Come back old ASAS - all is forgiven!

[EmissionGeneration]

It's been said many times already, but I agree that the new SAS is not aggressive enough on controlling the Ship. Best solution I would enjoy seeing is SAS parts you attach if you want it more aggressive, and not just the Pod Controls, probably about 3 parts with 3 different SAS force.

I also have to say I don't enjoy the over-ride system, I'd be much happier with the "F" key to temp disable or, or "T" to just turn it off for any turns you need.

Also, yes, I'm using fresh install, no Mods, blah blah blah...

[Yuming]

I feel it works fine for both spaceplanes and rockets. It made a noticeable improvement with spaceplanes, and worked just as well holding a heading. It is a little un aggressive for large rockets, but careful piloting fixes that for me. Overall, it was an improvement. I may have different opinions on the rocket part as my "large" launcher is pretty small compared to some, but the aggressiveness is really the only thing I notice, and I'm sure I will get used to it.

[Baenki]

I right now have no access to my pc and ksp so i can not test my idea, but i saw a video that makes the engines seem to have little upforce pulling the nose of the rocketaway from the prograde marker in orbit.

My idea: has somebody tried to invert the engines in vab if use more than one.

For example: put one on normaly and rotate the other one by 180º. I think they schould cancel the forces of each other. One pulling up, one pushing down. Could somebody test that? 2 engines could make your rocket spin in this configuration. Maybe make it so, that one pulls left, the other one to the right.

The idea behind is this: http://imgur.com/1zOSbQg.png

Sorry this is drawn on my phone.

Circles = engines

Red = force of the engines

Blue = resulting force

Edited July 25, 2013 by Baenki

[Shyanmar]

@Tiron:

Yeah. I think the ideal solution would be to keep the new SAS module and reintroduce the old one as well. I don't even mind if it uses electrical charge now, but with a rocket it's simply necessary, especially with the larger variety.

[KSK]

I really dislike the way the SAS seems to treat your velocity vector on the navball as your heading, rather than the spot at which you let off the controls; that is my main gripe with it at present.

Um, at the risk of being ornery, physically, that sounds like the way it ought to be. Or at least it should be in space. The velocity vector is your heading - it's the direction and speed your ship is going in. The attitude of your ship, i.e. the direction its pointing in at any given time, may or may not the same thing at all.

Of course accurate physics may not make for a fun game. I'm also commenting without having played 0.21 yet so I don't actually know how the new system plays.

[BeefTenderloin]

43 pages and 25,989 views.

I think this says something.

[bsalis]

Well it looks like i'm late to this.. 44 pages now BeefTenderloin!

As an avid SSTO Spaceplane enthusiast, including some VTOLs, the nerfed torque and heavy reaction wheels impacted me more, but I can deal with that.

As for the ASAS (now called SAS-enabled), welp... first question is, emotions aside, is it meant to hold a heading? If it meant to hold a heading, then it does a miserable job of that. If you like that or not is a different matter. Like != Correct.

[Koren]

I have issues too launching rockets but at first I didnt even realized than SAS and ASAS was totally different things.

And later I seen that it was required to have some battery. And later... etc. I discover the game again I think it still works but I don't have what it needs to work. Someone said it needs CPU, I didn't find it I'll try again tonight.

[BlueSubstance]

My airplane is wobbeling as hell! I can't even land it with SAS on.. I miss the old System a bit.

[Mephane]

As for the ASAS (now called SAS-enabled), welp... first question is, emotions aside, is it meant to hold a heading? If it meant to hold a heading, then it does a miserable job of that. If you like that or not is a different matter. Like != Correct.

Quote from the dev blog post, emphasis mine:

One of the first changes you will notice, is that when the SAS is active. It will not fight with the main controls. Any time you press a button to pitch, yaw or roll. The controller for that axis will deactivate. The other two controllers will remain active and attempt to assist with the control of the plane.

This allows you to pilot to a new stable course, and have the sas automatically "lock" back on to that position. However, the new controller is not designed to be a full autopilot. It will just help to maintain stable control.

The new system take into account not only the rotational velocity, but the momentum of the craft. This helps solves cases where the ship is rotating slowly, but is very massive, and requires full control force for a meaningful slow down of the spin.

Also check out the video posted in the blog post. It is clearly meant to keep a craft's attitude just like the old one, but now more smoothly, and allowing player input to override each axis individually, then keeping the new attitude, when the player releases control, as the new target attitude.

The thing is, it appears to try to do exactly what is described, but fail to apply sufficient torque, using only a small fraction of what is available (thus the straightforward "moar reaction wheels" solution fails). My guess is a simple error in the calculation of the amount of torque it needs to apply, might just be a single comma in off by one place somewhere.

[Pixel of Life]

Quote from the dev blog post, emphasis mine:

Also check out the video posted in the blog post. It is clearly meant to keep a craft's attitude just like the old one, but now more smoothly, and allowing player input to override each axis individually, then keeping the new attitude, when the player releases control, as the new target attitude.

The thing is, it appears to try to do exactly what is described, but fail to apply sufficient torque, using only a small fraction of what is available (thus the straightforward "moar reaction wheels" solution fails). My guess is a simple error in the calculation of the amount of torque it needs to apply, might just be a single comma in off by one place somewhere.

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!

[Ben88]

I have absolutely no problems with the new SAS. Where pretty much all of my planes pre-0.21 were a bitch to fly unless I used ASAS to cement the heading to a spot (which didn't always work and if it did always nearly tore the plane in half), which is definitely not the way it should have to be - it's a goddamn plane, after all - now almost every plane I build handles very nicely. Even if I oversteer and get into a spin the recovery speed is amazing.

The only thing where it currently annoys me is rockets. The large ASAS supposedly works as usual, at least if one goes by its description, but even with reaction wheels added and additional decoupler stabilization my larger rockets usually wobble like hell, especially on takeoff it's extremely hard to keep it pointed upwards. Once I get to sufficient altitudes for a gravity turn the problem minimizes, but until then it's hell. I can still get 100t payloads into a very low orbit, but it's a huge hassle every time. Same goes for smaller rockets, but they're usually easier to keep under control. I think we basically need two versions of SAS - one for planes and one for rockets.

Exact same feelings here.

Great for planes, not so much for old fashioned rockets.

[robly18]

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.

It has four reaction wheels, but that's because it's asymetrical. If it was symetrical, the pod's SAS would have sufficed.

If SAS was off, that rocket would work more like a ferris wheel.... A downward going ferris wheel.

Apologies for the big picture by the way.

Edited July 25, 2013 by robly18

[Mephane]

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.

It has four reaction wheels, but that's because it's asymetrical. If it was symetrical, the pod's SAS would have sufficed.

If you had read at least part of these 30 pages, you'd have seen that there are people who claim it works fine, and people who claim otherwise. There may very well be a bug that applies only under specific conditions which none of us has any idea about. Trust me, sometimes bugs can be triggered by the most unrelated thing you would never think of, software is very complicated like that. The fact that so many experience it not to function properly (with "properly" being defined as the way it was described in the dev blog post I linked and quoted above*), it certainly is not an issue of all these people only doing something wrong with their rockets.

*Come to think about the blog post again: there was so much rejoicing at the video demonstration. People would not have cheered if what had been shown were a replacement of the ASAS "attitude lock" with a mere "soft kill rotation" functionality.

Edited July 25, 2013 by Mephane

[sanity]

yes but its only using maybe 10% of its available torque, so you should be able to have stability with 1/10th of the amount of SAS units currently equipped, and that rocket is tiny, if you were to scale that rocket up you would need to scale up the ammount of sas units you need as well, the sas unit requirement for stable launches grows at an enormous rate as your rocket size increases,

[carazvan]

Can we please have both sides of the argument start giving some more constructive feedback;

neither: "OMG it's broken" or "it's perfect" are really adding to what we need to find here

For everyone who is having issues like me try to narrow them down (for ex:

when I have my heading at 45* coming from 30* the rocket moves back to 30* <- this gives something that can actually be tested

For those saying everything is ok try to simulate the issues others are facing (same situation) and find reasonable explanations instead of going oh it just works perfectly.

(I've seen so many add more reaction wheels posts to people who were pointing out that only 1/5 to 1/4th of the available torque was actually being used by the ASAS)

From my tests so far it seems also like ASAS works better for spaceplanes in level flight and Rockets in Vertical so try to get away from those headings when doing examples.

[Khoal]

I don't really understand the problem, I launched a 43 tons payload and landed it safely on the mun (at the third time.. was going way to fast, damn nuclear rockets...):

Flew perfectly fine, the only, real, problem was that the spaceplane hangar ****ed my struts and fuel lines up, not funny to refuel rockets while moving...

Jeb and Bill died because of a lack of fuel; the screenshot though was taken afterwards... I detected some zombies o.O

Ps: Sorry for the large image...

[dlmarti]

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.

[sanity]

I Have the store version,

i think mine is working, to some degree, anyway, if i launch the kerbal X (baby's First Rocket) it will be perfectly stable with the new SAS system, but if i try to launch any more advanced or larger rockets, the SAS does not use more than about 5% of its capability, even when the rocket drifts massively off course it just continues to use that 5% to try and fight to stay aligned with the point where i engaged it, even it my rocket is 90 degrees of the engaged course it will still try to correct with only 5% of the maximum available torque, the obvious result being that my rocket falls off its course if i don't constantly input control to keep them on course and babysit them all the way up, so ultimately the SAS has no effect, i might as well just fly the whole flight manually since i have to do that anyway now that the sas wont use the control surfaces, gimballing or torque that it has to full effect

it is however very useful for planes, increasing stability and allowing easier keyboard flight in most cases, but absolutely useless for any rockets except the most basic training whel equiped rockets

[EmissionGeneration]

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.

[BlueSubstance]

The damn new Stabilization Wheel thing didn't help ANYTHING. I had to fly my Large Rocket MANUALLY. Squad, the new function is a nice idea, but it doesn't work in any way.

[Mulbin]

No.. I have the store version, ASAS is useless.

Uploading some vids at the moment to give more useful feedback.

[Chrisd857]

(Store version)

So far (after playing only a few hours) I happen to like the new system. One, I actually know what each thing does, with specified control computers and reaction wheels (and a combined one), instead of an all-powerful ASAS and a seemingly useless SAS. Two, pod torque is not insanely over-powered now. Three, the whole shebang uses electricity now for extra realism. I haven't had any exceptional stability problems.

On a side note, lag on looking at an ocean planet is now low enough that I can actually fly planes sensibly.

[skyace65]

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.

no there are some people on steam who say it works fine. well I'm going back to testing what people are saying on how to make rockets control better.

also, I did a test with a small rocket. 1 engine, 1 fuel tank, 1 command module, 1 battery, 3 wings, 3RCS, and it took 3 or 4 reaction wheels to properly keep it on course. assuming that there is no bug, this NEEDS a rebalance, I shouldn't have to slap on 3 reaction wheels to the tiniest rocket possible to make it fly strait

EDIT: okay after more testing turns out I only needed 2 reaction wheels not counting the command pod to make it fly as it should(doing so many tests I feel like glados) that is still way too much.

Edited July 25, 2013 by skyace65

[23278]

If I was building standard rockets, I could deal with this. However, I like to build a lot of style into my designs and the occasional slight asymmetrical craft. Getting them into orbit went from difficult to now near impossible, and they WILL NOT hold a heading in space anymore.