Fly a stock plane with only engine thrust

[SRB]

A while ago, I tried flying the stock Stearwing A300 jet up to about 7000 m on SAS, then turning the SAS off and trying to fly and land it using only differential engine thrust just like https://en.m.wikipedia.org/wiki/United_Airlines_Flight_232: throttle up to pitch up, throttle down to pitch down, turn only one engine off at a time to yaw/roll towards the off engine, no use of ailerons, elevators, rudder, flaps, or slats, and after SAS went off you could only deploy the gear once, and never retract it again. I could fly to some degree, but never land.

Ladies and gentlemen, that is your challenge: to land an A300 back @ KSC with only engine power.

Good Luck!

[vyznev]

Just to check, can we adjust the thrust limiters on the engines in flight instead of just toggling them on and off?

[SRB]

55 minutes ago, vyznev said:

Just to check, can we adjust the thrust limiters on the engines in flight instead of just toggling them on and off?

We can, but I never tried doing that.

[vyznev]

Well, I gave this a quick try. After all, I've been flying some pretty crappy planes lately, so how much harder could this be?

 

Spoiler

Turns out, it's pretty hard. 

The plane is indeed surprisingly controllable in flight, but especially the pitch response is very slow (and strongly linked to roll), and you're coming in pretty fast no matter what you do. So there's really no time for any major last-second adjustments. To have any kind of chance of a survivable landing, you'll have to line up for the approach way far out. And of course, the runway in stock KSP doesn't come with any kind of approach indicators, so you'll pretty much have to eyeball it.  

Ps. One detail I noticed is that the stock Stearwing A300 has some lift spoilers that are automatically linked to both the brakes and the thrust reversers via action groups. (The thrust reversers are group 6, FWIW.) For maximum realism, you might want to either remove those linkages or just set the control authority of those spoilers to zero when the simulated hydraulics failure occurs. OTOH, it probably doesn't matter too much, and in any case landing that thing with no aerodynamic pitch, yaw or roll control is still a darn impressive accomplishment.

Edited May 11, 2019 by vyznev

[SRB]

I didn’t know about the spoilers

thanks for trying the challenge, and IIRC the high landing speed and inability to make quick corrections is also what doomed UAL232.

[silks]

Have you tried quickloading this plane while flying? The wings disintegrate.

Crashed twice so far but will try again later.

[SRB]

2 hours ago, silks said:

Have you tried quickloading this plane while flying? The wings disintegrate.

Crashed twice so far but will try again later.

I never tried quickloading.

Thanks for trying!

[swjr-swis]

9 hours ago, ErinBensen said:

then turning the SAS off

Just a FYI: this is not enough to replicate the UAL0232 situation. In KSP, turning off SAS only disables attitude hold assistance, but control surfaces will still react as normal to your inputs. You will need to:

1. go into the tweakables of all control surfaces and disable pitch/roll/yaw control
2. set authority to all control surfaces to 0 to also disable action group activation (since you can't edit action groups in flight)
3. set authority to 0 on the Mk3 cockpit reaction wheels

The A300 appears to have a symmetry issue, at least on my install: I had to lower the thrust limiter on the left/port engine to 97% to nullify the tendency to bank to right/starboard. It's not a fixed value though, it depends on airspeed and altitude.

This asymmetry actually helps a bit in steering, because you can play with just that engine's limiter to either side of the equilibrium value to make the plane bank either way, and use the throttle purely for pitch and altitude hold. Like @vyznev mentions though, it's a slow way to steer the plane, and you have to account for the very slow spool down/up speed of the engine (iow the plane's attitude changes lag quite a bit to the input), so set up your approach from very far out.

A method to bank a good bit faster is to shortly toggle the thrust reverser of the engine on the side you want to bank to, then toggle the other one when you want to level out. Downsides: it's a lot less accurate and you have to account for some altitude loss due to the sudden drop in net forward thrust, so spooling up and gaining a bit of pitch up before the maneuver is highly recommended. Pretty sure this is not an option in a real life plane.

I tried flying the A300 'pre-crippled' as described above (including the limiter port-side), and it's actually not that hard to fly. Getting it properly aligned at enough distance to minimize the need for last-minute corrections is the hardest part, mostly because of the lack of good indicators/instruments in stock KSP. It reacts pretty benignly to thrust leveling.

My only other problem was I couldn't get it up to 7000m in level flight this way. So for me to enter a valid attempt I'll need the starting altitude requirement to drop to 5500m.

[vyznev]

18 minutes ago, swjr-swis said:

Just a FYI: this is not enough to replicate the UAL0232 situation. In KSP, turning off SAS only disables attitude hold assistance, but control surfaces will still react as normal to your inputs. You will need to:

1. go into the tweakables of all control surfaces and disable pitch/roll/yaw control
2. set authority to all control surfaces to 0 to also disable action group activation (since you can't edit action groups in flight)
3. set authority to 0 on the Mk3 cockpit reaction wheels

Alternatively, turning SAS off, resetting trim (alt+X) and keeping you hands off the QWEASD keys ought to do it. (Well, except for that one pair of spoilers that I mentioned earlier that's linked to the brakes and the thrust reversers via action groups.) I assume this is what the OP had in mind.

(Also, if you've used custom action groups to manually deploy any control surfaces, you'll need to reset those. On the Stearwing, pressing 1 will retract all the flaps on the back of the wings, but there's also a pair of leading-edge slats toggled by the 5 key (IIRC), and I think the abort key (= backspace) directly toggles the spoilers. But if you just don't use any of those action groups to start with, then it won't be an issue.)

Edited May 11, 2019 by vyznev

[SRB]

Yes, I meant turning off SAS, neutralizing trim, and not touching the QWSWED keys.

Yes, the altitude can be reduced to 5500 m.

Also, can you use action groups to disable some, but not all controls like https://en.m.wikipedia.org/wiki/American_Airlines_Flight_96 or https://en.m.wikipedia.org/wiki/Eastern_Air_Lines_Flight_935, using the remaining controls and engines to land?

[vyznev]

8 hours ago, ErinBensen said:

Also, can you use action groups to disable some, but not all controls like https://en.m.wikipedia.org/wiki/American_Airlines_Flight_96 or https://en.m.wikipedia.org/wiki/Eastern_Air_Lines_Flight_935, using the remaining controls and engines to land?

Sort of. You can deploy control surfaces via action groups, which locks them to their maximum extension and thus makes them no longer respond to steering inputs. So you could certainly make an action group to simulate e.g. the full rudder deflection on the AA flight 96. But as far as I know there's no way in stock KSP to lock a control surface at zero deflection, other than via the right-click menu. (You could probably do it using mods like kOS, but that kind of feels like overkill.)

Edited May 12, 2019 by vyznev

[SRB]

4 hours ago, vyznev said:

Sort of. You can deploy control surfaces via action groups, which locks them to their maximum extension and thus makes them no longer respond to steering inputs. So you could certainly make an action group to simulate e.g. the full rudder deflection on the AA flight 96. But as far as I know there's no way in stock KSP to lock a control surface at zero deflection, other than via the right-click menu. (You could probably do it using mods like kOS, but that kind of feels like overkill.)

I know that the “extend” action group command fully deploys a control surface and makes stop responding to control inputs.  But what does the “retract” command do?

[vyznev]

9 minutes ago, ErinBensen said:

I know that the “extend” action group command fully deploys a control surface and makes stop responding to control inputs.  But what does the “retract” command do?

I believe the "retract" command just stops the effect of a previous "extend" command (or the "start extended" option in the VAB/SPH right click menu), returning the control surface to its normal state where it's not constantly extended and reacts to steering inputs (unless configured not to do so). There's also a "toggle" command that just extends the control surface if it's retracted, and retracts it if it's extended.

Edited May 12, 2019 by vyznev

[vyznev]

Second attempt. This time I managed to keep the kerbals alive! 

 

[silks]

Missed the runway completely but landed in one piece. Thank you for a fun challenge @ErinBensen.

 

[vyznev]

I know it's been a while, but I finally completed this challenge (and another, more recent one at the same time). 

The trick seems to be to fly far enough from the KSC to give you plenty of time to gain control and line up the approach. It took me about half an hour to fly this plane back to KSC and land it (the video is considerably speeded up), and that was just about enough time to actually get it done relatively safely. Fast maneuvers using this control method are strongly contraindicated.  

FWIW, this was my third attempt this time. (The first one fell short and ended up landing in the water — and, a bit surprisingly, survived almost intact! — while the second one came down too hard and blew up.) I know I made at least two earlier attempts last year, though there might have been more.