Come back old ASAS - all is forgiven!

[Pixel of Life]

I went through HarvesteR's Aeris 3A benchmark test thingy and here are the results. This was with B9 Aerospace, Kerbal Live Feed and TAC Fuel Balancer installed.

In the base test, I did notice the nose dropping tendency, however, it only occurs below 140 m/s - Any faster and the plane will just fly straight or start climbing slightly. So can't do that on full throttle. This works like it's supposed to.

Trying again, with a completely fresh install, no mods: Again, the plane starts dropping the nose below 140 m/s, and flies straight or climbs if flying faster.

Then, finally, the test with the SAS on: Level flight, SAS on - the aircraft holds its current attitude, although it pitches down just a tiny bit before the SAS can deflect the control surfaces enough to stop the pitching motion. I then pitched up to 30 degrees, and it held that attitude as well, however, it did drop the nose quite a bit before settling into place, thus forcing me to pitch the plane up to more than 30 degrees (35-40) to counter that drop.

Now, the weird stuff: First, I can report the same problem that many others have reported as well - when one controller is disconnected, the remaining 2 are not capable of keeping the plane still on their corresponding axes. I noticed this when I tried rocking the wings back and forth. The nose would slowly creep downward every time I applied roll input. The more I rolled the plane to the sides, the more it would drop the nose. The same happened with yaw inputs. Also, I too noticed the "JetEngine" and "turboFanEngine" folders in KSP's Parts folder.

I've said this before, but it seems like the SAS just isn't aggressive enough - By the time it has deflected the control surfaces enough, the plane has already overshot the desired attitude. Because of said slow response, it is incapable of countering even slightly asymmetric thrust in orbit, or keeping the nose up while maneuvering in the atmosphere. To me it looks like it is not a bug - they have just programmed the SAS to be too smooth!

Edited July 25, 2013 by CaptainKorhonen

[Torham234]

Now, the weird stuff: First, I can report the same problem that many others have reported as well - when one controller is disconnected, the remaining 2 are not capable of keeping the plane still on their corresponding axes. I noticed this when I tried rocking the wings back and forth. The nose would slowly creep downward every time I applied roll input.

The exact same thing I captured on my video, about 4 posts up.

[Captain Sierra]

There was a suggestion way earlier in this thread (around page 27 I think, it got lost with all the ranting) to have multiple bite levels.

Mild: As it is now.

Aggressive: The old all or nothing SAS.

Moderate: a blend of both, strong enough to hold even big rockets in place but not shake them to bits.

Make this toggled by another key. We don't use TAB for anything in KSP keybinds right? Set it to that or maybe to Y.

[ComradeGoat]

OK, here's an updated video of my problem for those apparently wanting to attribute a sudden 10º shift in heading to orbital precession. To eliminate that possibility, I rolled the ship 90º. I also selected a much higher orbit.

Interestingly the effect seems less pronounced in a higher orbit (no, it's really not orbital precession I'm seeing). This video is much shorter. I thrust from stationary, the ship pitches up about 5º and hods it there (briefly - it runs out of fuel at this point), and then when the engines cut out it returns to the set heading.

The effect is smaller further away from Kerbin (also, there's less lag, so probably a different physics delta?), and can almost be lived with, but it's still annoying.

[Weatherman159]

For those who want it, here is the craft file. https://docs.google.com/file/d/0B_BVD9gAcd6BTS1laDhPN0llLW8/

And this is what it is for. Look closely before asking if I am missing something.

I just pulled it to orbit. Pretty much straight up with SAS. None of your problems detected, so it must be installation/system specific? I'm running Win 8, store copy.

EDIT: To clarify: It does lean if you toggle SAS on on the pad. What I did was take off and then immediately turn SAS on, which makes it fly straight up.

Edited July 25, 2013 by Weatherman159

[FleetAdmiralJ]

The sepatrons are slightly offcenter (to get the pod clear of rocket flying at mach 2) but that SHOULD cause it to lean toward the camera because of where they are. It leans left, westward, therefore they are not causing the issue but you can download the craft file above and test it with and without.

OK, I downloaded the craft file and flew it, and I've come up with a hypothesis.

Flying it with no changes at all came up with the same problem - it would immediately start leaning. In my case, it leaned toward the south. I went back and added a pair of Launch Stability Enhancers to it and flew it again (twice, actually, to make sure). It may have leaned a couple of degrees, but nothing like it was before.

I'm wondering if what was happening before, without the launch stability enhancers, it was leaning slightly on the launch pad, and once you turned on the SAS, it kept that "lean" when you launched it.

[Feldynn]

I don't know if this has already been noted, but the KSP/Parts folder is NOT empty: it contains the JetEngine and turboFanEngine folders.

...Harv?

Above from a few pages back and below from more recently...

<snip>..

Also, I too noticed the "JetEngine" and "turboFanEngine" folders in KSP's Parts folder.

Last night with 0.21 was my first ever flight in some stock space planes so I don't have anything to compare to but I like how the new system works with them. I checked after some initial test flights / crashes and found I had a number of empty folders as well as 3 files in both JetEngine and TurboFanEngine folders. Removing all of that seemed to make a small difference with space planes for the better, with the files some of the smaller ones (Aries 3A and Ravenspear Mk1 I think) were quite twitchy but were a bit less twitchy / more manageable without them.

I've said this before, but it seems like the SAS just isn't aggressive enough - By the time it has deflected the control surfaces enough, the plane has already overshot the desired attitude. Because of said slow response, it is incapable of countering even slightly asymmetric thrust in orbit, or keeping the nose up while maneuvering in the atmosphere. To me it looks like it is not a bug - they have just programmed the SAS to be too smooth!

For me planes seemed pretty decent but I was only testing in a relatively low altitude area, I think you might be right about the new system simply not being aggressive enough at least for rockets. I did launch an old Mun rocket I copied over from the 0.18 demo, landed on the new moon, planted a flag and returned to Kerbin without too much of a problem (landing actually seemed a little easier with the new system!), but it did feel like I was having to constantly correct the heading during the flight much more so than I think perhaps I should have been.

I understand that it's not going to be the same degree of "heading lock" that the old ASAS had but it did seem that the new system was perhaps a touch too loose or vague when it came to keeping a heading, maybe it is to do with my design because I didn't have RCS on anything I launched last night (electricity wasn't a problem though, watched that like a hawk)? The new system doesn't appear to have quite the same consistency as the old one, it seems to work better on some craft than others and differently again depending on the atmosphere /vacuum as well the amount of thrust

[RepoRogue]

My copy is not a Steam copy, however my issue is also different. My rockets do drift off course significantly, however my main problem is with violent plane wobble at low altitudes.

[regex]

I understand that it's not going to be the same degree of "heading lock" that the old ASAS had but it did seem that the new system was perhaps a touch too loose or vague when it came to keeping a heading

The problem is that it doesn't even bother to try correcting back to the original heading, it drifts to a heading it wants to keep and stays there until the burn is complete, at which point it returns to the original heading. I could easily live with it if it even tried to correct back to the original heading, and I'm talking about a craft that can be thrown all over the place by torque alone through user input during a burn; there is no lack of control authority.

If the SAS isn't for heading lock on rockets, then what is it for?

[Diewoe]

No Problems for me, perfect controls and no wobble on planes. Now as far as Rockets they still get the bendy wobbles every now and again but no where NEAR as bad as before. What i believe is happening is bad designs from before are only amplifying bad effects with the new sas. Don't say " it worked perfectly fine with 0.20 because the Sas did not work perfectly fine in 0.20. You built a craft to compensate for bad sas design now go back to the drawing board and build a new craft that works perfectly with the new and much improved sas design. Not trying to be a jerk just stating the obvious.

It's like walking with one shoe that has a 3 inch lift sole because one leg was shorter than the other, then when you get that fixed you wonder why your other leg is now shorter when wearing those same shoes lol.

Edited July 25, 2013 by Diewoe

[Endeavour]

The new ASAS is not an autopilot anymore. It doesn't hold a heading like an autopilot would. Rather, it dampens rotations so you can better control. If you want something to keep you pointed at 15 degrees, you need an autopilot.

[RocketPilot573]

So, has anyone figured out a solution yet? I'm getting bored not being able to fly rockets or planes.

[Pixel of Life]

No Problems for me, perfect controls and no wobble on planes. Now as far as Rockets they still get the bendy wobbles every now and again but no where NEAR as bad as before. What i believe is happening is bad designs from before are only amplifying bad effects with the new sas. Don't say " it worked perfectly fine with 0.20 because the Sas did not work perfectly fine in 0.20. You built a craft to compensate for bad sas design now go back to the drawing board and build a new craft that works perfectly with the new and much improved sas design. Not trying to be a jerk just stating the obvious. its like walking with one shoe that has a 3 inch lift sole because one leg was shorter than the other, then when you get that fixed you wonder why your other leg is now shorter when wearing those same shoes lol.

No, you see, the SAS is doing an absolutely brilliant job at stabilizing things - but the problem here is not instability, it's the fact that the SAS is so sloppy that we cannot be certain of where our ships will be pointing when we stop giving control inputs or put the pedal to the metal when crossing a maneuver node. It's unpredictable.

[Sardonyx]

Gladly. It's here:

http://www.sarahlizzy.com/Labrys.craft

It behaves well enough at low altitude and low speed. Take it to orbit and play with it there.

1 to init.

2 toggles rockets.

3 toggles jet and intakes.

ETA: 4 toggles solar panels

Ok, I took two flights with it. I can't claim to be super experienced so hopefully you'll forgive my flying (video shortly).

What I noticed is that once in orbit it does tend to lose control at the slightest input. On the second flight, I ran out of gas and glided for a good amount of time. It wasn't on purpose but I think it shed the most light on the issue. I wanted to pull the nose level and hold it there to extend glide time for testing. I was able to pull it up with some difficulty and SAS was able to hold it. However, as soon as I input something like Roll, the SAS release the Pitch hold and the nose starts falling. When I let go of the roll, it continues to drift downward in pitch as if that's the proper level.

In orbit I had some issues when doing rolls where I would spin out even with SAS enabled.

Take a look at the video and see if that matches what you experienced.

http://youtu.be/5iLK_ksBydg

http://youtu.be/R4xkomm7Zy0

All in all I think there is an issue with the way it's doing hold. It seems like it's releasing control of all axes when an input on one occurs and doesn't resume them to their previous level when input is removed. Based on their description I feel like it's supposed to hold the other axes stable while releasing control on the channel you provide in put to.

Edited July 25, 2013 by Sardonyx

[draeath]

SAS isn't an autopilot, it's supposed to help kill rotation.

Put yourself into a free spin. Turn on SAS - does it slow (and stop, eventually) your spin? Then it works.

[4 IN 1]

After fooling around 0.21 a bit I find the change of the SAS behavior doesn't cause any problems for me on the rocket part, the SAS work as intented, spaceplane is a bit different, stock spaceplane works fairly well, but I do get a bit of pitching motion when thrusting forward, all in all I find the new SAS much more "gentle" then the old one, and I am happy with it.

SAS isn't an autopilot, it's supposed to help kill rotation.

Put yourself into a free spin. Turn on SAS - does it slow (and stop, eventually) your spin? Then it works.

TBH you have to be very careful that your craft don't spin too much that it max out your control vector because the new SAS will always fail at this rate, then again, if you managed to built yourself a rocket that you need to max out your control to keep it stable, you better off just learn how to put a rocket together from the start.

Edited July 25, 2013 by 4 IN 1

[RocketPilot573]

SAS isn't an autopilot, it's supposed to help kill rotation.

Put yourself into a free spin. Turn on SAS - does it slow (and stop, eventually) your spin? Then it works.

Again, HarvesteR has said that its supposed to hold attitude. It's not doing that.

[Pixel of Life]

SAS isn't an autopilot, it's supposed to help kill rotation.

Put yourself into a free spin. Turn on SAS - does it slow (and stop, eventually) your spin? Then it works.

So you're saying that a stabilization system magically becomes an autopilot if it stops movement more quickly? That makes no sense.

We are not asking for an autopilot. We just want a SAS that keeps the craft in whichever attitude we left it in.

[Sardonyx]

1) I can't wait to see that rocket.

2) My update was a hard install so the settings.cfg has been overwritten (I'm sure). My joystick (I do have one) has not been plugged in or configured on my stock backup KSP since updating. There is no way that's the problem.

For those who want it, here is the craft file. https://docs.google.com/file/d/0B_BVD9gAcd6BTS1laDhPN0llLW8/

And this is what it is for. Look closely before asking if I am missing something.

I didn't action group the solar panels (for if you reach orbit; and sorry).

Abort triggers parachute, fires sepatrons, and decouples the command pod (with running kill tallies, I'm into abort sequences to save kerbals now XD)

I downloaded the file and gave it a try. I experience the same issue. It seems the ship doesn't have enough control authority to correct the attitude even with manual controls. I changed the engine to one that's vectorable and it seems to be stable. Here's the craft file for that modification: https://docs.google.com/file/d/0B93L9D7Q-uYpY0lsQ3pmNEpkdDA/edit?usp=sharing

ETA: A second set of winglets at the top of the first stage also helps immensely. However, as the fuel burns away the ship becomes top heavy and really wants to point prograde. Even with the winglets I can't fight that with manual controls.

Edited July 25, 2013 by Sardonyx

[MrPopcup]

I agree! when I downloaded KSP today, I launched my saved ships immediately and I saw my rockets ( SRB's on sides with liquid engine in the middle ) are slightly going off-center ( during vertical launch ) and I was forced to keep it back on center, it's not the same ASAS, I ended up adding about 4 more ASAS to kill this random stuff, keep in mind that my rocket is perfectly balanced with no issues! It was perfect in 0.20, please fix this squad! at least make it hold it's heading like older version but without the wobble thingy!

Edited July 25, 2013 by MrPopcup

[Sapphire]

SAS isn't an autopilot, it's supposed to help kill rotation.

Put yourself into a free spin. Turn on SAS - does it slow (and stop, eventually) your spin? Then it works.

It's not supposed to be the old SAS, it's supposed to be a hybrid of ASAS and SAS.

[FleetAdmiralJ]

I downloaded the file and gave it a try. I experience the same issue. It seems the ship doesn't have enough control authority to correct the attitude even with manual controls. I change the engine to one that's vectorable and it seems to be stable. Here's the craft file for that modification: https://docs.google.com/file/d/0B93L9D7Q-uYpY0lsQ3pmNEpkdDA/edit?usp=sharing

See, when I downloaded it, I thought about that too, but putting a gimbaling engine on it really didn't help, but putting launch things to make sure it stayed straight until launch did help.

[Therion_Prime]

I started from scratch and the new ASAS works absolutely fine.

Flew to the mun and back without a problem. In fact it was the most relaxed flight ever.

I really like the new ASAS.

[Sean Mirrsen]

Might as well post this here as well:

The issue is entirely replicable with the stock Ares A3. You just need to take it to higher altitude. Here is my attempt to capture the problem. Basically it seems that every time pleyer does any input on the steering, the trim set by ASAS is reset. Watch closely at 1:35. The craft is holding the set heading, but the moment I touch the roll key the nose dives down to the prograde marker.

That's not a new problem, and definitely not an ASAS problem. It's a design problem, stemming from the limited amount of control surfaces on the craft. When you pitch the Aeris up at high altitudes, you can notice how it does not have enough lift to easily maintain that angle, and both its canards and elevons are occupied holding it up like that. So when you roll, you occupy them doing the roll - and since there's nothing else left to control it, the craft pitches down. FAR offers good solutions to this, by splitting up ailerons and elevons so that you can have pitch-control surfaces that won't fail you when you're doing roll.

[Tankbuster32]

when I design planes, I usually put the center of lift behind the center of mass to keep the plane balanced when low on fuel. Now if they aren't almost perfectly aligned the ASAS will do nothing! also if I trim the plane to keep it stable, the ASAS will just negate the trim and lower the nose rapidly! Also it won't keep my rockets pointing where i want them!