wrong when several controllers worked the same traffic are fixed. If your radar client crashed, the aircraft you had assumed were given up immediately, so you came back to traffic that belonged to nobody: they now stay yours for five minutes, and are simply still there when you reconnect. And the assigned values shown in an aircraft label (level, speed, runway, scratchpad) are now sent the way the reference network sends them, which also means you see your own entries confirmed back rather than only your neighbour seeing them.

A mistyped PAUSE now says so instead of doing nothing

16 August 2026 · Fixed · ATC Commands

PAUSE and RESUME take no argument, and until now a line like PAUSE NOW did nothing at all: no reply, nothing in the logs, no way to tell it from a dropped connection. It now answers "PAUSE takes no argument. Type PAUSE or RESUME on its own." and leaves the simulation exactly as it was.

The spelling itself has not been loosened. PAUSE NOW still does not pause anything, and there is still exactly one way to write each command. What changed is that the refusal is audible, which is how every other server command already behaved: a malformed session line answers with its usage, and a delete with no target says why it did nothing.

Aircraft labels are filled in when you join a session in progress

16 August 2026 · Fixed · Wire Protocol

Joining or rejoining a session left aircraft labels blank while a colleague's showed the assigned level, speed and scratchpad, and they only filled in once somebody typed a value again. The server was sending the label contents before it had told your radar client that the aircraft existed, so there was nothing to put them on. The aircraft are now announced first, and the labels arrive filled in.

A transferred track now stays with the controller who accepted it

16 August 2026 · Fixed · Wire Protocol

Transferring an aircraft to another controller used to complete and then come undone: the target turned green on the receiving scope and about a second later showed as assumed on neither scope. The server was passing on the sending controller's release of the track, which their radar client sends automatically once the transfer is accepted, and every scope read it as the last word. A release is now acted on only when it comes from the controller who currently holds the track, so a transfer stays where it was handed to. Giving a track up outright is unaffected and still clears it everywhere.

Turboprops keep climbing instead of levelling off partway up

15 August 2026 · Fixed · Flight Model

Turboprops used to run out of climb well below the altitude they can actually reach, and simply stop. An ATR 72 or a Dash 8 handed a climb to a normal cruising level would settle somewhere on the way and stay there for as long as you left it. That happened because the model worked out how much power the engine still has at altitude with a relation that falls away about twice as fast as the published one, so above roughly 10,000 feet the aircraft was being given far less power than its engines would really still be producing.

The relation is now the one aircraft performance textbooks give for a turboprop, and the difference is large: at 20,000 feet the engine keeps about 62 percent of its sea-level power rather than 43 percent. Scored against every type's own published climb table, the propeller fleet went from 45 percent of book performance to 81 percent, and the number of conditions where an aircraft could not climb at all fell from 47 to 30.

Temperature reaches these aircraft again as well. On a hot day a turboprop now loses climb thrust the way it should, about 9 percent at ISA+20; previously the two halves of the old relation cancelled out and the temperature made no difference whatsoever.

This is not the whole story. Some turboprops still stop short, because almost all of them are also flying a climb speed schedule that speeds up with altitude, which is what a jet does and not what a turboprop does. That is being worked on separately. Jets, helicopters and piston aircraft are untouched.

Most commands can now be typed as a single character

15 August 2026 · Added · ATC Commands

Twenty-one commands gained a symbol spelling, nineteen of them a single character, listed in the new Symbol column of every table in the command reference. It is a small kit rather than a set of abbreviations: / is the vertical, , is the speed, a + or - is a direction where one in space is meant, * sends the aircraft somewhere (a fix, a list of fixes, or a heading), and most symbols standing alone cancel. So *+090 /070 ,220 is the same clearance as tr 090 d 070 s 220, and the two mix freely in one line with the short forms.

Nothing you type today changes. The plain and short spellings work exactly as before, and the symbol forms are there for anyone fast enough that typing is what limits them.

Three things worth knowing before you use them. A space separates commands and a comma separates parameters inside one, so *MIQ,ROKIL is a route across two points while *MIQ, ROKIL with the space is refused outright. / carries no direction: /090 takes the aircraft to FL090 from wherever it is, where c 090 typed at an aircraft above FL090 is refused as a climb that would descend, so type the letter form where you want the server watching. And s now means two things, told apart by the number of digits: four digits is a squawk code (s1000) and one to three is a speed (s250). The one thing that costs is a leading zero, since s0250 is squawk 0250 rather than 250 knots.

Session traffic now arrives at the rate you asked for

15 August 2026 · Fixed · Spawn

Starting a replay with a rate delivered more traffic than the number said: asking the bundled Munich evening for 45 aircraft an hour produced about 63, and every bundled session ran between 1.37 and 1.73 times denser than requested. The rate is now measured against the stretch of the session that actually plays, so the number you type is the number you get. Sessions started without a rate are unchanged: they still run at their original pace. The figure session list prints has moved with it, so a rate at or near the printed baseline now leaves the session as it was authored. A session too short to have a measurable cadence declines a rate and says so, instead of applying one that means nothing.

The documentation is reorganised and goes into more detail

15 August 2026 · Changed · Documentation

Every command has its own page now, grouped by what the instruction does, with one overview listing all 42 of them and their short forms. A new page lists the 213 aircraft types the server flies and says, for each one, how much of its performance data was read off a published document rather than derived.

Where the pages only skimmed, they now go into it: what an arrival decides for itself once it is cleared for an approach, why a flap setting or a gear selection can come back reduced and which limit stopped it, and what the server fills in when you leave part of a hold out. Changes to the simulator reach these pages as they ship.

Aircraft told to climb and speed up now do both

15 August 2026 · Fixed · Flight Model

An aircraft that is given a higher speed while it is climbing now trades some of its climb rate for the acceleration and reaches the speed. Before, it held the speed it already had for as long as the climb lasted and put everything into altitude, so a speed instruction issued during a climb could go unanswered for minutes at a time.

A steady climb, where the aircraft is already at its own climb speed, is affected very little. What changed is how the aircraft divides the power it has when it is asked to gain height and speed at once.

Speed instructions below 100 knots are no longer ignored

14 August 2026 · Fixed · ATC Commands

A speed instruction below 100 knots used to be read back and then quietly dropped: the aircraft acknowledged the speed you asked for and flew 100 anyway. Nearly half the fleet was affected, because 93 of the 213 types land slower than that. Every type now carries its own minimum, which is the speed it lands at, so a Cessna 172 will fly the 70 knots you ask it for. Where a request really is too slow for the aircraft, the reply comes back with the speed it can manage and the reason attached, instead of a readback that does not match the tag. The same limit applies in the other direction for jets, which will no longer accept a speed far below their approach speed.

Aircraft on a published descent angle no longer sink below it in a headwind

14 August 2026 · Fixed · Flight Model

An aircraft following a charted descent angle, such as an LPV or LNAV/VNAV approach, worked out its rate of descent from its speed through the air rather than its speed over the ground. Into a headwind those differ, so the aircraft descended faster than the chart called for and settled below the path. It now uses the same speed over the ground that the angle itself is drawn in.

Two related limits changed with it. The vertical limits the autopilot works within while tracking a glideslope are now angles rather than fixed rates of descent, so what they allow follows the approach path instead of working against it: the old fixed figure was most restrictive exactly when an aircraft was fastest and needed the descent most, and on the steepest published approaches it was tighter than the path itself. And an aircraft following a charted angle is now held to the descent rate its own type can produce, instead of to one figure that was never measured against any of them.

Commands combine in any order, and a typo no longer half-executes a clearance

14 August 2026 · Changed · ATC Commands

You can now put any commands on one line in whatever order suits you, and every verb combines with every other. Previously the server worked through a line in one fixed order and stopped at the first thing it did not expect there, dropping that command and everything after it without saying so: c250 tr090 s280 sq4321 climbed the aircraft and lost the turn, the speed and the squawk.

Eleven commands also had to be alone on their line, among them the route, the procedure and approach clearances, and the hold. They combine now, so an arrival and the level and speed to fly it on go together: dct MIQ star LANDU2A 26L d100 s250. A direct, a route, a SID, a STAR and a transition join up into one longer plan in the order you typed them. A vector and the approach it sets up can share a line too, and the approach stays armed.

When something on a line cannot be read or cannot be flown, nothing on that line happens and the reply names the part that stopped it. Two instructions of the same kind, such as two altitudes or a turn and present heading, are refused the same way. A command that was understood but had nothing to do, such as cancelling a hold on an aircraft that is not holding, still leaves the rest of the line standing.

Server commands work with an aircraft selected, and delete uses your selection

14 August 2026 · Changed · ATC Commands

Your radar client puts the selected aircraft's callsign in front of everything you type, and the server used to read that as part of the command. So pause, resume, the session commands and spawn all failed while anything was selected, and you had to deselect, type, and click the aircraft again. They now accept a callsign in front of them and ignore it.

One command uses it instead of ignoring it: with an aircraft selected, <CALLSIGN>, delete removes that aircraft. delete <CALLSIGN> and delete ALL work as before.

One goes the other way. Deselect before you file a report: what you type after report is free text, so DLH123, report ready for descent cannot be told apart from something you meant to say to that aircraft, and filing it would bury what you wrote inside an archive. With nothing selected it works as always.

Watch the scope when you pause with an aircraft selected. The line reads as though it were addressed to that aircraft, but the freeze is still global, and the reply says so.

A callsign only counts as an addressee when an aircraft is actually flying under it, so an ordinary sentence with a comma in it cannot set off a command.

And when the callsign belongs to no aircraft, you now get told so. That covers a typo, and it covers the case where the aircraft you had selected landed or was removed a moment earlier while your client still showed it selected: previously the command simply did nothing, with no reply of any kind. It also covers an ordinary clearance with a digit wrong in the callsign, which has been disappearing silently for as long as the server has existed.

Turboprops and piston aircraft climb on their engines' shaft power

13 August 2026 · Changed · Flight Model

Propeller aircraft now work out their thrust the way a propeller actually produces it: from the shaft power the engine delivers at that altitude, divided by the speed the aircraft is flying. Until now they carried a single thrust figure that already had a speed baked into it, and then had a jet's speed correction applied on top, so the speed was counted twice.

What you will see is that most turboprops climb less eagerly above about 10,000 feet than they used to, and that light piston aircraft climb slightly better down low. The Merlin and the Jetstream shift the most. The old figures were optimistic: on 74 of 101 types the hidden speed was lower than the speed those aircraft actually climb at, so they were getting more thrust than their engines can produce.

Jets are untouched. So are helicopters, which need a different model again and keep the one they had.

The controller handbook is now a website instead of a PDF

13 August 2026 · Added · Documentation

Everything a controller needs is now at https://docs.aurora.schwingenschloegl.online: client setup, the full command reference, what the aircraft do on their own while you watch them, and the running list of what has changed. Link it once; it stays current.

It replaces the PDF that was rendered and sent around by hand. A copy somebody received in July described a speed model that had been replaced since, and there was no way for them to know that or to get a newer one. The site is rebuilt and republished whenever anything on it changes, so there is only ever one version of it.

The site is in English. The German handbook is retired rather than translated: a translation that falls behind is worse than none, because you cannot tell from the page which of the two is out of date.

Aircraft fly their approach at a landing weight instead of a take-off weight

12 August 2026 · Fixed · Flight Model

Approach speeds were worked out from the weight an aircraft departs with, which assumed it still had the whole flight's fuel on board when it landed. It now flies the approach at the weight it actually arrives at, so approach speeds are lower and the difference grows with the length of the flight.

Long-haul types change most: a Boeing 747-400 that used to fly final at 156 knots now flies it at 138, and the heavy fleet moves 9 to 18 knots. Short-haul types move a few knots — an A320 goes from 136 to 132. Turboprops barely move at all.

Expect slightly tighter spacing on final behind heavy arrivals than you are used to, and aircraft that configure a little later, because a lighter aeroplane needs less flap at the same speed.

Every command now has a short form, and ten duplicate spellings are gone

12 August 2026 · Changed · ATC Commands

Every command can now be typed short: the verb shortened to a prefix or its initials, and the space before a number dropped. c070 is c 070, s250- is smax 250, ca cancels the approach, p is present heading, tlb30 turns thirty degrees left. Both spellings work and can be mixed in the same line, so nothing has to be relearned — the short column is an offer, not a replacement. The full table is in the README and in the controller handbook.

Ten places where two or three spellings meant the same thing now have exactly one. th is gone (a bare t already means the shortest way), as are fph, cancel ils (it is cancel app, which is what it does — it cancels a localizer clearance too), hold cancel, the fl prefix in c fl100, the singular flap, flaps clean and flaps up (use flaps 0), stow, the sb and spdbrk abbreviations (the verb is speedbrake), and the spelled-out left / right in a hold, which is now l / r as it is everywhere else. Typing a retired spelling gets a refusal rather than silence.

One thing to know when combining commands on one line: the order is not free, and a command in the wrong place is dropped together with everything after it. Put them in the order the reference table lists them, or type one clearance per line. That limitation is older than this change and is being worked on.

Jet thrust now follows a published model for altitude, speed and heat

9 August 2026 · Changed · Flight Model

How much thrust a jet has left at altitude is now taken from a published engine-design model instead of a formula the server had carried without a source. Two things change that you can watch from the scope. Aircraft accelerate more gently in the cruise, where they used to gain speed briskly enough to look impatient. And warm weather now behaves the way a real engine does: a jet holds its full thrust up to a temperature the engine is rated to, and loses thrust above it. The server did model some heat effect before, but as a gradual loss starting from the first degree, which is not how these engines are built.

The model was checked closely against the two Boeing 737 and Airbus A320 engine families, where the manufacturer publishes thrust at cruise, and it lands within a few percent of them. It was checked more loosely against twenty further engines whose published figures do not translate cleanly, because the server carries one thrust rating per aircraft type while the same physical engine ships at several ratings. Propeller, helicopter and one early business-jet type are unchanged; their own model comes in a later change.

Aircraft no longer lose their speed climbing to a high level

7 August 2026 · Fixed · Flight Model

An aircraft given a climb to a level it could not comfortably reach used to hold the climb rate and give up its airspeed instead, without limit. It kept climbing, kept slowing, and ended up crawling along far below the speed at which it could fly at all — a 747 sent to FL350 settled at under 20 knots and stayed there, still indicating a climb.

Aircraft now do what a real one does: the climb rate gives way, not the speed. A type that cannot sustain the rate simply climbs more slowly, and one that reaches its ceiling levels off there and holds a normal speed.

This affected large aircraft on long climbs most visibly, but it reached every type; light aircraft asked for a high level ended up in the same state.

Arrivals now join the glideslope on it instead of a thousand feet under it

5 August 2026 · Fixed · Flight Model

Aircraft cleared for an ILS approach used to start down as soon as the glideslope came into view, which on a long final meant beginning the descent up to a thousand feet below the beam and then flying the whole approach low — crossing the threshold about a hundred feet under the path. They now hold their platform altitude until the beam actually reaches them, and join it within a few tens of feet.

Everything keyed to height came along with it. Landing gear and the final flap setting were going out roughly two miles early, because an aeroplane flying low reaches any given height above the airport sooner than it should; the configuration sequence now happens where it belongs. Controllers watching an arrival on final should see it settle onto the approach and stay there rather than drift under it.

Arrivals now configure and lower their gear on the approach by themselves

5 August 2026 · Changed · Flight Model

Aircraft on an approach now work through their own configuration as they slow down, and lower the landing gear at about 2000 ft above the airport. Until now nothing in the simulator ever extended gear on its own, so every arrival flew its final approach clean — noticeably slippery, and slower to lose speed than the real aircraft would be.

The added drag means arrivals decelerate more readily on final, so a speed instruction late on the approach takes effect sooner than it used to.

Aircraft held above the speed at which their gear may be operated do not lower it at all. This is not confined to light types: several business jets and turboprops may not put the wheels down until well below the speed a fast final approach is flown at, so an aircraft vectored in quickly stays fast — deliberately, because it shows the situation rather than resolving it with something no crew would do.

An aircraft that arrives high on the approach is now judged on the drag it is actually using to recover, including its landing gear and speedbrake. It was previously assessed with the gear up and no speedbrake even once it had started configuring, so approaches it could still fly were refused — and once refused, they stayed refused.

Manual gear and speedbrake instructions are respected: the aircraft moves those levers when its own approach plan changes, and never re-asserts a position against the controller.

Aircraft stop refusing approaches they could still fly

3 August 2026 · Fixed · Flight Model

An aircraft above the glidepath now captures it whenever the deviation is within one dot, the tolerance a real approach is flown to, instead of within a fixed 200 ft. Further out that fixed height was far too tight: near the edge of glideslope coverage it admitted barely half a dot, so aircraft the published scale puts comfortably on the beam were turned away.

An approach is now declined only when the height above the path can no longer be lost before 1000 ft above the airport, which is the point at which a real crew goes around. Until now anything more than 200 ft high inside 6 nm was declined outright, as was anything past full deflection at any distance, so a controller could get an "unable" for an aircraft that one descent clearance would have sorted out with ten miles to spare.

How much height an aircraft can still lose is now worked out from the type itself — its drag, weight and configuration — instead of one figure applied to the whole fleet. A light single and a widebody get different answers to the same situation, which is what a controller would expect of them.

Aircraft now extend flaps at the speeds their own type calls for

3 August 2026 · Changed · Flight Model

Flap selection follows two properties of the individual aircraft: the speed at which its current configuration stops being comfortable, and the placard it has to stay under. Because both differ by type, the approach now looks different from type to type — a turboprop stays clean noticeably longer, a narrowbody configures a little earlier through the middle settings, and fourteen types gain a landing flap setting that was previously out of reach.

Approach speeds and profiles are unchanged: aircraft arrive at the same speeds, at the same points, with the same number of configuration changes.

Approaches are no longer refused for being high when they are on the glidepath

2 August 2026 · Fixed · Flight Model

Every distance the server measured was short — by up to 3 percent across central Europe, and always in the same direction. Nothing looked broken, because a distance that is slightly too small still reads like an ordinary distance.

On an approach it was not harmless. The glidepath height was computed from that short range, so an aircraft flying the correct path read as too high — about 10 feet per mile out, which is 180 feet at the far end of the glideslope coverage. Aircraft joined the glidepath from above when they were on it, and an arrival genuinely between 140 and 200 feet high inside 6 miles was refused the approach with an "unable" reply for something that could have been flown.

Aircraft further out were also given speeds slightly too low for where they actually were, and now hold about 5 knots more at the same point.

The five arithmetic errors behind this are corrected. Distances now agree with an independent reference, and the checks that guard them assert accuracy rather than merely that a number came back.

July 2026

Aircraft slow down realistically when the engines are at idle

27 July 2026 · Fixed · Flight Model

How quickly an aircraft can lose speed depends on how much its engines still push while the thrust levers are at idle. Every aircraft was treated the same, and credited with far more residual push than a real engine gives — so aircraft coasted on power they did not have, and were slower to slow down than a crew would expect.

Idle thrust is now worked out for each aircraft type, and it changes with altitude the way a real engine does. Propeller aircraft see the biggest difference: a propeller at idle brakes rather than pulls, where before an ATR 72 was credited with as much residual push as a jet.

Expect deceleration on descent and on the approach to be more willing, turboprops most of all. One effect is worth knowing about at the scope: an aircraft told to descend and slow down at once, which cannot do both fully, used to level off sharply to bleed the speed. It now holds on to more of its descent while doing so, because the airframe alone takes off more speed than it was previously credited with.

New REPORT command captures the logs behind a problem

26 July 2026 · Added · ATC Commands

Typing REPORT <what happened> on any frequency now bundles the server's diagnostic logs covering the current session into a single archive and confirms the filename back on your frequency. The operator is notified and can review it after the session, so a problem you noticed mid-session becomes something that can actually be investigated.

Report while the problem is still in front of you. The archive is cut at the moment you type the command, so it holds the minutes that led up to what you saw — a report the next day describes traffic that no longer exists. Name the aircraft in the description if there is one; the command takes plain text and nothing else.

Aircraft handoffs between controllers now work in Aurora

20 July 2026 · Fixed · Wire Protocol

Handing an aircraft off from one controller to another used to do nothing on the receiving scope — the aircraft never moved across, whatever the auto-accept setting. Handoffs now reach the receiving controller: with auto-accept on it is taken over automatically, and with it off it flashes for a single-click accept. Once accepted, track ownership transfers and every connected controller's ATC list updates to show who holds it.

One controller's faulty client can no longer freeze everyone's radar

20 July 2026 · Fixed · Wire Protocol

A single controller whose client sent a malformed, oversized message could stall the whole session: aircraft appeared frozen on every scope and commands stopped registering for 20–40 seconds at a time, recurring. The server now rejects oversized messages the moment they arrive, so one misbehaving client can no longer affect anyone else. Normal commands and messages are completely unaffected — the limit sits far above anything a controller ever sends.

Squawk ident now flashes the target on the scope

6 July 2026 · Fixed · ATC Commands

The id command was acknowledged but did nothing on the radar. Now id flashes the target on the scope (a real SSR ident) and does so promptly — it reaches the scope at once instead of lagging up to 5 seconds, and the same immediacy now applies to assigned squawk codes. Only the canonical form id is accepted (no longer ident), and the readback correctly shows "Squawking ident." instead of echoing a direct clearance the controller never gave.

Radar icons no longer snap nose-to-north in multi-controller sessions

5 July 2026 · Fixed · Wire Protocol

With two or more controllers connected, every radar target used to briefly rotate its symbol nose-to-north in unison about once every 30 seconds. The 30-second position resync was sending each aircraft once per connected controller, and the duplicate identical positions made the radar client fall back to a default north heading. The resync now sends each aircraft once, so the icons hold their correct orientation.

Direct-to a fix on the route now keeps the rest of the route

5 July 2026 · Changed · ATC Commands

Giving dct <fix> to an aircraft whose cleared route or transition already contains that fix now flies it direct to the fix and continues the rest of the route, instead of dropping everything after it. It's a shortcut along the cleared path. Directing to a fix that isn't on the route still flies a plain direct, as before.

June 2026

A330-200 and ATR 72-600 join the fleet

29 June 2026 · Added · Aircraft

The server now supports the Airbus A330-200 and the ATR 72-600. Controllers can spawn them in Aurora, and FR24-derived sessions that previously dropped these aircraft at generation time now keep them.

Four Embraer E-Jet variants join the fleet

29 June 2026 · Added · Aircraft

The server now supports the Embraer 175 (both wingtip variants) and the new-generation E190-E2 and E195-E2. Controllers can spawn them in Aurora, and FR24-derived sessions that previously dropped these aircraft at generation time now keep them.

Five Boeing 737 variants join the fleet

29 June 2026 · Added · Aircraft

The server now supports the 737 MAX 8, MAX 9, 737-900/900ER, 737-700 and 737-400. Controllers can spawn them in Aurora, and FR24-derived sessions that previously dropped these aircraft at generation time now keep them.

Randomised sessions no longer dump traffic on resume

25 June 2026 · Fixed · Spawn

Pausing the simulation now also halts spawning for randomised pool sessions, exactly as it already did for fixed replays. Previously a pause kept generating traffic in the background, which thawed into a burst on resume; now the planned cadence pauses with everything else and continues from where it left off.

Deleted aircraft clear from the radar instantly

24 June 2026 · Fixed · Wire Protocol

Deleting an aircraft, or ending a whole session, now removes its target from the radar at once. Previously the last aircraft of a delete — and a whole-session delete — could linger on the scope for up to 30 seconds before vanishing, because the removal was held behind the position-update rate limit. The removal now bypasses that limit, the same way a newly spawned aircraft already does.

End a session cleanly with SESSION END

24 June 2026 · Added · Spawn

The server now lets a controller wind a session down gracefully. session end <airport> stops spawning new arrivals but leaves everything already airborne in place, so the remaining traffic can be worked down and the session finished tidily — the clean counterpart to session delete, which still clears all the traffic at once.

Once ended, the session "drains": session status shows how many aircraft are left, and as soon as the last one has landed or left the airspace the session ends on its own. A new session can't be started while one is still draining — work the remaining traffic down (or session delete it) first.

Traffic now arrives with pilot names and persons on board

21 June 2026 · Added · Spawn

Session and random-pool arrivals now carry a pilot-in-command name and a persons-on-board count, both visible in the flight-plan dialog. The pilot name matches the aircraft's registration nationality — a German-registered jet gets a German name, a Spanish one a Spanish name — and stays the same for a given flight. Persons on board is scaled from the aircraft type's certified maximum seating down to a realistic load plus crew. Nothing for the controller to do; the flight strips just read richer.

Pause and resume the whole simulation

19 June 2026 · Added · ATC Commands

Controllers can now freeze the entire session with PAUSE and continue with RESUME on any frequency — useful for briefings or breaks. Every target stops in place (groundspeed 0, no drift) and continues exactly where it left off, with no time catch-up and no position jump. The pause is instance-wide on purpose, since separation is coupled across positions and airports.

Pre-built sessions now use realistic per-type cruise levels

19 June 2026 · Changed · Spawn

The bundled arrival sessions and the randomised pool now file each flight at a realistic cruising level for its aircraft type and stage length — short hops stay low, long-haul climbs into the high thirties — instead of a flat FL150 on every strip. Sessions with too little real traffic (under 5 arrivals per session-hour) are no longer offered, since a near-empty replay isn't worth flying.

Disconnected controllers now leave the ATC list immediately

15 June 2026 · Fixed · Wire Protocol

When a controller disconnects, they now disappear from the other controllers' ATC lists right away. Previously a disconnected controller lingered in everyone else's list for several minutes until a timeout finally cleared them — making it look like they were still online. Reconnecting under the same callsign is unaffected.

Direct-to a fix off to the side now turns on cleanly

15 June 2026 · Fixed · Flight Model

When an aircraft is told to proceed direct to a fix that lies well off its current heading, it now turns smoothly onto the direct track and rolls out on course. Previously it swung well past the track and S-curved back over a couple of minutes before settling. Aircraft already on a published route or an ILS approach were never affected.

Joining mid-session now shows the assigned values already set

14 June 2026 · Fixed · Wire Protocol

A controller who connects after a colleague has already assigned headings, altitudes, speeds, or approach clearances now sees those values in the aircraft labels straight away, instead of blank labels until each field is re-entered. The shared radar picture is complete the moment you connect, not only for values assigned after you joined.

Assigned values now show in every controller's labels

14 June 2026 · Fixed · Wire Protocol

When more than one controller works the same airport, the values one controller assigns — heading, cleared altitude, speed, and approach clearance — now appear in the other controller's aircraft labels. Previously each controller saw only the live aircraft values (real position, altitude, speed) and never the figures a colleague had entered, so the shared radar picture was incomplete. The server now mirrors these label annotations between connected controllers automatically; no change to how you enter them.

Session traffic shows up reliably with strips and pilot callups

14 June 2026 · Fixed · Wire Protocol

When starting a traffic session, arriving aircraft could fail to appear correctly: only the first aircraft would show up, its strip and radar target would then freeze and drop off, and in some cases the whole traffic feed would stall with nothing new arriving. The simulated pilot sector-entry callups (for example "FL110, inbound LANDU") were also missing from your frequency.

This is fixed. Session traffic now appears reliably — every arrival shows up with its full flight strip and route, holds a smooth radar target, and calls up on your frequency as it crosses into the sector. New arrivals also populate their strip a moment faster. No action needed; just start a session as usual.

FR24 sessions now enter at their gates like the random pool

8 June 2026 · Changed · Spawn

Fixed FR24 replay sessions now spawn arrivals exactly like the randomised pool: each aircraft appears at its STAR entry-fix gate at the published handover altitude and speed, heading inbound, then holds heading until you vector it. One per-airport profile now defines where and how traffic enters the sector for both modes. Under time-compression (a higher aircraft/hour than the session baseline) arrivals sharing a gate are pushed apart so they no longer spawn on top of each other. And each arrival's first call now comes from the aircraft itself, naming its entry fix — "DLH123, FL120, inbound BETOS" — instead of a SERVER line reporting a speed.

Aircraft readbacks now show the aircraft callsign instead of SERVER

8 June 2026 · Fixed · Wire Protocol

When an aircraft reads back or acknowledges a controller's command on a frequency, the message in the radar client's chat window is now attributed to the aircraft's callsign instead of appearing as a message from SERVER. System messages — spawn confirmations, command rejections, session status — continue to show as SERVER, as before. Controllers no longer have to guess which aircraft a frequency readback came from.

Sessions can now replay at their original cadence

7 June 2026 · Changed · Spawn

Starting a curated FR24 session no longer requires an aircraft-per-hour figure. Leave it off — session start eddm <scenario> — and the session plays back exactly as it was recorded: the real landing sequence at the real timing, neither sped up, slowed down, nor thinned out. Supplying an aircraft-per-hour still scales the session as before (higher than the session's baseline compresses the timeline, lower thins the traffic). Fixed sessions also no longer apply a small random wobble to each arrival's timing — the recorded timing is already realistic, so it now plays through untouched.

Sessions now stop cleanly when all controllers leave

7 June 2026 · Fixed · Spawn

A running traffic session is now fully torn down when every controller disconnects and the idle-clean grace period elapses — pending spawns are cancelled along with the existing aircraft. Previously the server kept spawning new session traffic into the empty sim after the cleanup, leaving unattended aircraft that nothing would remove.

Aircraft registrations now file without the hyphen

6 June 2026 · Fixed · Wire Protocol

Aircraft registrations now appear in the flight plan's registration field without the nationality-separator hyphen, matching the ICAO format — a registration that read D-AIZP now files as DAIZP. This applies both to session traffic and to the SPAWN command's REG override; if you type a hyphen, the server removes it automatically. The bundled EDDM session was updated to match, so its strips read the same way.

Continuous randomised arrival traffic with selectable rate profiles

6 June 2026 · Added · Spawn

The server can now run an endless, randomised arrival flow: SESSION START EDDM RANDOM 26 60 injects traffic inbound at the four Munich entry gates (ROKIL, LANDU, BETOS, NAPSA) at the requested rate, drawn from a pool of real flights. Unlike the fixed scenarios, raising the rate raises the flow cleanly — a per-gate in-trail separation keeps aircraft from piling up. Each arrival appears on its real approach line at the published handover altitude and speed, then holds heading until the controller vectors it from the gate.

An optional trailing mode shapes how the traffic is spread over time: linear (steady, the default), poisson (realistic random bunching), rush (15-minute busy/quiet waves), bank (a dense 5-minute hub push then a quiet stretch), and ramp (builds up to the requested rate over 15 minutes). Start with the runway direction (26 / 08) and aircraft-per-hour; SESSION DELETE / SESSION STATUS work unchanged.

Radar targets no longer teleport across the map

5 June 2026 · Fixed · Flight Model

Radar targets no longer occasionally jump thousands of nautical miles and snap back — the long-standing "phantom jump" is gone. The cause was a numerical-precision error in how the server projected a gently-turning aircraft's position for its periodic full-state update: a very small turn rate produced an enormous turn radius where the spherical geometry lost precision and threw the target far off the map. The projection now uses a numerically stable method that holds a clean track at any turn rate. The simulated traffic itself was always correct — only the occasional broadcast copy was affected.

Radar targets no longer teleport across the map

1 June 2026 · Fixed · Wire Protocol

Aircraft targets in Aurora occasionally jumped thousands of nautical miles away and snapped back a few seconds later. The cause was a message-decoding bug that corrupted the periodic full-state refresh the server sends; on a busy session it surfaced as targets flicking across the screen. The decoder now reads only the bytes it actually received, so the full refresh stays intact, and a wire-output guard drops any impossible position jump before it can reach the radar client.

Controller altitude and speed clearances now override procedure limits

1 June 2026 · Changed · Flight Model

An explicit descend, climb, or speed clearance is now authoritative: it overrides a published or session-route altitude/speed restriction instead of being clamped to it. Session-spawned arrivals that previously appeared to ignore a descend clearance — holding their level because a route restriction silently won — now obey the controller. With no clearance set, aircraft still follow their procedure limits by default. The published glideslope/approach path and the 250-knot-below-FL100 rule still apply.

Server stays responsive under heavy traffic

1 June 2026 · Fixed · Performance

Under busy sessions — many aircraft with several controllers transmitting at once — the server used to force a full memory cleanup on every network message, which could consume a whole CPU core and delay position updates and ATC commands, so aircraft appeared to freeze or ignore instructions. That forced cleanup is gone; the server now stays responsive as traffic climbs.

May 2026

High-speed below FL100 now needs an explicit clearance

31 May 2026 · Changed · ATC Commands

Below FL100 aircraft are now held to 250 KIAS by default, matching the ICAO rule. An aircraft arriving fast levels off at FL100, slows to 250, then continues down. To keep one fast through FL100, give it an explicit speed clearance above 250 (s 300 or smin 280) — that counts as a high-speed approval. A speed set when an aircraft is spawned does not waive the limit, so spawned traffic settles to 250 below FL100 on its own.

Smoother, more realistic altitude level-offs

31 May 2026 · Changed · Flight Model

Aircraft now capture their cleared altitude the way a real autopilot does — the vertical speed eases off smoothly and steadily, and the aircraft settles exactly on the target. Previously the last ~100 ft of a climb or descent crawled at a near-stationary rate, so the radar tag could sit at a stuck "-01" for a minute or more before the level finally seated. Level-offs now finish in a dozen-odd seconds with the vertical speed visibly winding down to zero, matching how a real altitude capture looks on the tag.

Climb and descend now respect their direction

31 May 2026 · Fixed · ATC Commands

The server now refuses a climb command to a level below the aircraft and a descend command to a level above it. Previously either verb simply moved the aircraft toward the typed altitude, so a descend to a higher level actually climbed. The aircraft now holds its level and the controller gets a short rejection note instead. A climb or descend to the current level still means "maintain".

Radar targets no longer flick backward on Aurora

30 May 2026 · Fixed · Wire Protocol

Radar targets in Aurora no longer jump backward along their track for a single update before snapping forward again. When the server caught a target's position up between updates it could occasionally project it the wrong way; it now only ever projects forward. Controllers see smooth, monotonic tracks with no rewind.

New clr trans command flies approach transitions to the vectoring handoff

27 May 2026 · Added · ATC Commands

Controllers can now clear an aircraft onto an approach transition with clr trans <runway> via <fix> — e.g. clr trans 26R via ROKIL. The aircraft flies the published feeder route from the chosen entry fix (the start of the transition, or a mid-transition fix to enter partway) to the point where the procedure hands off to radar vectoring, then holds present heading awaiting your vectors and a follow-up clr ils. Distinct from clr ils, which captures the localizer geometrically from the current position.

Session inbounds spawn on an approach profile and fly their route

25 May 2026 · Changed · Spawn

Session traffic now spawns ~40 nmi out on a single approach profile (FL120 for jets, 6000 ft for light GA, type-appropriate speed) and tracks its filed route fix-to-fix toward the runway, turning onto each waypoint like a real inbound before holding for the controller's vectors. Previously every session aircraft flew straight ahead on its spawn heading regardless of its route. The SESSION START command drops its tier parameter (now SESSION START <ICAO> <scenario> <aph>), spawn order follows the real landing sequence instead of a random shuffle, and SESSION STOP is now SESSION DELETE.

Session routes now IFPS-validated via autorouter

24 May 2026 · Changed · Spawn

The session synth now generates routes via autorouter.aero instead of Flight Plan Database. Routes come back IFPS-validated — Eurocontrol would actually accept them if filed — and per-runway SID/STAR constraints feed each flight's real FR24 runway through as a routing hint. An EDDM 26L arrival from Italy now gets the STAR that actually serves 26L from the south instead of a geometric fallback. The bundled EDDM West-operation session has been regenerated under the new pipeline; all 65 spawns carry real, filable routes with correct SID/airway/STAR sequences for their respective origins and arrival runways.

The Munich evening session replays a real afternoon of arrivals

18 May 2026 · Changed · Spawn

The Munich evening session has been rebuilt from a real afternoon of arrivals at EDDM. Callsigns, aircraft types, origins, scheduling and routes mirror what controllers saw on that actual day, so the playback feels like a slice of the real operation rather than a generic scenario. Each arrival now flies its real filed flight plan on the strip — SID name at the head, en-route waypoints with airway labels between them, STAR name at the tail — and follows that route from spawn onward until the controller takes over with clr star or vectors. Spawn positions sit on the drawn route line at a realistic distance from the field, with the heading aimed at the next waypoint, so the inbound looks like an aircraft already established on the inbound track rather than a generic blob arriving from a placeholder direction.

Flight strips no longer show the destination twice

17 May 2026 · Fixed · Wire Protocol

The Aurora flight strip no longer shows the destination airport twice for session-spawned arrivals — previously the route column duplicated the destination ICAO (e.g. DCT ROKIL ROKIL2A EDDM with EDDM also in the destination column). Routes now render as DCT <entry-fix> <STAR-name> only, with the destination in its own column.

The bundled eddm/evening session has been removed; session list eddm currently returns no scenarios. FR24-synthesised replacement sessions land in the next release.

Flight strips appear without a manual lookup

17 May 2026 · Fixed · Wire Protocol

Aurora's flight-strip list now populates automatically — the server pushes flight plans to every connected controller the moment an aircraft is spawned, and brings a freshly-connected controller up to speed with the full fleet inventory the moment they log in. No more waiting on the operator to manually query each callsign before strips appear.