Come back old ASAS - all is forgiven!

[Mephane]

I want to remind everyone of a simple fact: if attitude can be kept manually, a ship definitely has enough attitude control systems (gimbal+torque+rcs+control surface); if the same ship fails to fly straight with ASAS on, the ASAS is at fault, not the craft's design.

So before you add more reaction weels and wings first try manually whether you can work against any possible imbalance or whatever it is making you drift off course. For example. if it pulls right on its own, yet pressing A manages to completely counteract that, it means the ship already has all it needs to fly straight.

Edited July 25, 2013 by Mephane

[ComradeGoat]

To put it simply,

New SAS (IAS) works neartly perfectly for planes and spaceplanes.

Only a lot of those of us having problems are having problems with planes, not rockets.

[23278]

It wouldn't be so bad, and I could almost deal with it, were it not for the rolling that gets induced with every input.

[ComradeGoat]

Exactly! It does sort of hold the heading, but it feels extremely lazy. Looking at the control surfaces, it's probably using less than 20% of the maximum angle of those. And it seems that it's using trim to hold the plane still, not the actual controls like the old ASAS did. That's the problem - the trim is too slow!

That's not all of it. See the video I posted some pages back where it will hold two different headings: one under power, one not. It *is* eventually applying enough torque to hold a heading, but it's the wrong heading!

[ComradeGoat]

I skipped about 30 pages because I didn't feel like reading this all the way. However, I did some testing and I can assure you that, as long as you add a few reaction wheels, the system works just fine.

Maybe if you'd read those 30 pages, you'd have seen proof that it most certainly does not work "just fine", including this video I posted where it will hold a heading until thrust is applied, then pitch up, and hold a new heading around 13° off the old one, until thrust disappears when it resumes the old heading. I know why it pitches up: the plane has a slight vertical imbalance (which increases as it burns fuel, unavoidable really), which it has more than enough turning force to counter (and indeed it does), but it won't hold a heading.

[Alien1099]

I am beginning to suspect that the store and steam releases were not identical.

I can clearly demonstrate that SAS does nothing on my install, yet I see others say it works fine.

I have the steam version, btw.

I'm using the Steam version. The SAS is working great for me as far as I can tell. I flew a space plane around the space center last night and I put a rather large rocket into orbit.

Edited July 25, 2013 by Alien1099

[purpletarget]

I am beginning to suspect that the store and steam releases were not identical.

I can clearly demonstrate that SAS does nothing on my install, yet I see others say it works fine.

I have the steam version, btw.

I have the Store Version, and it does nothing for me, so both versions are acting up.

Are the two of you using probes? The SAS won't do anything unless you have a SAS enabled part on the rocket, and Probes don't come with it by default.

You also need enough control authority in your rockets to give SAS something to work with.

If you can demonstrate that there's something amiss, then file a bug report, or posts a craft file, vid, or at least a pic.

Also I suggest going through Harvester's SAS verification post to make sure your install isn't pooched by mods, old data, bad configs etc.

[karolus10]

I'm very impressed with abilities of new SAS (There is no comparison with ASAS) but I also found it's inability to made very small corrections with SAS being always on-line, F key had still use for quick disengaging of SAS, to do some fine-tunning .

Also it's worth to test planes with reaction wheels disabled, and see how control surfaces work.

EDIT_1: I would like to had SAS computers with remote guidance units (these large, Stack probes) very badly...

Edited July 25, 2013 by karolus10

[sanity]

everyone should do a test using the stock kerbal-X,

add a ASAS module and report the result,

does the sas work or fail

if it fails in what way, describe the failure?

this should give us a baseline to determine whether its failing all the time on every install type or the SAS is just improperly calibrated for larger more massive rockets

[stavvy]

I have the store version.

I can't be certain as to whether the ASAS system is actually "broken" on some installs, but what I have I do like most of the time. I mostly build rockets, so I have very little input on planes.

While launching, I do see some movement toward the prograde indicator (if I have ASAS locked off of the prograde) without me telling it to do so, which makes me have to adjust it manually, if only slightly. While some might pull their hair out at this, I consider the alternative (large, complex rockets wobbling themselves apart from engine gimbal compensation) to be worse. Perhaps the strength of the reaction wheels' corrections need to be increased during burns/launch?

While in orbit or in transit, I have not had many issues with ASAS. I have experimented with what it does with RCS compared to what it used to do, which is eat through a 2.5m tank of monopropellant in about 2 minutes (exaggeration, I know, but it's for effect ). Mostly I've seen very few instances where I've had to move the craft over to a maneuver node, for instance. While burning, it seems to do a fairly good job at keeping my heading where it needs to be. Any small issues were solved by turning RCS on.

One thing I love about it is the ability to control the craft while the ASAS system is on. That alone makes me like this system more than the old one.

I won't point fingers at anyone having a lot of issues and say they're being unreasonable or anything. I'm sure there are issues with the new system that need to be solved for many people. That's why it's release 0.21, not 2.1. For my purposes, it's been mostly positive, with a few minor caveats. I'll keep experimenting with larger and more complex craft, and post feedback.

[purpletarget]

Maybe if you'd read those 30 pages, you'd have seen proof that it most certainly does not work "just fine", including this video I posted where it will hold a heading until thrust is applied, then pitch up, and hold a new heading around 13° off the old one, until thrust disappears when it resumes the old heading. I know why it pitches up: the plane has a slight vertical imbalance (which increases as it burns fuel, unavoidable really), which it has more than enough turning force to counter (and indeed it does), but it won't hold a heading.

Your SAS is holding it's heading fine, and returning to it as well. The nose up, like you said is from the imbalance on your thrust, no surprise there. And the SAS doesn't aggressively go hard over trying to compensate for stuff like a dog with a bone anymore, but it is correcting, and is coming back to settle on the original heading.

However, at your altitude and orbital velocity, your horizon is going to rotate away from you around 12 degrees per minute, and your video is almost that long. It looks as if your ship is returning to the same absolute attitude over the time elapsed on the video, and the horizon has dropped the expected amount. Given the inaccuracies inherent to the navball, the other degree or so could just be instrumentation/reading error.

[ComradeGoat]

If you can demonstrate that there's something amiss, then file a bug report, or posts a craft file, vid, or at least a pic.

See the post directly above yours. That is a bug, plain and simple.

[Sean Mirrsen]

I've tried the Aeris test proposed by Harvester, and I'm seeing a curious occurrence. See if anybody else can replicate it.

Take the Aeris 3A stock craft out. Thrust to full, take off from the strip, gear up. Get about 3000 meters up and fly straight and level. SAS off, standard coarse controls. This is the start of the test.

Now, try to flip the craft over. Just use the S key or pull hard on your joystick to try and throw the Aeris cockpit-over-tail. If your game is working right, you will fail - the Aeris is stable enough to just pull up hard and stabilize, and no amount of abusing the pitch controls will get it to flip over.

Level out again, and enable SAS. Now try to flip over again, but in two different ways.

First, just pull up hard (or jam S) and don't let go. This should produce the same effect as previously - i.e. the Aeris will pull up, near vertically, and rapidly gain altitude without flipping out.

Now try again, but instead of just pulling up continuously, pull up hard and let go when the craft is about 45 degrees off course. The exact value of deviation isn't entirely precise, but you have to let go when the craft is in "full upswing", severely off course and with some rotational momentum. In my experience, what happens is a total course reversal. Against its own aerodynamic capability (as established by previous tests), the Aeris will flip over and point exactly at the retrograde marker, keeping itself wedged there before slowly, ever so slowly returning to the point where you let go of it. This doesn't come up often in normal flight, since you either make minute corrections or go full pitch, but I think it hints at a problem in the ASAS logic.

I'll do a full reinstall of the game when I get home and see if it changes anything.

[Mulbin]

Some vids of some of my rockets being abused by the new system

I've tried stacking loads of extra reaction wheels to these to no avail.

PLEASE NOTE THE SHIPS ARE 100% STOCK, no mods installed, fresh install of KSP

No matter how many SAS parts I try adding to these the ASAS barely moves gimbals flaps etc at all. It's not a problem with the rocket, they fly fine manually, just the ASAS really not bothering to use any of the power available. It was supposed to scale the amount of pitch etc according to what was needed... it actually just always uses a tiny amount insufficient to do anything useful (other than make your manual controls sluggish).

V2 - this has an ASAS (whatever its now called) hidden at the base (before people flame that it isn't there). It clearly has anough torque as it holds the ship at certain points, then just goes "meh, I'll point this way for a bit"

No idea what the SAS is trying to do on this one...

Mercury Redstone. A fairly small, perfectly symmetrical rocket with a pod and ASAS... which veers it off course. Then makes no attempt to correct.

http://www.youtube.com/watch?v=TexTTdFkFxo

Mercury Atlas pod and concealed ASAS- oddly this one the ASAS holds quite well at first, onlly it insists on holding it several degrees off course (you can see my try to manually put it vertical again... and it sinks back to a it's strange leaning angle). however once I make gravity turn it goes bad - if you look at the indicators in the bottom left corner you can see it really isn't bothering to try and assist the rocket much at all... this is where the problem lies, it's just too lazy and weak.

[mknote]

Except there's no PRECISION to it at all. The fact that it lets it keep sliding on after you've let go is a PROBLEM. it is a FAULT. it prevents you from being able to point exactly where you want to point, because you have to somehow manually get on that point and then stop there, which is extremely difficult at best. Sure, it holds once you've gotten stopped, but it doesn't MATTER because it's absurdly difficult to get it to the attitude you WANT it to hold.

I know I'm replying to an older-ish post, but...

I don't see this as a problem or a fault. It's something that you can easily account for (I know because I have in the 2 hours that I've played 0.21), is intended by the devs, and really doesn't cause that much problem. Leave it as is, I say. Others have posted things that might be legitimate issues, but I don't think this is one.

[Francesco]

and check the KSP/Parts folder to see if there's anything in there (on a stock install it should be empty).

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?

[Sardonyx]

Would someone who's having trouble be willing to post a craft they're having trouble with and a description of what happens? I'd be willing to test it out and see if the same thing happens on a clean install on my system.

[Ted]

Please note that it appears that FAR is essentially nerfing the SAS to a non-functioning state. So, if using FAR, you're going to be having some SAS issues.

[CNB]

I'm just throwing fuel on the fire here but I think some people are missing the point.

I don't think it's working for one person and not working for another. This is not a "bug" on some versions etc etc. It's simply that some people have different expectations as to what the system is. In all the pre-release dev videos I've watched it acts exactly as it does in the released version, I just don't think it was made clear as to how it was going to differ from the original and not enough examples were given.

The point being in the dev videos we saw the Ravenspear flying around and locking into position when it was moved DURING ASAS; which it does. However that is flying low in a thick atmosphere with lots of control surfaces. So even the most minute of deflections applies lots and lots of torque to the vehicle. In contrast flying the A4 in the upper atmosphere using the liquid rocket motor with ASAS is virtually pointless; human control is far more effective because the ASAS can't get hold of the ship.

The biggest problem is when you are flying craft throttled in a vacuum, even partially. As MANY people have pointed out the system simply does not apply enough of the available torque to combat heading drift. This seems to me to be a symptom of the the fact the new system is "soft" to counter wobble but now not aggressive enough to tame more demanding craft that require high levels of torque to change heading.

The reason for the mixed feedback is because a lot of people simply don't come across any problems; especially if you fly balanced and symmetrical craft. In that case ASAS is basically just there for reassurance and helps hold you steady if your making a longer burn and that's fine. I have not a wink of trouble landing (for example) the Kerbal X lander on the Mun with the new system. The real shame is that complex and ambitious craft are now absolute monsters to fly.

I use a VTOL shuttle that can run interplanetary transfers, landings and ascents in a single stage (to duna etc) In 0.20 I used the ASAS on Kerbin to test the balance of the craft watching the flight controls compensate depending on ground speed, fuel distribution and throttle. I then used the information to improve the balance of the craft and eventually ended up with something that I loved and was a joy to fly. This is absolutely impossible with the new system and I think that's a real shame.

It's not impossible to make complex usable craft it's just that it is far more difficult and time consuming. They also require unjustified amounts of attention during more trivial parts of flight. Don't tell people to "be a better pilot" or "design better ships" because they're annoyed they have to adjust their heading every 5 seconds during a 10 minute burn. ESPECIALLY when the ship obviously has more than enough torque to compensate. If anyone has seen (or indeed played lol) Desert Bus by Penn and Teller then you will understand why this system is NOT good. I can see what the devs were trying to do and I like the idea but it would be a bad move to keep it as it is now and kill a lot of the enjoyment from the game for many people.

[boxman]

I think I understand this issue more now after watching some of the videos in this thread.. They should just make it so that it locks on more like the old asas when tapping F or turning SAS off and then back on. I think the main reason I didnt notice this issue was because I have a habit of stopping most of the momentum before enganging SAS.

[Nicok]

Please note that it appears that FAR is essentially nerfing the SAS to a non-functioning state. So, if using FAR, you're going to be having some SAS issues.

Is it normal to have those problem with a fresh install and no mod at all ?

[theflyingfish]

Some vids of some of my rockets being abused by the new system

-megasnip-

That V2 clearly does not have enough control authority, and is very unstable. As it goes back and forth, you can easily see the SAS in the bottom corner trying to control it, but doesn't have enough authority.

I suggest you try Harvester's Aeris 3A test. Take a video of that and tell us if SAS is broken.

[sanity]

the ASAS works alright for planes, except during landing, but it works not at all for rockets

we need a rocket ASAS not a plane ASAS

[ComradeGoat]

I don't really understand the problem, I launched a 43 tons payload and landed it safely on the mun

The vessel I'm having most trouble with weighs less than 10 tonnes on the runway.

[Broco]

Simple solution: Keep the new SAS, give us back the old one, too (including the old avionics package).

Flying spaceplanes with the new SAS is possible but it's a freaking horror.

I mean seriously guys, I built a spaceplane with the old system in it and it worked like a charm, could easily get into orbit. Now from 6000 meters up, the new SAS ****s it up because it slightly turns the plane.

If I try to correct the course it completely kills the SAS and the nose goes down. If I try to correct that it steers sideways. And so on.

Completely desastrous.

Going back to 0.20. Screw the new parts, they don't even slightly compensate for the broken SAS.