Repository navigation
Releases: bochen2029-pixel/a32x-cbt
Release list
A32X 0.1.6 (pre-release)
The seventh Windows pre-release of the A32X flight simulator: FlyByWire's real A32NX flight-control computers
running at 50 Hz on a JSBSim A320 of ours, hand-flown in the browser from Dallas Fort Worth (KDFW), with the
systems trainer beside it. Also live at https://flightsim.bochen.xyz.
Unofficial. Not affiliated with Airbus, FlyByWire Simulations or Aerowinx. Not for type training.
Downloads
a32x-sim-0.1.6-win64.zip: unzip anywhere and double-clickA32X.exe(Microsoft Edge's app mode),
or openweb/index.htmlin Chrome or Edge. No installer, no network. The pilot's guide opens from
guide/index.html.a32x-sim-0.1.6-source.zip: its Corresponding Source (GPLv3, section 6).SHA256SUMS: the two archives' SHA-256.A32X-pilots-guide-0.1.6.pdf: the pilot's guide, typeset (12 pages), from the tagged commit's
docs/pilots-guide.md; the same guide ships in the zip asguide/index.html.
New since 0.1.5
- Fixed: the radio altimeters' options no longer break the page (0.1.5's known issue). The page now places
the world by the aircraft's true height and shows what the instruments read: a failed radio altimeter is an
amber RA flag on the HUD and the PFD. - VR for a Quest headset, the operator's request: a VR button at the bottom right puts the view in true
stereo, seated at the captain's eyepoint; the number keys still move the view. It needs a WebXR browser and an
https page, so try it in the Quest's own browser at https://flightsim.bochen.xyz. Untested on a headset:
none was at hand. Not yet: the Quest's controllers mapped to the stick and thrust levers, and an in-world HUD
and debrief. - The pieces of a landing experiment (off unless a scenario asks for them): an outside driver can trigger the
flare (flare=external), throughlaw_harness --pipe(lockstep JSON frames: go, flare, quit) or the page's
?lockstep=1with a nose camera (N) and frames for it. A driver-flown recording says# crew: jev, and the
gate refuses it. One finding, measured through the pipe: a fixed flare lands stable only when begun between
about 30 and 35 ft, a quarter of a second either way. - A check of the stall at Mach, against the ELAC's own alpha-max table: our A320 is inside at Mach 0.45 and
0.60, outside at 0.70 and 0.78. At cruise its wing stalls 0.4° past the alpha max the law allows a pilot.
Left open, its cause narrowed to the lift peak at Mach; no flight changed.
The gate across releases
No flight changed since 0.1.5: every digest and every scripted ILS line is the same. Recordings flown on 0.1.5
fly the same here; law_harness --rejudge, built from this release, confirms it for five.
Known limits
- Our A320 is inside 77 of its 80 targets. The three outside are held-out checks left open: the pitch deep past
the stall, and the stall at Mach 0.70 and 0.78. - The transonic pitching moment and the drag past the drag rise are estimates: no public A320 source, and both
act only beyond the A320's maximum operating Mach, where the high-speed protection holds it. - Thrust above 10,000 ft is not fitted: the only public climb model disagrees with ANP's own by 14 to 28 %.
- MAX REV's setting (0.7 of the way from idle to take-off thrust) is still provisional.
A32X.exeis unsigned: Windows SmartScreen warns on a downloaded copy (More info, Run anyway).- Tested in Chrome and Edge on Windows only.
- A fast, full back-stick pull with full flaps can still stall the aircraft, which a real A320's
protection prevents: the A32NX protection acts on an alpha lagged by 2 s (flybywiresim/aircraft
discussion #11009). Flare progressively. - No cockpit displays yet (an overlay stands in until P8) and no FMS or autopilot (P9).
The simulator is GPLv3; the trainer inside it is MIT. Every licence and notice is in the archive.
"Airbus", "A320" and "A32NX" belong to their owners.
SHA-256
cc698e7526f4248d752d82447e945364de10729b67498725028055430a5bcae3 a32x-sim-0.1.6-win64.zip
b842c8275a2d508ecec114d9d2f0fe0bd5395138594562ace89196438a6dfbd6 a32x-sim-0.1.6-source.zip
Both are built from the tagged commit, deterministically: the same commit gives the same bytes. The
guide's PDF is typeset separately (pandoc and XeLaTeX), from the same commit's docs/: SHA-256
e8a853ab7a159e2b069fe8ec9e1589ecaff7db7bd39bddbcc168c270a0733196.
A32X 0.1.5 (pre-release)
The sixth Windows pre-release of the A32X flight simulator: FlyByWire's real A32NX flight-control computers
running at 50 Hz on a JSBSim A320 of ours, hand-flown in the browser from Dallas Fort Worth (KDFW), with the
systems trainer beside it. Also live at https://flightsim.bochen.xyz.
Unofficial. Not affiliated with Airbus, FlyByWire Simulations or Aerowinx. Not for type training.
Downloads
a32x-sim-0.1.5-win64.zip: unzip anywhere and double-clickA32X.exe(Microsoft Edge's app mode),
or openweb/index.htmlin Chrome or Edge. No installer, no network. The pilot's guide opens from
guide/index.html.a32x-sim-0.1.5-source.zip: its Corresponding Source (GPLv3, section 6).SHA256SUMS: the two archives' SHA-256.A32X-pilots-guide-0.1.5.pdf: the pilot's guide, typeset (12 pages), from the tagged commit's
docs/pilots-guide.md; the same guide ships in the zip asguide/index.html.
New since 0.1.4
- The debrief's aircraft, whole and moving. In the debrief view (K) the aircraft now has its fuselage
outlined, and its gear, flaps, slats, control surfaces, spoilers, fans and reverser doors move and hide
as they did in flight: gear up shows no gear. Both were reported from screenshots. - Take-off thrust at a high airfield, fitted to Bartel and Young's take-off thrust model, which agrees
with the ANP database within 3.5 %: at 8,000 ft our static thrust read 7.6 % high and now does not.
Every A320 flight moves a little; a scripted ILS approach touches down as before. - Radio altimeters that can fail and read wrong (an instructor's option, off unless asked for): a
failed radio altimeter sends a failure warning, a biased one reads high or low, and the scripted crew
flies its flare and retard by what they read, falling back to the second radio altimeter when the first
fails. Both failed on approach: A32NX's computers go to direct law, as the aircraft's would. - A check of the pitch deep past the stall, against NASA's GTM-based TCM report: held out, and our
A320 is outside it (the moment keeps its pre-stall slope too far). Left open, with its cause written
down; no flight changed.
The gate across releases
Every A320 flight moves slightly with the thrust fit, so recordings flown on 0.1.4 replay a little
differently here; law_harness --rejudge, built from this release, says whether five still pass.
Known limits
- Known issue, found after release: in the page, the new radio-altimeter options (
fail=ra1,fail=ra2,
ra_bias=) break the view. The page places the aircraft by the radio altimeter's reading, so a failed
pair leaves the view broken and a biased one draws the world from the wrong height. Without those options
nothing changes, andlaw_harnessis unaffected. Fixed in 0.1.6. - Our A320 is inside 75 of its 76 targets; the one outside is the pitch deep past the stall (above).
- Thrust above 10,000 ft is not fitted: the only public climb model disagrees with ANP's own by 14 to 28 %.
- MAX REV's setting (0.7 of the way from idle to take-off thrust) is still provisional.
A32X.exeis unsigned: Windows SmartScreen warns on a downloaded copy (More info, Run anyway).- Tested in Chrome and Edge on Windows only.
- A fast, full back-stick pull with full flaps can still stall the aircraft, which a real A320's
protection prevents: the A32NX protection acts on an alpha lagged by 2 s (flybywiresim/aircraft
discussion #11009). Flare progressively. - No cockpit displays yet (an overlay stands in until P8) and no FMS or autopilot (P9).
The simulator is GPLv3; the trainer inside it is MIT. Every licence and notice is in the archive.
"Airbus", "A320" and "A32NX" belong to their owners.
SHA-256
90f4ff4c3cb93eaf575e6b48ce64cbe5e43577fde6bccbda10588192bbc79c60 a32x-sim-0.1.5-win64.zip
8c685f913c5ba0ab079d69305738b5f026fc0af860d55b4c69bcb187df563d91 a32x-sim-0.1.5-source.zip
Both are built from the tagged commit, deterministically: the same commit gives the same bytes. The
guide's PDF is typeset separately (pandoc and XeLaTeX), from the same commit's docs/: SHA-256
19c9e44d33d86487d1b359f21a04cffa36d3b8aaa868f1cbb9594b61c34de3d9.
A32X 0.1.4 (pre-release)
The fifth Windows pre-release of the A32X flight simulator: FlyByWire's real A32NX flight-control computers
running at 50 Hz on a JSBSim A320 of ours, hand-flown in the browser from Dallas Fort Worth (KDFW), with the
systems trainer beside it. Also live at https://flightsim.bochen.xyz.
Unofficial. Not affiliated with Airbus, FlyByWire Simulations or Aerowinx. Not for type training.
Downloads
a32x-sim-0.1.4-win64.zip: unzip anywhere and double-clickA32X.exe(Microsoft Edge's app mode),
or openweb/index.htmlin Chrome or Edge. No installer, no network. The pilot's guide opens from
guide/index.html.a32x-sim-0.1.4-source.zip: its Corresponding Source (GPLv3, section 6).SHA256SUMS: the two archives' SHA-256.A32X-pilots-guide-0.1.4.pdf: the pilot's guide, typeset (12 pages), from the tagged commit's
docs/pilots-guide.md; the same guide ships in the zip asguide/index.html.
New since 0.1.3
- Watch a saved flight again. Ctrl+O, the "open a recording" button, or dropping a saved recording on
the page flies it again through the flight core, start to stop; the debrief and its timeline open at
the end, and the page sets the replay's verdict against the one the file was saved with, line for line,
naming any band that changed. A recording from another build loads, and the page says so. Asked to save
a replay, the page hands back the file as it was opened, so a replayed flight can never be passed off
as a new one. - The drag of CONF 1+F, CONF 1 and CONF 2, fitted to the ANP database's drag-to-lift ratios: our A320
is inside all 72 of its targets. Take-offs and departures change a little (the take-off gradient moves
closer to ANP's); approaches flown in CONF FULL fly exactly as they did in 0.1.3.
The gate across releases
Approaches flown on 0.1.3 in CONF FULL replay identically on 0.1.4. Anything flown through CONF 1, 1+F or 2
moves a little; law_harness --rejudge, built from this release, says whether five recordings still pass.
Known limits
- MAX REV's setting (0.7 of the way from idle to take-off thrust) is still provisional.
A32X.exeis unsigned: Windows SmartScreen warns on a downloaded copy (More info, Run anyway).- Tested in Chrome and Edge on Windows only.
- A fast, full back-stick pull with full flaps can still stall the aircraft, which a real A320's
protection prevents: the A32NX protection acts on an alpha lagged by 2 s (flybywiresim/aircraft
discussion #11009). Flare progressively. - Unfitted so far: thrust above 10,000 ft, the stall above Mach 0.3, transonic pitching, ground effect;
high-speed protection is approximate. - No cockpit displays yet (an overlay stands in until P8) and no FMS or autopilot (P9).
The simulator is GPLv3; the trainer inside it is MIT. Every licence and notice is in the archive.
"Airbus", "A320" and "A32NX" belong to their owners.
SHA-256
1e84a05f1964454e412a6e47fa000e2aef69efe0c797ef3c1adcc1d227027247 a32x-sim-0.1.4-win64.zip
fb789eb305829cdb6524e8b36815f33695aacc6e64de9cf731c3bf6a7a46ea37 a32x-sim-0.1.4-source.zip
Both are built from the tagged commit, deterministically: the same commit gives the same bytes. The
guide's PDF is typeset separately (pandoc and XeLaTeX), from the same commit's docs/: SHA-256
a105a24159d8764bdef35fd05f6ee3b769ef2edcee8dcd89a2ef3b92adbd3429.
A32X 0.1.3 (pre-release)
The fourth Windows pre-release of the A32X flight simulator: FlyByWire's real A32NX flight-control computers
running at 50 Hz on a JSBSim A320 of ours, hand-flown in the browser from Dallas Fort Worth (KDFW), with the
systems trainer beside it. Also live at https://flightsim.bochen.xyz.
Unofficial. Not affiliated with Airbus, FlyByWire Simulations or Aerowinx. Not for type training.
Downloads
a32x-sim-0.1.3-win64.zip: unzip anywhere and double-clickA32X.exe(Microsoft Edge's app mode),
or openweb/index.htmlin Chrome or Edge. No installer, no network. The pilot's guide opens from
guide/index.html.a32x-sim-0.1.3-source.zip: its Corresponding Source (GPLv3, section 6).SHA256SUMS: the two archives' SHA-256.A32X-pilots-guide-0.1.3.pdf: the pilot's guide, typeset (12 pages), from the tagged commit's
docs/pilots-guide.md; the same guide ships in the zip asguide/index.html.
New since 0.1.2
- The gate's judge ends the approach at the stop. Before, if you stopped and then taxied or rolled on
before saving, the movement after the stop counted against the rollout band and could turn a stable
approach NOT STABLE, in the page and when replayed. Now nothing after the stop is the approach's. Fly the
gate on this release or later. - The debrief's timeline, in the manner of Tacview: a bar along the debrief's foot coloured by the
gate's bands, with the flight's moments marked. Drag the playhead, step a second with,and.(ten
with<and>), play with Space; the aircraft is drawn where it was and the camera follows; a readout
gives height, distance, speed against Vapp, sink, beams and bank; clicking a moment in the panel jumps
there. - The engines' spool-up, fitted: from flight idle to 95 % of take-off thrust in 4.05 s, inside the 5 s
certification bound (14 CFR 33.73(b)) by the fitter's margin for one-sided limits. Ground idle (4 % of the
rating, NASA's APEX measurements) and the reversers' efficiency (0.31, NASA/TM-2000-210300) are now cited.
Our A320 is inside all 69 of its targets. Take-offs accelerate a little faster; approaches touch down as
before. - The README leads with the simulator, with links to fly it, download it and read the guide.
Known limits
- MAX REV's setting (0.7 of the way from idle to take-off thrust) is still provisional.
A32X.exeis unsigned: Windows SmartScreen warns on a downloaded copy (More info, Run anyway).- Tested in Chrome and Edge on Windows only.
- A fast, full back-stick pull with full flaps can still stall the aircraft, which a real A320's
protection prevents: the A32NX protection acts on an alpha lagged by 2 s (flybywiresim/aircraft
discussion #11009). Flare progressively. - Unfitted so far: the stall above Mach 0.3 and transonic pitching; high-speed protection is approximate.
- No cockpit displays yet (an overlay stands in until P8) and no FMS or autopilot (P9).
The simulator is GPLv3; the trainer inside it is MIT. Every licence and notice is in the archive.
"Airbus", "A320" and "A32NX" belong to their owners.
SHA-256
b28254a1460a19180c3fd7b6cc7272ea9401de85e5c1a3e3f47271b4e5af4a5a a32x-sim-0.1.3-win64.zip
2ef204224bd8bc6cf72904de9f8a6b7ac4b1f02277febded08a67d35bc243299 a32x-sim-0.1.3-source.zip
Both are built from the tagged commit, deterministically: the same commit gives the same bytes. The
guide's PDF is typeset separately (pandoc and XeLaTeX), from the same commit's docs/: SHA-256
9d985166e2220fce55c4d29536c2b5975057ce9eed5b724044828f51dd18fae1.
A32X 0.1.2 (pre-release)
The third Windows pre-release of the A32X flight simulator: FlyByWire's real A32NX flight-control computers
running at 50 Hz on a JSBSim A320 of ours, hand-flown in the browser from Dallas Fort Worth (KDFW), with the
systems trainer beside it. Also live at https://flightsim.bochen.xyz.
Unofficial. Not affiliated with Airbus, FlyByWire Simulations or Aerowinx. Not for type training.
Downloads
a32x-sim-0.1.2-win64.zip: unzip anywhere and double-clickA32X.exe(Microsoft Edge's app mode),
or openweb/index.htmlin Chrome or Edge. No installer, no network. The pilot's guide opens from
guide/index.html.a32x-sim-0.1.2-source.zip: its Corresponding Source (GPLv3, section 6).SHA256SUMS: the two archives' SHA-256.A32X-pilots-guide-0.1.2.pdf: the pilot's guide, typeset (11 pages), from the tagged commit's
docs/pilots-guide.md; the same guide ships in the zip asguide/index.html.
New since 0.1.1
- Ground idle. On the ground the engines settle to ground idle, below the approach idle they hold in
the air, coming in over 5 s as they spool. A flight that starts parked on the runway now spools its
take-off up from there. - Reverse on the landing roll. The scripted crews land as Airbus's procedure has it: MAX REV with the
main gear down, reverse idle from 70 kt, levers at idle and reversers stowed by taxi speed. The A320's
landing roll from the 17C ILS shortens from 1,250 m in 35 s to 1,049 m in 30 s; touchdowns are
unchanged. By hand, Shift with PgDn (D-pad up on a gamepad) takes the levers into reverse, as in 0.1.1. - The FADEC interlock. While a reverser's doors move, either way, its engine holds idle, as the
A320's FADEC does: push the levers forward while the doors stow, about two seconds, and there is no
thrust until they are in. - Reverse on the view. REV per engine on the overlay and boxed in each N1 dial, amber while the doors
move and green when deployed; the CFM56's four reverser doors per nacelle on the model; and the debrief
logs when the reversers deploy and stow, with speed and distance. - The build in the status line. The page names the build it is; a page built from uncommitted
changes, or one that cannot name its commit, says so in amber: "not for the gate". - Fixes: each fan now turns with its own engine's N1 (they were swapped), and the cockpit's upper
display row sits lower, so the glareshield no longer hides the EWD's N1 row or the tops of the PFDs. - The pilot's guide, its pictures retaken on this build: the status lines name it, and the debrief
shows the reverse roll. - The gate across releases. Approaches flown on 0.1.1 still count for 0.1.2 if
law_harness --rejudge,
built from this release, passes them: every touchdown figure is unchanged, and only the peak load factor
in the 2 s after contact can move by 0.01 g, the reversers coming out within it.
Known limits
- The ground idle, reverse efficiency and MAX REV settings are provisional (0.03, 0.3, 0.7 of the rating)
until cited; the engines' spool-up (5.8 s to 95 % of take-off thrust) is still outside its certification
bound. A32X.exeis unsigned: Windows SmartScreen warns on a downloaded copy (More info, Run anyway).- Tested in Chrome and Edge on Windows only.
- A fast, full back-stick pull with full flaps can still stall the aircraft, which a real A320's
protection prevents: the A32NX protection acts on an alpha lagged by 2 s (flybywiresim/aircraft
discussion #11009). Flare progressively. - Unfitted so far: the stall above Mach 0.3 and transonic pitching; high-speed protection is approximate.
- No cockpit displays yet (an overlay stands in until P8) and no FMS or autopilot (P9).
The simulator is GPLv3; the trainer inside it is MIT. Every licence and notice is in the archive.
"Airbus", "A320" and "A32NX" belong to their owners.
SHA-256
390b30c6a274aebbf08e0bb8bbc350453b670d3eb80caa2fdf5e0bc4dc319703 a32x-sim-0.1.2-win64.zip
2225532dc4378a7dc2e5f5e3d6ed4b7730276afa89618dcf670c53be1a66e211 a32x-sim-0.1.2-source.zip
Both are built from the tagged commit, deterministically: the same commit gives the same bytes. The
guide's PDF is typeset separately (pandoc and XeLaTeX), from the same commit's docs/: SHA-256
34b8f86b959a1ebb012ec9d9822f6602fe7adf2779043003740894fe3bd2d319.
A32X 0.1.1 (pre-release)
The second Windows pre-release of the A32X flight simulator. Same shape as 0.1.0: FlyByWire's real A32NX
flight-control computers running at 50 Hz on a JSBSim A320 of ours, hand-flown in the browser from Dallas
Fort Worth (KDFW), with the systems trainer beside it. Also live at https://flightsim.bochen.xyz.
Unofficial. Not affiliated with Airbus, FlyByWire Simulations or Aerowinx. Not for type training.
Downloads
a32x-sim-0.1.1-win64.zip: unzip anywhere and double-clickA32X.exe(Microsoft Edge's app mode),
or openweb/index.htmlin Chrome or Edge. No installer, no network. The pilot's guide opens from
guide/index.html.a32x-sim-0.1.1-source.zip: its Corresponding Source (GPLv3, section 6).SHA256SUMS: the two archives' SHA-256.A32X-pilots-guide-0.1.1.pdf: the pilot's guide, typeset (11 pages), from the tagged commit's
docs/pilots-guide.md; the same guide ships in the zip asguide/index.html.
New since 0.1.0
- A pilot's guide, inside the zip and as a typeset PDF: running the page and its tiers, the controls
and cameras, flying the ILS to 17C, the gate's bands and its five-approach rule, the verdict and the
debrief, saving a flight and handing it in. - The A320 past its stall, and its drag, fitted to cited data: the short period in landing flaps
(MIL-F-8785C), the highest alpha a full nose-up elevator holds (Foster et al., AIAA 2005-5933), lift
past the stall (NASA's Generic Transport Model), and the drag of the landing flaps and gear (EASA's
public ANP database, DLR). It is inside 68 of its 69 targets, all but the engines' spool-up. Landings
move slightly: touchdowns up to 1.5 m shorter, sink up to 6 fpm higher, every stabilised-approach band
still met. - Recordings name the build that flew them (a
# build:line). The gate judges five hand-flown
approaches only with thelaw_harnessof that same build, andlaw_harness --rejudgechecks whether
five flown on an earlier release still pass on this one. - Reverse thrust by hand, on the ground, on provisional engine data: the thrust levers now have a
reverse range. The scripted crew does not use it yet, and ground idle comes in a later release. - The mouse: drag to look round the flight deck or swing the orbit and debrief cameras; the wheel
zooms; a double click resets.
Known limits
A32X.exeis unsigned: Windows SmartScreen warns on a downloaded copy (More info, Run anyway).- Tested in Chrome and Edge on Windows only.
- A fast, full back-stick pull with full flaps can still stall the aircraft, which a real A320's
protection prevents. Fitting the airframe to cited data narrowed the swing (its worst alpha 20.4° →
18.1°) but cannot stop it: the A32NX protection acts on an alpha lagged by 2 s, which keeps it swinging
after a fast entry. Asked upstream as flybywiresim/aircraft discussion #11009. Flare progressively. - The engines spool from idle to 95 % of take-off thrust in 5.8 s, where certification asks 5 s from
flight idle; to be fitted once flight idle and ground idle are told apart. - Unfitted so far: the stall above Mach 0.3 and transonic pitching; high-speed protection is approximate.
- No cockpit displays yet (an overlay stands in until P8) and no FMS or autopilot (P9).
The simulator is GPLv3; the trainer inside it is MIT. Every licence and notice is in the archive.
"Airbus", "A320" and "A32NX" belong to their owners.
SHA-256
8f295562f8b1b1ab59e7b9c82cd084cc668ca3ca317086460df30e92d52876d7 a32x-sim-0.1.1-win64.zip
570fdd027fb53866a969c2bd2756521b573d1d5e566151adc12d8a570ed4353a a32x-sim-0.1.1-source.zip
Both are built from the tagged commit, deterministically: the same commit gives the same bytes. The
guide's PDF is typeset separately (pandoc and XeLaTeX), from the same commit's docs/: SHA-256
1b0f6862d9e8145b205eeee50ebd9657ade60a36dafe8038f13c93ff7167702c.
A32X 0.1.0 (pre-release)
The first Windows release of the A32X flight simulator, a pre-release: an A320 whose fly-by-wire is
FlyByWire's real A32NX flight-control computers (two ELACs, three SECs, two FACs, two FCDCs), running at
50 Hz on a JSBSim airframe of ours fitted to 57 cited A320 targets, flown by hand in the browser from
Dallas Fort Worth (KDFW). It carries the systems trainer (ECAM, flows, graded drills) beside it.
Unofficial. Not affiliated with Airbus, FlyByWire Simulations or Aerowinx. Not for type training.
Downloads
a32x-sim-0.1.0-win64.zip: unzip anywhere and double-clickA32X.exe, which opens the simulator
in Microsoft Edge's app mode (Edge ships with Windows 10 and 11), or openweb/index.htmlin Chrome or
Edge. No installer, no network. A joystick or gamepad is the way to fly (press a button on it to use
it; M maps, C calibrates); the keyboard is the fallback (H lists the keys).trainer/index.htmlis the
systems trainer.a32x-sim-0.1.0-source.zip: its Corresponding Source (GPLv3, section 6): this repository at the
release's commit, the exact A32NX and JSBSim sources compiled into it, and how to rebuild every
shipped byte.SHA256SUMS: the two archives' SHA-256.
What it does now
- The A32NX computers fly the airframe in normal, alternate and direct law; the instructor's failures
take them there (?scenario=...;fail=...). The scripted test crew lands every degraded case but the
double hydraulic losses (green with blue, green with yellow). - In the clean configuration the flight-envelope protections hold the A320's bank, load-factor and
low-speed angle-of-attack limits; the exceptions are listed below. - KDFW's seven runways from public data, an ILS on each, wind and turbulence; a view drawn as a 1980s
full-flight simulator drew it, with a GPU tier (sky, haze, the Dallas country and skyline) over a
floor that runs on software rendering. - The gate's judge in the page: a hand-flown ILS is scored against stabilised-approach bands, and the
recording you save (Ctrl+S) replays natively to the same verdict.
Known limits
A32X.exeis unsigned: Windows SmartScreen warns on a downloaded copy (More info, Run anyway).- Tested in Chrome and Edge on Windows only.
- A fast, full back-stick pull with full flaps can stall the aircraft, which a real A320's
protection prevents. The cause is under work (the airframe's pitch response with full flaps meets
the ELAC's lagged alpha); normal approaches and flares stay well inside the protection. - Unfitted so far: the stall above Mach 0.3 and transonic pitching; high-speed protection is approximate.
A full push reaches −17.7° of pitch against the A320's −15°, and a released bank settles at 26.5°
rather than 33°. - No cockpit displays yet (an overlay stands in until P8) and no FMS or autopilot (P9).
The simulator is GPLv3; the trainer inside it is MIT. Every licence and notice is in the archive.
"Airbus", "A320" and "A32NX" belong to their owners.
SHA-256
5713d4faa25479bb00804785e55e4b4989c117accb01b0585169fcbb9ace5225 a32x-sim-0.1.0-win64.zip
81b1fa87c1f035e28645e5537c5b862001f288523c50ebe2601576bb2edd0d65 a32x-sim-0.1.0-source.zip
Both are built from the tagged commit, deterministically: the same commit gives the same bytes.