The new (and improved??????) SAS system.

[DChurchill]

After playing around with this for a couple hours, I think design is certainly a large role. But I'm not sure there's not a bug in here somewhere. At one point this rocket would not leave the pad. I thought it was maybe something I was doing, so I went to .20.2 And tried it. 1.71 TWR. Lifted off no problem. I went back into .21 and it was OK then, too. And this is basically the same rocket that I was trying to control before with no luck at all.

Now it is controllable, but the nose slowly wanders in the direction of down (after I start the gravity turn) and I can see RCS trying to push the nose back up. And then it sometimes wanders in the up direction.

I dunno what to think now.

Edited July 25, 2013 by DChurchill
Wrong screenshot

[Sapphire]

Ok I've changed my mind. The large ASAS module works great. Love it. My vote was based on that alone and I didn't even look at the The reaction wheels are total crap. The inline advanced stabilizer almost ignores my input. With advanced inline stabilizer, wings and RCS installed, I could NOT keep it from following it's own course. Mashing 'S' to try to bring the nose up did next to nothing.

To everyone saying the new SAS is great, have you tried the smaller reaction wheels or just the large ASAS?

I think the new ASAS system is terrible (for rockets, at least. I like it on planes), and I haven't tried the big 2.5m one yet. I have an OCTO probe core, a 1.25m reaction wheel, and a 1.25m inline stabilizer, all attached to a symmetrical probe, being pushed in space by an LV-T30, and the thing is slewing all over the place on my burns. I have to constantly shut off the burn, aim the thing back at the maneuver node, then burn again, then rinse and repeat. I'll give the big 2.5m one a try.

[Xeldrak]

Well, I realy cannot see your problem. I've build this thing here to test the new SAS. Note the single radialy mounted engine, this thing should drift off, but as long as it has electricity this thing will stay on target. If I disable all reaction wheels and enable RCS it will stay on target using RCS-fuel.

[Quantum Jim]

I really like the new ASAS! Finally, I can get a ship to orbit relatively accurately without MechJeb (yes, I can do it by hand, but I'm inaccurate). And it doesn't use nearly as much mono-propellent as the old ASAS.

Edited July 25, 2013 by Quantum Jim

[DChurchill]

Redacted.

Edited July 25, 2013 by DChurchill
irrelevant

[TheHengeProphet]

After playing around with this for a couple hours, I think design is certainly a large role. But I'm not sure there's not a bug in here somewhere. At one point this rocket would not leave the pad. I thought it was maybe something I was doing, so I went to .20.2 And tried it. 1.71 TWR. Lifted off no problem. I went back into .21 and it was OK then, too. And this is basically the same rocket that I was trying to control before with no luck at all.

Now it is controllable, but the nose slowly wanders in the direction of down (after I start the gravity turn) and I can see RCS trying to push the nose back up. And then it sometimes wanders in the up direction.

I dunno what to think now.

*snip*

This.

I made a few simple rockets to test SAS functionality, and it simply lacks the gall to do a heading lock. Rockets and planes alike drift off course, and the SAS does minimal correction (enough to prevent it from flipping out, provided you don't touch the controls).

Another issue is that any time you press a control (WASDQE), it is effectively like turning the SAS off. The old aeronautics package worked better than this, which is what I always used to use on my rockets and planes, because the rigidity of the old ASAS got in the way, but now I think it all has gone soft...

I want a middleground vote, but without one, I vote it does not work as it should.

[DChurchill]

After all my sturm-und-drang, I'm come to the conclusion of: meh. I like the fact that it doesn't eat RCS like it's candy, but I hate the fact that it wanders, or at least seems to be hard to fine tune without a ton of fiddling. (On the smaller diameter. I didn't seem to have nearly the problem with the large diameter stuff.)

Honestly, I think hand controlling a rocket at launch and for sustained burns is unrealistic and NOT fun but tedious, and that's what the new SAS is going to require. Pass. I'm gonna wait until MechJeb is fixed for .21. It's a godsend for large stations, though.

I've stayed up far too long as it is. Night, all.

[Naelo]

Hmm, I feel like "it sucks" is a horrible thing to call it, considering C7 put a lot of work into this.

I don't like it very much either, but it does work.. sorta.

Also, isn't this the 5th million thread about it? Can we please only have one?

EDIT: Maybe we just need to rethink the way we design ships with SAS?

It was kinda a "plug and play" solution before, maybe if you put a bit of work into placing SAS modules, it works better than the old system.

I haven't tried yet, just a thought.

[Danger Will Kerbinson!]

My take is that it works great at the controlling wobble with docked craft, saving rcs and the freaking out engine gimbals on ascent aspect. My gravity turns are much smoother now but they are squishy, they drift along axes that I have not made any input to and like to rebound back the way I just came if I don't babysit constantly.

The whole thing feels spongy, it won't hold a heading even in a vacuum for any length of time under thrust and just trying to get it perfectly aligned to a node mark to begin with is tedious.

The basic operation style is a great start but there needs to be a button press to enable a course lock sort of mode, the drift just sucks. Count me in on the waiting for MechJeb to be updated before I bother playing more crowd. Smart A.S.S. will be a godsend for the tedium of constant course correction on even short burns.

Edit: Pre-emptive response. Yes I have power, yes I have a pod with SAS and have tried loading up extra reaction wheels, no it still does not hold a heading for crap, even on a simple, small, balanced ship. Clean install, no mods, pilot is not a dumbass nor intoxicated.

Edited July 25, 2013 by Danger Will Kerbinson!

[Xeldrak]

After flying my first mün-mission in .21 I'd say we will simply have to adapt to the new system. My new rockets will have more RCS - it's realy usable now and not just for docking.

I'd say - try flying with the new ASAS for a week. Mayme some minds will change...

[DChurchill]

My take is that it works great at the controlling wobble with docked craft, saving rcs and the freaking out engine gimbals on ascent aspect. My gravity turns are much smoother now but they are squishy, they drift along axes that I have not made any input to and like to rebound back the way I just came if I don't babysit constantly.

The whole thing feels spongy, it won't hold a heading even in a vacuum for any length of time under thrust and just trying to get it perfectly aligned to a node mark to begin with is tedious.

The basic operation style is a great start but there needs to be a button press to enable a course lock sort of mode, the drift just sucks. Count me in on the waiting for MechJeb to be updated before I bother playing more crowd. Smart A.S.S. will be a godsend for the tedium of constant course correction on even short burns.

Edit: Pre-emptive response. Yes I have power, yes I have a pod with SAS and have tried loading up extra reaction wheels, no it still does not hold a heading for crap, even on a simple, small, balanced ship. Clean install, no mods, pilot is not a dumbass nor intoxicated.

Glad I'm not alone. Spongy is a great word. I just don't want to have to fiddle with it. I'm fine with it if that's how it ends up, because I can use MJ, but for me, it's less than ideal.

Now I really need to go to bed.

[Sof]

I haven't had any problems with SAS. First rocket I made, not even sure about what had changed with the SAS so I just built a design from older versions.

Held the heading straight all the way for all stages of the rocket. Landed on the Mun perfectly first time.

People have just gotten too used to cheating with mechjeb that they get confused when the game prompts them to actually play.

[boxman]

Judging by this poll this issue seems to only affect a minority of players...

[Danger Will Kerbinson!]

Judging by this poll this issue seems to only affect a minority of players...

Yeah, you are right... Its only a about 1 in 4 so probably it should be just dismissed and we can all move on. Thanks for bringing some much needed perspective to this issue.

[Aphox]

Judging by this poll this issue seems to only affect a minority of players...

41% is hardly a minority. I think a large part of the problem is different people with different play-styles were all expecting different behavior to emerge from the updated system and a lot of people didn't get what they were expecting, and got a system that takes time to get used to and isn't as exact as it used to be. I'm not saying there isn't a problem, but these things people are perceiving as "problems" may be intended behavior. Obviously, that's not going to stop it from being changed if there's enough evidence that it's a real problem.

52 of 127 is nowhere near 1/4.

Edited July 25, 2013 by Aphox
I'm a retard ignore my math

[stupid_chris]

52 of 127 is nowhere near 1/4.

Aphox, as much as I love you as a moderator, I'm forced to correct your math here. It's 52 out of 181, and that's nearly 1/4th :l

[Aphox]

Aphox, as much as I love you as a moderator, I'm forced to correct your math here. It's 52 out of 181, and that's nearly 1/4th :l

Dude who cares if I'm a moderator I can be a retard too. I applaud you for calling me out.

And you're right. For some reason I forgot to add the two together. It's about 28%, which I personally still wouldn't call minority. 1/4 is a pretty big chunk of a statistic.

[Temeter]

Judging by this poll this issue seems to only affect a minority of players...

Right, because 30% of all players is pretty much nobody. Furthermore, the poll is compeletely useless. It isn't a matter if the sas is perfect or complete bs. The new system is a great improvement in some cases (wobbly rockets, docking, landing) and almost pointless in other cases - just try to fly in a plane arount the world or making a 30 minute burn.

Edited July 25, 2013 by Aphox
Try to be less inflammatory. This isn't a place to start arguments.

[stupid_chris]

Dude who cares if I'm a moderator I can be a retard too. I applaud you for calling me out.

That's why I love you as a mod

[Sapphire]

If it changes anything (it will, by like 0.1%), I voted for "no problems". Then I started orbital maneuvers and long burns, and so started getting problems.

[boxman]

Right, because 30% of all players is pretty much nobody. Furthermore, the poll is compeletely useless. It isn't a matter if the sas is perfect or complete bs. The new system is a great improvement in some cases (wobbly rockets, docking, landing) and almost pointless in other cases - just try to fly in a plane arount the world or making a 30 minute burn.

-snip-

Why put words into my mouth?? I was just saying that based on this poll that this issue seems to affect a minority of players, this does not mean it should not be fixed. And I do not agree with the wording of the poll either, but this is the only poll we got right now.

And I think it is you who are behaving stupid now attacking me for just stating that it seems like it works for majority.

And BTW I had a nearly 30 minute flight earlier today which was nearly around the planet and the plane kept the heading all the way.

And what does this have to do with opinions??? I guess we who have it working fine should not be able to even respond?

The fact is that for me it works at least as good as the old ASAS judging by how it has worked so far.

I never EVER called anyone a crybaby. And I would be on this forum and bitch about this issue myself if I had experienced it.

Edited July 25, 2013 by Aphox
Snipped inflammatory material in quoted post.

[Danger Will Kerbinson!]

I was once asked by a lawyer trying to give me the Matlock moment in a deposition about my previous statment that the suspect was 6 or 7 hundred feet from me "does it surprise you to learn that this distance you have indicated is ACTUALLY an 8th of a mile?"

He ended the questioning when I asked in return "Does it surprise YOU to learn that 660 feet IS an 8th of a mile?"

Ill accept your appology that was not directed at me though.

[boxman]

41% is hardly a minority. I think a large part of the problem is different people with different play-styles were all expecting different behavior to emerge from the updated system and a lot of people didn't get what they were expecting, and got a system that takes time to get used to and isn't as exact as it used to be. I'm not saying there isn't a problem, but these things people are perceiving as "problems" may be intended behavior. Obviously, that's not going to stop it from being changed if there's enough evidence that it's a real problem.

52 of 127 is nowhere near 1/4.

From what I understood in the other thread someone recreated a rocket that could not even keep a steady heading on launch, and oddly enough for another user it worked fine.. So might even be some odd bug that makes it behave differently depending on your system.. This is why I am so curious about this problem since it only seems to affect some users while most people it seems to work more or less as intended.

[Guardian_Gryphon]

I think a large part of the problem is different people with different play-styles were all expecting different behavior to emerge from the updated system and a lot of people didn't get what they were expecting, and got a system that takes time to get used to and isn't as exact as it used to be. I'm not saying there isn't a problem, but these things people are perceiving as "problems" may be intended behavior. Obviously, that's not going to stop it from being changed if there's enough evidence that it's a real problem.

It is also possible that the pattern aligns with people who use Joysticks, and have no deadzone set. Previously this was not an issue; SAS locked you out, and any random control input spikes from joysticks got lost in all the higglety-pigglety of existing problems with the RCS wobble and whatnot. Now that the new SAS tries to be input friendly? If you have input spikes it has no way whatsoever of distinguishing those from intended manuevers, and KSP's controls set deadzones to 0 by default...

I know that while I dislike issues to do with the 'Spongyness' of the system, it is still currently usable for me; but I did a clean re-install, started a clean profile, and set 10-15% deadzones on all my flight axes.

Before? I had far more serious veering issues. Now? It's just the annoying sponginess. So it's quite possible that this is the intersection of installation-based bugs, of the system still being unrefined, of Joystick deadzones not being set, *and* of either directly SAS related bugs, or behavior that is 'working as intended' but just so generally disliked that it deserves some small tweaking.

Personally, in Squad's shoes? I'd push a patch tomorrow that sets a small default deadzone on flight axes, and;

A; Tweaks the SAS to actually lock to selected headings rather than the velocity vector

B; Tweaks the strength of compensation it uses at varying times and in varying situations, to eliminate sluggishness.

I'd be willing to stake a pretty penny that doing just those would fix all issues for 90% of users.

[Superluminaut]

If your SAS will not hold its heading.

A. Add a command pod or add what use to be known as the small ASAS.

B. Set your deadzone.