Come back old ASAS - all is forgiven!

[C7Studios]

Just to let everyone know, I've read through the entire thread here as soon as I got up this morning. It's been hard to post and reply with the forums being under such heavy load. Sufficient to say, we're working on finding out the source of your problems, and coming up with a solution to it.

[Gopherbashi]

I had been running into issues with the new SAS keeping on a heading last night, so decided to run some tests today. Clean install from the store, no mods, running on Windows. I've never had a joystick installed, and kept my hands off the keyboard/mouse the whole time.

Ship has Mk1 command pod, remote guidance unit (so I wouldn't have to kill Jeb), and inline stabilizer. It has no tailfins and uses the LV-T30 engine, so no gimballing. Any control (as I understand SAS) will therefore have to come from rotation wheels and RCS (when enabled). It also has two SRBs attached so it can reach orbit (one on each side of the main payload, in an east-west orientation).

I launched 10 times with SAS disabled and 10 times with SAS enabled. I took my final heading/elevation readings when the SRBs cut out, as the SAS was always able to bring me back vertical after they detached.

With SAS disabled, the rocket tended to go either west (270°, 284°, 234°, 237°) or southeast (133°, 109°, 142°, 121°, 156°, 114°). Final elevations ranged from 75-87°, and didn't seem to follow a pattern with direction. This may have something to do with the SRBs being on the east and west sides of the rocket.

With SAS enabled (but RCS disabled), the rocket tended to go west (7/10 trials - 262°, 265°, 265°, 282°, 276°, 259°, 264°). However, in two of the trials, SAS decided to take me southeast (119°, 96°). Only in one trial did I really manage to go straight up (just a tiny bit east, elevation at 89° or better). All other elevations ranged from 84-87°.

In addition, I launched 5 times with both SAS and RCS enabled. In four of the tests, I ended on a heading of 270° and 88° elevation. RCS remaining from my 100-unit tank ranged from 98.36-99.13. In the other test I went straight up, with no detectable heading, 90° elevation, and 99.51 RCS remaining.

Overall, it seems to be less of over/under-correcting and more of a westward bias by the SAS. It is physically possible to fly this thing vertically (as shown by two of the tests, and I've done this manually too). It's possible that this has to do with the aging speed of my computer - SAS might not be able to check all the required parameters before it's required to respond, and those small errors accumulate over time.

[Wayfare]

Did some more testing on quickly-built rockets and design definitely seems to be a factor. SAS doesn't make use of a craft's full control authority when it needs to (whereas the old ASAS did it all the time, which caused its own problems with wobble and such). This leads to a sluggish response and cases where the SAS simply doesn't fight hard enough to compensate for drift caused by slight instabilities in design, even if they're so minor that old ASAS handled them perfectly - or, indeed, manual flight without an excessive amount of input. On smaller, lighter craft - both airplanes and rockets - this isn't too much of an issue. But big rockets beware...

[ComradeGoat]

Just to let everyone know, I've read through the entire thread here as soon as I got up this morning. It's been hard to post and reply with the forums being under such heavy load. Sufficient to say, we're working on finding out the source of your problems, and coming up with a solution to it.

Thank you! Thank you! Thank you! <3

[Wayfare]

Just to let everyone know, I've read through the entire thread here as soon as I got up this morning. It's been hard to post and reply with the forums being under such heavy load. Sufficient to say, we're working on finding out the source of your problems, and coming up with a solution to it.

Sorry we broke your baby, C7

Let us know if you need any specific data.

[acealeam]

Just to let everyone know, I've read through the entire thread here as soon as I got up this morning. It's been hard to post and reply with the forums being under such heavy load. Sufficient to say, we're working on finding out the source of your problems, and coming up with a solution to it.

Thanks a ton! Im sure once it's fixed it will work marvelously.

[DaveSteadman]

My rockets are rolling all over the place... the new ASAS sucks.

I just landed my first space plane. Ever... the new ASAS rocks.

Does the big reaction wheel have any torque? It has no text...the new ASAS sucks.

RCS fuel usage is far more sensible... the new ASAS rocks.

When will my rocket hold a heading? When I toggle the ASAS or cancel out a move? I don't get it...the new ASAS sucks.

[Pixel of Life]

So seeing as the dev team is now working to solve the issue, I think this thread has more than enough information for them to get a rough idea of what's causing it. May I propose something - close this thread and open up a new one if and only if the devs need more information. I'm fairly certain this thread, and the whole SAS ****storm contributes to a large fraction of the server load.

[Therion_Prime]

How about making a test:

Load the Kerbal X and add a 2,5m ASAS, a 2,5m battery (both right below the capsule) and 8 OX-STAT solar Panels around the ASAS.

For me it performes like a charm.

[RepoRogue]

I expirienced the ASAS Bug too in my steam version. It is full vanilla and got updated by steam.

I tested a big rocket with about 1.05 TwR on launch. It was tillting over with enabled SAS.

Because I have two accounts (Steam and KSP Store)

I downloaded the version from KSP Store and installed it.

Tested the same rocket and it flew straigt up.

So I think everybody who updated his version needs a fresh install.

I'm on my second fresh install, and there is no change.

[Sapphire]

Just to let everyone know, I've read through the entire thread here as soon as I got up this morning. It's been hard to post and reply with the forums being under such heavy load. Sufficient to say, we're working on finding out the source of your problems, and coming up with a solution to it.

Much appreciated, C7!

[acealeam]

I'd like to add though, while it's not good for rockets, it's great for planes. Not long burn times, or anything, but it makes landing a lot cleaner.

[regex]

Just to let everyone know, I've read through the entire thread here as soon as I got up this morning. It's been hard to post and reply with the forums being under such heavy load. Sufficient to say, we're working on finding out the source of your problems, and coming up with a solution to it.

That's great to hear! Good luck and godspeed.

[check]

The 'problem' with the new SAS was definitely one of the first things I noticed. And after a LOT of experimenting, and reading through all 563 comments on this thread, I've come to this conclusion:

-SAS does a pretty good job at killing rotation if you have enough control surfaces

-SAS does NOT seem to be using the control surfaces to their full potential, requiring you to add more control surfaces than you normally would to get the same effect

-SAS kinda, sorta holds attitude but not really.. What I mean is that there IS definite drift in the attitude, especially apparent after you start a gravity turn after launch. If you have enough control surfaces (by that I mean winglets, gimbaled engines and RCS), the drift seems to lessen. But the fact that you can manually correct for the drift with less control surfaces means that SAS is NOT using the control surfaces to its full potential.

-SAS does appear to attempt to hold attitude better if you tap 'F' when you've reached your desired attitude after rotating, rather than at the point where you let go of a rotation key. But this looks to me the new SAS settles on a new attitude based on where it has to do the minimum amount of corrections to kill rotation. The more balanced your craft, the less noticeable.

Just to let everyone know, I've read through the entire thread here as soon as I got up this morning. It's been hard to post and reply with the forums being under such heavy load. Sufficient to say, we're working on finding out the source of your problems, and coming up with a solution to it.

Sweet.

Edited July 25, 2013 by check

[pokeman]

Same problems for me. The SAS seems to have trouble using enough of it's control in a timely enough fashion to stick to a heading. Even when it seems to be holding it has issues telling exactly what heading you told it to stay at, so it will stop but not where you wanted. Or even worse it will decide to just let your craft drift slowly off course despite the fact that you're hands off and the torque RCS etc are only at 10% of their potential...

Played with stock KSP on a totally fresh install, both stock craft and my own, with the same results.

IMO the times that it works seem to be when you have a fairly simple, perfectly symmetrical craft with a fair amount more control than it actually needs, then that 10% is able to do the job (sometimes )

[Fleb]

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.

I tried your rocket, and yeah the new SAS really doesn't like it . Almost as if it's oblivious to whats happening. I made a couple of changes though, and managed to just get it into orbit. One thing I noticed about the new SAS whilst flying your rocket, during the gravity turn you have to keep the rocket pointing in the direction you want, until it settles down, otherwise it just tries to snap back towards prograde. Like was suggested, maybe some kind of toggle that switches between modes would be helpful.

http://www.youtube.com/watch?v=29aXUqtbnWY

[reilem]

I am fed up with seeing people complaining. Can I see some hands for people who actually like it?

My first impressions:

"Oh cool it's built in to the ship no more extra dead weight"

"Wow i can actually make my gravity turns smoothly now instead of: HERP 20 DEGREES woblewoblwewoblewoble DERP 45 DEGREES woblewoblewoble HERP 75 DEGREES"

"My space station now no longer uses all it's RCS in a futile attempt to kill itself stabilize"

"I can make my gravity turn without my ship going on a spin like a balerina on LSD"

I don't understand what people don't like about it, yes it may be a little bugged in space planes and i'm sure the devs will explain/fix that feature in the near future. But seriously, just make sure that your ship isn't out of balance and use wing flaps on the side of your rockets during take off (because this is when your ship has the most mass and requires the most assistance if there are any problems).

I made a rocket that was slightly out of balance and these are my conclusions: The new SAS - in an attempt to prevent overturning and setting off a chain reaction of wobble - will now no longer use 100% of available reaction control. This is a good thing, if your ship is balanced it won't need 100% reaction control so this has removed all the wobble on my previous ships where as a previous fix i would have to disable gimballing on 90% of my engines so that the retarded ASAS wouldn't break the entire thing.

If your craft is out of balance you are just going to have to learn to steer it yourselves, because that is not what ASAS is designed for, neither in KSP or in real life.

[Daisy Blossom]

I suspect it has more to do with what rockets people are flying... looking at a lot of the people with problems it is larger, high part count, complex and beautiful rockets. The people having no trouble are those happy with a purely symetrical, low part craft. If it is the case that it's only effecting the ambition craft then it's really going to suck the creativity out of the game and leave us all flying dull stacks of dustbins with the aesthetics of a car crash as currently that's all the ASAS can cope with - I suspect the devs just test with the stock ships which don't even scratch the surface of what can bu built in KSP.

Nope. I had zero problem launching my giant, 200ton to LKO lifter last night that had a total part count around 800. I am guessing this is an isolated bug; the new SAS works great for me.

[SpaceGremlin]

I would hate to see the guidance on real rockets wander as much as the new ASAS.

I like the concept behind the new ASAS (it is intended to hold a heading, btw). It just needs to be fine tuned to hold a heading more precisely instead of wandering around like it currently does.

edit: I mean attitude, not heading.

[frisch]

FWIW, the new version seems much nicer to me too. It might be nice to have it pop up a feedback tuning panel, but not required.

[reilem]

I would hate to see the guidance on real rockets wander as much as the new ASAS.

I like the concept behind the new ASAS (it is intended to hold a heading, btw). It just needs to be fine tuned to hold a heading more precisely instead of wandering around like it currently does.

edit: I mean attitude, not heading.

Actually real rockets do use this system of SAS. The turning is done by a computer and the SAS corrects mistakes made by the turning. In order to simulate this in game you must adjust your heading more by yourself, which is what i think the devs wanted to achieve. ;D

[regex]

There is an issue with the SAS that a significant portion of the userbase is experiencing, and SQUAD have acknowledged the fact that it is not working as intended for that portion, and that they are looking for solutions.

http://forum.kerbalspaceprogram.com/showthread.php/41835-Come-back-old-ASAS-all-is-forgiven%21?p=540433&viewfull=1#post540433

http://forum.kerbalspaceprogram.com/showthread.php/41835-Come-back-old-ASAS-all-is-forgiven%21?p=542605&viewfull=1#post542605

P.S.: I'm not a hater. I just want that awesome SAS I saw in the pre-release videos.

[Tiron]

Actually, it's not so much 'hating' as 'it doesn't do the one thing we most need it to to do'. It's SPECTACULAR in every way except that it's merely a motion damper (and a slightly slow one), instead of an attitude hold. Based on pre-release material, it's perfectly capable of functioning as an attitude hold in a manner similar to Mechjeb's Killrot...only much, much, much better in every possible way. It's just not doing that at the moment, making trying to reach and maintain a precise heading difficult in the extreme.

I'm praying desperately it's a bug, but I just woke up.

[reilem]

There is an issue with the SAS that a significant portion of the userbase is experiencing, and SQUAD have acknowledged the fact that it is not working as intended for that portion, and that they are looking for solutions.

http://forum.kerbalspaceprogram.com/showthread.php/41835-Come-back-old-ASAS-all-is-forgiven%21?p=540433&viewfull=1#post540433

http://forum.kerbalspaceprogram.com/showthread.php/41835-Come-back-old-ASAS-all-is-forgiven%21?p=542605&viewfull=1#post542605

P.S.: I'm not a hater. I just want that awesome SAS I saw in the pre-release videos.

Oh i see thank you for that, i just hope that people realise that it will be fixed, because tbh the old system was a load of arse and really don't want it back.

[-RanZ-]

I'm glad we have another thread about this.